Fittingness: Essays in the Philosophy of Normativity [1 ed.] 2022937555, 9780192895882

Citation preview

Fittingness

Fittingness Essays in the Philosophy of Normativity Edited by

C H R I S T O P H E R HOWA R D AND R .  A .  R OW L A N D

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 2022 The moral rights of the authors have been asserted First Edition published in 2022 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: 2022937555 ISBN 978–0–19–289588–2 DOI: 10.1093/oso/9780192895882.001.0001 Printed and bound in the UK by Clays Ltd, Elcograf S.p.A. 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.

Contents List of Contributors

1. Fittingness: A User’s Guide Christopher Howard and R. A. Rowland

vii

1

SE C T IO N O N E :   T H E NAT U R E A N D E P I S T E M O L O G Y O F  F I T T I N G N E S S 2. The Deontic, the Evaluative, and the Fitting Selim Berker

23

3. Against the Fundamentality of Fit Thomas Hurka

58

4. What Is Evaluable for Fit? Oded Na’aman

80

5. Fitting Emotions Justin D’Arms

105

6. Intuitions of Fittingness Philip Stratton-­Lake

130

SE C T IO N T WO :   F I T T I N G N E S S , R E A S O N S , N O R M AT I V I T Y 7. Reasons and Fit Garrett Cullity

151

8. Value-­First Accounts of Normativity R. A. Rowland

176

9. Feasibility and Fitting Deliberation Nicholas Southwood

200

10. In Defence of the Right Kind of Reason Christopher Howard and Stephanie Leary

221

vi Contents

SE C T IO N T H R E E :   F I T T I N G N E S S A N D VA LU E T H E O RY 11. Value and Idiosyncratic Fitting Attitudes Conor McHugh and Jonathan Way

245

12. Well-­Being as Fitting Happiness Mauro Rossi and Christine Tappolet

267

13. The Things We Envy: Fitting Envy and Human Goodness Sara Protasi

290

14. Response-­Dependence and Aesthetic Theory Alex King

309

SE C T IO N F O U R :   F I T T I N G N E S S A N D R E SP O N SI B I L I T Y 15. Fittingness as a Pitiful Intellectualist Trinket? Michael McKenna

329

16. Blame’s Commitment to Its Own Fittingness Rachel Achs

356

17. Making Amends: How to Alter the Fittingness of Blame Hannah Tierney

380

Index

405

List of Contributors Rachel Achs, University of Oxford Selim Berker, Harvard University Garrett Cullity, Australian National University Justin D’Arms, Ohio State University Thomas Hurka, University of Toronto Christopher Howard, McGill University Alex King, Simon Fraser University Stephanie Leary, McGill University Conor McHugh, University of Southampton Michael McKenna, University of Arizona Oded Na’aman, Hebrew University of Jerusalem Sara Protasi, University of Puget Sound Mauro Rossi, Université de Montréal R. A. Rowland, University of Leeds Nicholas Southwood, Australian National University Philip Stratton-­Lake, University of Reading Christine Tappolet, Université de Montréal Hannah Tierney, University of California, Davis Jonathan Way, University of Southampton

1 Fittingness A User’s Guide Christopher Howard and R. A. Rowland

1.1  Introduction This volume explores the nature, roles, and applications of the notion of fittingness in contemporary normative and metanormative theory. The fittingness relation can be glossed as the relation in which a response stands to a feature of the world when that feature merits, or is worthy of, that response. It is thus the relation in which each of our responses stand when I admire an admirable effort, you laud a laudable performance, Beri believes a credible proposition, and Dhitri desires a desirable outcome. Likewise, it is the relation that fear stands in when its object is fearsome, that love stands in when its object is lovable, and that blame stands in when its object is blameworthy. Across these cases, the relevant attitudes or responses are merited by, and hence fitting with respect to, their objects.1 In the late nineteenth and mid-­twentieth centuries, this normative notion of fittingness occupied a prominent place in the theoretical toolkits of the period’s most influential ethical theorists. Then, up until the early aughts, discussion of the relation all but disappeared from the discourse of ethical theory. Today, the notion has regained prominence, promising to enrich the theoretical resources of contemporary normative theorists and taking centre stage in many debates in attitude’ normative and metanormative philosophy. For example, the ‘fitting-­ analysis of value is now perhaps the most well-­known and most discussed account of value and the notorious ‘wrong kind of reason problem’ (WKR problem) has emerged from discussion of this analysis. And there is now also great momentum behind the ‘fittingness-­first’ research programme, which tries to understand all of normativity ultimately in terms of fittingness. Still, despite its historical significance and the recent revival of interest, there has been no central discussion of the notion of fittingness to date. The present volume aims to fill this gap. The chapters to follow cover a range of topics including the nature and epis­ tem­ ol­ ogy of fittingness, the relation(s) between fittingness and reasons, the 1  For a similar gloss on fittingness, see Howard (2018).

Christopher Howard and R. A. Rowland, Fittingness: A User’s Guide In: Fittingness: Essays in the Philosophy of Normativity. Edited by: Christopher Howard and R. A. Rowland, Oxford University Press. © Christopher Howard and R. A. Rowland 2022. DOI: 10.1093/oso/9780192895882.003.0001

2 Fittingness normativity of fittingness, fittingness and value theory, and the role of fittingness in theorizing about responsibility. This introduction surveys these issues and ­highlights the chapters in which they’re discussed. We conclude with a brief discussion of issues to do with fittingness that aren’t covered extensively or at all by the contributions to the volume in order to indicate avenues for further research. This highlights our view of the volume as a conversation starter rather than as a comprehensive guide to, or exhaustive analysis of, all things related to fit. This volume’s chapters collectively if tacitly support the hypothesis that the notion of fittingness has great theoretical utility in grappling with a range of normative matters. We thus think that further study and application of the relation is called for, and we hope this volume helps to motivate this.

1.2  The Nature and Epistemology of Fittingness It’s sometimes supposed that the normative domain divides neatly into two broad kinds of normative categories: deontic ones (e.g. rightness, requiredness, and permissibility) and evaluative ones (e.g. goodness, badness, and betterness). The question of which of these fittingness falls under has been a matter of debate, though it’s often assumed that the relation is deontic (see, e.g. Rabinowicz and Rønnow-­Rasmussen 2004). Thus, fitting-­attitude analyses of value are sometimes held to involve a commitment about the relation of priority between the right and the good, taking the former to be prior to the latter. In his contribution to the volume, however, Selim Berker challenges this picture by arguing that fittingness is neither deontic nor evaluative but of its own normative kind with its own ­distinctive nature (Chapter  2). The result is a call to restructure historically entrenched thinking about the normative domain itself and our inquiry into it; this domain has more texture than is often thought, with fittingness occupying a sui generis and ineliminable place within it. An upshot, according to Berker, is that the traditional question of priority between the right and the good should be reframed as a question about the priority between the right, the good, and the fitting. Although fittingness is often taken to be a deontic category, the idea that the relation is sui generis within this category, and unanalysable in other normative terms (whether deontic or evaluative), has strong historical roots. Indeed, starting with Brentano (1889/1969), there’s a venerable tradition of taking fittingness not only to be unanalysable in other normative terms, but of holding that all other normative entities, whether deontic or evaluative, can be analysed in terms of it (see also Ewing 1947). Within the last decade there’s been a resurgence of interest in this ‘fittingness-­first’ approach to normativity, with several authors advancing different versions of it (Chappell 2012; McHugh and Way 2016, forthcoming; Cullity 2018; Howard 2019). This is partly a response to the rise in the early aughts of the ‘reasons-­first’ approach to normativity, which holds that normative

A User ’ s Guide  3 reasons are the fundamental normative entities in terms of which all others can be explained (Scanlon 1998; Schroeder 2007; Skorupski 2010; Parfit 2011; Rowland 2019). Many contemporary fittingness-­firsters have argued that their view has important advantages over this rival reasons-­first view, particularly given its fittingness-­based account of value which, unlike a reasons-­based account, seems to avoid the famous ‘WKR problem’ (more on which below; see also McHugh and Way’s Chapter 11). However, despite its seeming advantages, the fittingness-­first approach isn’t without its critics. In his contribution to the volume (Chapter 3), Thomas Hurka considers and rejects recent arguments in favour of fittingness-­ first and develops new arguments against it. Instead, Hurka defends the view that there is no single, normatively fundamental entity in terms of which all others can be accounted for and argues for a ‘two-­concept’ view, according to which rightness and goodness are equally normatively basic, unable to be explained in terms of each other or in terms of normative entities of any other kind. Regardless of whether fittingness is normatively first or constitutes a sui generis normative kind, other questions about its nature are pressing. One such question concerns what kinds of things are evaluable for it, i.e. what kinds of things are assessable as fitting or unfitting. Common to the contemporary and historical literature is the idea that attitudes, or intentional mental states, are assessable for fittingness, where this includes states like beliefs, emotions, desires, and intentions. And some have held that actions, too, are assessable for fittingness (see Section 1.6 below). Oded Na’aman’s contribution to the volume, however, proposes a radically liberal conception of what’s fit-­evaluable, according to which not only actions and attitudes are fit-­assessable but also physiological conditions such as headaches and heartrates (Chapter  4). According to Na’aman, this view not only aligns with our actual evaluative practices but is also supported by powerful theoretical considerations. Although the fit-­evaluability of certain types of human response is up for debate, it’s generally regarded as uncontroversial that emotions, at least, can be assessed as fitting or unfitting. Equally uncontroversial is the idea that the fittingness of every fitting emotion corresponds (extensionally) to some evaluative quality as possessed by the emotion’s object. It’s fitting to love someone, for instance, just in case they’re lovable, to deplore something just in case it’s deplorable, to contemn someone just in case they’re contemptible, and to adore something just in case it’s adorable.2 Moreover, it’s widely held that the fitting intensity of an emotion is constrained by the degree of the evaluative quality to which it’s a fitting response. For instance, how much adoration it’s fitting to feel towards Alan is constrained by how adorable Alan is (for discussion, see esp. Maguire 2018 and Berker’s Chapter  1). Thus, in terms made famous by Justin D’Arms and Daniel

2  For discussion, see, e.g. Brandt (1946), Schroeder (2010), Way (2012), and Howard (2018).

4 Fittingness Jacobson (2000), whether a token emotion is fitting (or how fitting it is) turns not only on whether it’s fitting in shape, i.e. whether it’s a response to an evaluative quality that renders emotions of its type fitting, but also on whether it’s fitting in size, i.e. whether its intensity is proportionate to the degree of the quality in ­question. In his contribution to the volume, Justin D’Arms proposes that the ­connections between the fittingness of certain emotions and the presence of ­certain evaluative qualities can be explained by appeal to the natures of the relevant emotions themselves (Chapter  5). Further, D’Arms suggests that we can gain insight into certain evaluative qualities by interrogating the natures of the emotions whose fittingness those qualities correspond to. For example, we can learn what sorts of non-­evaluative features make for shamefulness, or ground our judgements about this evaluative quality, by investigating the nature of shame. Turning from issues to do with the nature of fit and its explanatory role within the normative domain, we might also wish to know how we can know what’s ­fitting to what, or how we can gain knowledge of facts involving fittingness more generally. In principle, any moral epistemology might be applicable here, though of course some may be more plausible than others depending on our view about the metaphysics of fittingness. In his Foundations of Ethics, Ross suggests in several places that at least certain fittingness-­involving propositions—­for instance, that admiration is fitting towards what’s intrinsically good—­are self-­evident and that our knowledge of such a priori truths (and how we acquire it) is analogous to our knowledge of a priori truths in other domains, e.g. mathematics. In his contribution to the volume (Chapter 6), Philip Stratton-­Lake explores the prospects for this epistemology of fit, applying the specific rationalist epistemology of the a priori that he defends in other work (Stratton-­Lake 2016). Stratton-­Lake argues that this approach to the epistemology of fittingness is at least as promising in accounting for our knowledge of facts involving fit as it is in accounting for our knowledge of other kinds of normative facts, for example, facts about oughts and values.

1.3  Fittingness, Reasons, Normativity In the last several decades of normative philosophy, normative reasons have been all the rage. In What We Owe to Each Other, Scanlon famously suggests that the reason relation can’t be analysed or accounted for in more basic terms, but that it can only be glossed in terms of ‘favouring’ such that reasons ‘count in favour’ of what they’re reasons for. Unsurprisingly, Scanlon’s claim that facts about reasons are unanalysable resulted in many philosophers attempting to analyse them. And in recent years, several philosophers have attempted to explain facts about ­reasons in terms of facts about fittingness specifically (see, e.g. Danielsson and Olson

A User ’ s Guide  5 2007; Chappell 2012; Sharadin 2015; McHugh and Way 2016; Whiting 2022; and Howard 2019). One motivation for this, beyond the aspiration to analyse a philosophically interesting category, is to explain a putatively necessary connection between fittingness and reasons, viz. that facts that contribute to the fit of a response seem also to provide pro tanto reasons for it. For example, facts that contribute to a person’s admirability, and thus the fittingness of admiring them, seem also to provide reason to admire them. Likewise, facts that contribute to something’s fearsomeness, and thus the fittingness of fearing it, seem also to provide reasons to fear it. The hope is that a fit-­based analysis of reasons could provide an explanation of this and other interesting links between the two relations. Of course, a reasons-­based analysis of fit might be similarly explanatory, and some authors have attempted precisely this (e.g. Schroeder 2010; Rowland 2019). In his contribution to the volume, however, Garrett Cullity joins the former camp, advancing a new account of reasons in terms of fit (Chapter 7). Cullity argues that his fit-­based account of reasons is extensionally and explanatorily superior to existing accounts of the relationship between the two relations. An alternative approach to explaining the connections between fittingness and reasons appeals not to an analysis of one in terms of the other, but rather to ana­ lyses of each in terms of some third factor. One version of this approach appeals specifically to facts about value, or goodness, in order to explain facts about ­reasons and fit. A virtue of this approach is that it stands to explain not only facts about the relationship between fittingness and reasons, but also the connections between each of these and value. And there do seem to be connections here that call for explanation. For example, something’s being valuable seems both necessary and sufficient for its being fitting to value, and at least sufficient for there being reasons to value it. And value-­based accounts of each of reasons and fit might explain these connections. For instance, a value-­based account of fittingness might explain why something has value if and only if it’s fitting to value by appeal to the hypothesis that a thing’s being fitting to value consists in its being intrinsically valuable to value, plus the substantive, axiological claim that something is valuable if and only if it’s intrinsically valuable to value it (Hurka 2001). A value-­based account of fittingness might thus explain the above connection between what’s valuable and what’s fitting to value, by explaining fittingness directly in terms of the value of valuing responses that stand in this relation. Alternatively, a value-­based approach might explain facts about fittingness and reasons not in terms of facts about the value of responses that stand in these relations but rather in terms of facts about the value of the objects of such responses. For instance, a value-­based account of reasons that explains facts about reasons to value things fully in terms of facts about the value of those things might explain why there are always reasons to value valuable things (see, e.g. Orsi 2013b). Common to these different value-­based views is a commitment to the idea that facts about reasons and fittingness can be explained directly in terms of facts

6 Fittingness about the value of either the responses that stand in these relations or the objects of those responses. Also common to these views is that they seem to face powerful criticisms (see, e.g. Way 2013, Howard 2018, Rowland 2019, Kiesewetter 2022, and McHugh and Way’s Chapter 11). However, as R. A. Rowland argues in their contribution to the volume, a distinct, indirect value-­based view is also in the offing (Chapter 8). On this view, both facts about reasons and fit are explained by facts about the value of being guided by certain normative standards. Rowland’s chapter introduces and explores a version of this view, argues that it seems able to avoid the problems that direct value-­based views face, and compares the view against competing ones that would reverse the order of explanation, explaining facts about value in terms of facts about reasons or fit (more on which in Section Three of the volume; see esp. McHugh and Way’s Chapter 11). Discussions concerning the nature of reasons, value, and fittingness, and the relationships between these normative properties, are by this point familiar in the normativity literature. And, as was indicated above, it’s increasingly common today to hold that fittingness plays an important role in accounting for facts about reasons and value. In his contribution to the volume, Nicholas Southwood puts fittingness to work in a further, less explored but complementary way, arguing for a fit-­based account of a property that’s central to normative theorizing, viz. feasi­ bil­ity (Chapter 9). Building on recent work, Southwood develops and defends an account of feasibility in terms of fitting deliberation specifically.3 On this view, what makes an action feasible, roughly, is its being a fitting subject of deliberation about what to do. In addition to providing a novel account of a theoretically and practically important property, Southwood’s contribution thus demonstrates yet another way in which fittingness might be profitably appealed to in explaining properties of clear and central normative significance. In the wake of Scanlon’s reasons-­based account of value—­the ‘buck-­passing’ account of value—­came the notorious WKR problem (D’Arms and Jacobson 2000; Rabinowicz and Rønnow-­Rasmussen 2004). On the buck-­passing account, something has value if and only if there are reasons to value it. The WKR problem is the problem that, in some cases, there seem to be reasons to value things that have no value—­for example, if a demon threatens to kill me unless I value a saucer of mud (Crisp 2000).4 These reasons are of the ‘wrong kind’ to figure in the buck-­passing account: they’re reasons to value something that don’t also make the thing valuable, and hence constitute counter-­examples to the view. ‘Right-­kind’

3  Southwood introduces his fit-­based account of feasibility in Southwood (2022). 4  In addition to there being cases in which there seem to be reasons to value something of no value, there are also cases in which there seems to be no reason to value something that is of value (see, e.g. Bykvist 2009 and Reisner 2015). Putative counter-­examples of this second sort are sometimes taken to be instances of the WKR problem, but are other times characterized as exemplifying the ‘wrong kind of value’ problem. For critical discussion of this latter problem, see esp. Orsi (2013a), Elliott (2017), and Rowland (2019: ch. 7).

A User ’ s Guide  7 reasons to value something are reasons that do make the thing valuable and hence figure properly in the account. This is the origin of the right-­/wrong-­kind reason distinction in recent normative philosophy (for more on which, see esp. Gertken and Kiesewetter 2017). One response to the WKR problem is to try to distinguish between reasons of the right and wrong kind in a way that doesn’t make the buck-­passing account circular, and to revise the account such that it references only right-­kind reasons. A second response is to reject the counter-­examples that constitute the problem, i.e. to deny that putative wrong-­kind reasons for attitudes are genuine reasons for those attitudes. Following Jonathan Way (2012), call this latter response ‘WKR skepticism’. For a time, WKR skepticism was a dominant view (Skorupski 2010; Parfit 2011; Way 2012; Rowland 2015). Wrong-­kind reasons make good or beneficial the attitudes they seem to favour. But according to many philosophers, it’s not sufficient for a fact’s being a reason for an attitude that it makes the attitude good to have (see, e.g. the long line of skeptics about pragmatic reasons for belief). However, in recent years, many authors have come not only to reject WKR ­skepticism, but to defend an ethics of attitudes on which wrong-­kind reasons are the only genuine, authoritative reasons for attitudes there are (see esp. CôtéBouchard and Littlejohn 2018; Rinard 2019; and Maguire and Woods 2020). These philosophers are thus ‘RKR skeptics’: they deny that right-­kind reasons are genuine ­reasons for attitudes. In their contribution to the volume, Chris Howard and Stephanie Leary argue that RKR skepticism fails to capture intuitions about which attitudes we authoritatively ought to (or may) adopt and that existing arguments that right-­kind reasons aren’t reasons, or that they’re at best formally normative, are unsound; hence, we should accept that right-­kind reasons are genuine reasons for attitudes, i.e. that they can contribute to determining which attitudes we authoritatively ought to have (Chapter  10). On the popular and plausible assumption that right-­kind reasons are facts that contribute to the fittingness of the attitudes they favour,5 it follows that fittingness itself is an authoritatively ­normative relation. Hence, Howard and Leary’s argument doubles as an argument for the authoritative normativity of fit.

1.4  Fittingness and Value Theory The plausible idea that right-­kind reasons for attitudes contribute to the fittingness of the attitudes they favour suggests a promising solution to the WKR problem. Rather than explaining facts about value in terms of facts about reasons for valuing, we might instead explain the former facts in terms of facts about the 5  There are too many authors to cite here, but see, inter alia, Danielsson and Olson (2007), Chappell (2012), D’Arms and Jacobson (2014), Sharadin (2015), McHugh and Way (2016), and Howard (2019).

8 Fittingness fittingness of valuing. On this fit-­based account of value, for something to be good or valuable is for it to be fitting to value. As noted above, this view, unlike a reasons-­based account, doesn’t seem to face the WKR problem. For although the fact that a demon will kill you unless you value a saucer of mud may give you a reason to value the mud, this fact doesn’t make the mud fitting to value since it’s not a fact in virtue of which the mud merits, or is worthy of, a valuing attitude. Fittingness-­based accounts of value thus look to have an advantage over reasons-­based accounts in answering the WKR problem (cf. Rowland 2017). However, reasons-­based views may seem to have an advantage over fit-­based accounts when it comes to answering a different problem, viz. the problem of partiality (Ewing 1939; Bykvist 2009). Suppose there are two possible outcomes: one in which your friend is saved and a stranger dies and one in which the reverse occurs. And suppose also that, beyond this difference, all else is equal. Then assuming that being valuable or good (simpliciter) is an agent-­neutral evaluative notion, the two outcomes seem equally good. But intuitively, the balance of ­reasons supports your preferring the former. In addition, and crucially, it can seem fitting for you to have this preference (cf. Olson 2009). This data rules out a natural reasons-­based account of betterness on which one outcome’s being better than another consists in the balance of reasons supporting just anyone’s preferring it. Likewise, it rules out a fit-­based view on which an outcome’s being better than another consists in its being fitting for just anyone to prefer it. This is the problem of partiality. As Conor McHugh and Jonathan Way make clear in their contribution (Chapter  11), reasons-­based views have a plausible answer to this problem. Relying on the well-­founded distinction between agent-­neutral and agent-­relative reasons, proponents of reasons-­based views might restrict the reasons in terms of which they analyse facts about what’s better than what to agent-­neutral reasons specifically, claiming that one outcome’s being better than another consists in the balance of agent-­neutral reasons supporting just anyone’s preferring it (Stratton-­ Lake and Hooker 2006; Rowland 2019). This revised view is consistent with our intuitions about cases like the above: since the outcome where your friend lives is equally as good as the outcome where the stranger does, the balance of agent-­ neutral reasons supports being indifferent, but the balance of all the reasons, including agent-­ relative ones, supports your preferring the former. Hence, reasons-­based accounts have an elegant solution to the problem of partiality which draws on an independently motivated distinction between neutral and rela­tive reasons; hence, such views may seem to have an advantage over fit-­based views in answering this problem. However, McHugh and Way argue in their contribution that a broadly similar solution is available for fit-­based views, one which relies not on distinguishing between types of fittingness—­neutral and relative—­ but rather between types of valuing attitudes. Thus, McHugh and Way argue, insofar as their solution is equally as plausible as that offered by proponents of

A User ’ s Guide  9 reasons-­based accounts, reasons-­based views have no advantage over fit-­based accounts in answering the problem of partiality. In addition to figuring in an attractive metaphysical account of the nature of goodness (and betterness), fittingness can also be put to work in first-­order value theory, as Mauro Rossi and Christine Tappolet demonstrate in their contribution (Chapter  12). Rossi and Tappolet marshal fittingness specifically to provide a first-­order account of goodness-­for, or well-­being, according to which what’s basically good for people is fitting happiness. Rossi and Tappolet argue that this view avoids the standard objections to more traditional happiness-­based theories, while maintaining the intuitive, close connection between well-­being and happiness. One important feature of their proposal is a view of happiness as consisting in a complex of affective states including emotions, moods, and sensory pleasures, which, they claim, can themselves be assessed as fitting or unfitting; hence, the fit-­evaluability of happiness itself. In addition to contributing to the debate about what makes lives go well, Rossi and Tappolet thus also take a meaningful stand on the issue of what kinds of states can be assessed for fit. In addition to being profitably deployed to investigate different varieties of goodness, fittingness has also traditionally been appealed to in order to explicate more specific evaluative properties, such as that of being lovable, despicable, adorable, pitiful, delightful, and amusing. Indeed, as indicated above, and going as far back as Brandt (1946), it’s been widely held that such properties are at least equivalent to, if not analysable in terms of, the fittingness of various types of human response (see Berker’s Chapter  1 and D’Arms’s Chapter  5). Importantly, fit-­based analyses of these properties are compatible with the possibility that their analysanda are never instantiated, and hence that the responses in terms of which they’re analysed are never fitting. Historically, even realists about various other evaluative properties have found this thought tempting regarding certain ‘negative’ evaluative properties—­in particular, that of being enviable. Indeed, it may seem that envy, no matter the circumstances, could never be a merited response. However, in her contribution to the volume (Chapter 13), Sara Protasi resists this, arguing not only that envy can be fitting—­and hence that the property of being enviable can be instantiated—­but that it often is. Indeed, Protasi argues, envy’s central concern—­relative positioning—­is connected intimately and sys­tem­at­ic­ al­ly with our flourishing as human beings. Among the specific evaluative properties that might be explicated by appeal to fit are aesthetic properties, such as that of being delicious, charming, and beautiful. Fitting attitude theories of various aesthetic properties have been suggested in recent years by philosophers including, among others, Daniel Jacobson (2011) and Keren Gorodeisky (2021). Alex King, in her contribution to the volume (Chapter  14), argues that such theories can be helpfully classified as a kind of response-­dependence theory, where a response-­dependence theory of a property F claims that a thing’s F-ness depends on its bearing a certain relationship to a

10 Fittingness certain sort of human response (see also D’Arms’s Chapter  5; cf. Berker’s Chapter  1). According to fitting attitude theories, the relevant relationship to human responses is normative: something has a certain aesthetic property when, and because, a certain response is merited by—­or fitting with respect to—­the relevant thing. But on a different kind of response-­ ­ dependence theory— dispositionalism—­the relationship in question is purely descriptive. For example, a dispositionalist theory of a certain aesthetic property might claim that something has the property when and because, as a purely descriptive matter, humans are disposed to respond to it in a certain way. Historically, response-­ dependence views of aesthetic properties have skewed dispositionalist (see, e.g. Hume 1740/1975 and Kant 1790/2000). However, King argues that, in fact, ­fitting attitude theories have several important advantages over dispositionalist views. She highlights the versions of fitting attitude theories she takes to be most promising for explaining aesthetic value, but also raises several new ­challenges for theories of this kind. King’s contribution thus constitutes a ­thoroughgoing assessment of the prospects for response-­dependence views of aesthetic properties.

1.5  Fittingness and Responsibility Fittingness has a chequered history in the moral responsibility literature. It features prominently and helpfully in Joel Feinberg’s ‘Justice and Personal Desert’, for example, but is characterized (only a year earlier) as a ‘pitiful intellectualist trinket’ by P.  F.  Strawson in his seminal ‘Freedom and Resentment’. In the last decade, however, fittingness has assumed an increasingly visible and important role in work on moral responsibility. David Shoemaker (2017), for example, has recently defended a normative response-­dependence theory of blameworthiness, according to which someone is blameworthy if and only if, and because, they’re fitting to blame. This contrasts with (a possible interpretation of) Strawson’s view that amounts to a dispositionalist response-­dependence account, and with other normative response-­dependence theories of blameworthiness which specify a distinct normative relation that blame must stand in to its object for its object to be blameworthy (see, e.g. Wallace 1994). The final chapters of this volume investigate and discuss this and other ways in which fittingness may be a the­or­et­ic­al­ly useful concept in theorizing about moral responsibility. Michael McKenna’s contribution to the volume (Chapter  15) interrogates Strawson’s belittling remarks about fittingness, argues that a normative in­ter­pret­ ation of Strawson’s view of responsibility is required, and that in fact Strawson ought to have appealed in his own account to the very notion of fittingness that he seems to deride. In the course of making this case, McKenna also defends a Feinberg-­inspired account of desert as a species of fittingness and offers a novel

A User ’ s Guide  11 account of what differentiates deserved responses from merely fitting ones, drawing out the implications of this analysis for debates concerning free will and moral responsibility: although deserved blame might require the satisfaction of a strong freedom requirement, merely fitting blame does not. Normative response-­ dependence theories of blameworthiness that appeal to desert over and above mere fittingness may in this way be more metaphysically committal than those that appeal only to fit. Beyond the task of elucidating the nature of blameworthiness, a second and similarly important project, central to discussions of moral responsibility, is to investigate the nature of blame itself. One question is whether there’s anything important enough in common to blame responses to justify our theorizing about blame as such. As Rachel Achs makes clear in her contribution to the volume (Chapter  16), this question is pressing given the impressive diversity of blame responses. For example, one might blame by withdrawing care, verbally scolding, silently stewing, or via a critical subtweet. What, if anything, is distinctive of all these responses such that they’re worthy of investigation as kinds of  blame? Insofar as our theorizing about blame presupposes a single uniform subject, an answer to this question seems necessary to vindicate it. And, according to Achs, the notion of fittingness can be fruitfully appealed to here, in order to provide such an answer. On Achs’s view, blame essentially involves a kind of reflexive endorsement, i.e. a commitment to its own fittingness and to its being fitting directly on the basis of the target’s having done something wrong, and this is why blame, as such, is worthy of investigation as a unified phenomenon. In addition to identifying an essential feature of blame, Achs argues that this view also helps to explain certain other of blame’s marks—­namely, its directedness, and the felt character of blame common to many of its varied manifestations. Recent work in the ethics of attitudes has drawn attention to a puzzle concerning the fitting duration of attitudes (see esp. Marušić 2018, 2020; Na’aman 2021). Roughly: it seems fitting for certain attitudes to fade with time, for example, grief and regret. But it can seem that, in many cases, the facts that make these attitudes fitting persist. But if the facts that make an attitude fitting persist, then how can it be fitting for the attitude to fade? In short: why shouldn’t the attitude stay fitting forever? This is the puzzle, that of explaining how certain attitudes can fittingly fade over time. This puzzle applies widely, but can seem especially pressing in the case of blame (Callard 2018; Hieronymi 2001; Na’aman 2020). According to a common view, agents are blameworthy when and because they’ve culpably performed a wrongful act. But if an agent culpably performs a wrongful act, then it will always be true that they did. Hence, if an agent is blameworthy in virtue of culpably performing a wrongful act, then they’re blameworthy forever. Hence, if it’s fitting to blame the blameworthy, blame is forever fitting. But this seems ­intuitively false, and so something must give. In her contribution to the volume (Chapter  17), Hannah Tierney proposes a novel solution to this puzzle,

12 Fittingness spe­cif­ic­al­ly as it applies to blame. Tierney rejects the view that culpable wrongdoing suffices to make agents blameworthy over time, and instead proposes a reparative view of diachronic blameworthiness, on which agents stay blameworthy, and so fitting to blame, only insofar as their reparative duties to their victims go unfulfilled. Notably, although Tierney’s proposal stands to explain why fitting blame needn’t stay fitting forever, it entails that it could, insofar as blameworthy agents fail to satisfy their reparative duties.

1.6  The Future of Fit The discussion of fittingness in the twenty-­first century is still in its infancy. In closing this introduction, we want to highlight several potential avenues of further research concerning the nature of fittingness and its possible applications in philosophical theorizing. Some of these issues are touched on by contributions to the volume, but have yet to be fully pursued; others are wholly uncharted.

1.6.1  Fittingness and Correctness One important question concerns the relationship between fittingness and correctness. Many contemporary authors gloss if not analyse the fittingness of an attitude as a matter of the attitude’s satisfying a standard of correctness that is internal to or constitutive of it (McHugh and Way 2016; Schroeder 2010; Sharadin 2015). Some historical precedent for equating fittingness with a kind of correctness comes from Brentano (1889/1969), who uses the language of correctness in formulating his fitting attitude account of value: for something to be of value is for it to be ‘correct’ to love (16, 100). Brentano clarifies that ‘[o]ne loves or hates correctly provided that one’s feelings are adequate to their object—­adequate in the sense of being appropriate, suitable, or fitting’ (48). And it’s independently intuitive to think that there’s a close link between fittingness and correctness. It seems correct to admire admirable people, but incorrect to admire evil demons and deplorable dictators; and it may seem incoherent to claim, for example, that friendship is desirable but incorrect to desire. Furthermore, for any attitude, it seems quite natural to call it ‘correct’ insofar as it’s fitting However, there is more work to be done in clarifying the relationship between fittingness and correctness. For one, there are several senses in which an attitude might be ‘correct’. For example, an attitude’s being ‘correct’, in one sense, amounts to its satisfying a standard or norm, but ‘correct’ can also mean accurate or true. And indeed, although many philosophers have suggested that fittingness amounts to correctness in the first sense, some also suggest, instead, that fittingness is correctness in the second sense (e.g. Tappolet 2011; Rossi and Tappolet, this volume).

A User ’ s Guide  13 On this latter view, sometimes called the ‘alethic view’ (Rosen 2015), for an a­ ttitude to be fitting is for it to correctly, i.e. accurately, represent its object. But this view is rejected by many if not all those who hold that fittingness is correctness in the first sense, i.e. in the sense that amounts to norm satisfaction. One reason for this is that it’s not clear that attitudes that are correct in the norm-­satisfying sense always represent their objects accurately (or at all), and so the view that ­fittingness is correctness in the norm-­satisfying sense is at odds with the alethic view (Schroeder 2010; McHugh and Way 2016).6 But even among those who reject the alethic view and hold that fittingness is correctness in the sense of norm satisfaction, there is also a debate about whether the norms in question are ­internal to, constitutive of, or external to, the attitudes they seem to govern (on this, see D’Arms’s Chapter 5 and Howard and Leary’s Chapter 10). Considering the relationship between fittingness and correctness also leads to a further question about the nature of fit, viz. whether the relation is gradable. Goodness and badness are gradable properties—­something can be more or less good or bad—­and so too are the weights of reasons—­a reason can be more or less weighty. According to some, fittingness is also gradable, e.g. it can be more fitting to desire one thing than to desire another (Howard 2019). But according to ­others, fittingness is not a gradable property and is instead an all-­or-­nothing status, like requirement or permission (McHugh and Way 2016, this volume; Maguire 2018). Whether fittingness is gradable has implications for how it relates to non-­gradable deontic properties such as permissibility (see Berker’s Chapter 2 and Hurka’s Chapter 3) and may also be relevant to how we should formulate fitting attitude accounts of comparative evaluative properties, such as that of being better or worse. The possible equivalence of fittingness and correctness is relevant here since, on the face of things, correctness seems to be an all-­or-­nothing status rather than a gradable one: something can’t be more or less correct (cf. Wedgwood 2013). So, settling how fittingness relates to correctness may also involve settling whether fittingness is gradable.

1.6.2  Fitting Action Most discussions of fittingness in the last twenty years have focused on the fittingness of attitudes. But might actions be fit-­assessable too? As Hurka’s Chapter 3 and Stratton-­Lake’s Chapter 6 discuss, the idea that actions are fit-­assessable was common among philosophers writing about fit in the mid-­twentieth century: C. D. Broad (1930) claimed that we could understand Ross’s prima facie duties in terms of fittingness and Ross embraced this view in his Foundations of Ethics.

6  For further criticism of the alethic view, see especially Svavarsdóttir (2014) and Naar (2021).

14 Fittingness And indeed, it can seem hard to deny that actions are fit-­assessable. It seems ­fitting to praise the praiseworthy, for instance, and to discuss what’s worthy of discussion (Gertken and Kiesewetter 2017). Likewise, it seems fitting to punish those who merit punishment, and unfitting to reward them. Furthermore, one of the most influential accounts of the right-­/wrong-­kind reason distinction, due to Mark Schroeder (2010), holds that this distinction arises in any domain that’s constitutively governed by a standard of correctness (see also Sharadin 2015). On this view, since there’s a constitutively correct way to set the table for a royal dinner and to execute the Queen’s Gambit, there can be right-­kind reasons to set the royal table in a certain way and to move one’s chess pieces. So, assuming that right-­kind reasons for a response explain or indicate the response’s fit, it follows that at least some types of acts can be fitting. But what might it take for an act to be fitting? If we accept a Schroeder-­style view, then our answer should differ depending on the type of act at issue, given that different types of act are constitutively governed by their own, differing standards of correctness. However, this may ultimately seem like a reason to reject the Schroeder-­style analysis. For as Howard and Leary suggest in Chapter 10, it seems that an act could be correct by its own standard, but unfitting—­for example, a ­correctly executed pirouette might be unmerited and hence unfitting in the ­context of a funeral. An alternative view is that the fittingness of acts is similar to the fittingness of attitudes in the following way. The fittingness of an attitude is unaffected by facts about the good or bad consequences of having the attitude: the good consequences of admiring an evil demon can’t make the demon fit to admire since they don’t make them admirable, or worthy of admiration. Likewise, we might think that the fittingness of an act is determined by considerations other than those to do with the good or bad consequences of performing the act. Instead, we might say that the fittingness of an action is similarly determined entirely by facts about whether the action is merited by the situation, or by certain features of it. This seems to be the view espoused by Broad (1930/1956: 221). But this view faces some difficulties, one of which is pointed out by Ross (1939/1949: 81–2). Contrary to Broad, Ross saw no reason why the goodness of an act’s outcome couldn’t affect whether the act is fitting to a situation. After all, one feature of your situation might be that if you were to act in a certain way, you’d produce a good outcome. So, why couldn’t this feature of your situation even in principle merit the performance of the act? On Ross’s view, then, the goodness of an act’s outcome is among the factors that can contribute to its fittingness. If this view is right, then there is an important difference between the fittingness of acts and attitudes: the goodness of performing an act can contribute to its fit, whereas the fittingness of an attitude is unaffected by its goodness. One issue for the Rossian view is this. Consider acts of praise and blame. It seems fitting to praise only the praiseworthy and to blame only the blameworthy.

A User ’ s Guide  15 Yet the view that the goodness of an act’s consequences can affect the act’s fit seems in tension with these claims. For example, if this view is right, then presumably there are possible scenarios in which it’s fitting to praise someone who is not praiseworthy, viz. scenarios in which the consequences of doing so are sufficiently good. Worries like this may lead us back to a Broadian view on which the fittingness of actions is similar to the fittingness of attitudes in being unaffected by considerations to do with their good consequences. Hence, the question of what it might take for an action to be fitting remains unsettled. And indeed, it remains open whether acts are fit-­assessable at all.7 This area is thus ripe for further research.

1.6.3  New Applications The chapters in the fourth part of this volume explain how fittingness can be appealed to in order elucidate properties and attitudes of central significance to the moral domain, viz. responsibility and blame. And other work on fittingness, including Alex King’s contribution to the volume (Chapter 14), has explored the utility of fittingness for understanding the normativity of aesthetic properties. A further avenue for future fittingness research concerns whether fit can be ­pro­duct­ive­ly appealed to in investigating various other domains and their normativity. Nicholas Southwood’s contribution to the volume articulates an account of feasi­bil­ity in terms of fit (Chapter 9). One of the main drivers behind Southwood’s project is to understand the nature and role of feasibility in politics since feasibility is an especially politically significant property. But fittingness might have other applications in political philosophy too. For example, some political theorists have suggested that political normativity is importantly distinct from moral normativity (Williams 2005; Rossi and Sleat 2014; cf. Maynard and Worsnip 2018). Since the normativity of fittingness isn’t always moral, one possibility is that fit could be appealed to in elaborating this hypothesis. For instance, some suggest it can be fitting to be amused by immoral jokes insofar as they’re amusing (D’Arms and Jacobson 2000); and it might be fitting to envy someone even if envying them is overall morally bad (see, e.g. Protasi’s Chapter 13). Hence, given that the normativity of fit seems already to extend beyond moral normativity, we might postulate a form of fit-­related normativity that extends beyond the moral in the political realm. Indeed, some moves in this direction have already been made by political theorists attracted to related ideas. For example, Michael Walzer (1980) appeals to the ‘fit’ between a state’s government and its community to explain why, on his view, the moral benefits of humanitarian intervention inside a nation state’s territory don’t always override states’ rights to political sovereignty. 7  For reasons to doubt that actions per se can be fitting, as opposed to intentions to perform them, see McHugh and Way (2022, forthcoming); see also Rowland.

16 Fittingness Fit might be put to use to explain concepts and properties in social philosophy, too. For instance, we might appeal to it in order to understand certain identities. People who have a gender identity at odds with the gender they were assigned at birth often explain that they experience the latter gender as not fitting them, or judge that it’s not fitting to treat them as having that gender (e.g. Serano 2016: 226; Weiss 2018). One possibility is that the notion of fit at issue here is purely descriptive—­corresponding to the relation that obtains, for instance, when a puzzle piece fits into place—­and is hence distinct from the notion of fit that’s the subject of this volume’s investigation, i.e. the notion that can be paraphrased in terms of merit and worthiness. But another possibility is that the sense of ‘fit’ that figures in claims like the above in fact refers to our target relation, and could thus be usefully appealed to in order to shed light on gender concepts and our judgements about them (Rowland forthcoming). These are just some of the avenues of future fittingness research that seem interesting and important to us. We hope this introduction, and the diverse and wide-­ranging contributions to this volume, will encourage readers to think more about fittingness and its potential to shed light on various normative matters, across a variety of domains. We’re excited to see what the future of fit might hold.

References Brandt, Richard (1946). ‘Moral Valuation’. Ethics 56: 106–21. Brentano, Franz (1889/1969). The Origin of Our Knowledge of Right and Wrong. Trans. Roderick Chisholm. Routledge and Kegan Paul. Broad, C. D. (1930/1956). Five Types of Ethical Theory. Routledge and Kegan Paul. Bykvist, Krister (2009). ‘No Good Fit: Why the Fitting Attitude Analysis of Value Fails’. Mind 118 (469): 1–30. Callard, Agnes (2018). Aspiration: The Agency of Becoming. OUP. Chappell, Richard Yetter (2012). ‘Fittingness: The Sole Normative Primitive’. Philosophical Quarterly 62 (259): 684–704. Côté-Bouchard, Charles and Clayton Littlejohn (2018). ‘Knowledge, Reasons, and Errors about Error Theory’. In Robin McKenna and Christos Kyriacou (eds), Metaepistemology: Realism and Antirealism. Palgrave. Crisp, Roger (2000). ‘Review of Kupperman, Value . . . and What Follows’, Philosophy 75: 458–9. Cullity, Garrett (2018). Concern, Respect, and Cooperation. OUP. Danielsson, Sven and Jonas Olson (2007). ‘Brentano and the Buck-Passers’. Mind 116 (463): 511–22.

A User ’ s Guide  17 D’Arms, Justin and Daniel Jacobson (2000). ‘The Moralistic Fallacy: On the “Appropriateness” of Emotions’. Philosophy and Phenomenological Research 61 (1): 65–90. D’Arms, Justin and Daniel Jacobson (2014). ‘Wrong Kinds of Reason and the Opacity of Normative force’. Oxford Studies in Metaethics 9: 215–44. Elliott, Aaron (2017). ‘Reasons, Dispositions, and Value’. Philosophers’ Imprint 17 (3): 1–18. Ewing, Alfred Cyril (1939). ‘A Suggested Non-Naturalistic Analysis of Good’. Mind 48 (189): 1–22. Ewing, Alfred Cyril (1947). The Definition of Good. Routledge and Kegan Paul. Gertken, Jan and Benjamin Kiesewetter (2017). ‘The Right and the Wrong Kind of Reasons’. Philosophy Compass 12: 1–14. Gorodeisky, Keren (2021). ‘On Liking Aesthetic Value’. Philosophy Phenomenological Research 102 (2): 261–80.

and

Hieronymi, Pamela (2001). ‘Articulating an Uncompromising Forgiveness’. Philosophy and Phenomenological Research 62 (3): 529–55. Howard, Chris (2018). ‘Fittingness’. Philosophy Compass 13 (11): 1–14. Howard, Chris (2019). ‘The Fundamentality of Fit’. In Russ Shafer-Landau (ed.), Oxford Studies in Metaethics, Volume 14. OUP. Hume, David (1740/1975). Treatise of Human Nature, ed. L. A. Selby-Bigge. OUP. Hurka, Thomas (2001). Virtue, Vice, and Value. OUP. Jacobson, Daniel (2011). ‘Fitting Attitude Theories of Value.’ In Edward N. Zalta (ed.), The Stanford Encyclopedia of Philosophy (Spring 2011 Edition), http://plato.stanford. edu/archives/spr2011/entries/fitting-attitude-theories/. Kant, Immanuel, 1790 [2000], Critique of the Power of Judgment  (Kritik der Urteilskraft). Trans. Paul Guyer and Eric Matthews. CUP. Kiesewetter, Benjamin (2022). ‘Are All Practical Reasons Based on Value?’. In Russ Shafer-Landau (ed.), Oxford Studies in Metaethics, Volume 17. OUP. Maguire, Barry (2018). ‘There Are No Reasons for Affective Attitudes’. Mind 127 (507): 779–805. Maguire, Barry and Jack Woods (2020). ‘The Game of Belief ’. Philosophical Review 129 (2): 211–49. Marušić, Berislav (2018). ‘Do Reasons Expire? An Essay on Grief ’. Philosophers’ Imprint 18: 1–21. Marušić, Berislav (2020). ‘Accommodation to Injustice’. In Russ Shafer-Landau (ed.), Oxford Studies in Metaethics 15: 263–83. Maynard, Jonathan Leader and Alex Worsnip (2018). ‘Is There a Distinctively Political Normativity?’ Ethics 128: 756–87. McHugh, Conor and Jonathan Way (2016). ‘Fittingness First’. Ethics 126 (3): 575–606. McHugh, Conor and Jonathan Way (2022). ‘All Reasons Are Fundamentally Reasons for Attitudes’. Journal of Ethics and Social Philosophy 21: 151–174.

18 Fittingness McHugh, Conor and Jonathan Way (forthcoming). Getting Things Right: Fittingness, Value, and Reasons. OUP. Na’aman, Oded (2020). ‘The Fitting Resolution of Anger’. Philosophical Studies 177: 2417–30. Na’aman, Oded (2021). ‘The Rationality of Emotional Change: Toward a Process View’. Noûs 55 (2): 245–69. Naar, Hichem (2021). ‘The Fittingness of Emotions’. Synthese 199: 13601–19. Olson, Jonas (2009). ‘Fitting Attitude Analyses of Value and the Partiality Challenge’. Ethical Theory and Moral Practice 12: 365–78. Orsi, Francesco (2013a). ‘Fitting Attitudes and Solitary Goods’. Mind 122 (487): 687–98. Orsi, Francesco (2013b). ‘What’s Wrong with Moorean Buck-Passing’. Philosophical Studies 164 (3): 727–46. Parfit, Derek (2011). On What Matters, Volume 1. OUP. Rabinowicz, Wlodek and Toni Rønnow-Rasmussen (2004). ‘The Strike of the Demon: On Fitting Attitudes and Value’. Ethics 114 (3): 391–423. Reisner, Andrew (2015). ‘Fittingness, Value, and Trans-World Attitudes’. Philosophical Quarterly 65 (260): 464–85. Rinard, Susanna (2019). ‘Believing for Practical Reasons’. Noûs 53 (4): 763–84. Rosen, Gideon. (2015). ‘The Alethic Conception of Moral Responsibility’. In R. Clarke, M.  McKenna, and A.  M.  Smith (eds.), The Nature of Moral Responsibility: New Essays. OUP. Ross, W. D. (1939/1949). Foundations of Ethics. Clarendon Press. Rossi, Enzo and Matt Sleat (2014). ‘Realism in Normative Political Theory’. Philosophy Compass 9: 689–70. Rowland, R. A. (2015). ‘Dissolving the Wrong Kind of Reason Problem’. Philosophical Studies 172 (6): 1455–74. Rowland, R. A. (2017). ‘Reasons or Fittingness First?’ Ethics 128 (1): 212–29. Rowland, R. A. (2019). The Normative and the Evaluative: The Buck-Passing Account of Value. OUP. Rowland, R.  A. (forthcoming a). ‘Reasons as Reasons for Preferences’. American Philosophical Quarterly. Rowland, R.  A. (forthcoming b). ‘Gender Incongruence and Fit’. Australasian Philosophical Review. Scanlon, T. M. (1998). What We Owe to Each Other. Harvard UP. Schroeder, Mark (2007). Slaves of the Passions. OUP. Schroeder, Mark (2010). ‘Value and the Right Kind of Reason’. In Russ Shafer-Landau (ed.), Oxford Studies in Metaethics, Volume 5. OUP. Schroeder, Mark (2012). ‘The Ubiquity of State-Given Reasons.’ Ethics 122: 457–88. Serano, Julia (2016). Whipping Girl. Seal Press.

A User ’ s Guide  19 Sharadin, Nathaniel (2015). ‘Reasons Wrong and Right’. Pacific Philosophical Quarterly 96 (2): 1–29. Shoemaker, David (2017). ‘Response-Dependent Responsibility; or, A Funny Thing Happened on the Way to Blame’. Philosophical Review 126 (4): 481–527. Skorupski, John (2010). The Domain of Reasons. OUP. Southwood, Nicholas (2022). ‘Feasibility as Deliberation-Worthiness’. Philosophy and Public Affairs 50: 121–162. Stratton-Lake, Philip (2016). ‘Intuition, Self-Evidence, and Understanding’. In Russ Shafer-Landau (ed.), Oxford Studies in Metaethics, Volume 11. OUP. Stratton-Lake, Philip and Brad Hooker (2006). ‘Scanlon versus Moore on Goodness’. In Terry Horgan and Mark Timmons (eds.), Metaethics after Moore. OUP. Strawson, Peter (1962). ‘Freedom and Resentment’. Proceedings of the British Academy 48: 187-211. Svavarsdottir, Sigrún (2014). ‘Having Value and Being Worth Valuing’. Journal of Philosophy 3: 84–109. Tappolet, Christine (2011). ‘Values and Emotions: Neo-Sentimentalism’s Prospects’. In C. Bagnoli (ed.), Morality and the Emotions. OUP. Wallace, R. Jay (1994). Responsibility and the Moral Sentiments. Harvard UP. Walzer, Michael (1980). ‘The Moral Standing of States: A Response to Four Critics’. Philosophy and Public Affairs 9 (3): 209–29. Way, Jonathan (2012). ‘Transmission and the Wrong Kind of Reason’. Ethics 122 (3): 489–515. Way, Jonathan (2013). ‘Value and Reasons to Favour’. In R.  Shafer-Landau (ed.), Oxford Studies in Metaethics, Volume 8. OUP. Wedgwood, Ralph (2013). ‘Doxastic Correctness’. Aristotelian Society Supplementary Volume 87 (1): 217–34. Weiss, Suzannah (2018). ‘9 Things People Get Wrong about Being Non-Binary’. Teen Vogue, 15 February.   https://www.teenvogue.com/story/9-things-people-get-wrongabout-being-non-binary. Whiting, Daniel (2022). The Range of Reasons in Ethics and Epistemology. OUP. Williams, Bernard (2005). In the Beginning Was the Deed: Realism and Moralism in Political Argument. Princeton UP.

SECTION ONE

T HE NAT U RE A ND EPIST E MOLO GY OF F I T T I NG NE S S

2

The Deontic, the Evaluative, and the Fitting Selim Berker

2.1 Introduction The normative categories can be split up into several distinct families.1 The most familiar of these are the deontic and the evaluative categories, or ‘the right’ and ‘the good’ as they are commonly known. The members of each family of cat­egor­ ies (i) resemble each other in certain ways and (ii) bear certain relationships to each other that they don’t bear to non-­family members. Think here of the ties between required, permitted, and forbidden, on the one hand, and between good, better, and best, on the other. It is these interconnections that make it so natural to split up the deontic and the evaluative into separate families and then ask that well-­worn question, ‘Is the good prior to the right, or the right prior to the good?’ Where do the fittingness categories—­fitting, apt, warranted, admirable, fearsome, shameful, and the like—­fit into this familiar taxonomy? The overwhelming tendency among contemporary philosophers has been to shoehorn them into the standard bipartite framework. Most authors assume that the fittingness categories are deontic categories, and hence are forms of either requiredness or permissibil­ ity (although it is usually left unclear which). Other authors go the other way and assume that the fittingness categories are evaluative categories, and hence are types of goodness and badness, of value and disvalue. The purpose of this chapter is to argue that both of these tendencies are a mis­ take. The fittingness categories are neither a subclass of the deontic categories nor a cross section of the evaluative categories. Rather, they constitute a separate fam­ ily of normative categories, with their own distinctive interrelations, their own distinctive logic, and their own distinctive nature. There are at least three major 1  As is by now customary, I use ‘normative’ as an umbrella term referring to anything having to do with oughtness, value, virtue, merit, and so on. I use ‘category’ as a term of art that encompasses not just properties and relations but also entities, quantifiers, connectives, and anything else that is the worldly analogue of words or concepts, in the way in which the property being red is the worldly analogue of the predicate ‘is red’ (a piece of language) and the concept is red (a vehicle of thought). I use italics to designate categories, and I occasionally drop the copula ‘being’ when referring to properties and relations, if no confusion results.

Selim Berker, The Deontic, the Evaluative, and the Fitting In: Fittingness: Essays in the Philosophy of Normativity. Edited by: Christopher Howard and R. A. Rowland, Oxford University Press. © Selim Berker 2022. DOI: 10.1093/oso/9780192895882.003.0002

24 Fittingness families of normative categories, and we should resist the urge—­tempting though it is—­to treat the fitting as simply another variety of the deontic or the evaluative.2

2.2  The Deontic and the Evaluative It will help to begin with a review of the deontic and the evaluative categories, and of why it is customary to see them as forming distinct families. The deontic categories include at least the following: required/obligatory/mandatory permitted/permissible/allowed forbidden/impermissible/prohibited

ought to/should/must/have a duty to may/can (on a deontic reading)

optional/merely permitted (i.e. permitted but not required) The evaluative categories, by contrast, include at least the following: good better than bad worse than neutral/indifferent equal in value to on a par with (?)

strongly best (i.e. better than all the rest) weakly best (i.e. at least as good as all the rest) strongly worst weakly worst

at least as good as second best no worse than good enough at least as bad as better than most no better than Note that both families of categories might come in various ‘flavours’. For instance, maybe we can distinguish moral, legal, prudential, epistemic, all-­things-­considered, and so on varieties of the deontic categories. Note, also, that various distinctions can be made within a given category. For instance, maybe we can distinguish final and instrumental badness, attributive and predicative goodness, basic and derivative requirements, directed and undirected duties, and so on. Corresponding to each of these families of categories (properties, relations, entities, etc.) is a family of terms and a family of concepts. So in addition to de­ontic categories such as being required, there are deontic terms such as ‘required’ and deontic concepts such as required, and in addition to evaluative categories

2 

Where do reasons fit into my proposed three-part taxonomy? I turn to that issue in §2.7, below.

The Deontic, the Evaluative, and the Fitting  25 such as being good, there are evaluative terms such as ‘good’ and evaluative con­ cepts such as good. For the most part I shall conduct my inquiry in a meta­phys­ ic­al mode and formulate my claims in terms of categories rather than words or concepts, because I hold that ethics is primarily concerned with investigating the nature of, say, requiredness itself, not with investigating the nature of a certain word in English or of a certain vehicle of thought, and similarly for other normative disciplines. But readers who prefer to approach normative subjects in a semantic (or conceptual) key can feel free to translate most of my claims about categories into analogous claims about words (or concepts) via semantic (or conceptual) ascent. There are myriad structural and substantive differences between the deontic and the evaluative categories that justify grouping them together into separate families. I mention here five.3 The first difference: requiredeness and permittedness are duals of one another, but no such relation exists among the evaluative categories. This relation of duality takes several forms. On the one hand, requiredness and permittedness are extensional duals, in that the following biconditional—­or something close to it—­holds: B1.  ϕ-ing is required iff not-ϕ-ing is not permitted.4 But we do not have: B1ʹ.  x is good iff not-x is not E,

3  See also Smith (2005: §3) and Tappolet (2013). There is one commonly cited difference between the deontic and the evaluative on which I do not focus, because I think it is a mistake to take this to be genuine difference: I have in mind the claim that deontic categories such as ought and required are connected to voluntary choice or action in a way that evaluative categories are not. But this alleged distinguishing feature is spurious: there can be things we ought to believe or intend despite belief and intention not being under our direct or even indirect voluntary control (Feldman 2008; Hieronymi 2008; Chuard and Southwood 2009; McHugh 2017). 4  I add the ‘or something close to it’-qualifier to allow that B might not be the best way of formu­ 1 lating extensional duality. One variant formulation adds in a requirement that ϕ-ing be the sort of thing to which norms of the relevant flavour—­whether moral, or legal, or whatever—­apply:

B1*.  If ϕ-ing is subject to norms of variety x, then (ϕ-ing is requiredx iff not-ϕ-ing is not permittedx). (An advantage of this formulation: it allows moral error theorists to say both that nothing is morally required and that nothing is morally permitted.) A second variant formulation replaces the appeal to ϕ-ing’s negation with a quantification over ϕ-ing’s alternatives: B1**.  ϕ-ing is required iff every alternative to ϕ-ing is not permitted. (An advantage of this formulation: it allows one to countenance relations of extensional duality even if one holds, say, that the positive state believing that p is deontically assessable whereas the negative state not believing that p is not.) Analogues of every argument I go on to make for B1 could also be made with B1* and B1**, so I stick with the more familiar formulation in the main text.

26 Fittingness where ‘E’ refers to some other evaluative category. For one thing, we need not hold that the sorts of things that can be good are also the sorts of things that can be negated, so ‘not-x’ might not make sense for some relevant values of x. And for another, even when ‘not-x’ does make sense (as it arguably does when x is an action or a state of affairs), that x is good is compatible with not-x being good or bad or neutral: sometimes it is good both to do something and not to do it, and other times it is good to do something without its being actively bad or actively good to refrain from doing it. Because of their extensional duality, requiredness and permittedness bear certain logical relations to each other that have no analogue for the evaluative cat­egor­ies. But there is another form of duality that leads to an even deeper difference. Requiredness and permittedness are also definitional duals, insofar as they are interdefinable via their relation of duality, like so: D1.  ϕ-ing is required =df not-ϕ-ing is not permitted. D2. ϕ-ing is permitted =df not-ϕ-ing is not required.5 Some authors act as if we must choose between these two definitions: either requiredness is defined in terms of permittedness, or permittedness is defined in terms of requiredness, but we cannot have both. I myself am suspicious of this commonly held thought. There is a deep-­rooted symmetry between requiredness and permittedness, and to say that one is more fundamental than the other strikes me as unacceptably arbitrary, in the way in which it is unacceptably arbitrary to hold that possibility is more fundamental than necessity or disjunction more fun­ damental than conjunction. It is far better, I would say, to hold onto the intuitive thought that these two deontic categories are interdefinable—­or, more precisely, are interdefined—­and hence to hold that D1 and D2 are both true.6 And regardless of whether we accept both D1 and D2 or only one of them, the crucial point is that no analogous definition of goodness, badness, or any other evaluative category exists.7 I have been focusing on the duality of requiredness and permittedness, but this is just one among a number of similar relations that hold among the 5  I leave it open whether these definitions are semantic, conceptual, metaphysical, or some combination thereof (maybe each is a semantic definition that entails a corresponding metaphysical definition). For more on metaphysical definition, see Rosen (2015). 6  If doing so is in tension with the sorts of foundationalist assumptions usually made by philosophers when they theorize about definitions, then so much the worse for those assumptions, I would say. 7  What about strong and weak bestness? Aren’t they evaluative categories that are duals of one another? That is far from clear. To get them to be extensional duals, we need to make contentious assumptions either about the range of ϕ-ing’s alternatives or about how the value of not-ϕ-ing relates to the value of more specific ways of not-ϕ-ing. (The basic problem: how do we establish the right-toleft direction of if ϕ-ing has alternatives other than not-ϕ-ing?) Moreover, even if strong and weak bestness turn out to be extensional duals, it is not plausible that they are definitional duals: a thing’s being better than all the rest cannot be defined as its negation’s not being at least as good as all the rest.

The Deontic, the Evaluative, and the Fitting  27 traditional deontic categories of requiredness, permittedness, and forbiddenness (i.e. impermissibility). Forbiddenness is related to the other two like so: B2.  ϕ-ing is forbidden iff not-ϕ-ing is required. B3. ϕ-ing is forbidden iff ϕ-ing is not permitted. Moreover, although it is commonly overlooked and lacks a natural prefix-­less term in English, there is a fourth central deontic status—­ let us call it ‘unrequiredness’8—­that is the dual of forbiddenness and bears corresponding relations to requiredness and permittedness: B4. ϕ-ing is unrequired iff not-ϕ-ing is not forbidden. B5. ϕ-ing is unrequired iff not-ϕ-ing is permitted. B6. ϕ-ing is unrequired iff ϕ-ing is not required. B1–B6 give us the logical relations that make up the traditional square of op­pos­ ition. And underwriting these biconditionals are a raft of definitional intercon­ nections beyond D1 and D2, such as: D3.  D4. D5. D6.

ϕ-ing is forbidden =df not-ϕ-ing is required. ϕ-ing is required =df not-ϕ-ing is forbidden. ϕ-ing is forbidden =df ϕ-ing is not permitted. ϕ-ing is permitted =df ϕ-ing is not forbidden.9

We find nothing like this structure within the evaluative categories. A second difference: goodness and badness are non-­privative opposites of a certain sort, but no deontic categories are related in this way. Permittedness and forbiddenness are privative opposites (or complements): each is the absence or lack of the other. To be forbidden is to be non-­permitted, and to be permitted is to be non-­forbidden. By contrast, goodness and badness are what I call polar opposites: each is the inversely charged flipside of the other, not its mere lack or absence. To be bad is to be anti-­good, not to be non-­good, and to be good is to be anti-­bad, not to be non-­bad.10 No deontic categories are opposites of this sort. Forbiddenness is 8  Paul McNamara (2019) calls this fourth deontic status ‘omissibility’, but I have reservations about his terminology that will emerge in due course. 9  I omit here six additional definitions linking unrequiredness to the other three traditional deontic categories. So in total we have four deontic categories tethered together by six biconditionals and twelve definitions. 10  Here the prefix ‘anti-’ must here be understood as expressing inversion, as in ‘antihero’ or ‘anti­ climax’, not as expressing adversariality, as in ‘anti-aircraft’ or ‘anti-Semite’. Compare Ibram X. Kendi’s (2019) notion of an anti-racist, who is not merely someone who opposes racists, but rather is a kind of inverse racist who actively works to level the playing field among the races. Indeed, Kendi’s central distinction between an anti-racist and a non-racist is exactly the distinction I mean to be invoking in distinguishing anti-goodness from non-goodness.

28 Fittingness neither anti-­requiredness nor anti-­permittedness; rather, it is related to requiredness via D3 and to permittedness via D5. Just as duality is a distinctive relation that we find among the deontic categories but not among the evaluative categories, good­ ness and badness’s polar opposition is a distinctive relation that we find among the evaluative categories but not among the deontic categories. A third difference, which perhaps is the most obvious: goodness and badness are gradable, but the traditional deontic categories are not. The evaluative properties being good and being bad are gradable, which involves a package of three related features: first, they have comparative forms (being better than, being worse than, etc.); second, they have superlative forms (being best, best worst, etc.); and, third, they can be acted on by very, somewhat, and other grading modifiers (so there exist the properties being very good, being somewhat bad, etc.). By contrast, all of the deontic categories we have considered so far are not gradable. It is never the case that one action is ‘more required’ than another, or that one of the avail­able actions is ‘the most permitted’, or that another of the available actions is ‘somewhat forbidden’.11 A fourth difference, closely related to the previous one: in addition to the ­properties being good and being bad, there is a third, in-­between property being neutral,12 but no analogous neutral status exists for the deontic categories. The closest candidate here is being merely permitted (i.e. being permitted but not required), but that isn’t a true neutral, in-­between status. Goodness and badness fall on a (perhaps non-­scalar, perhaps multidimensional) spectrum with neutrality lying in between, but requiredness and forbiddenness do not fall on a spectrum with mere-­permittedness lying in between. If I can save someone’s life by giving them either antidote A or antidote B, we would never say, ‘Giving that person antidote A is neither required nor forbidden but in between,’ in the way in which we might say of a different situation, ‘That outcome is neither good nor bad but in between.’13

11  Two comments about gradability. First, my central claim here is about the gradability of categories—­of the properties themselves—­not about the gradability of words or other linguistic items. Gradability is a metaphysical phenomenon, in addition to being a semantic phenomenon that can be studied by linguists. Second, although philosophers and linguists often implicitly assume otherwise, being gradable is not the same as coming in degrees. Degreedness entails gradability, but gradability does not entail degreedness. If being F is gradable but being more F than allows cycles, then it will not make sense to reify a thing’s F-ness into degrees. (I owe this point to Angel Navidad.) If being F is gradable but being more F than is a massively partial ordering, so that there are a large number of cases in which some x is neither more F than, nor less F than, nor equal in F-ness to some y, then it will also not make sense to reify a thing’s F-ness into degrees. And so on. 12  Other common names for this status: ‘being indifferent’ and, especially in the case of attributive goodness, ‘being average’ (as in: ‘He’s neither a good nor a bad free throw shooter, but merely average’). 13  That the deontic categories are non-gradable and lack a neutral middle state might seem obvi­ ous, but these points are not always appreciated. In a study with over 1,000 citations, psychologists Fiery Cushman, Lianne Young, and Marc Hauser asked their subjects to rate a number of actions ‘on a scale from 1 to 7, with 1 labeled “Forbidden,” 4 labeled “Permissible,” and 7 labeled “Obligatory”’ (2006: 1083). Even if we set aside the worry that obligatoriness entails permissibility and interpret ‘Permissible’ here as meaning ‘Merely Permissible’, it is difficult to make sense of what Cushman et al. were asking their subjects. What does 3 on their scale signify? A little bit forbidden? What does 5 sig­ nify? Permissible but not quite obligatory?

The Deontic, the Evaluative, and the Fitting  29 Finally, a fifth difference: something’s deontic status depends on what the alternatives are in a way that its goodness or badness doesn’t. Permittedness and requiredness have a property that I shall call alternatives dependence: whether ϕing is permitted or required depends, in part, on the normatively relevant proper­ ties of all the alternatives to ϕ-ing.14 What matters is not just the case that can be made for ϕ-ing, but also how that case stacks up in comparison to the case that can be made for each alternative to ϕ-ing. So an action that is required in one situation might have the same (non-­comparative) properties and yet be forbidden in another, because certain additional courses of action are available. Goodness and badness, by contrast, are not alternatives dependent in this way.15 Let us take stock. We have uncovered five key differences between the core deontic properties required, permitted, forbidden, and unrequired and the core evaluative properties good, neutral, and bad:  

Required/permitted/ forbidden/unrequired

Good/neutral/bad

Duality? Polar opposition? Gradable? Neutral state? Alternatives dependent?

Yes No No No Yes

No Yes Yes Yes No

These differences show why requiredness and permittedness are not forms of goodness and goodness not a form of requiredness or permittedness, and similarly why forbiddenness and unrequiredness are not forms of badness and badness not a form of forbiddenness or unrequiredness. Thus we can rule out the most straightforward way of viewing deontic properties as evaluative properties, or vice versa. But we have left open the possibility of other, more complex pro­posals. For instance, what if we take requiredness to be a form of bestness? Would such a proposal show that the deontic categories and the evaluative categories are not separate families after all? No, it wouldn’t. To see why, let us consider one salient group of views on which it might be thought that requiredness turns out to be a kind of bestness, namely maximizing forms of consequentialism. It follows on such views that requiredness is co-­extensive with some manner of (strong) bestness, so that something like the following biconditional holds: 14 

Johann Frick’s (MS) term for much the same thesis is ‘comparativism’ (about a given category). It might seem that at least some first-order moral theories deny that an action’s deontic status is necessarily an alternatives-dependent matter; think here of the ‘absolutist’ view on which intentionally killing another innocent person is always impermissible, even if every available alternative would lead to billions of deaths. Personally I find absolutism so outrageously implausible that I am happy to set it aside. But for those who do not wish to rule it out, a version of my fifth difference can survive. Even if something’s deontic status is not always an alternatives-dependent matter, it typically is, but the same is not true of a thing’s goodness. 15 

30 Fittingness C1.  ϕ-ing is required iff ϕ-ing leads to the (strongly) best outcomes.16 But this biconditional on its own is not enough to establish that requiredness is a form of bestness, for biconditionals are cheap. Instead, we need something stronger. What might that stronger thesis be? I have argued elsewhere (Berker 2018, 2019) that we do best to see maximizing consequentialists as committed to: C2. Necessarily, ϕ-ing is required iff, and because, ϕ-ing leads to the best outcomes. However, C2 is not a view on which requiredness turns out to be an evaluative category after all; rather, it is a view on which requiredness turns out to depend on an evaluative category, and dependence is irreflexive. So on the most natural way of understanding maximizing consequentialism, it does not lead to a collapse of the deontic into the evaluative. What, though, about other ways of interpreting maximizing consequentialism? Suppose that instead of C2 we have one or more of the following: C3. The property being required = the property leading to the best outcomes. C4. The concept is required = the concept leads to the best outcomes. C5. The phrase ‘is required’ means ‘leads to the best outcomes’. Now to start with, I think these are all implausibly strong ways of interpreting the consequentialist’s commitments (Berker 2018: §3). But even if we were to accept one or more of these theses, I don’t think we would stop distinguishing deontic categories (or concepts, or terms) from evaluative categories (or concepts, or terms): we would simply come to see the former as a particularly interesting subset of the latter. The deontic and the evaluative would remain distinct in one sense—­they

A different sort of objection: if to be good is to be better than most of the items in some contextually salient comparison class (as some linguists propose), does this mean that goodness is alternatives dependent after all? No, for that comparison class need not consist in the alternatives to the thing being evaluatively assessed. Moreover, even if being good does turn out to be alternatives dependent for this reason, a version of my fifth difference remains, for being better than would still be an alternativesindependent matter: that x is better than y does not depend on the normatively relevant properties of the alternatives to x and y. 16  The details of this biconditional will of course vary depending on the variety of maximizing consequentialism at issue. (See Berker 2013: §2 for a survey of the possibilities; there I use the term ‘teleology’ where I now prefer ‘consequentialism’.) Henceforth by ‘best’ I mean ‘strongly best’ unless I specify otherwise.

The Deontic, the Evaluative, and the Fitting  31 are different families—­while not being distinct in another sense—­they overlap with one another.17

2.3  The Fitting Let us turn now to our main topic: the fitting. The fittingness categories include at least the following: the thin (or basic) fittingness categories: fitting/appropriate/apt/warranted/justified/merited/deserved/worthy/suitable/ ­proper/correct (?)/called for/to be done (or felt, or held) unfitting/inappropriate/inapt/unwarranted/unjustified/unmerited/undeserved/ un­worthy/unsuitable/improper/incorrect (?)/uncalled for/not to be done (or felt, or held) the thick (or derivative) fittingness categories: (un)acceptable/admirable/(dis)agreeable/contemptible/credible/deplorable/ (un)desirable/(un)enviable/(in)excusable/lamentable/(im)plausible/preferable/ (un)reliable/(ir)responsible/etc. awful/disgraceful/dreadful/shameful/useful/etc. (un)amusing/annoying/disgusting/(un)interesting/(un)surprising/etc. (un)attractive/(in)offensive/(un)persuasive/repulsive/suggestive/etc. bothersome/fearsome/gruesome/irksome/worrisome/etc. blameworthy/choiceworthy/noteworthy/praiseworthy/(un)trustworthy/etc. (un)funny/scary/tasty/etc. 17 

A similar result holds if we accept C6. It is part of the essence of being required that something which is required leads to the best outcomes together with E. If category x features an evaluative category in any part of its essence, then x is itself an evaluative category. But, once again, I think this possibility need not overly concern us, since neither C6 nor E strike me as particularly plausible. (Concerning C6: I do not see why our consequentialist thesis is part of the essence of being required on its own, rather than part of the collective essence of being required and being best, taken together, or not an essential truth at all, as Kit Fine and his followers hold. Concerning E: that a category features a logical [or natural] category within its essence does not thereby make that category a logical [or natural] one, so why should things be any different with evaluativeness? Why does the taint of the evaluative spread through essences, but the taint of the logical [or the natural] does not?) And whether a similar result holds if we have C7.  ϕ-ing is required =df ϕ-ing leads to the best outcomes will depend on whether we are working with a conception of definition on which it entails C2 or instead entails one of C3–C5.

32 Fittingness As with the deontic and evaluative categories, each of these categories might come in different ‘flavours’: moral appropriateness, epistemic warrantedness, and so on. Some authors use ‘fittingness’ only to refer to varieties of the above normative categories that hold in virtue of the internal or constitutive standards of the item being assessed. That is not how I am using the term. An inappropriate comment need not be inappropriate in virtue of violating ‘the constitutive standards of comments’ (whatever those might be) in order for its inappropriateness to qualify as a fittingness category—­to qualify as being concerned with how that comment, as it were, fits the situation. This metaphor of fit, of something matching or suit­ ing a given circumstance or object, links all of these categories together and makes it natural to view them as a unified group that it is handy to label with the term ‘fitting’. I have split up these categories into two groups, the thin (or basic) ones and the thick (or derivative) ones.18 The properties apt, appropriate, merited, deserved, called for, etc. are all, broadly speaking, forms of thin fittingness, in the way that obligatory, mandatory, ought, should, must, etc. are all, broadly speaking, forms of requiredness. Some of these may even be the same category picked out in differ­ ent ways: it is not clear there is a difference between being appropriate and being apt, in the way that it is not clear there is a difference between being required and being mandatory. But in other cases, there obviously are differences. For instance, to say that something is warranted is not the same as saying that it is deserved; believing some proposition might be warranted, given one’s evidence, but no proposition deserves to be believed (Howard 2018: 7). However, a similar thing is also true in the case of the deontic categories: to say that one ought to ϕ is not the same as saying that it is obligatory for one to ϕ. But this difference between ought and obligatory is no bar to classifying both as forms of requiredness—­to classify­ ing both as boxes, not diamonds, in the familiar notation. Similarly, the fact that being warranted and being deserved are not mutually entailing is no bar to classi­ fying both as forms of thin fittingness. ‘Warranted’ and ‘deserved’ can both be thin fittingness terms without being synonyms. The thick fittingness properties are analysable in terms of the thin ones, as was pointed out by Richard Brandt (1946) three quarters of a century ago. Brandt only explicitly applies his analysis to fittingness properties picked out by adjectives ending in ‘-able’/‘-ible’, but his analysis can be straightforwardly extended to all thick fittingness properties, like so:

18  My primary labels for these two groups are ‘thin’ and ‘thick’, but I use those terms with some trepidation, because it is controversial whether admirable, shameful, and the like are thick con­ cepts in the sense made popular by Bernard Williams (1985). Some authors deem them to be (Anderson 1993: 98; Tappolet 2004; Kyle 2020), others deem them not to be (Gibbard 1992; Suikkanen 2009: 778, n. 20), and yet others express uncertainty (Väyrynen 2021). I am happy to switch to a differ­ ent pair of labels if the doubters are right.

The Deontic, the Evaluative, and the Fitting  33 x is admirable =df it is fitting to admire x. x is contemptible =df it is fitting to have contempt for x. x is desirable =df it is fitting to desire x. x is reliable =df it is fitting to rely upon x.19 x is annoying =df it is fitting to be annoyed by x. x is attractive =df it is fitting to be attracted to x. x is fearsome =df it is fitting to be afraid of x. x is trustworthy =df it is fitting to trust x. I have formulated these analyses using the word ‘fitting’, but I just as easily could have used a number of other thin fittingness terms instead: for something to be admirable is for it to merit admiration, or for it to be worthy of admiration, or for admiring it to be apt, or for it to call out for admiration, or . . . Often our term in English for a given thick fittingness property is derived via suffixation from a noun or verb in English expressing the reaction whose fittingness is at issue, but not always. Sometimes the process of suffixation occurred in another language without the root being passed along to English as well, as in the case of ‘risible’ (‘that which merits laughter’, from the Latin ‘ridere’ meaning ‘to laugh at’) or ‘cul­ pable’ (‘that which deserves blame’, from the Latin ‘culpare’ meaning ‘to blame’). And in a few cases, we have no ready way to characterize the relevant reaction except via the thick fitting property being analysed. For instance, although ‘plau­ sible’ originally referred to that which it is fitting to applaud, that meaning is obsolete, and ‘plausible’ now refers to that which it is fitting to . . . well, find plausible. In an important and influential series of articles, Justin D’Arms and Daniel Jacobson call those who advocate analyses of the above sort ‘neo-­sentimentalists’ (D’Arms and Jacobson 2000a, 2000b, 2006a, 2006b, 2017; D’Arms 2005), but in my opinion that is a regrettable piece of terminology which should be avoided. Brandt-­style analyses of thick fittingness properties are entirely independent of the sentimentalist programme in metaethics, for two reasons. First, a similar analysis holds for various thick fittingness properties that have nothing to do with the sentiments, such as: x is punishable =df it is fitting to be punished for x. x is hirable =df it is fitting to hire x. x is credible =df it is fitting to believe x.

19  Note that this analysis shows that ‘reliable’ is a normative term that is part of the fittingness family. So when epistemologists treat ‘reliable’ as a non-normative term meaning ‘truth-conducive’, they are using ‘reliable’ as a technical term, not using it in its ordinary sense. (In everyday speech, to say that Boston’s subway system is unreliable is to normatively assess that subway system, not to say something about whether it promotes true belief.)

34 Fittingness x is useful =df it is fitting to use x (for such-­and-­such purpose). x is persuasive =df it is fitting to be persuaded by x. x is choiceworthy =df it is fitting to choose x. x is seaworthy =df it is fitting to sail x. Hence definitions of this sort are not distinctively sentimentalist in any way. Second, even those Brandt-­style analyses that do appeal to sentiments are com­ patible with facts about the fittingness of sentiments being fully grounded in facts having nothing to do with the sentiments, such as facts about the on-­average pro­ motion of pleasure and pain or facts about the import of the actions those senti­ ments dispose us to perform. But a metaethical view whose appeal to sentiments is itself explained in terms of something deeper is not one that is aptly described as sentimentalist, I would say.20 Not all ‘-ible’/‘-able’ adjectives in English are normative terms that pick out thick fittingness properties. Some are descriptive terms that pick out potentiality properties, such as: x is visible =df x is able to be seen. x is detachable =df x is able to be detached. And some ‘-ible’/‘-able’ adjectives can be used to denote either a thick fittingness property or a descriptive potentiality property. ‘That sentence is unprintable’ can mean ‘That sentence is not fit to be printed’ (because, say, it is too offensive), and it can mean ‘That sentence is not able to be printed’ (because, say, it is too long to write down using existing technology).21 Interestingly, negative prefixes work differently for some thick fittingness terms than others. Often the negative prefix acts as a wide-­scope negation: x is unacceptable =df it is not the case that (it is fitting to accept x). Or perhaps instead:

20  D’Arms and Jacobson also call properties amenable to Brandt-style analyses ‘response-dependent properties’, and others have followed them in this practice (Tappolet 2013; Shoemaker 2017), but in my opinion that is another regrettable piece of terminology which should also be avoided. To say that something is dependent on which of our responses are merited is very different from saying that it is dependent on our responses, and ‘response-dependent’ only makes sense as a label of the latter, not the former. No one would ever say that punishable offenses have a ‘punishment-dependent’ property because they merit punishment. 21  See Kjellmer (1986). There is also a third class of meanings for ‘-ible’/‘-able’ adjectives: a small number of them refer to a certain ready tendency. For example, ‘knowledgeable’ means ‘tends to read­ ily know things (about the domain in question)’, and ‘perishable’ means ‘tends to perish quickly’.

The Deontic, the Evaluative, and the Fitting  35 x is unacceptable =df x is the sort of thing that can be accepted, and it is not the case that (it is fitting to accept x).22 But sometimes the negative prefix acts as a narrow-­scope polarity-­flipper: x is undesirable =df it is fitting to be adverse to (i.e. to anti-­desire) x. Not all non-­desirable outcomes are undesirable; what makes an outcome undesir­ able is that it positively calls out for a negative orectic attitude, not that it fails to call out for a positive one. A large percentage of the normative terms that we use on a daily basis are thick fittingness terms formed via suffixes like ‘-able’, ‘-ible’, ‘-ing’, ‘-ive’, etc. and their analogues in other languages: these terms are everywhere, which makes it all the more surprising how under-­theorized they are. Moreover, thick fittingness terms and the properties they pick out are all over the place in philosophy: we regularly appeal to them in our theories (some examples: accountable, admissible, an­swer­able, attributable, culpable, eligible, excusable, liable, reasonable, reliable, choiceworthy), and we constantly use them to evaluative each other’s arguments and proposals (some examples: [im]plausible, [un]compelling, [un]convincing, fas­cin­at­ing, interesting, [un]promising, surprising, impressive, [un]persuasive, suggestive). But since thick fitting properties can be analysed in terms of thin fittingness properties, I will primarily be focusing on the thin ones in what follows.

2.4  Fittingness Is Not Deontic Here is a common thought: the thin fittingness properties such as fitting, apt, warranted, merited, and the like are themselves deontic properties, and in particular are a type of permittedness or requiredness. This thought is probably the most widespread assumption made about the typology of fittingness categories. For instance, it is simply baked into the literature on so-­called ‘fitting-­attitude the­or­ ies of value’, insofar as that label is standardly taken to apply equally well to views that analyse (or explain, or otherwise account for) goodness in terms of what one ought to value (or favour, or desire) and to views that analyse (or explain,

22  A reason to prefer this analysis over the former one: then we need not say that rocks and other things that cannot be accepted are unacceptable. Alternatively, that x is the sort of thing that can be accepted might be an enabling condition for or presupposition of unacceptability, without being part of its definition. But regardless of how we implement this thought, that some such qualifier is needed seems to be a general feature of negative prefixes that act as wide-scope negations, regardless of whether they are applied to normative or non-normative terms. Even though rocks are not afraid, it doesn’t follow that rocks are unafraid.

36 Fittingness otherwise or account for) goodness in terms of what it is fitting to value (or favour, or desire).23 I myself used to hold this common thought, but I now believe it to be mistaken. I present here five reasons to deny it. First of all, fittingness and requiredness have different guiding metaphors. (This is a warm-­up reason that on its own is not probative.) The guiding meta­ phors behind requiredness or oughtness—­which also happen to be linked to these terms’ etymological origins—­concern a person being bound, tied, or forced to do something. When we are required to do something, it is as if there is a magic bond to that thing which compels us to do it, and when we are permitted not to do that thing, no such bond is present. Compulsion, bondage, being forced: these are all natural metaphors to reach for when discussing requirements. But not so in the case of fittingness; here the guiding metaphors are quite different. One nat­ ural metaphor for fittingness: a puzzle piece locking into place. Another natural metaphor: a key fitting into a lock. These are powerful images that are easily asso­ ciated with all of the standard locutions used to pick out fittingness properties. When a certain emotional reaction is merited, it is the edge piece that fits the current situation. When some action is called for, it is the key that goes into the current lock. And so on. But neither of these metaphors—­the puzzle piece and the key—­are natural ones to invoke when discussing duties, requirements, per­ missions, and other uncontroversially deontic categories. Why would that be so, if fittingness were simply a form of permittedness or requiredness? A second, more substantive reason to deny our common thought is this. When the time comes to decide whether fittingness is a form of permittedness or instead a form of requiredness, we find that neither proposal quite works: in some cases, taking fittingness to be a type of permittedness seems too weak a proposal, and in other cases, taking fittingness to be a type of requiredness seems too strong. When I do something shameful, shame on my part is fitting or called for. When you hear a persuasive argument, it is fitting or appropriate for you to be persuaded by it. When a trustworthy person tells you something, it is fitting or apt for you to trust what they say. In none of these cases is it plausible to interpret the normative category at issue as being a form of permission. It is far too weak to say that we are simply permitted to be ashamed at our shameful conduct, to be persuaded by per­ suasive arguments, or to trust what trustworthy people tell us. To say that shame is fitting is not merely to say that being ashamed is allowed—­is not forbidden—­ but rather to say something more.

23  Authors that follow this terminological convention include Rabinowicz and Rønnow-Rasmussen (2004), Danielsson and Olson (2007), Bykvist (2009), Suikkanen (2009), Jacobson (2011), Tappolet (2013), Gertken and Kiesewetter (2017), and Schroeder (2021), among others. Many of these authors also explicitly tell us that fittingness is a deontic category (see, for example, Rabinowicz and RønnowRasmussen 2004: 391–2, 422; Bykvist 2009: 2–3; Jacobson 2011: §1; Tappolet 2013: 1792; and Schroeder 2021: §3.2).

The Deontic, the Evaluative, and the Fitting  37 So maybe thin fittingness is instead a form of requiredness. But then we run into the opposite problem in numerous other cases. When I go to the rescue ­centre and am confronted with thirty lovable cats waiting to be adopted, it is far too strong to say that I am required to love each cat. Similarly, as we go about our lives, we are not required to laugh at every risible situation, to use every useful tool, to sail every seaworthy vessel, to believe every credible proposition, to hire every hirable candidate, or to be appalled by every appalling politician. Now it might seem that I am raising a demandingness concern here, but that is not the central issue. Although cases in which it would be difficult or even impossible to have all the fitting responses open to us at a given time make vivid the worry I mean to be raising, that worry is not one that would disappear if we had unlim­ ited cognitive, emotional, and physical resources. Even if there is only one lovable cat at the rescue centre, it is still implausible to say that I am required to love it. When someone tells a cringeworthy joke, it is not compulsory for me to cringe, but that is not because cringing would expend crucial emotional resources that could be devoted elsewhere. If I happen upon the last seaworthy ship left in the world, I am under no obligation to sail it, even if I am bored out of my mind and have nothing better to do with my time. As a first-­order normative matter, the idea that fittingness takes the form of a requirement is highly dubious. My argument here has proceeded by (1) assuming a Brandt-­style analysis on which something is shameful when it is fitting to be ashamed of it, lovable when it is fitting to love it, cringeworthy when it is fitting to cringe at it, and so on; (2) assuming that fittingness, if it is indeed a deontic category, must take the same deontic shape in each instance of this analysis; and (3) finding some cases where taking that shape to be a form of permission seems too weak and other cases where taking that shape to be a form of requirement seems too strong. A natural strategy for resisting this argument would be to give up on providing a uniform analysis of the thick fittingness properties, in one of two ways. The first is to reject (1) by replacing Brandt’s analysis with a mixed analysis on which some thick fit­ tingness properties are given a Brandt-­style analysis but other thick fittingness properties are given an analysis that is the dual of his, so that we have: x is useful =df it is fitting to use x (for such-­and-­such purpose). x is shameful =df it is not fitting to not be ashamed of x. x is attractive =df it is fitting to be attracted to x. x is persuasive =df it is not fitting to not be persuaded by x. x is cringeworthy =df it is fitting to cringe at x. x is trustworthy =df it is not fitting to not trust x. The second way is to hold onto Brandt’s analysis across the board but to reject (2) by offering a mixed account of the form of thin fittingness that gets slotted into that analysis, so that for some thick fittingness properties it is a type of

38 Fittingness permission—­call it ‘p-­fittingness’—and for other thick fittingness properties it is a type of requirement—­call it ‘r-­fittingness’. Then we would have: x is useful =df it is p-­fitting to use x (for such-­and-­such purpose). x is shameful =df it is r-­fitting to be ashamed of x. And so on for the others. Both of these mixed proposals are deeply implausible. There is no evidence that the English suffixes ‘-ful’, ‘-ive’, ‘-worthy’, and the like function in these two importantly different ways.24 It would be one thing if one of these suffixes lent itself to a permissive analysis and another lent itself to a requiring analysis, or if one thin fittingness term (‘fitting’, perhaps) were more naturally construed as picking out a form of permittedness whereas another thin fittingness term (‘mer­ ited’, maybe) were more naturally construed as picking out a form of required­ ness. But that is not what we find, so the current proposal requires us to construe the relevant definitions to be metaphysical definitions about the nature of the properties themselves and then to posit that the normative language we have right now is severely out of lockstep with normative reality—­always a dangerous claim to make, especially if our language resists adjusting itself over time so as to take on board this alleged distinction in the way things are (as I suspect it will). The first proposal—­on which some thick fittingness properties are given an ana­ lysis that is the dual of Brandt’s—­also faces the problem that such an analysis, with its two extra negations, is implausible on its face. Intuitively, persuasiveness is about the fittingness or aptness of being persuaded by the object in question, not about the aptness of failing to be persuaded by it, and shamefulness is about whether shame itself is merited or called for, not about the normative status of abstaining from shame. Even worse, these mixed proposals do not succeed in addressing the very problem they were designed to address, for many of the cases we have been considering in which the requiring reading is inadequate are also ones in which the permissive reading is problematic. Let us go back to one of those lovable cats in the rescue centre. Yes, it is too strong to say that I am required to love that cat. But it is also too weak to say that the cat’s lovability simply consists in my being permitted to love it and nothing more. In calling this animal lovable, I am not just saying that it has qualities that make loving it permissible—­that make loving it not forbidden, in the way in which I am not forbidden from tying my right shoe 24  Nor is there evidence that ‘fitting’ is polysemous and has both a permissive and a requiring read­ ing. Indeed, the evidence very much points in the other direction. If ‘fitting’ were polysemous in the specified manner, then ‘It is fitting for me to be ashamed and for you to be angry’ should sound zeug­ matic; but it doesn’t. So ‘fitting’ fails the conjunction-reduction test for polysemy. It also fails many other tests standardly used to detect polysemy or ambiguity, such as the ellipsis test and the contradic­ tion test. (For more on these tests, see Sennett 2016.)

The Deontic, the Evaluative, and the Fitting  39 before my left one. Rather, I am saying something more about how the cat’s qual­ ities positively call out for or merit a certain affective response on my part. (Here I am falling back on using other fittingness locutions to describe the relation I am after that is neither a permission nor a requirement, but I suspect there is no hope of doing otherwise.)25 Reflection on this sort of case leads me to conclude that construing fittingness as a form of permittedness is always too weak. The appall­ ingness of appalling politicians is not exhausted by our being permitted to be appalled at them and their actions; rather, there is a further way in which our being appalled suits or fits what they have done. Useful tools are not simply ones we are allowed to use for the purposes for which they were built. Praiseworthy conduct is worthy of praise, but the fittingness of this praise consists in more than such praise being not prohibited. These problems with mixed proposals and with uniformly permissive pro­ posals might lead us to go back and revisit the prospects for a uniformly requiring proposal. Maybe the solution is to hold onto Brandt’s analysis of thick fittingness properties across the board and also to hold that fittingness is indeed a form of requiredness, but to add more structure to the way in which we connect fitting­ ness to requiredness, like so: ϕ-ing is fitting =df if one considers whether to ϕ, one is required to ϕ. The basic idea is that while we may not be required to believe every single credible proposition out there, perhaps we are required to believe every credible prop­os­ ition that we consider. Similarly, although we might not be required to be appalled by every appalling thing that someone has done, perhaps we are required to be appalled at an appalling action whenever we attend to it and its appallingness. However, there are at least three problems with this proposal. First, it doesn’t help with many of our original cases where construing fittingness as a permission seems too weak. So it’s fine for me not to be ashamed of my shameful conduct, provided I never consider whether to be ashamed of it? Second, although this proposal does help with some of the cases where construing fittingness as a requirement seems too strong, it doesn’t help with all of them. I am not required to sail every seaworthy ship I consider sailing, to hire every hirable candidate I consider hiring, to love every lovable pet I consider loving. Third, if this is all fit­ tingness comes to, then the normative pressure to ϕ that accrues when ϕ-ing is fitting completely disappears the instant one stops considering whether to ϕ. But surely something is going wrong if, for no reason at all, I stop admiring an ad­mir­able person a millisecond after I cease considering the question of whether to admire

25  Compare Scanlon (1998: 17) on explaining the reason relation in terms of the counts-in-favourof relation and vice versa.

40 Fittingness that person. Similarly, suppose I drop my belief in some credible prop­os­ition a split second after I bring my deliberations about whether to believe it to a close. It is simply not plausible that fittingness is a form of normative pressure that gets exerted only while one is considering some question. And it is also not plausible to revise the antecedent of our proposed analysis so that it is in the past rather than present tense, like so: ϕ-ing is fitting =df if one has considered whether to ϕ, one is required to ϕ. My having once considered some question, several decades ago, does not make me permanently required to respond in a certain way to some fitting thing. (Am I really required to right now believe every credible proposition I considered at some time or other while a teenager?) So perhaps we should instead say that the normative pressure that arises due to consideration of a question dissipates after a certain amount of time. But what could possibly determine that length of time? Does the normative statute of limitations expire after several minutes, several days, several years? In short, taking fittingness to be a requirement that kicks in only during deliberation makes it too ephemeral, taking fittingness to be a requirement that kicks in after any bout of deliberation makes it too permanent, and taking fittingness to be a requirement that kicks in during deliberation and hangs around for some specified amount of time forces us to make arbitrary distinctions. Thus we are left with a problem: neither permittedness nor requiredness is the right, as it were, fit for the fittingness categories. Does this mean that fittingness is an in-­between category that is logically stronger than permittedness and logically weaker than requiredness? I don’t think that is the right conclusion to draw. Compare the relation between goodness and permittedness/requiredness. It would be a mistake to hold that being good is logically stronger than being permitted and logically weaker than being required. For one thing, some good things are not even the type of entity that can be permitted or required, so being good does not entail being permitted. For another, there can be required things that fail to be good, as might happen when some obligatory action is the best of a bad lot, so being required does not entail being good. Rather than holding goodness to be a logically intermediary between requiredness and permittedness, we are better off holding goodness to be an altogether different normative category from required­ ness and permittedness that doesn’t bear any straightforward logical relations to either. That, I think, is the proper way to think of the relation between the thin fittingness categories and the deontic categories. It is not a logical truth that fit­ ting responses are the sorts of things that have a deontic status such as permitted or required, and it is not a logical truth that being fitting is entailed by being required and entails being permitted. Let us turn, then, to a third, equally powerful reason to deny that fittingness is a deontic category: whereas permittedness and requiredness have duals (namely,

The Deontic, the Evaluative, and the Fitting  41 each other), fittingness does not have a dual. We have no word in English for a normative category that is the dual of fitting, apt, appropriate, warranted, merited, correct, and the like. Unfitting is not the dual of fitting: it is, rather, fitting’s comple­ ment. Similarly, incorrect stands to correct as impermissible stands to permissible, not as required stands to permissible. Perhaps, it might be replied, fitting does have a dual, and it is simply an acci­ dent that English lacks a handy term to pick it out. Moreover, the reply continues, we can rectify this situation by introducing a new expression, ‘schmitting to ϕ’, meaning ‘not fitting to not ϕ.’ But this is a cheap way of trying to secure duality. By a similar move, we could argue that any property has a dual. For instance, we can no longer say that a crucial difference between required and good is that one has a dual whereas the other doesn’t, because we could always introduce a new expression that supposedly is the latter’s dual. But more importantly, this trick doesn’t even work. Adding ‘schmitting’ into our lexicon does secure extensional duality by brute force: it is schmitting to ϕ iff it is not fitting to not ϕ. But the pro­ posed linguistic innovation fails to get us definitional duality. Although we might perhaps have: It is schmitting to ϕ =df it is not fitting to not ϕ, it is not at all plausible that we have: It is fitting to ϕ =df it is not schmitting to not ϕ.26 This situation stands in contrast to the situation with permittedness and required­ ness, where each definition of the other via an application of duality is on the same footing. In short, permittedness and requiredness are interdefinable, but fit­ tingness and schmittingness are not. So when it comes to the deepest form of duality, a key difference remains even after we have introduced our new term, for fittingness—­unlike permittedness and requiredness—­has no definitional dual, and creating new words cannot change that fact. A fourth reason to hold that fittingness is neither a form of permittedness nor a form of requiredness is that whereas those two deontic statuses are non-­gradable, it is at least arguable that fittingness is gradable. The thick fittingness properties— ­desirable, interesting, attractive, tiresome, trustworthy, and the like—­are all

26  The implausibility of this definition is clearest if we take it to be semantic or conceptual in nature. Could it really be that one cannot know the meaning of ‘fitting’ or possess the concept fitting with­ out also possessing the concept schmitting, in the way in which one cannot know the meaning of ‘bachelor’ or possess the concept bachelor without also possessing the concept married? However, the needed definition is nearly as implausible if we take it to be a metaphysical definition. Surely there was not a crucial part of fittingness’s nature that you were unable to articulate until you were introduced to this new term, ‘schmitting’.

42 Fittingness un­deni­ably gradable. A given outcome can be the most desirable among a range of possibilities, a given point can be very interesting, a given person can be more attractive than another, and so on. It is a delicate question, though, whether the thin fittingness properties are gradable. Some seem not to be. For example, con­ sider correct. When Alan Dershowitz defended his apparently shifting views about the criteria for presidential impeachment by saying, ‘I am . . . far more cor­ rect now than I was then,’ what he said was not just legally questionable but also grammatically suspect.27 Other thin fittingness properties are commonly taken to be gradable. For example, consider justified. Epistemologists often tell us that epistemic justifiedness comes in degrees, and hence is gradable. Finally, a number of thin fittingness properties are borderline cases. For example, consider appropriate. Is it okay to say, ‘A tuxedo would be more appropriate for this event than a coat and tie’? I myself go back and forth in my assessment. However, if the thick fittingness properties are gradable, doesn’t that settle the issue of whether the thin fittingness properties are gradable? How could the thick ones be gradable while the thin ones are not? Barry Maguire (2018) has recently proposed a way of taking the thick fittingness properties to be gradable while holding that the thin ones are not. His suggestion is that we analyse, for example, desirability in a standard Brandt-­style manner but then analyse greater desirabil­ ity in terms of the fittingness of a greater level of desire, rather than in terms of some desire’s greater level of fittingness, so that we have: x is desirable =df it is fitting to desire x. x is more desirable than y =df it is fitting to desire x more than one desires y. Thus desirability can be gradable even if fittingness is not. However, there is a major problem with Maguire’s proposal: it doesn’t work for thick fittingness properties in which the act or attitude at issue is not gradable. Shall we say that x is more memorable than y when it is fitting to ‘more remem­ ber’ x than y? That x is more forgivable than y when it is fitting to ‘more forgive’ someone for x than for y? That x is more seaworthy than y when it is fitting to ‘more sail’ x than y? That x is more choiceworthy than y when it is fitting to ‘more choose’ x than y? Remembering, forgiving, sailing, and choosing are not gradable, so we cannot apply Maguire’s proposal in these cases.28 But memorability, forgiv­ ability, seaworthiness, and choiceworthiness are all gradable properties.29

27 https://www.cnn.com/2020/01/20/politics/dershowitz-impeachment-much-more-correctcnntv/index.html. 28  Nor is it plausible to adapt Maguire’s analysis to focus instead on some gradable property of remembering, forgiving, etc., so that, for example, we say that x is more forgivable than y when it is fitting to more readily forgive someone for x than for y. To do so is to conflate an analysis of the com­ parative form of being forgivable with an analysis of the comparative form of being readily forgivable. 29  Although I use choiceworthiness here as one of my examples, I am somewhat hesitant to do so. ‘Choiceworthy’ is not a word of ordinary English but rather appears (based on my research using

The Deontic, the Evaluative, and the Fitting  43 A second problem with Maguire’s suggestion is that it is not plausible that properties with an analysis of the form: x is F =df it is G to H x are gradable in terms of the embedded act or attitude of H-­ing, even when H-­ing is gradable. Suppose we are playing a game in which we are allowed to be (pre­ tend) angry at various people for doing various things, and we introduce a new term, ‘anger-­allowed’, such that: x is anger-­allowed =df it is allowed for us to be angry at x. Does it follow that being anger-­allowed is a gradable property whose comparative form is given as follows? x is more anger-­allowed than y =df it is allowed for us to be more angry at x than at y. No. Maybe the rules for how angry we are allowed to be at someone, when we are allowed to be angry at that person, have nothing to do with the rules specifying whom we are allowed to be angry at. (Perhaps I am allowed to be angry at you only when you step on a crack in the sidewalk, but how angry I am allowed to be varies with your age.) In fact, even if the rules for how angry we are allowed to be  at someone are intimately related to the rules for whether we are allowed to be angry at that person, the proposed definition of being more anger-­allowed is unsatisfactory. Suppose we are allowed to be angry at someone playing the game if they step on at least three cracks, and the more cracks they step on, the greater a level of anger we are allowed to have. If Gertrude steps on four cracks and Ichabod steps on five, our anger at Ichabod can be greater than our anger at Gertrude, but that fact about our permitted levels of anger does not make it the case that Ichabod is ‘more anger-­allowed’ than Gertrude. Indeed, I would say that nothing makes this the case, because anger-­allowed-­ness is not plausibly a gradable ­property, much less a gradable property whose comparative form is determined in the manner specified.30 But Maguire’s proposal predicts that it is a gradable property and that its comparative form is determined in precisely that manner.31 Google Books) to have been introduced into the language in 1847 by D. P. Chase in his translation of Aristotle’s Nicomachean Ethics. Since then ‘choiceworthy’ has become a part of philosopher-speak, but I think we should be wary of the term, and I find that philosophers often wobble between different ways of understanding what it comes to. 30  Note that properties determined by a threshold along some dimension are not for that reason gradable: someone with a billion dollars is not ‘more of a millionaire’ than someone with a million dollars and one cent (Moon 2017: 771–2). 31  Or, more precisely, this is what Maguire’s proposal predicts when it is combined with the view that fittingness is a form of permission and the natural assumption that a similar proposal holds for all

44 Fittingness Thus we should reject Maguire’s proposal. And given the failure of Maguire’s attempt to explain how thick fittingness properties can be gradable while the thin ones are not, the natural conclusion to draw is that it is because the thin fitting­ ness properties are gradable that the thick ones are as well: for x to be more desir­ able than y is not for it to be on/off fitting to have more desire for x than for y, but rather for it to be more fitting to on/off desire x than to on/off desire y. Thus we may infer that the thin fittingness properties are gradable after all, despite the equivocal nature of the linguistic evidence for the gradability of the correspond­ ing terms. By contrast, it is indisputable that the deontic categories being permitted and being required are non-­gradable. So we have another crucial difference.32 Finally, a fifth difference is that whereas permittedness and requiredness are alternatives dependent, fittingness is not alternatives dependent. Whether an act/ attitude is permitted or required depends on how that act/attitude stacks up in comparison to its alternatives. But we find no such thing for fittingness: whether something is fitting (or appropriate, or merited, or . . . ) depends on the relation­ ship between that thing and the situation, not on how that thing stacks up in comparison to its alternatives. Recall our guiding metaphor: whether the key fits the lock is determined by the relation between the lock and the key, not the rela­ tion between the lock, the key, and the other keys that could possibly be used instead. I do not need to consider the normatively relevant properties of the alter­ natives to admiring some trait of yours in order to determine that this trait of yours merits admiration and hence is admirable; all I need to consider is that trait and how it relates to an attitude of admiration. Similarly, when we assess whether a job candidate is hirable, we assess whether they are suitable for hiring in­de­pend­ent­ly of how hiring them would compare to hiring one of the other

types of permission. An analogous objection holds if instead we combine his proposal with the view that fittingness is a form of requirement. My arguments here can be supported by cross-linguistic considerations. In addition to a fittingness suffix ‘-tos’ that corresponds to ‘-able’ or ‘-worthy’ in English, ancient Greek contains a requirement suffix ‘-teos’ that has no direct analogue in English. So not only is there the ancient Greek word ‘haire­ tos’ meaning ‘choiceworthy’ or ‘that which it is fitting to choose’, but there is also the word ‘haireteon’ meaning ‘that which ought to be chosen’. However, whereas ‘hairetos’ is gradable, ‘haireteon’ is not. Thus ancient Greek appears to assume that it is the gradability of the applied normative category that determines the gradability of each of these adjectives, not the gradability of the act or attitude to which the normative category is applied. (My thanks to Jacob Rosen for the point.) 32  Does our argument for fitting’s gradability extend to correct as well, so that it turns out correct is gradable, and hence Dershowitz did not misspeak after all? I find that difficult to swallow. One solu­ tion would be to insist that although some of the thin fittingness properties are gradable, others of them are not, and it is the gradable ones that are featured in Brandt-style analyses. A second solution is to claim that ‘correct’ is what linguists call a ‘maximum standards absolute adjective’, on the model of ‘pure’ (Kennedy 2007): just as we cannot have two pure samples, one of which is purer than the other, perhaps we cannot have two correct claims, one of which is more correct than the other. But although that helps explain the infelicity of Dershowitz’s statement, it doesn’t explain why we are also reluctant to say that one incorrect claim is ‘more correct’ than another incorrect claim. A third solu­ tion is to deny that correctness is a thin fittingness property. None of these solutions strikes me as entirely satisfactory, and I am not entirely sure what to say about this issue.

The Deontic, the Evaluative, and the Fitting  45 candidates in the pool instead. Fittingness, suitability, appropriateness, aptness: these are not alternatives-­dependent notions.33

2.5  Fittingness Is Not Evaluative So thin fittingness is not a deontic category; in particular, it is not a type of per­ mittedness or requiredness. Could thin fittingness instead be an evaluative cat­ egory; in particular, could it be a type of goodness? Notice, off the bat, that this would be a very surprising conclusion. Although a number of authors claim that certain thick fittingness properties (such as being admirable and being disgusting) are evaluative properties,34 the claim that the thin fittingness properties (such as being fitting and being warranted) are evaluative properties is pretty much never made—­and for good reason, I would say. I can think of at least five reasons to doubt that thin fittingness is a form of goodness. First, fittingness and goodness have different guiding metaphors. Recall the guiding metaphors behind fittingness: a puzzle piece locking into place and a key fitting into a lock. These are not natural metaphors to invoke in the context of evaluative thought and talk. When consequentialists claim that certain states of affairs are good in themselves, I tend to imagine a cosmic thumbs up next to one of those states of affairs, or maybe a happy face hovering above it. What I don’t imagine is a key fitting into a lock or a puzzle piece snapping into place, and I trust that I am not idiosyncratic in this regard. But, as before, this is a warm-­up reason that is not probative on its own. Second, if fittingness is a form of goodness, then fitting-­attitude theories of  value have been entirely misconceived by their proponents. This research 33  What about meritedness? When we claim that some award is merited, aren’t we making a com­ parison between all the people or products or performances that are up for the award? Most likely yes, but that isn’t the comparison that matters when we are assessing whether meritedness is alternatives dependent: what matters is whether we are comparing the normatively relevant properties of giving the award to x to the normatively relevant properties of giving the award to y instead, and that latter comparison is not one we are making. When the award in question takes places in the context of a contest with only one winner, then x will merit the award, if indeed it does, in virtue of being better along some dimension or dimensions than each of its competitors, but how x stacks up in comparison to its competitors is a different issue from how the action of bestowing the award to x stacks up in comparison to its alternatives—­and even then, how x stacks up to its competitors is only relevant because of the nature of the contest at issue, not because of the nature of meritedness itself. 34  For instance, D’Arms and Jacobson do so in the articles cited earlier. They sometimes offer a quick argument for this conclusion: thick fittingness properties are evaluative, D’Arms and Jacobson tell us, because thinking that something has one of these properties ‘is thinking it good or bad in a way’ (2017: 251; I have harmlessly translated their argument in terms of concepts into an equivalent one in terms of properties). I worry, though, about whether we really do think this unless we are using the phrases ‘good in a way’ and ‘bad in a way’ so capaciously that even requiredness counts as good in a way and forbiddenness as bad in a way. There are also a number of cases where neither applies: is deeming something to be surprising deeming it to be good or bad in a way? Finally, ways of being good or bad need not themselves be evaluative properties but rather in many cases are the non-evaluative grounds of evaluative properties: having no ink left is a way for a pen to be bad, but having no ink left is not itself an evaluative property.

46 Fittingness programme is almost always pitched as a way of analysing (or reducing, or explaining) goodness in terms of other normative categories. But if fittingness is a form of goodness, then fitting-­attitude theories of value end up analysing (or reducing, or explaining) one form of goodness in terms of another form of good­ ness. This is another reason that is not probative on its own, since philosophers can misconceive the nature of their research programmes, but it should give us pause. Third, and more significantly, whereas goodness has a non-­privative, polar opposite (namely, badness), thin fittingness does not have a non-­privative, polar opposite. ‘Unfitting’ means ‘not fitting’ (or, perhaps, ‘not fitting, and the type of thing that could be fitting’), not ‘anti-­fitting’. So unfittingness is merely a lack of fittingness, not the presence of anti-­fittingness, whatever that might be. Similarly, for something to be undeserved is for it to be non-­deserved, not for it to be anti-­ deserved, and for something to be unmerited is for it to be non-­merited, not for it to be anti-­merited.35 Fourth, and equally significantly, there is no property that stands to being fitting and being unfitting as being neutral stands to being good and being bad. As we noted earlier, there is a neutral status that lies in between being good and being bad: some outcomes are neither good nor bad but in between, some knives are neither good nor bad but in between, and so on. But there is no neutral status in between being fitting and being unfitting: it is not the case that some emotional reactions are neither fitting nor unfitting but in between, that some awards are neither merited nor unmerited but in between, that some beliefs are neither justified nor unjustified but in between, and so on.36 35  Moreover, it is not just that we lack a term for the polar opposite of each of these thin fittingness properties, but rather we have no grasp on what such a thing would be, so inventing a new term— ‘anti-fitting’, ‘inverse-merited’, etc.—will not help. 36  Perhaps it might be objected that there are, in fact, some cases in which a fit-evaluable act or attitude is not fitting and also not unfitting. Probably the best candidate here is a desire for something that is neutral in value, such as a desire that there be parsley on the moon (Nagel 1970: 45) or that the total number of atoms in the universe be prime (Kagan 1998: 37). Could such desires be neither fitting nor unfitting? No, for at least two reasons. First, everyone must agree that an incredibly strong desire for a neutral state of affairs is unfitting, but it is altogether unclear how we can explain this result if moderate and weak desires for neutral things are neither fitting nor unfitting. If some moderate or weak desires lie in the grey zone between fittingness and unfittingness because their object’s neutrality gives them nothing to match or anti-match, why does ramping up the desire’s strength eventually make this fail to be the case? (One is tempted to say that then the desire’s strength no longer fits the neutrality of its object, but to say this is to go back on the claim that the desire lacked both fittingness and unfittingness before its strength was increased.) Second, a proposal on which desiring the neutral is neither fitting nor unfitting is naturally paired with a view on which blaming someone for some­ thing neutral is neither fitting nor unfitting, admiring someone for something neutral is neither fitting nor unfitting, being amused by a joke that is neither good nor bad is neither fitting nor unfitting, and so on. But these latter verdicts are beyond the pale: if I admire you for having decided to tie your right shoe before your left one, surely my admiration is unfitting (unwarranted, unmerited, undeserved: pick your term). So a proposal that ties a desire’s lack of fittingness or unfittingness to the neutrality of its object predicts that we will find many more cases in which something is neither fitting nor unfit­ ting than we can plausibly claim to find. And more generally, any proposal that gives up on the idea that unfittingness is simply non-fittingness is left with the difficult task of explaining why it is so hard to find a single plausible (non-vague) case in which blame is neither fitting nor unfitting, admiration is neither fitting nor unfitting, etc. For these reasons, I think it is better to hold that instead of the cases

The Deontic, the Evaluative, and the Fitting  47 Fifth, whereas goodness is gradable, it is at least arguable that fittingness is not gradable. Earlier we saw that there is some room, however slight, for arguing that the thin fittingness properties—­especially being correct—­are not gradable, even though we ultimately concluded that they are in fact gradable. However, if the thin fittingness properties are forms of goodness, and we know this, then it should be completely obvious that they are gradable. So the equivocal nature of the evi­ dence for the gradability of the fitting, the appropriate, the proper, the correct, and the like gives us some reason to doubt that these are all forms of goodness. Thus perhaps it is understandable why thin fittingness properties are not often claimed to be evaluative in nature. But what about the claim that at least some thick fittingness properties are evaluative? Don’t advocates of fitting-­attitude the­ or­ies of value make that very claim? And wouldn’t such a claim subvert my asser­ tation that about the distinctiveness of the fittingness categories? No, it wouldn’t. If anything, fitting-­attitude theories of value threaten the dis­ tinctiveness of the evaluative, not the fitting, for at most such analyses show that evaluative categories are a type of fittingness category, not that fittingness cat­egor­ ies are a type of evaluative category. Moreover, I doubt they show even that. Just as we do best to interpret maximizing consequentialism as a thesis about how a certain sort of bestness is explanatorily prior to a certain sort of requiredness, where this priority relation requires the two categories to be distinct, I think we do best to interpret fitting-­attitude theories of value as being concerned with an explanatory relation between two separate categories: a type of fittingness with regard to attitudes and a type of value with regard to objects. For fitting-­attitude theorists, it is because certain attitudes are fitting that a given thing has the value it does, where that ‘because’ is irreflexive.37 Fitting-­attitude theories of value are compatible with the evaluative being distinct from the fitting, in precisely the way in which maximizing consequentialism is compatible with the deontic being dis­ tinct from the evaluative.

2.6  Going Too Far? I have just argued for the hypothesis that there are at least three separate families of normative categories: the deontic, the evaluative, and the fitting. Some of the under discussion being ones in which we have some neutral thing, n, such that desiring n is neither fitting nor unfitting, instead what we have is the following: desiring n is unfitting, but being averse to n is also unfitting. 37  Furthermore, even if I am wrong that we should interpret fitting-attitude theories of value in this way, so that instead they should be understood as proposing something stronger such as an identity between properties, concepts, or word meanings, recall that it was open to maximizing consequentialists to interpret their position in that stronger way as well, yet no one would for that reason stop distin­ guishing the deontic from the evaluative. So we still have the following claim, which in the end is what I am most concerned to establish: we have as much reason to distinguish the fitting from both the deontic and the evaluative as we have to distinguish the deontic from the evaluative in the first place.

48 Fittingness central differences we have found between the core categories in each family are as follows:

Duality? Polar opposition? Gradable? Neutral state? Alternatives dependent?

Required/permitted/ forbidden/unrequired

Good/neutral/bad

Fitting/unfitting

Yes No No No Yes

No Yes Yes Yes No

No No Probably yes No No

In a way, my argument here is overkill: any one definitive difference between each pair of families is enough to establish their distinctness. Moreover, there are dis­ tinguishing features we have uncovered that are not captured by this chart; for example, one of our primary reasons for rejecting the identification of fittingness with either permittedness or requiredness was the difficulty we encountered choosing between these two options. So I think the overall case for my hypothesis that the fittingness categories are their own special group of normative categories is strong. However, I recognize that this hypothesis faces some significant challenges. One sort of challenge involves pointing to canonical deontic or evaluative categories that it would appear I am committed to classifying as fittingness categories, as a way of showing that my way of dividing up the normative landscape has gone too far. For instance, I can imagine an objector replying to my tripartite division between the deontic, the evaluative, and the fitting as follows: ‘What about right and wrong? If proper and correct are thin fittingness properties, then surely right is a thin fittingness prop­ erty as well. But right is a canonical deontic property! So something must have gone wrong somewhere.’ I have two options for how to reply to this challenge, either of which strike me as a fine way to go. The first is to deny that rightness is a thin fittingness property (perhaps by also denying that correctness is a thin fittingness property). In defence of such a move, it is striking that the opposite of ‘right’ in English is ‘wrong’, whereas every other thin fittingness term’s opposite is formed in English by add­ ing a negative prefix (‘unfitting’, ‘inapt’, ‘unmerited’, ‘improper’, etc.). So something different is going on with rightness and wrongness, at least lin­guis­tic­al­ly if not conceptually and metaphysically. The second option, which I prefer, is to accept that being right is a thin fitting­ ness property but to hold that it was a mistake for people to assume that rightness and wrongness are deontic categories. So on this proposal—­which I recognize is extremely controversial—‘wrong’ and ‘forbidden’ are not synonyms, and ‘right’ and ‘wrong’ are not deontic terms. That might sound like crazy talk, but over time

The Deontic, the Evaluative, and the Fitting  49 I have come to think that this suggestion is not as outlandish as it might initially seem. For one thing, it is telling that authors who assume rightness to be a deontic category sometimes get confused about whether rightness is a form of permission or of requirement.38 Moreover, although authors occasionally tell us that ‘right’ is ambiguous between a reading on which it means ‘required’ and a reading on which it means ‘permitted’ (a semantic thesis for which, as far as I can tell, there is  no linguistic evidence), these authors never make an analogous distinction between a reading of ‘wrong’ on which it means ‘forbidden’ and a reading on which it means ‘unrequired’, which is a puzzling gap given that wrongness is rightness’s complement. It is also telling that although we can distinguish the monadic property being wrong from the dyadic relation wronging someone, there is no analogous distinction for being forbidden: there is no directed notion of doing something forbidden with regard to someone; but if ‘wrong’ and ‘forbidden’ were synonyms, it would be puzzling why this would be so. Finally, there is strong linguistic evidence that we do not use ‘right’ as a deontic term in everyday parlance, even when we mean to be invoking a distinctively moral form of assessment. Is ‘Do the right thing!’ just a way of saying ‘Do the permitted thing!’ for some suitable form of permission? When I tell someone, ‘You were quite right to criticize him in the way you did,’ is this equivalent to saying, ‘You were quite permitted to criticize him in the way you did’? Once we take off our moral-­philosopher hats and look afresh at how we use the term ‘right’ on a daily basis, even in moral contexts, I think the idea that rightness is not a deontic category becomes quite plausible—­or, at least, not as outrageously implausible as it might seem at first blush. So that is how I would handle one sort of going-­too-­far challenge. Here is another challenge of that sort: ‘What about permissibility? Isn’t being permissible amenable to the same Brandt-­style analysis that you endorsed for other normative “-ible”/“-able” adjectives in English, in that for ϕ-ing to be permissible is for it to be fitting to permit ϕ-ing? But that would make permissibility a thick fitting­ ness property, whereas you yourself took it to be a canonical deontic property. So something has gone wrong.’ This is another important challenge; indeed, it is probably the going-­too-­far worry that most concerns me. Here my preference is to deny that the property in question is in fact a fittingness property. As we use the term now, ‘permissible’ does not mean ‘fit to be permitted’, and the property it picks out is not a property that concerns the fittingness of a certain act, the act of permitting someone to do (or think, or feel) something. To say that ϕ-ing is per­ missible is to normatively assess ϕ-ing itself, not to normatively assess the granting of a permission to ϕ. Thus ‘-ible’ in ‘permissible’ is not functioning as a fittingness suffix as it does in ‘contemptible’ or ‘credible’. It might seem that I am making a

38  My favourite example of this: in the preface to Principia Ethica, G. E. Moore equates ‘a right action’ with ‘a duty’ (1903: iv), but presumably it is more plausible to equate a right action with a permitted action (Ross 1930: 3–4).

50 Fittingness bold linguistic hypothesis here, but in fact ‘permissible’ behaves quite differently from standard thick fittingness terms. For instance, ‘contemptible’, ‘credible’, and the other ‘-ible’/‘-able’ fittingness terms are gradable, but ‘permissible’ is not. Similarly, for most thick fittingness terms it is possible to construct a nearly synonymous term using the suffix ‘-worthy’, but not so for ‘permissible’: ‘contempt-­worthy’ is close in meaning to ‘contemptible’, ‘belief-­worthy’ close to ‘credible’, ‘admiration-­worthy’ close to ‘admirable’, etc., but ‘permission-­worthy’ is jarringly far in meaning from ‘permissible’. Although almost every normative adjective ending in ‘-ible’ or ‘-able’ is a thick fittingness term, ‘permissible’ is a rare exception.

2.7  Other Families? My focus so far has been on establishing that there is a third family of normative categories beyond the most familiar two. A natural question arises, though: are there yet more families other than these? Two families immediately spring to mind. First, there are the aretaic categories: those concerning virtue and vice. Second, there are the reason-­related categories: those concerning reasons, reason, and reasoning.39 I want to set aside the aretaic categories; whether they constitute a fourth family is too large a topic to settle here. However, it is worth considering the reason-­related categories, because they might be thought to give rise to an objection to what I have argued so far. This objection can be spun in one of two ways. First, it might be claimed that the reason-­related categories are a third family of normative categories beyond the deontic and the evaluative which have been widely studied for several decades now, and that the fittingness categories are best identified with some subset of the reason-­related categories; for instance, maybe being fitting is the same as being supported by a certain sort of reason. Thus, the objection goes, my thesis about the distinctness of fittingness as a normative category only appears surprising when we focus exclusively on the deontic and evaluative categories and ignore the reason-­related categories. A second way of spinning this objection is not to claim that the reason-­related categories are a separate family outside the arc of the deontic and the evaluative, but rather to claim that the reason-­related categories are a subset of the deontic categories which I have neglected to mention so far. Thus, the objection continues, my arguments that fittingness is not deontic were incomplete, because those arguments were directed against the possibility that fittingness is a form of permission or requirement, not against the possibility that it is a way of being reason supported. 39  A third further family that springs to mind is what some—­including my past self (Berker 2013: 382, n. 64)—call, following Michael  J.  Zimmerman (2002: 554), ‘the hypological categories’: those categories relating to responsibility, blameworthiness, and praiseworthiness. I now prefer to think of the hypological as a tiny and not particularly cohesive subset of the fitting.

The Deontic, the Evaluative, and the Fitting  51 Both of these ways of spinning the objection rely on an assumption I deny: namely, that the reason-­related categories are a unified family of normative cat­egor­ies. In fact the reason-­related categories crosscut the various families of cat­egor­ies we have considered so far. Some of them are straightforward deontic categories of the sort we have already been discussing. For instance, there being sufficient reason to ϕ is a type of permittedness, and there being decisive reason to ϕ is a type of requiredness; after all, the two are duals of each other, are non-­ gradable, are alternatives dependent, etc.40 Other reason-­related categories are fittingness categories of a kind already mentioned. For instance, being reasonable when it is applied to acts and attitudes (as opposed to, say, persons) is a thick fit­ tingness property of the standard ‘-ible’/‘-able’ sort. Finally, there are favouring relations such as being a reason to and having a reason to. Some authors classify these last two relations as deontic categories, but I do not see why they should be. Perhaps the thought here is that these favouring relations are just pro tanto versions of requiredness, which is certainly something suggested by the somewhat older terminology of ‘pro tanto duties’ (as a way of updating W. D. Ross’s unfortunate phrase ‘prima facie duties’ so as to avoid the suggestion that what is at issue is whether certain things are ‘at first glance’ duties). But conceiving of reasons in that way is confused, and I think it is a good thing that philo­sophers have largely moved away from the terminology of pro tanto duties and obligations. ‘Pro tanto’ literally means ‘as far as that goes’, so, taken literally, a pro tanto duty is something that considered on its own would be a duty: when no other pro tanto duties are present, it is one’s duty sans phrase, as Ross used to say. But if that is what a pro tanto duty is, then the category should not be gradable, for it cannot be the case that one thing is more-­a-­duty-­sans-­phrase-­when-­considered-­ on-­its-­own than another. However, the being a reason to relation is gradable, so ‘pro tanto duty’ is a misleading label for it, and we cannot conclude that this rela­ tion is a deontic category on the basis of such terminology. Even if it is true that a reason when present on its own determines one’s overall duty,41 that is not what a reason is, but at most something that follows from what a reason is (together, perhaps, with what a duty is). On an alternate and by now more common way of understanding the label ‘pro tanto’, we should not take that label literally but instead should understand it as a synonym for ‘contributory’. Hence the practice of referring to a consideration as ‘a pro tanto reason to ϕ’: this phrase cannot mean ‘as far as that goes a reason to ϕ’

40  Thus my arguments against fittingness being a form of permittedness (or requiredness) are also arguments against analysing fittingness in terms of what there is sufficient (or decisive) reason of the right kind to do. 41  And even that is in doubt: for instance, it does not appear to be true of epistemic reasons for belief (when one has a very weak epistemic reason to believe that p and no other epistemic considerations bear on the matter, one is not epistemically required to believe that p), and it might be that there are reasons that can only be present when other reasons are present (Dancy 2004: 19–20).

52 Fittingness (or else we must already have a handle on the sort of thing it is when present on its own, so what work is the ‘pro tanto’-qualifier doing?), but rather must mean something like ‘a contributory reason to ϕ’. But contributes towards what? If we mean ‘contributes towards ϕ-ing’s requiredness’, then perhaps we can deem the relation in question to be a deontic category. However, that is another confused way of conceiving of what reasons are. ϕ-ing’s requiredness is not the sort of thing that a reason can contribute towards or count in favour of. A reason to ϕ favours ϕ-ing, not ϕ’s requiredness: it is acts and attitudes themselves that are favoured, not facts about their normative status. Talk of ‘contributory reasons to ϕ’ should be understood as shorthand for ‘considerations that contribute towards the case in favour of ϕ-ing’, and it is not clear why reasons so construed should be viewed as deontic in nature. As a substantive normative matter, it might well be that what one is required or ought to do is determined by the overall balance of contribu­ tory reasons for and against the act in question and its alternatives, but this is no more reason to deem the reason relation to be a deontic category than the conse­ quentialist’s claim that what one is required or ought to do is determined by the comparative goodness of the resulting outcomes is a reason to deem goodness to be a deontic category. So I would deny that being a reason to is a deontic relation. Indeed, it appears, if anything, to have all the properties we have found evaluative categories to have. Being a reason to lacks a dual. It has a polar opposite: namely, being a reason against. Both it and its polar opposite are gradable: one consideration can favour some act more than another does, and a given fact can disfavour some attitude more than another does.42 There is a neutral relation that lies in between being a reason to and being a reason against: namely, the relation a consideration bears to an act or attitude when it counts neither in favour of nor against that act or atti­ tude. Finally, being a reason to is alternatives independent: whether r is a (perhaps outweighed) reason to ϕ doesn’t depend, I claim, on the normatively relevant properties of the alternatives to r and to ϕ-ing. These features of the being a reason to relation mean that in the end it doesn’t much matter for my purposes whether that relation qualifies as deontic, because I can take my arguments against fittingness being a form of goodness and redeploy them as arguments against the suggestion that something’s being fitting is the 42  In taking these relations to be gradable, we must be cautious. In English, the locution ‘is a reason to’ is not gradable: although we speak of reasons having a certain feature—­namely, their weight or strength—­that is gradable (as in: ‘That reason is stronger than this one’), such talk is not the same as saying that reasons themselves are gradable. In order to conclude that reasons are gradable, we must make the further claim that a reason is not something separate from its weight but rather consists in that weight in a given direction and nothing more, in the way in which a vector is not something separate from its magnitude but rather consists in that magnitude in a given direction and nothing more. However, this additional claim does seem plausible, especially given the widespread view that ‘r is a reason to ϕ’ and ‘r favours ϕ-ing’ are two ways of saying the same thing—­which is naturally paired with a view on which ‘r1 is a stronger reason to ϕ than r2 is’ and ‘r1 favours ϕ-ing more than r2 does’ are also two ways of saying the same thing.

The Deontic, the Evaluative, and the Fitting  53 same as its having some particular type of pro tanto reason in its favour. If that is what fittingness comes to, we would expect it to have a polar opposite; but it doesn’t. We would also expect there to be a neutral status that lies in between fit­ tingness and its opposite; but we find no such thing. Finally, it should be com­ pletely obvious that fittingness is gradable; but it isn’t.43 In short, being fitting has a variety of structural features that distinguish it from being reason supported.44 So even if I am wrong when I insist that pro tanto reasons are not a deontic category, it still would not follow that fittingness is a deontic category because it is a certain way of being favoured by a certain sort of pro tanto reason.

2.8 Conclusion Philosophy is like all disciplines: once we have a theoretical framework that works well for us in one area, we have an unfortunate habit of trying to apply that frame­ work everywhere else, even when it doesn’t fit. I have been arguing that such a situation has occurred with the normative categories. We are all so familiar with the evaluative and the deontic categories that we assume the fittingness categories must conform to one or the other of these comfortable options. But we should stop trying to choose whether to cram the round peg of fittingness into either the square hole of requiredness or the triangular hole of goodness. Fittingness is its own thing, not a type of goodness or a variety of permittedness. Moreover, fittingness (broadly construed) is everywhere, both in daily life and throughout philosophy. The appropriate and the merited, the commendable and the deplorable, the reasonable and the justified, the compelling and the sug­gest­ ive, the liable and the excusable, the choiceworthy and the noteworthy—­these are all species of fittingness. Fittingness is not some niche topic to be investigated only by the coterie of metanormative theorists seeking to explain everything normative in terms of the fitting, and ‘fitting’ is not a technical term that is the private reserve of neo-­sentimentalists. Any moral theorist who thinks that the blameworthy is that which is worthy of blame and the preferable is that which it is fitting to prefer—­and really, what are you doing not thinking these things, if you profess to 43  More precisely: it should be completely obvious that either fittingness itself is gradable or it has this feature, its strength or weight, that is gradable. (See previous footnote.) But this disjunction is also not obvious. 44  There is an additional structural problem with the claim that being fitting is the same as being supported by a reason of the right kind that has no parallel in my argument against identifying fitting­ ness with goodness. entails , but does not entail : that’s why a person who is admirable to some extent need not be admirable full stop, a story that is amusing to some extent need not be amusing full stop, etc. And if we try to address this problem by identifying being fitting not with being supported by a reason of the right kind but instead with being supported by a sufficiently strong reason (or collection of reasons) of the right kind, then being fitting will no longer be gradable.

54 Fittingness understand the words ‘blameworthy’ and ‘preferable’?—should be interested in the nature and normative significance of fittingness. Taking the fitting to be a distinct family of normative categories leaves many questions about those categories unsettled. But it also frees us up to see those categories for what they really are when we theorize about them and in terms of them, and to appreciate the ways in these categories are particularly well suited for certain domains where many other normative categories seem out of place—­ the realm of aesthetics and the realm of the emotions being two obvious ex­amples. Fittingness, meritedness, aptness, and the like do not need to earn their keep by being reduced to value, or oughtness, or reasons, and perhaps it is precisely because they cannot be so reduced that they are the right categories to reach for when normatively assessing, say, aesthetic experiences or emotional reactions. I opened by noting how the two more familiar families of normative categories are usually labelled ‘the right’ and ‘the good’, especially when moral theorists ask that perennial question, ‘Is the good prior to the right, or the right prior to the good?’ Although most of us became acquainted with this question through the work of John Rawls, it was used as a way of structuring the theoretical options in ethics—­as a way of dividing up the terrain between what we now call ‘consequentialists’ and what we now call ‘deontologists’—for most of the twentieth century. It is also a question that we can ask in normative disciplines other than ethics: for example, in epistemology we can ask, ‘Is the (epistemically) good prior to the (epistemically) right, or the (epistemically) right prior to the (epistemically) good?’ However, if my arguments in this chapter are correct, we need to revise the way in which this traditional question is phrased. What is at issue are the priority relations between three families of categories, not two. Moreover, since ‘right’ might end up being a fittingness term, we are better off labelling the deontic cat­egor­ies as ‘the required’, not ‘the right’. So instead of asking, ‘Is the good prior to the right, or the right prior to the good?’ what we should be asking is, ‘What are the priority relations between the required, the good, and fitting?’45

45 Many thanks to Rachel Achs, Robert Audi, Garrett Cullity, Conor McHugh, Eli Pitcovski, Michael Rabenberg, Mark Richard, Christine Tappolet, Jonathan Way, Quinn White, and especially this volume’s editors, Chris Howard and Richard Rowland, for written feedback on earlier drafts of this chapter; to Sasha Arridge, Simon-Pierre Chevarie-Cossette, Justin D’Arms, Cora Diamond, Jeremy David Fix, Daniel Fogal, Thomas Hurka, Ram Neta, Joseph Raz, Andrew Reisner, Jacob Rosen, Susanna Siegel, Philip Stratton-Lake, Ralph Wedgwood, and Kevin Zhang for correspondence and discussion; to audiences at online talks hosted by Tel-Hai Academic College, Texas A&M University’s Group for Analytic Metaphysics and Epistemology, the University of Neuchâtel, University of Oxford’s Moral Philosophy Seminar, and Uppsala University for their probing comments and questions; and to the participants at the Fit Fest, the Northeast Normativity Workshop, the Practical and Epistemic Normativity Workshop at the University of Helsinki, and my graduate seminar on Normative Categories at Harvard University in the spring of 2020, at which I presented earlier iterations of the arguments from this chapter. Because of space concerns, I was not able to take up every excellent suggestion made on these various occasions, but I hope to do so in future work.

The Deontic, the Evaluative, and the Fitting  55

References Anderson, Elizabeth. 1993. Value in Ethics and Economics. Cambridge, Mass.: Harvard University Press. Aristotle. 1847. The Nicomachean Ethics, translated by D.  P.  Chase. London: Whittaker and Co. Berker, Selim. 2013. ‘Epistemic Teleology and the Separateness of Propositions.’ Philosophical Review 122: 337–93. Berker, Selim. 2018. ‘The Unity of Grounding.’ Mind 127: 729–77. Berker, Selim. 2019. ‘The Explanatory Ambitions of Moral Principles.’ Noûs 53: 904–36. Brandt, Richard. 1946. ‘Moral Valuation.’ Ethics 56: 106–21. Bykvist, Krister. 2009. ‘No Good Fit: Why the Fitting Attitude Analysis of Value Fails.’ Mind 118: 1–30. Chuard, Philippe, and Nicholas Southwood. 2009. ‘Epistemic Norms without Voluntary Control.’ Noûs 43: 599–632. Cushman, Fiery, Liane Young, and Marc Hauser. 2006. ‘The Role of Conscious Reasoning and Intuition in Moral Judgment: Testing Three Principles of Harm.’ Psychological Science 17: 1082–9. Dancy, Jonathan. 2004. Ethics without Principles. Oxford: Oxford University Press. Danielsson, Sven, and Jonas Olson. 2007. ‘Brentano and the Buck-Passers.’ Mind 116: 511–22. D’Arms, Justin. 2005. ‘Two Arguments for Sentimentalism.’ Philosophical Issues 15: 1–21. D’Arms, Justin, and Daniel Jacobson. 2000a. ‘The Moralistic Fallacy: On the “Appropriateness” of Emotions.’ Philosophy and Phenomenological Research 61: 65–90. D’Arms, Justin, and Daniel Jacobson. 2000b. ‘Sentiment and Value.’ Ethics 110: 722–48. D’Arms, Justin, and Daniel Jacobson. 2006a. ‘Anthropocentric Constraints on Human Value.’ Oxford Studies in Metaethics 1: 99–126. D’Arms, Justin, and Daniel Jacobson. 2006b. ‘Sensibility Theory and Projectivism.’ In The Oxford Handbook of Ethical Theory, edited by David Copp, 186–218. Oxford: Oxford University Press. D’Arms, Justin, and Daniel Jacobson. 2017. ‘Whither Sentimentalism? On Fear, the Fearsome, and the Dangerous.’ In Ethical Sentimentalism: New Perspectives, edited by Remy Debes and Karsten  R.  Stueber, 250–67. Cambridge: Cambridge University Press. Feldman, Richard. 2008. ‘Modest Deontologism in Epistemology.’ Synthese 161: 339–55. Frick, Johann. MS. ‘Dilemmas, Luck, and the Two Faces of Morality.’ Gertken, Jan, and Benjamin Kiesewetter. 2017. ‘The Right and the Wrong Kind of Reasons.’ Philosophy Compass 12/5: 1–14. Gibbard, Allan. 1992. ‘Thick Concepts and Warrant for Feelings.’ Proceedings of the Aristotelian Society, Supplementary Volumes 66: 267–83.

56 Fittingness Hieronymi, Pamela. 2008. ‘Responsibility for Believing.’ Synthese 161: 357–73. Howard, Christopher. 2018. ‘Fittingness.’ Philosophy Compass 13/11: 1–14. Jacobson, Daniel. 2011. ‘Fitting Attitude Theories of Value.’ Stanford Encyclopedia of Philosophy (Spring 2011 Edition), edited by Edward N. Zalta. https://plato.stanford. edu/archives/spr2011/entries/fitting-attitude-theories/. Kagan, Shelly. 1998. Normative Ethics. Boulder, Colo.: Westview Press. Kendi, Ibram X. 2019. How to Be an Antiracist. New York: Random House. Kennedy, Christopher. 2007. ‘Vagueness and Grammar: The Semantics of Relative and Absolute Gradable Adjectives.’ Linguistics and Philosophy 30: 1–45. Kjellmer, Göran. 1986. ‘Legible but Not Readable: On the Semantics of English Adjectives in -ble.’ Studia Neophilologica 58: 11–38. Kyle, Brent G. 2020. ‘The Expansion View of Thick Concepts.’ Noûs 54: 914–44. Maguire, Barry. 2018. ‘There Are No Reasons for Affective Attitudes.’ Mind 127: 779–805. McHugh, Conor. 2017. ‘Attitudinal Control.’ Synthese 194: 2745–62. McNamara, Paul. 2019. ‘Deontic Logic.’ Stanford Encyclopedia of Philosophy (Summer 2019 Edition), edited by Edward  N.  Zalta. https://plato.stanford.edu/archives/ sum2019/entries/logic-deontic/. Moon, Andrew. 2017. ‘Beliefs Do Not Come in Degrees.’ Canadian Journal of Philosophy 47: 760–78. Moore, G. E. 1903. Principia Ethica. London: Cambridge University Press. Nagel, Thomas. 1970. The Possibility of Altruism. Princeton, New Jersey: Princeton University Press. Rabinowicz, Wlodek, and Toni Rønnow-Rasmussen. 2004. ‘The Strike of the Demon: On Fitting Pro-Attitudes and Value.’ Ethics 114: 391–423. Rosen, Gideon. 2015. ‘Real Definition.’ Analytic Philosophy 56: 189–209. Ross, W. D. 1930. The Right and the Good. Oxford: Oxford University Press. Scanlon, T.  M. 1998. What We Owe to Each Other. Cambridge, Mass.: Harvard University Press. Schroeder, Mark. 2021. ‘Value Theory.’ Stanford Encyclopedia of Philosophy (Fall 2021 Edition), edited by Edward  N.  Zalta. https://plato.stanford.edu/archives/fall2021/ entries/value-theory/. Sennet, Adam. 2016. ‘Ambiguity.’ Stanford Encyclopedia of Philosophy (Spring 2016 Edition), edited by Edward  N.  Zalta. https://plato.stanford.edu/archives/spr2016/ entries/ambiguity/. Shoemaker, David. 2017. ‘Response-Dependent Responsibility: Or, A Funny Thing Happened on the Way to Blame.’ Philosophical Review 126: 481–27. Smith, Michael. 2005. ‘Meta-ethics.’ In Oxford Handbook of Contemporary Philosophy, edited by Frank Jackson and Michael Smith, 3–29. Oxford: Oxford University Press. Suikkanen, Jussi. 2009. ‘Buck-Passing Accounts of Value.’ Philosophy Compass 4/5: 768–79.

The Deontic, the Evaluative, and the Fitting  57 Tappolet, Christine. 2004. ‘Through Thick and Thin: Good and Its Determinates.’ Dialectica 58: 207–21. Tappolet, Christine. 2013. ‘Evaluative vs. Deontic Concepts.’ In International Encyclopedia of Ethics, edited by Hugh LaFollette, 1791–9. Oxford: Wiley-Blackwell. Väyrynen, Pekka. 2021. ‘Thick Ethical Concepts.’ Stanford Encyclopedia of Philosophy (Spring 2021 Edition), edited by Edward  N.  Zalta. https://plato.stanford.edu/ archives/spr2021/entries/thick-ethical-concepts/. Williams, Bernard. 1985. Ethics and the Limits of Philosophy. Cambridge, Mass.: Harvard University Press. Zimmerman, Michael J. 2002. ‘Taking Luck Seriously.’ Journal of Philosophy 99: 553–76.

3 Against the Fundamentality of Fit Thomas Hurka

3.1 Introduction Many British philosophers of the late nineteenth and early twentieth centuries were minimalists about the ethical concepts, believing there’s just a small number of fundamental ones. Some thought there’s only one such concept. Thus Henry Sidgwick thought the primary ethical concept is that of what one ought, in a categorical sense, to do, and he therefore analysed ‘good’ in terms of ‘ought’, as what one ought to desire (1907: 112, 381). In Principia Ethica G. E. Moore, too, recognized just one basic concept, though this was now ‘intrinsically good’ (plus its contrary ‘bad’) and he analysed the right, or what one ought to do, as what will result in the most good possible (1903: 147). He was later persuaded by Bertrand Russell (1904: 330) to abandon this analysis and thereafter recognized two basic concepts, ‘good’ and ‘ought’ or ‘right’. Though he still thought one ought always to do what will produce the most good, this was now a synthetic, not analytic, truth (1912: 172–73, 181; 1942: 574–77). That there are these two basic concepts was also the view of H. A. Prichard, E. F. Carritt, and W. D. Ross, as reflected especially in the latter’s The Right and the Good (1930). It has separate discussions of ‘right’ and ‘good’, each arguing first that the concept is unanalysable and then addressing what falls under it. These philosophers all thought there are duties other than the duty to promote the good that can conflict with and sometimes outweigh it. But they shared Moore’s later view that there are two basic ethical concepts, one deontic and one evaluative. Others, however, reverted to the view that there’s a single fundamental concept, though this was now the different one of fittingness. This was first proposed, albeit tentatively, in C. D. Broad’s Five Types of Ethical Theory, published in the same year as The Right and the Good. Broad proposed understanding what Ross called ‘prima facie duty’, and some call an ‘ought other things equal’, in terms of fittingness to an aspect of your situation (1930: 164). Thus that you ought other things equal to keep your promises means that, if you promised to do x, doing x fits that aspect of your situation that consists in your having promised to. And what you ought all things considered to do is what’s most fitting to your situation as a whole, or considering all its aspects. Broad also said ‘it might be held . . . that Thomas Hurka, Against the Fundamentality of Fit In: Fittingness: Essays in the Philosophy of Normativity. Edited by: Christopher Howard and R. A. Rowland, Oxford University Press. © Thomas Hurka 2022. DOI: 10.1093/oso/9780192895882.003.0003

Against the Fundamentality of Fit  59 “X is good” means that it is fitting for every rational being to desire X’ (1930: 278). He therefore suggested, without definitely endorsing, the view that the single concept of fittingness underlies both ‘ought’ and ‘good’, or both deontic and evalu­ative ethical judgements. This view was defended more fully by A. C. Ewing, in a 1939 paper and then in The Definition of Good (1947). Ewing adopted Broad’s account of the prima facie ‘ought’ in terms of fittingness to an aspect of your situation, calling it simply the ‘ought of fittingness’. And he defined the good as ‘a suitable object of pro-­ attitudes’ (1939: 9) or a ‘fitting object of a pro-­attitude’ (1947: 152). His analysis of ‘good’ was pluralistic, since he thought different pro-­attitudes can be fitting to different objects, for example welcoming for its own sake to some and admiration to others (1939: 11–13; 1947: 153–56, 166–67). But he took ‘good’ always to involve some pro-­attitude’s being fitting. He also suggested, though more ­cautiously, a fittingness-­based analysis of the to him distinct concept of moral obligation, which he associated with issues about blame. That you failed to do what you ought in the fittingness sense to do needn’t, he held, make you liable to blame; for that you must have violated an obligation. But he thought the claim that you’re obligated to do x might be read as saying that you ought in the fittingness sense to do x and that, if you don’t, it will be fitting for others to blame you (1939: 14; 1947: 168–69). Ewing later withdrew this suggestion (1959: 91) but he continued to hold that the one concept of fit underlies both deontic judgements about what you ought to do and evaluative judgements about what’s good. The view that ‘fitting’ is the one basic ethical or, more generally, normative concept or property has recently been revived and defended (Chappell 2012; McHugh and Way 2016; Cullity 2018; Howard 2019). In this chapter I raise three doubts about it, ones questioning whether fittingness can play this role and suggesting instead that deontic and evaluative concepts or properties are both distinct from each other and irreducible to anything more fundamental. But I do so with frequent references to earlier discussions of this issue by Sidgwick, Moore, Prichard, Ross, Broad, and Ewing, which are often illuminating; I defend the two-­concept view of the later Moore, Prichard, and Ross against the one-­concept, fittingness-­ first view of Broad and Ewing. I consider, however, only uses of ‘fitting’ in ethics rather than in normativity more broadly, as in claims about the fittingness of beliefs or, in non-­ethical contexts, of emotions. This is partly because if ‘fitting’ isn’t fundamental in ethics it can’t be so more generally, and partly because what makes for fittingness in these other contexts is often so different that it’s unclear a single concept is being used. Finally, like the historical figures above I focus on the concepts ‘ought’, ‘good’, and ‘fitting’ rather than, as some recent writers do, the corresponding properties. I assume that for the most part claims about the ­concepts and the properties are intertranslatable.

60 Fittingness

3.2  Fittingness: Deontic vs. Evaluative Reducing two concepts to a single fundamental one is more difficult the more the two differ, and this is especially so if they have contradictory properties. Then what’s needed for the one reduction may be incompatible with what’s needed for the other. There are several important differences between ‘ethically ought’ and ‘good’ that raise this difficulty for the attempt to analyse both in terms of fittingness, but I’ll mostly discuss one emphasized by the historical philosophers mentioned above. It concerns the two concepts’ relation to an ‘ought’ implies ‘can’ condition taken to require voluntary control, or control by the will. These philosophers thought the ethical ‘ought’ is governed by a fairly strong such condition, so it can be true that you ought to do an act, either other things equal or all things considered, only if you can do it, in at least the sense that if you were motivated to do it you could or would. But ‘good’ involves no similar condition. If a sunny day makes people happy, their happiness is good even if no one was responsible for the sunshine; unavoidable pain is likewise bad. For Sidgwick, Ross, and others a central difference between deontic and evaluative ethical ­judgements is that the former require voluntary control while the latter do not. This creates a problem for a fittingness-­based view in one or other of these contexts. Consider first a fitting-­attitudes analysis of ‘good’. That ‘good’ isn’t governed by a strong ‘ought’ implies ‘can’ condition is captured in part if what the analysis says is fitting is desire or some pro-­attitude rather than any act or choice. Someone who can’t affect external events can nonetheless want others to be happy and be saddened if they suffer. Analysing ‘good’ in terms of fitting desire therefore allows it to range more widely than voluntary action, or to characterize outcomes no one could produce. But what about the fittingness concept itself? Is the claim that an attitude is fitting governed by a strong ‘can’ condition or not? The difficulty here arises most clearly for Sidgwick’s analysis of ‘good’. He defined the good as what one ought to desire, initially using the same ‘ought’ that figures in deontic judgements about acts (1907: 112). But in an early edition of The Methods of Ethics he recognized that this sense isn’t really available here. Noting that ‘irrational desires cannot always be dismissed at once by voluntary effort,’ he said the analysis of ‘good’ can’t use ‘ought’ in ‘the strictly ethical sense’, but only in ‘the wider sense in which it merely connotes an ideal or standard’ (1890: 110–11). But this raises the questions whether, if the wider ‘ought’ isn’t governed by a strong ‘can’ condition, it really is ethically deontic, and how exactly it differs from the ‘good’ it’s being used to analyse. How does saying one ought in the wider sense to have a desire differ from saying that the desire is good? The wider ‘ought’ is supposed to express an ‘ideal’, but the concept of an ideal seems evaluative, referring to something with an exemplary degree of goodness. (The chapter in Principia Ethica giving Moore’s substantive account of the good is

Against the Fundamentality of Fit  61 called ‘The Ideal’.) Ross made this point in Foundations of Ethics. If we say someone ‘ought’ to feel sorrow when his character makes that impossible, he wrote, we can mean only that his not feeling sorrow ‘is a bad thing’ (1939: 45). If that’s right, however, Sidgwick’s wider ‘ought’ isn’t just a minor variant of the deontic ‘ought’ but is in the opposed, evaluative category and in effect equivalent to ‘good’. He was therefore using not one basic concept but two, and, far from being in­form­ ative, his analysis of ‘good’ was circular, since it defined the goodness of outcomes in terms of the goodness of desires. It didn’t analyse the evaluative in terms of the deontic but explained value in terms of value. One could try analysing ‘good’ conditionally, as what one ought, in the full deontic sense, to desire if one can. But Sidgwick didn’t consider this possibility, and in my view he was right not to. On a view like his the concept ‘good’ gives directives, and it should give them to everyone, so everyone ought to desire what’s good. If it’s analysed conditionally, however, it says nothing to those who don’t now have and can’t make themselves have the relevant desire; for them claims using ‘good’ are idle. In consequence a conditionally analysed ‘good’ doesn’t have some implications an ethical view like Sidgwick’s needs. That you ought to desire x is meant to imply, in conjunction with a general demand to take effective means to your ends, that you ought, if you can, to do what will promote x. But this conclusion won’t follow, for those who can’t desire x, if ‘good’ is analysed conditionally. The ‘ought’ that governs acts requires only ‘physically can’, not the stronger ‘phys­ic­ al­ly and given his motivation can’; it isn’t restricted to those who can have some desire. It therefore won’t follow in its full generality from a ‘good’ that’s addressed only to those who can have a desire. It won’t be possible to say of someone who can’t now desire x that he acted wrongly in not promoting x when he physically could have. The same difficulty about ‘ought’ implies ‘can’ arises for fittingness-­first views. If the concept of fit is to underlie or be equivalent to that of prima facie duty or ought other things equal, it too must be governed, at least in that context, by a strong ‘can’ condition. But it can’t involve that condition when it’s used to analyse ‘good’, since we don’t in the relevant way control our desires. Ewing for one recognized this. Granting that ‘we cannot alter our emotions at a moment’s notice’, he said we can nonetheless ‘speak of the emotions a man “ought” to have in the sense of “the suitable emotions for him to have”’ (1947: 165, also 133, 150–51; 1939: 13–14). But if ‘fitting’ in the analysis of ‘good’ doesn’t require voluntary control whereas deontic fittingness does, Ewing too was using not one basic concept but two, with the difference between them masked by the ambiguous words ‘fitting’ and ‘suitable’. And we can again ask how the concept of fit he used to define ‘good’ differs from ‘good’ if the latter concept, too, makes no assumptions about control. Was he, too, analysing value in terms of value? It may be replied that ‘fitting’ has content not found in ‘good’ and differs from it in that way; for one, it’s explicitly relational as ‘good’ is not. ‘Fitting’ does have distinctive content, which I discuss in the next section. But this doesn’t prevent it

62 Fittingness from being in the same general category as ‘good’; it can just be an evaluative concept with specific features. If so, Ewing’s analysis doesn’t reduce the evaluative to something non-­evaluative; it explains one evaluative concept in terms of another. Another response agrees that ‘fitting’ is evaluative and never subject to a strong ‘can’ condition, but proposes a conditional analysis of the deontic ‘ought’ in terms of ‘fitting’. More specifically, it says an act is one you ought to do if the act is fitting and in your voluntary control, or one you could or would do given the requisite motivation. This conditional analysis is the reverse of the one suggested for Sidgwick’s account of ‘good’, since it’s of the deontic in terms of the evaluative rather than vice versa, and therefore not open to the same objection. A conditional analysis narrows the extension of the concept it’s defining as compared to that of the one used to define it, but that’s not a problem here since ‘ought’ does apply more narrowly than ‘fitting’ or ‘good’, namely only to acts one can do. But this response leaves standing the question whether an analysis of ‘good’ in terms of ‘fitting’ isn’t circular, and beyond that is problematic as an account of ‘ought’. The proposed analysis makes the ethical or normative content of a deontic judgement using ‘ought’ the same as that of an evaluative judgement using something like ‘good’. The deontic judgement just adds, as a further empirical condition external to that content, a requirement of voluntary control. That doesn’t seem right: the requirement of control seems internal to the normative component of the ethical ‘ought’ rather than an outside, non-­normative addition to it. A view that makes it just an addition, or makes the normative content of ‘ought’ merely evaluative, doesn’t reflect this fact. To equate ‘ethically ought’ with an evaluative concept plus a separate ‘can’ is to omit something vital (compare Chappell 2012: 691–92). The difficulty about ‘ought’ implies ‘can’ is less obvious for fittingness-­based views like Broad’s and Ewing’s than for Sidgwick’s, since it’s unclear whether ‘­fitting’ requires voluntariness or not; this unclarity serves precisely to obscure the issue. The same is true of views in which the one basic concept is that of a normative reason. Despite its prominence in recent moral philosophy, talk of ‘reasons’ is notoriously ambiguous. The best-­known ambiguity is between explanatory or motivating reasons on one side and normative or justifying ones on the other, but another concerns the ‘can’ condition. Does the claim that you have a normative reason to do x assume that you can voluntarily produce x if you’re motivated to or not? It seems that in some contexts it does and in others not. If x is an act and ‘normative reason to x’ plays a similar ethical role to ‘ought other things equal to x’, voluntariness does seem to be assumed. But in an analysis of the good as what there’s ­normative reason to have a pro-­attitude to, it can’t be: we can’t produce attitudes in ourselves just because we want to. So ‘reason’, even in just its normative use, seems to express two different concepts governed by different conditions. And we can again ask how, if an analysis of ‘good’ in terms of attitudes can’t assume

Against the Fundamentality of Fit  63 voluntary control, saying an attitude is one there’s reason to have differs from ­saying the attitude is good. To this point I’ve followed Sidgwick, Ross, and others in assuming a ‘can’ condition that requires control by the will. But some philosophers have proposed a weaker condition requiring only the ability to respond, whether through the will or not, to reasons or other normative factors (Hieronymi 2009; McHugh 2017). This weaker condition is needed, they argue, to make sense of deontic claims about belief—‘you ought to believe p given your evidence’—emotion, and desire, none of which are subject to voluntary control. And unlike a voluntariness condition, one requiring only responsiveness to reasons isn’t violated when ‘ought’, ­‘fitting’, or ‘reason’ is used to analyse ‘good’. That ‘ought’ can be applied to beliefs and emotions doesn’t mean it then expresses the same concept as the ethical ‘ought’ used of acts or that the ‘can’ required in the former contexts is all that’s needed in the latter. The word may just be ambiguous. And that the ethical ‘ought’, which is the only one I’m considering, involves a stronger voluntariness condition has long been held and has considerable intuitive appeal. If we ask why it’s not true that you ought to cure cancer tomorrow, the answer is surely that you couldn’t do so even if you wanted; any inability to respond to reasons here rests on an inability to will. Though he defends a reasons-­ responsiveness condition for epistemic ‘oughts’, Conor McHugh agrees that the ‘ought’ applied to acts and central to ethics requires voluntary control (2017: 2758); it makes a stronger demand. And that control can’t be assumed in any concept said to underlie ‘good’. There are other accounts of the difference between deontic and evaluative concepts, ones not involving any ‘can’ condition, but they have similar implications. One view (e.g. Cullity, this volume) says deontic judgements are directive, or give recommendations, as evaluative judgements do not; the latter may be ‘merely classificatory’ (Nagel 1970: 109). But this account creates the same type of difficulty for a fittingness-­first view. If ‘good’ isn’t directive, the ‘fitting’ used to analyse it can’t be directive either. How then can it capture what on this view is an essential feature of ‘ought’, one internal to its normative content rather than just an external addition? Rather differently, Selim Berker (this volume) argues that deontic and evaluative properties and concepts have opposed logical features. The core deontic elements, ‘obligatory’ (or ‘simply ought’) and ‘permitted’, don’t admit of degrees, have a dual—­each is the other’s dual—­and don’t have a contrary. In contrast, the core evaluative elements, ‘good’ and ‘evil’, do admit of degrees, don’t have a dual, and have each other as contraries. It’s again hard to see how a unitary fittingness concept can underlie ones with these contradictory features.1

1  Berker himself identifies fittingness properties and concepts as a third category distinct from the deontic and the evaluative. I argue implicitly against this, at least in the ethical context, in section 3.3.

64 Fittingness It’s widely held that there’s an important difference between deontic and evalu­ ative ethical concepts, and though there are competing accounts of what that difference is they all create difficulties for a fittingness-­first approach, since the features needed for a reduction of one type of concept to fittingness can conflict with what’s needed for a reduction of the other.

3.3  Deontic Fittingness: Thin or Thick? Whether fittingness can be fundamental depends in part on its content, or what it is for a thing to be fitting. If ‘fitting’ is an ethical concept, then like other such concepts it denotes a property that supervenes on some other, non-­ethical property that is its ground. Just as anything good is made good by some non-­ethical property such as its being pleasant, so anything fitting is made so by some other property on which its fittingness supervenes. There are then three issues, which discussions of fittingness haven’t to date distinguished: about the ethical concept or property of fittingness itself, about the non-­ethical properties that can be its ground, and about the relation between the two. This last issue concerns how far ‘fitting’, if ethical, is a thin, and how far a thick, concept. An entirely thin ethical concept places no restrictions on what can, or can make a thing, fall under it. ‘Good’ is close to entirely thin. Perhaps only states of affairs can be good, but beyond that any non-­ethical property whatever can, as far as the concept goes, ground goodness. (If some properties are more plausibly good-­making than others, that’s a substantive, not conceptual, matter.) ‘Ought’ may be slightly less thin, since it applies only to acts a person can do. But it, too, places no limits on which properties can make an act one you ought to do and so allows many competing deontic views. An ethical concept is entirely thick, in contrast, if it completely specifies its ground, which then in conjunction with the non-­ethical facts fully determines its extension. Few if any ethical concepts are entirely thick, but many are what we can call partly thick. These concepts restrict their admissible grounds to properties of a certain type or in a certain area but don’t specify a particular property in that area, leaving that for ethical determination. ‘Courage’ is partly thick, since it allows only acts that accept harm or the risk of harm for a sufficiently important purpose to count as courageous but leaves other questions open. These include whether this purpose must in fact justify accepting the harm or you need only believe it does; whether the purpose must involve good consequences or can be merely symbolic; and which types of consequence are good. Another partly thick concept is distributive justice. Whether a distribution, say of income, is just can depend only on some property it has as a distribution rather than, say, on the motives of those who produced it. But the concept of justice doesn’t specify which property this is: it could be equality, proportionality to merit, or something

Against the Fundamentality of Fit  65 historical. Those who share the concept can therefore disagree about what, ­substantively, is just. Fittingness can’t be an entirely thick concept. If it were, any claim based on a false belief about what makes things fitting would contradict itself, since it would assume a ground the concept excludes. (Alternatively, claims by people with differing views about what makes for fittingness wouldn’t contradict each other, since they would use different concepts.) The credible options, if ‘fitting’ is ethical, are that it’s thin and that it’s partly thick. Neither, I’ll argue, allows fittingness to be fundamental. The reasons, however, are different in the deontic and evaluative contexts; I begin with the former. Imagine first that a fundamental deontic fittingness concept is thin, so, like ‘ought’, it’s unanalysable, involves a strong ‘can’ condition, and places no restrictions on the properties that can make an act fall under it. In one situation the ground of fittingness may be just that an act keeps a promise and in another just that it promotes happiness, with no commonality needed between the two. The question here is how this thin fittingness concept differs from the ‘ought’ it’s supposed to be more fundamental than. On the view we’re considering ‘fitting’ applies to the same acts as ‘ought’ and does so on the same basis; it also has no additional non-­ethical content. How then does calling an act fitting to your situ­ ation differ from saying it’s one you ought in your situation to do, and how can the first claim explain the second? The two could in principle be distinct. ‘Fitting’ could be one thin deontic concept and ‘ought’ another, with the first somehow underlying the second. Then a property like keeping a promise would first make an act fitting and then that would make it one you ought to do. But it’s hard to see what function fittingness would serve here or what substantive difference positing it would make. If ‘fitting’ in its deontic use is thin, it seems indistinguishable from ‘ought’ and so can’t be more basic. The alternative is that ‘fitting’ is partly thick, and there are independent reasons to believe this. Fittingness is an explicitly relational or two-­place concept, since any x that’s fitting is fitting to some y; what makes x fitting must presumably also involve a relation to y. This distinguishes ‘fitting’ both from the one-­place ‘good’ and from ‘ought other things equal’, which can sometimes depend on an act’s intrinsic character. And there are further restrictions on the types of relation that can ground fittingness if the connotations of the everyday word carry over even partly to its philosophical use. (If they don’t, why use the word?) In everyday usage ‘fitting’ connotes some relation of matching or, more broadly, complementarity. Fitting is what a round peg does in a round hole but a square peg doesn’t; it’s what your clothes do if they’re roughly the same size and shape as you. An ethically relevant fittingness-­relation can’t be exactly the same as these, but it can be of the same general type. The idea will then be that ‘fitting’ is partly thick because it requires its ground to be some relation of matching or com­ple­men­tar­ity. It’s a further substantive question which specific relation this is and therefore

66 Fittingness which specific items are fitting. But it’s part of the concept that what makes x fitting to y must be some non-­ethical complementarity between x and y. The same idea is suggested by philosophical claims such as that fitting is ‘the relation in which a response stands to an object when the object merits—or is worthy of—that response’ (Howard 2018: 2; 2019: 216). This requires the relation that makes x fit y to involve, more specifically, x’s being a ‘response’ to y, and further restrictions are suggested by the term ‘merit’. Merit can’t here be a distinct concept from ‘fitting’, since that would make it and not fit fundamental. It must be an equivalent to ‘fitting’ that brings out more clearly some of its connotations. But merit is closely connected to the concept of desert, and that often rests on matching. Thus moral desert is satisfied when good things happen to good people or bad things to bad ones. This again suggests that ‘fitting’ is a partly thick concept whose ground must involve some form of non-­ethical matching or com­ple­men­ tar­ity. That fittingness is ethically basic then implies that all ethical facts, both deontic and evaluative, rest on some such relation. Anything obligatory or good is so because of some complementarity between an x and a y. Assume then that deontic fittingness is in this way partly thick and combine it with Broad’s equation of prima facie duty with fittingness to an aspect of your situation. The resulting view fits a number of Ross’s prima facie duties, such as the duty to keep promises. Here the act that you ought other things equal to do matches, or is identical to, one you promised to do, so a relevant relation does seem present. The view may also fit the duty of gratitude, where you match another’s benefit to you with a return benefit to her, and perhaps the duty to compensate, though here there’s the reverse complementarity of a benefit that makes up for a previous harm. But these are all duties of a specific type, ones triggered by a previous act—­of promising or harming by you, or of benefiting by another—­to which the act they require ‘responds’. But many other duties, including arguably the most important ones, don’t require any triggering act; they’re standing duties that apply to everyone at all times. One is the duty to promote good consequences such as happiness, as in Ross’s duty of beneficence. What informative com­ple­ men­tar­ity does this duty rest on? To say an act fits the fact that in your situation it will promote happiness is just to repeat that it will promote happiness; no serious relation to the situation is invoked. Broad himself recognized this. Having early in Five Types of Ethical Theory grounded all moral duties in fittingness, he later in the book said an act’s being right depends on two factors: its fittingness to the total course of events as modified by it, and, separately from that, its utility, or the value in its consequences (1930: 218–20). Here the duty to promote the good wasn’t a matter of fittingness but was distinct from it. One could say that an act with good consequences is fitting, not to the situation in which it’s performed, but to the value of those consequences, since it involves a positive orientation to, or effort that will lead to, that value. This makes the act’s fittingness similar to that of a fitting attitude, which likewise turns on its orientation to its object. But it

Against the Fundamentality of Fit  67 invokes a complementarity relation quite unlike that of a promised act to the promise to do it, with a different type of match to a different type of object, one not present in the current situation. This raises the question whether a single thick fittingness concept really is being invoked in the two cases. (This is also a difficulty for the larger programme of reducing both ‘ought’ and ‘good’ to ‘fitting’: if the complementarities in the two contexts are different, is the reduction really to one concept?) And a further such relation is needed for the duty of non-­ maleficence, or duty not to cause harm, for example not to cause pain. On the one hand, that another person may suffer pain grounds a prima facie duty to prevent the pain, as part of the duty to promote good and prevent bad consequences. But in deontological views like Ross’s it also grounds a separate and stronger duty not to directly cause the pain even if that will result in less pain overall, as when torturing one person will prevent two others from being tortured. This means there are two prima facie duties related to another’s pain, one to prevent it and the other not directly to cause it, which requires two forms of fittingness based on two different relations to the same thing. The second duty could perhaps be grounded in the anti-­complementarity of positively promoting a negative value, but on its own that wouldn’t capture the agent-­relativity of non-­maleficence, which tells you, not to minimize the total amount of harming in the world, but to ensure that you don’t harm. Incorporating that in the underlying relation would make for an additionally distinctive form of complementarity. The general point is that, given the many prima facie duties in a theory like Ross’s, any attempt to ground them all in a partly thick fittingness concept requires a plurality of different com­ple­ men­tar­ity relations, and the question is whether they have enough in common to constitute a single concept, in particular one that explains the duties rather than one we merely posit, along with whatever complementarity seems needed, whenever we detect what’s really primary, an ‘ought other things equal’. A further difficulty for treating deontic fittingness as partly thick would arise even if all prima facie duties rested on a single non-­ethical complementarity. It concerns how the elements of a thick concept relate to each other, an issue on which there are two competing views. A reductive view of thick concepts takes them all to be analysable into a thin concept plus, what’s separable from it, some non-­ethical restriction on its ground. Then to call an act courageous, for example, is to say it’s one you ought or that it would be good to do because of something related to its accepting of harm. On this view fittingness, if partly thick, combines a thin concept with a separable claim about its ground, so to call an act fitting to an aspect of your situation is to say it falls under some thin concept because of how it complements that aspect. The question then is what the relevant thin concept is. It could be one of the familiar concepts, in this case ‘ought’, so an act is fitting if it’s one you ought to do because of some complementarity. But then ‘fitting’ isn’t fundamental; it’s just a restriction of the more basic ‘ought’. The alternative is that the thin concept is a

68 Fittingness distinctive fittingness one, so an act’s complementing an aspect of your situation gives it an irreducible thin property of fit, which then makes it one you ought to do. But this returns us to the question of how a thin deontic ‘fitting’ differs from the equally thin ‘ought’ it’s supposed to underlie. If it has no additional content, what separates it from ‘ought’? A rival view says thick concepts can’t be reductively analysed because their ethical and non-­ethical parts aren’t separable but interpenetrate each other, making each such concept an indissoluble whole. If fittingness were in this way irreducible it could be fundamental, but the view that it’s so is in several ways problematic. The main argument for the irreducibility of thick concepts, due to John McDowell (1981) and repeated by Bernard Williams (1985) and others, assumes in effect that all such concepts are entirely thick. It objects that, if the elements of a thick concept were separable, we could know its extension while grasping only its descriptive part and having no idea of its evaluative point. This, it says, is impossible: applying a thick concept requires evaluative judgement. But this argument is irrelevant to partly thick concepts, whose non-­ethical parts leave important ethical issues open and don’t fully determine their extension, which depends in part on evaluation (Elstein and Hurka 2009). Since ‘fitting’ must be just partly thick if disagreements about it are to be possible, the argument doesn’t apply to it. In addition, seeing ‘fitting’ as irreducibly thick is in some tension with the claim that it’s conceptually basic. That claim is reductive, saying all other eth­ ic­al concepts involve, at bottom, the single one of fittingness. But then the view that ‘fitting’ is irreducibly thick invites the question why other concepts such as courage and justice can’t be so too. Rejecting a reductive analysis of ‘fitting’ sits uncomfortably with the claim that all other ethical concepts are reducible; it also precludes explaining what’s distinctive about the concept by giving a non-­ethical specification of its required ground. To summarize: if deontic fittingness is thin it’s unclear how it differs from ‘ought’, whereas if it’s partly thick the different complementarities needed for the different ‘oughts’ make it questionable that it’s a unified concept. If thick concepts are reducible, the first issue about differing from ‘ought’ recurs, while the view that fittingness is irreducibly thick is unsupported by arguments and at odds with other fittingness-­first claims.

3.4  Evaluative Fittingness: Thin or Thick? When we turn to the evaluative context the difficulties facing the view that ‘fitting’ is fundamental are different, because the concept’s relation to ‘good’ is different. In the deontic case ‘fitting’ applies to the same items as ‘ought’, so when an act is fitting, say to an aspect of your situation, the same act is one you ought other things equal to do. But when a state of affairs is good, what’s fitting is something

Against the Fundamentality of Fit  69 distinct from it, a desire for or pro-­attitude to it. There therefore isn’t the same obstacle to treating fittingness here as thin: it can’t be objected that it’s hard to see how a thin evaluative ‘fitting’ differs from ‘good’ since it differs at least in what it applies to. It’s perfectly coherent, and not redundant, to say a desire’s falling under one thin concept, ‘fitting’, makes its object fall under the different thin concept ‘good’. On the resulting view, however, nothing informative can be said about why a desire for a given object is fitting. If desires for pleasure and for knowledge are both fitting, what makes the one fitting is just its being a desire for pleasure and what makes the other so is just its being a desire for knowledge. There’s nothing common to the relations between desire and pleasure and between desire and knowledge that explains why both make for fittingness; both just do. Though there’s no inherent difficulty in analysing ‘good’ using a thin fittingness concept, doing so is in tension with the general reasons to understand ‘fitting’ as having non-­ethical content, those coming from the connotations of the everyday word and of philosophical equivalents such as ‘merit’. They strongly suggest that the concept is partly thick. Moreover in this context there’s a non-­ethical matching relation that desires like those for pleasure and knowledge do share and that can make them all fitting. It just doesn’t allow ‘fitting’ to be fundamental. This relation figures in a substantive claim about the good made by several philo­sophers of our period, especially Hastings Rashdall, Moore, and Ross, and associated by many with the good of virtue. It says that if some initial states are intrinsically good or evil, certain attitudes to them are also intrinsically good and other attitudes to them evil. More specifically, if something is intrinsically good, desiring, pursuing, or taking pleasure in it, what some call loving it, for its own sake is a further intrinsic good and for many an instance of virtue. Thus if another’s pleasure is good, wanting, pursuing, or being pleased by it as an end is an additional good and involves the virtue of benevolence. If another’s pain is evil, hating it for itself, that is, desiring or seeking its absence or being pained by its presence, is a good and compassionate. But the contrary attitudes of loving an evil and hating a good, as in wanting another’s pain or being pained by his pleasure, are evil and vicious. Given an initial set of intrinsic values, this claim generates a further set of higher-­level intrinsic values (Hurka 2001). And underlying the claim is a simple idea about complementarity: that attitudes whose orientation matches the value of their object are good while ones whose orientation opposes that value are evil. A positive attitude to a positive value is good, as is a negative attitude to a negative one, but a positive attitude to a negative value or a negative attitude to a positive one is evil. The matching relation here isn’t naturalistic, since one of its terms is ethical: the goodness or evil of an object. But it also isn’t ethical, since by itself the relation makes no normative claim and gives no normative guidance. It doesn’t say matching attitudes are good or ones we ought to promote; it just reports, descriptively, a relation of match. But

70 Fittingness that relation figures in a further ethical claim about the goodness of matching attitudes and the evil of unmatching ones. The same relation can make for fittingness, so an attitude is fitting when its orientation matches the value of its object, and as so used it can explain the fittingness of many diverse attitudes. In the deontic context different com­ple­men­ tar­ity relations were needed for different duties, but here the one matching relation can explain the fittingness of desires for pleasure, knowledge, virtue, just­ ice, and more, in fact for anything good. The relation also fits the use of terms like ‘merit’. An item usually merits a response because of some merit or demerit, some good or bad quality it has; thus you can merit or deserve happiness for the intrinsic good of virtue or income for the instrumental good of contributing to others. This is reflected if what makes an attitude merited is some positive or negative value in its object.2 At the same time, fittingness as so understood can’t be used to analyse ‘good’. Since what makes an attitude fitting is its relation to some independent value, the fittingness presupposes ‘good’ and can’t be more basic. Though involving a relevant complementarity, it can’t be fundamental. The above points allow two more specific understandings of ‘fitting attitude’. On one it’s a partly thick ethical concept whose non-­ethical content requires a matching of orientation to value as its ground. On the other, which I find on balance preferable, it’s not ethical or normative but just asserts, descriptively, that a match of this kind obtains. I start with the first understanding. If ‘fitting’ here is partly thick, then to call an attitude fitting is to say it falls under some thin concept because its orientation matches its object’s value. This concept could in principle be a thin fittingness one, so when a desire’s orientation matches a value the desire is unanalysably fitting. But it’s hard to see what point this would have if ‘fitting’ can’t be used to analyse ‘good’, given that the match presupposes ‘good’. A more plausible view exploits the substantive claim above and takes the thin concept to be ‘good’, so to call an attitude fitting is to say it’s good because its orientation matches a value. Moore suggested an analysis of this type when he said, ‘by saying that different emotions are appropriate to different kinds of beauty, we mean that the whole which is formed by the consciousness of that kind of beauty together with the emotion appropriate to it, is better than if any other emotion had been felt in contemplating that particular beautiful object’ (1903: 190). This analysis of ‘fitting attitude’ parallels a similar one of the deontic ‘fitting’ in terms of ‘ought’ and means there’s a second reason why ‘fitting’ can’t be evaluatively fundamental: as well as presupposing ‘good’, in its relation of match, it also reduces to ‘good’. But Christopher Howard has made three objections to this analysis. 2  Uses of ‘merit’ where this condition isn’t satisfied, as in the claim that there being seven specks of dust on a desk merits the belief that there are seven (Howard 2019: 225), surely misuse the concept. Facts about dust can’t merit anything.

Against the Fundamentality of Fit  71 He argues, first, that analysing ‘fitting attitude’ in terms of ‘good’ gives the wrong result in non-­ethical cases: though a colleague’s getting a promotion makes it fitting to envy her, it doesn’t make envying her good. He also argues that the analysis has an unwanted implication in ethical cases. If an attitude’s being fitting involves its being good, we ought to do what will bring about the attitude, which includes bringing about its necessary conditions. But then that compassion, especially for really existing pain, is fitting implies that we ought to bring about its conditions, which include the pain. Yet compassion’s being fitting to pain doesn’t imply that we ought, even prima facie, to cause pain. Finally, he says the analysis gets the order of explanation wrong. It implies that an attitude’s being good helps explain why it’s fitting, whereas in fact its being fitting explains why it’s good (2018: 4–5; 2019: 227–29). The analysis can, however, answer all these objections. It can say, first, that non-­ethical cases like the one about envy are irrelevant to an analysis that’s only of ethical fittingness, because the complementarity they involve is so different. If envy of another’s success can indeed be called fitting—­which I myself doubt—­its fittingness may rest on nothing more than its involving a true belief about its object. That’s completely different from what makes desire fitting to another’s happiness or compassion to her pain. The second objection, though serious, doesn’t face only this analysis; it arises for any view that values virtues like compassion. Consider the view that compassion is an underivative good, one whose value has no further explanation, in terms of fittingness or anything else. It, too, seems to imply that we ought to bring about compassion and anything necessary for it, including pain. And in fact there are a number of contexts where, if an x exists, there’s a reason or duty to bring about some y, but there’s no reason to bring about the conjunction (x and y) or to create x so y can exist later. If you’ve made a promise, you ought other things equal to keep it, but there’s no duty to (both make and keep promises). If you’ve decided to conceive a child you should ensure that it will be happy, but many philosophers think there’s no duty to (decide to conceive a happy child and then do so), or to procreate. Desire-­fulfilment theories of the good say that if someone desires x, it’s good for her to get x, but many deny that if you create a desire in her, say through advertising, and then fulfil it you’ve made her better off. Virtues like compassion provide further instances of this pattern, as does desert. There should therefore be some general explanation why, in all these cases, the reason-­giving force of the y doesn’t extend back before the existence of the x. Whatever this is, a reduction of ‘fitting attitude’ to ‘good’ can use it to deny that compassion’s being fitting to pain is any reason to cause pain. Finally, the analysis allows the right explanations. If ‘fitting’ requires some non-­ ethical complementarity as its ground, the claim that an attitude is good because it’s fitting says the attitude is good because of some complementarity. That an act was wrong because it was murder is explanatory even though murder is by

72 Fittingness def­in­ition wrongful killing; it says the wrongness depends somehow on the act’s being a killing. That an attitude is good because it’s fitting likewise gives an at least partial specification of the value’s ground. And though it would be odd to say, what’s in one sense true, that an attitude is fitting because it’s good, it’s equally odd to say an act was murder because it was wrong. In both cases the ‘because’ is most nat­ur­al­ly read as citing the ground of the partly thick judgement, so the concept’s non-­ethical part is taken to explain its ethical part rather than, what would be less natural, the ethical part explaining the whole.3 Though a reduction of ‘fitting attitude’ to ‘good’ can answer Howard’s objections, the alternative understanding of the phrase is to me more attractive. It denies that ‘fitting’ in ‘fitting attitude’ is any kind of ethical concept. Instead it’s merely descriptive, saying only that an attitude stands in a non-­ethical com­ple­ men­tar­ity relation to its object and drawing no normative conclusion from this fact, which it simply reports. When what’s fitting is an attitude, the relevant relation involves a match between an orientation and a value, and one possibility is that ‘fitting’ here connotes this specific relation, so the claim that an attitude is fitting means it matches in this particular way. But the concept may instead be less determinate, allowing different complementarity relations in different contexts. Then if envy can be fitting—­though I’ve questioned this—­its being so may consist in its involving a true belief about its object, and a different relation may be involved in a belief ’s being fitting. But on both versions of this view the claim that an attitude is fitting is just descriptive, not ethical or normative. It can of course be ethically significant, if we add that attitudes whose orientation matches their object’s value are good. But the claim itself has no ethical content. This reading of ‘fitting attitude’ avoids all Howard’s objections. There are no unwanted implications about ‘good’ if a fittingness-­claim says nothing about ‘good’. There may be a reason to cause pain if it’s added that fitting attitudes are good, but that’s a difficulty about the addition, not about fittingness itself. And an attitude’s being fitting explains its being good, since the matching grounds the goodness, but the attitude’s goodness doesn’t explain the matching. This merely descriptive reading is simpler than the reductive one, which to me makes it preferable, but what’s important is that both are available. On one, ‘fitting attitude’ expresses a partly thick ethical concept that presupposes ‘good’, while on the other it has no ethical content. On neither is it both ethical and irreducible, and on neither can it be ethically fundamental. It doesn’t follow that there can’t, as a third option, be an irreducible thin evaluative fittingness. But the question, parallel to one about deontic fittingness, is what reason there is to posit this concept when the other two options are available and reflect, as the third doesn’t, the non-­ethical connotations of ‘fitting’. In fact the availability of the two options

3  For a parallel response to a similar objection see Cullity (this volume, section 7.7).

Against the Fundamentality of Fit  73 suggests an explanation of why some philosophers have been led to think ­fittingness is fundamental. What there clearly is in the evaluative context is a non-­ethical matching ­relation between an attitude and a value. This relation is entirely different from the unanalysable ethical fittingness these philosophers posit, and the suggestion is that they may have slid from a tacit recognition of the non-­ethical relation to, what doesn’t follow, an explicit affirmation of the more problematic ethical one. That there’s a kind of fit that can’t be evaluatively basic hardly implies that there’s one that can. They could then slide from this supposedly fundamental evaluative fittingness concept to, what’s even more questionable, an equally fundamental deontic one. Whether they made these transitions or not, the availability in the evaluative context of the non-­ethical relation is a reason to doubt that fit is there fundamental, a reason that supplements the even stronger reasons for doubt in the deontic case.

3.5  Gradable vs. Non-­Gradable Deontic Concepts My final doubt concerns only the claim that ‘fitting’ is the basic deontic concept. It doesn’t question a fittingness-­based analysis of ‘good’, though I’ve made other objections to that, nor is it relevant only to fittingness. It applies equally to other potentially basic deontic concepts such as prima facie duty, ought other things equal, and normative reason. These are all gradable concepts, ones that admit of degrees. An act that’s fitting can be more or less so; thus it can be more or less fitting to a given situation than some other act. Prima facie duties, oughts other things equal, and normative ­reasons can likewise differ in their degrees of strength, so one is stronger than, or outweighs, another when they conflict, or outweighs a third that the second does not. These gradable concepts can be used to define further deontic ones that aren’t gradable, such as those of what’s most fitting, your strongest prima facie duty, or what you have most reason to do. The latter concepts don’t themselves admit of degrees; something either just is most fitting or just isn’t. But they’re the superlatives of more fundamental concepts that do, and therefore derive from something gradable. These concepts are also all implicitly conditional. Thus to say you ought other things equal to keep a promise to do x is to say you ought to do x given that you promised to. Conditional oughts like this admit of degrees but can be used to define a further concept that doesn’t, that of what you ought all things considered or on balance to do. This concept is still conditional, however, since it’s that of what you ought to do given all the facts, or given all aspects of your situation. In contrast with all these deontic concepts is that of what you ought or have a duty simply or unconditionally to do. This alternative concept doesn’t admit of degrees; it’s either just true that you ought simply to do an act or just false. Nor is

74 Fittingness it defined in terms of a concept that does. It’s not the superlative of some ­comparative but makes a simple unconditional claim: not that you ought to do an act given certain facts but just that you ought to do it. Since it’s neither itself gradable not defined in gradable terms, I’ll call it entirely non-­gradable. Those who use this simple concept usually take its ground to be something gradable. Thus on many views what you ought, in the simple or unconditional sense, to do is always what’s most fitting or what you have most prima facie duty or reason to do. For these views the two types of concept, though distinct, are connected, and the connection will be conceptual if one of the two is defined in terms of the other, which then is primary. There are two possibilities here, each illustrated in one of Ross’s books. In The Right and the Good Ross took the simple or entirely non-­gradable concept, which he called ‘duty proper’ or ‘absolute duty’, to be primary and defined his gradable concept of prima facie duty in its terms. More specifically, his best definition of prima facie duty equated it with a ‘tendency’ to make an act your duty proper, so a prima facie duty to do acts of type F means an act’s being F tends to make it one you ought simply to do, or can contribute to its being one you ought simply to do (1930: 28–29). He illustrated this idea of tendency with an analogy to physical forces: qua subject to one force, an object may tend to move in one direction, while qua subject to another it tends to move in a different direction, with the one it actually moves in determined by the two forces’ relative strengths. A prima facie duty to do F acts may likewise tend to make an act simply right while a competing duty not to do G acts tends to make it wrong, with the act’s final deontic status depending on which of these duties is stronger. But here the gradable ‘prima facie duty’ is defined by its contribution to the simple ‘duty proper’.4 In Foundations of Ethics Ross changed his view, perhaps without fully realizing he was doing so. Influenced by Broad, he now said you have a prima facie duty to do an act of type F if doing an F act fits, or is suitable to, an aspect of your situ­ ation (1939: 51–53, 79–82). Unlike Broad in his later discussion, Ross didn’t contrast an act’s fittingness with its utility. He thought the prima facie duty to promote the good, too, involves fittingness to an aspect of your situation, now that it’s one in which good can be promoted (1939: 81), so all prima facie duties were understood in terms of fit. Far from defining ‘prima facie duty’ in terms of ‘duty proper’, his new account of the prima facie made no reference to duty proper. And though it did define prima facie duty in terms of fittingness, that concept, too, is gradable. Whereas The Right and the Good defined a gradable deontic concept by its

4  Note that for this definition to be informative, ‘duty proper’ has to express the simple concept. Defining ‘prima facie’ by a tendency to make an act the one most supported by prima facie duties would be circular, given the latter concept’s dependence on the former.

Against the Fundamentality of Fit  75 relation to an entirely non-­gradable one, Foundations of Ethics took a gradable concept to be primary. This would seem to require relating the simple concept to the gradable one, and Foundations of Ethics did say we can identify rightness with ‘the greatest amount of suitability possible in the circumstances’ (1939: 53). But the concept of the most suitable act isn’t the duty proper of The Right and the Good. It’s the superlative of a comparative and so derives from something gradable; it’s not at all the entirely non-­gradable concept of the earlier book. It’s in fact hard to see how the simple deontic concept could be defined or built up from a gradable one, as it must be if the latter is taken to be fundamental. How, starting with fittingness, prima facie duty, or normative reason, all of which admit of degrees, could one arrive, by purely conceptual means, at a deontic concept with no element of degrees? Whereas The Right and the Good showed how a gradable deontic concept can be defined in terms of the simple one, it’s hard to see how, starting with a gradable concept, one could ever generate an entirely non-­gradable one like duty proper. Yet an ethical view needs a concept like that if it’s to give determinate practical guidance. To say only that an act is most fitting or what you have most prima facie duty or reason to do isn’t yet to tell you specifically to do it. It’s not to give you that simple directive, as the unconditional ‘ought’ or ‘duty’ would. It’s not to do what an ethical view centrally should do. Consider another view that uses only gradable concepts, so-­called scalar consequentialism. Consequentialist ethical views first rank acts by the amount of good they’ll produce: this one will produce more, this one less, perhaps this one the most possible. The most familiar such views then make a cut on this ranking and say we ought, in the simple sense, to do only acts above this cut. In maximizing consequentialisms an act is above the cut only if no other act will produce more good; in satisficing ones an act need only produce a sufficient amount of good. But some philosophers defend a view that makes no cut and no claim using the simple ‘ought’: it just says some acts will produce more good, some less, and some perhaps the most (Slote 1985: Ch. 5; Norcross 2006). The resulting scalar consequentialism has the merit, they say, of not making an arbitrary decision about where to make a duty-­defining cut. At the same time, though, it seems not to give determinate guidance. That one act will do more good and another less may be an interesting fact but doesn’t itself tell you to do the first act, nor does it license blame if you don’t. That would require the further claim that you ought, in the simple sense, to do the act. Without that claim scalar consequentialism gives no clear directives. The same is true of views that use only gradable deontic concepts or ones defined in their terms. That an act is most fitting, your strongest prima facie duty, or what you have most reason to do doesn’t in itself tell you to do it; that requires a further claim using the simple concept. Nor can that concept be generated from

76 Fittingness gradable ones; from claims that don’t tell you simply to do an act we can’t get to one that does. But if the simple concept can’t be generated from a gradable one, a view that takes a gradable deontic concept to be fundamental can’t give de­ter­min­ ate guidance. It can generate a ranking of acts but not say what we’re to do with it. Essentially this argument has been made by John Broome (2015). He takes the English word ‘ought’ to express only what I’ve called the simple or entirely non-­ gradable concept. Though it wouldn’t affect his main conclusion to recognize a gradable ‘ought’, as in ‘ought other things equal’, he thinks ‘ought’ has just the simple use. And he takes the main gradable deontic concept to be that of a normative reason, given its prominence in recent philosophy. His own view analyses the gradable ‘reason’ in terms of the non-­gradable ‘ought’, so you have a normative reason to do F acts if an act’s being F contributes to a ‘weighing explanation’ of why you ought (simply) to do it. This parallels The Right and the Good’s analysis of prima facie duty in terms of a tendency to make for duty proper, or makes a similar claim in different language.5 Broome contrasts his view with a ‘reasons-­first’ view that takes the gradable concept of a normative reason to be primary and any other deontic concepts to derive from that. This view can talk of what you have ‘sufficient’ or ‘most’ reason to do, but neither of these, he argues, is equivalent to our ordinary concept of what you ought simply to do, since that ‘is not the concept of the superlative of a gradable property’ (2015: 96). Nor can the reasons-­first view arrive at this concept if it starts with only a gradable one. Part of his charge against the view, then, is that it isn’t faithful to our everyday ethical concepts, but behind that is a more philosophical one. ‘We cannot do without oughts in an account of normativity,’ he writes, since ‘the whole point of normativity is to determine what we ought to do, ought to believe, and so on’ (2015: 93). A view that can’t arrive at simple oughts, in other words, can’t do what a normative view should do: give determinate guidance. A related argument was made in our period, though in mostly unpublished correspondence of Prichard’s. He was the original source of Ross’s concept of prima facie duty and, in this period, of the general idea of a gradable deontic concept. He assumed this type of concept in ‘Does Moral Philosophy Rest on a Mistake?’ (2002: 7–20) and explicitly defended it in later Oxford lectures (2002: 77–83); Ross acknowledged a major debt to him on this score (1930: v). Unlike Ross in The Right and the Good, however, Prichard seems to have taken the grad­ able concept, his term for which was ‘a claim’, to be primitive or unanalysable (Letter to Collingwood of 10 February 1933), and this coloured his reaction to Ross’s book. Soon after its publication he wrote to Ross saying he now believed 5  Much of Broome’s discussion is ontological, about the ought and reason properties rather than concepts. But one of his main arguments against a reasons-­first view (see later) is that it doesn’t fit our everyday ought concept.

Against the Fundamentality of Fit  77 ‘that what there is is not the act which I ought to do but acts which I ought to do  (i.e. your prima facie duties not your duty sans phrase)’ (Letter to Ross of 9 September 1931). Later he said he was ‘disposed to think that your prima facie duties are the only duties there are’, so the most we can say about an act is that it’s your strongest prima facie duty (Letter to Ross of 28 January 1938), and that ‘your “my duty sans phrase” is, really that of a man’s duties which he most ought to do’ (Letter to Ross of 14 July 1932; 2002: 287). He was in effect proposing a merely scalar deontology, one using only concepts that are either gradable or defined in gradable terms and no entirely non-­gradable one like Ross’s duty proper. He recognized, I believe, that if we start with a gradable concept like ‘prima facie duty’ we can’t derive, by purely conceptual means, any simple or entirely non-­gradable one; if we start with degrees, we’ll end with something at least based on degrees. Nor could he accept adding the simple ‘duty proper’ to ‘prima facie duty’ as a further, independent primitive. To him the idea that there are two distinct senses of ‘ought’ and ‘duty’ was ‘difficult’; there would also have to be a principle linking the two, which he thought can’t be given (Letter to Paton of 31 January 1938). His reasoning here may have been as follows. We can certainly say ‘Your duty proper is always to do the act most supported by prima facie duties,’ but if the two concepts are independent this principle will be synthetic rather than analytic, where, intuitively, he may have thought, it must be at least partly analytic, or partly conceptually based. On a strong reading of The Right and the Good the definition of prima facie duty makes the principle entirely analytic, so, by conceptual necessity, your duty proper is always to do what’s most supported by prima facie duties. On a weaker reading, some acts can be most supported by prima facie duties but not your duty proper, say because other acts, though less supported, are also sufficiently supported. (This would be a satisficing account of duty proper.) But even on this weaker reading it’s analytic that a prima facie duty to do F acts means an act’s being F counts towards its being your duty proper, and analytic that whether an act is your duty proper depends somehow on the prima facie duties that apply to it. But there would be no such conceptual connections if ‘prima facie duty’ and ‘duty proper’ were independent, and it’s arguably that which Prichard couldn’t accept. If ‘duty proper’ can’t be analytically linked to ‘prima facie duty’, he concluded, it should be dropped. Whereas Prichard seems to have been content with a merely scalar deontology, I’ve argued that that view is problematic. As Broome says, it doesn’t fit our everyday ethical thought, which uses the simple ‘ought’ and in which that entirely non-­ gradable concept is arguably more prominent than any gradable one. (Note that Prichard and Ross had to introduce a gradable deontic concept to a philosophical literature that before them was unaware of it; Sidgwick, Rashdall, and Moore all made arguments that ignored the possibility of any such concept.) In addition, the merely scalar claim that an act is most fitting, most supported by prima facie duties, or what you have most reason to do doesn’t tell you specifically to do it or

78 Fittingness give you determinate guidance. If they’re treated as basic, gradable concepts like ‘fitting’ can’t fulfil the main task of a deontic concept. They can, however, ­contribute to that task if they’re analysed in terms of the simple ‘ought’ that then is the primary such concept. For yet another reason, fittingness can’t be fundamental.

References Broad, C. D. 1930. Five Types of Ethical Theory. London: Routledge & Kegan Paul. Broome, John. 2015. ‘Reason versus Ought.’ Philosophical Issues 25: 80–97. Chappell, Richard Yetter. 2012. ‘Fittingness: The Sole Normative Primitive.’ Philosophical Quarterly 62: 684–704. Cullity, Garrett. 2018. Concern, Respect, and Cooperation. Oxford: Oxford University Press. Elstein, Daniel  Y. and Thomas Hurka. 2009. ‘From Thick to Thin: Two Moral Reduction Plans.’ Canadian Journal of Philosophy 39: 515–36. Ewing, A. C. 1939. ‘A Suggested Non-Naturalistic Analysis of Good.’ Mind 48: 1–22. Ewing, A. C. 1947. The Definition of Good. London: Routledge & Kegan Paul. Ewing, A.  C. 1959. Second Thoughts in Moral Philosophy. London: Routledge & Kegan Paul. Hieronymi, Pamela. 2009. ‘Two Kinds of Agency.’ In Mental Actions, 138–62. Edited by L. O’Brien and M. Soteriou. Oxford: Oxford University Press. Howard, Christopher. 2018. ‘Fittingness.’ Philosophy Compass 13/11: 1–14. Howard, Christopher. 2019. ‘The Fundamentality of Fit.’ In Oxford Studies in Metaethics 14, 216–36. Edited by R. Shafer-Landau. Oxford: Oxford University Press. Hurka, Thomas. 2001. Virtue, Vice, and Value. New York: Oxford University Press. McDowell, John. 1981. ‘Non-Cognitivism and Rule-Following.’ In Wittgenstein: To Follow a Rule, 141–62. Edited by S.  H.  Holtzman and C.  M.  Leich. London: Routledge & Kegan Paul. McHugh, Conor. 2017. ‘Attitudinal Control.’ Synthese 194: 2745–62. McHugh, Conor and Jonathan Way. 2016. ‘Fittingness First.’ Ethics 126: 575–606. Moore, G. E. 1903. Principia Ethica. Cambridge: Cambridge University Press. Moore, G. E. 1912. Ethics. London: Williams and Norgate; reprinted Oxford University Press 1965. Moore, G. E. 1942. ‘A Reply to My Critics.’ In The Philosophy of G. E. Moore. Edited by P. A. Schilpp, 533–687. Chicago: Northwestern University Press. Nagel, Thomas. 1970. The Possibility of Altruism. Oxford: Clarendon Press. Norcross, Alastair. 2006. ‘The Scalar Approach to Utilitarianism.’ In The Blackwell Guide to Mill’s Utilitarianism, 217–32. Edited by H. R. West. Oxford: Blackwell.

Against the Fundamentality of Fit  79 Prichard, H.  A. 2002. Moral Writings. Edited by Jim Macadam. Oxford: Clarendon Press. Prichard, H.  A. Correspondence. The Prichard Papers. The Bodleian Library. University of Oxford. Rashdall, Hastings. 1907. The Theory of Good and Evil. 2 vols. London: Oxford University Press. Ross, W. D. 1930. The Right and the Good. Oxford: Clarendon Press. Ross, W. D. 1939. Foundations of Ethics. Oxford: Clarendon Press. Russell, Bertrand. 1904. Review of Principia Ethica. Independent Review 2: 328–33. Sidgwick, Henry. 1890. The Methods of Ethics, 4th ed. London: Macmillan. Sidgwick, Henry. 1907. The Methods of Ethics, 7th ed. London: Macmillan. Slote, Michael. 1985. Common-Sense Morality and Consequentialism. London: Routledge & Kegan Paul. Williams, Bernard. 1985. Ethics and the Limits of Philosophy. Cambridge, Mass.: Harvard University Press.

4 What Is Evaluable for Fit? Oded Na’aman

4.1 Introduction Our beliefs, intentions, desires, regrets, and fears are evaluable for fit—­they can succeed or fail to be fitting responses to the objects they are about.1 Can our headaches and heartrates be evaluable for fit? The common view says ‘no’. I will argue: sometimes, yes. To get a sense of what I have in mind, consider Darwin’s description of a frightened man: The frightened man at first stands like a statue motionless and breathless, or crouches down as if instinctively to escape observation. The heart beats quickly and violently, so that it palpitates or knocks against the ribs . . . the skin instantly becomes pale, as during incipient faintness . . . That the skin is much affected under the sense of great fear, we see in the marvelous and inexplicable manner in which perspiration immediately exudes from it. This exudation is all the more remarkable, as the surface is then cold, and hence the term a cold sweat . . . The hairs also on the skin stand erect; and the superficial muscles shiver. In connection with the disturbed action of the heart, the breathing is hurried. The salivary glands act imperfectly; the mouth becomes dry . . . One of the best-­marked symptoms is the trembling of all the muscles of the body; and this is often first seen in the lips. From this cause, and from the dryness of the mouth, the voice becomes husky or indistinct, or may altogether fail. (Darwin 1872/2009: 290–291)

My initial case against the common view asks: given that fear is evaluable for fit, why not the various bodily episodes and sensations that accompany fear? I then consider and resist various answers to this question.

1  I am indebted to an extremely helpful discussion with participants of the Fit Fest Workshop, held in May 2021, most of whom contributed their own chapters to this volume. I also benefited from discussion of the chapter at a conference on ‘Reasoning and Agency’ held at Tel Aviv University and online in August 2021. For written comments and extensive conversation, I am grateful to Rachel Achs, Selim Berker, Chris Howard, and Alex Prescott-­Couch. Oded Na’aman, What Is Evaluable for Fit? In: Fittingness: Essays in the Philosophy of Normativity. Edited by: Christopher Howard and R. A. Rowland, Oxford University Press. © Oded Na’aman 2022. DOI: 10.1093/oso/9780192895882.003.0004

What Is Evaluable for Fit?  81 The common view is usually not expressed in terms of fit, but in terms of rationality, justification, or reasons: the view is that sensations and bodily episodes are not evaluable by standards of rationality, justification, or reasons. I will eventually argue that the notion of fit-­evaluability better captures the phe­nom­ enon in question, but for the sake of argument I begin by adopting the notion of rational evaluability. Often, the common view makes its appearance when authors contrast states that are rationally evaluable with certain bodily episodes and sensations. Scanlon (1998: 20) contrasts belief and intention with hunger, tiredness, and distraction; Moran (2001: 114) contrasts desires we can reason to with hunger or fatigue; Boyle (2011: 22) says we can reason to belief but not to pain; Brady (2018: 81) contrasts rational emotions with experiences of coldness, tiredness, hunger, nausea, and irritation; and Neta (2018: 289) contrasts conspiring, concluding, resenting, and fearing with feeling tired, craving Doritos, and having an itch on your elbow. This is a small selection; there are many, many more examples of the common view. That hunger, tiredness, and pains, like other bodily episodes and sensations, are not rationally evaluable is considered a truism not worth arguing for. The common view about what is rationally evaluable has shaped contemporary discussions of fittingness, which focus on fitting attitudes to the exclusion of other possible candidates for fit-­evaluation.2 I will challenge the assumption that sensations, feelings, and bodily episodes can at most be caused by fit-­evaluable attitudes and suggest that, like attitudes, they are fit-­evaluable when and because they are explained by fit-­evaluable narratives. I begin, in section 4.2, by drawing the distinction between rationally evaluable items and non-­rational items. I recount what is often said in order to distinguish the kinds of items that fall in each category. Then, in section 4.3, I offer an initial case against the common view: I argue that according to our evaluative practices, an accelerated heartrate is rationally evaluable when associated with one’s fear. In section 4.4, I consider and reject the objection that I misdescribe our evaluative practices. In section 4.5, I argue that the common view is informed by ques­tion­ able theoretical assumptions. In particular, it seems to be widely assumed that whether a physical or mental phenomenon is ever fit-­evaluable is determined by the type of phenomenon it is. In section 4.6, I suggest an alternative whereby the same type of phenomenon can be fit-­evaluable on one occasion but not on another. What explains the difference, I argue, is that only in the former occasion the phenomenon is explained by the agent’s fit-­evaluable narrative as an element of emotion.

2  To be sure, some of the historical literature on fittingness seems to assume that actions (as well as attitudes) are fit evaluable. See, for example, Broad (1930) and Ross (1939) (I thank Chris Howard for this point and for the references). However, McHugh and Way (forthcoming) is a great example of the contemporary tendency to privilege attitudes. I say more about this later.

82 Fittingness

4.2  The common view Philosophers distinguish (1) things that are subject to, or assessable by, norms of rationality and justification and (2) things that are not. I will refer to things falling under (1) as rationally evaluable items and to things falling under (2) as non-­ rational items. How to analyse the distinction is controversial and various accounts of it have been offered, but there is a general agreement on what falls on each side of the distinction. Consider the following lists, divided according to what I will call the common view of the distinction: LIST 1. Rationally evaluable items: Believing an online-­meeting starts at 11 a.m., dreading the meeting, intending to join the meeting, joining the meeting, coming to the conclusion that the meeting had started at 10:30 a.m., regretting that you didn’t double check the meeting time, and apologizing for your late arrival. LIST 2. Non-­rational items: Feeling tired, having an itch on your chin, being hungry, experiencing a warm glow, feeling nauseous or dizzy, having a headache, digesting, sensing your heartrate accelerating, and your heartrate accelerating. LIST 1 includes paradigmatic examples of rationally evaluable kinds, such as actions (joining the meeting), beliefs (believing the meeting starts at 11 a.m.) and intentions (intending to join the meeting). The list also includes emotions (dreading the meeting, regretting you didn’t double check the meeting time). The view that at least some emotions are rationally evaluable is somewhat more controversial than the view that intentions and beliefs are rationally evaluable, and yet it is widely accepted. LIST 2 includes items that are almost universally thought to be non-­rational, such as hunger, headaches, and heartrates. My first goal is to dispute the way the distinction is normally applied, as illustrated by LIST 1 and LIST 2; I will argue that hunger, headaches, and heartrates can sometimes be rationally evaluable. My second goal is to make sense of the distinction in light of my first thesis. But first, consider what is often said of our evaluative practices in order to distinguish the kinds of items in LIST 1 from those in LIST 2. I summarize seven generally agreed upon and closely related characteristics of rationally evaluable items. With regard to each I explain how it leads to the common application of the distinction: a. Rationally evaluable items are items to which we normally apply norms of rationality and justification. It makes sense, according to our evaluative practices, to ask whether a person’s belief that the meeting starts at 11 a.m.

What Is Evaluable for Fit?  83 and her intention to join the meeting are rational and justified. We thus evaluate beliefs and intentions according to norms of rationality and justification. We also ask such questions and make such evaluations with respect to emotions, such as guilt, admiration, anger, pride, etc. However, it does not make sense to evaluate a person’s headache or heartrate by applying to them norms of rationality and justification, so headaches and heartrates are non-­rational. b. Rationally evaluable items are items for which we normally seek and offer reasons-­for-­which explanations. You might intelligibly ask why I dread the meeting in the specific sense of asking for the reason in light of which I dread the meeting.3 But it would not be intelligible to ask me for the reason in light of which I am tired or dizzy. Though my tiredness and dizziness can, in principle, be explained, they cannot be rationalized or justified (Neta 2018: 289). So rationally evaluable items are items that can be given an explanation of a specific kind; they may be given a reasons-­for-­which explanation. c. Rationally evaluable items are items we normally expect to align with the agent’s judgements. Rationally evaluable items are answerable to the agent’s judgements: their presence or absence is impacted by the presence or absence of the relevant evaluative judgements (Moran 2017: 144). My belief that the meeting starts at 11 a.m. should change when, after joining the meeting, I come to the conclusion that the meeting had started at 10:30 a.m. Similarly, I should change my intention to join the meeting if I judge that there is no reason for me to attend it after all. By contrast, it is not a rational failure on my part that my headache persists despite my judgement that it is bad nor do we expect my heartrate to drop simply because I cannot explain its sudden acceleration. Sometimes this point is put by saying that rationally evaluable items are ‘judgement sensitive’: they depend on a rational agent’s judgements about normative reasons (Scanlon 1998: 20). d. Rationally evaluable items are items we can normally reason to. Rationally evaluable items need not result from reasoning but they can, in principle, be reasoned to, and if challenged one should be disposed to provide the 3  I prefer the term reasons-­for-­which to motivating reasons. The term motivating reasons is often used to refer to the explanatory role of reasons as motivators of actions. A so-­called motivating reason motivates an agent to act, as opposed to a normative reason, which counts towards the justification of the action. Attitudes, however, are not formed on the basis of motivations. That the bear is dangerous is not the reason that motivates my fear of the bear, it is the reason in light of which I fear the bear. It is also the reason in light of which I run away. So the term reasons-­for-­which captures the explanatory role of reasons in the formation of both actions and attitudes. Moreover, I might have motivations that explain my action or attitude but are not the reasons in light of which I perform an action or form an attitude. My shyness can motivate my decision not to ask a question without being a reason in light of which I decide not to ask a question. So the term reasons-­for-­which is adequately broader but also adequately narrower than the term motivating reasons. I therefore think it more precisely picks out the relevant reason-­explanation of rationally evaluable items. Thanks to Selim Berker for discussion.

84 Fittingness reasoning that leads to them (e.g. Smith 2005; McHugh 2017). My apology upon realizing that I was late to the meeting need not result from a process of practical reasoning, but if someone asked why I apologized I should be able to provide the reasoning that leads to my apology, or to my decision to apologize. Pamela Hieronymi captures this point by saying that ‘an intention to φ embodies one’s answer to the question of whether to φ’, which makes one vulnerable to ‘questions and criticisms that would be satisfied by reasons that (one takes to) bear positively on whether to φ’ (Hieronymi 2009: 138–139). When a person is tired or hungry there can be no similar pressure because such states cannot be reasoned to. To be sure, one can reason to the intention to bring about or prevent states of tiredness and hunger, but one cannot reason to tiredness and hunger directly. Conor McHugh makes a similar point about the relation between belief and the reasoning that supports it: If your visual experience causes you to acquire the belief that the wall is white, but you are in no way disposed to reason from the belief that the wall looks white to the belief that it is white—­should your belief that it is white be challenged, say—­then you don’t count as basing the latter belief on the former, and thus as responding to the putative reason given by its content. (McHugh 2017: 2757)

e. Rationally evaluable items are normally attributable to the agent. Rationally evaluable items reveal the agent’s evaluative point of view and therefore reveal something about the agent’s mind or self (Hieronymi 2014: 16). My dread of the meeting reveals that I view it as threatening and my intention to join the meeting reveals that I view it as worth joining despite the threat I take it to pose. Similarly, although I say that the dog is not dangerous, my fear can reveal that I view it as dangerous indeed. Moreover, even if I sincerely judge that the dog poses no danger, the fact that my fear persists indicates a failure of rationality due to a conflict within my evalu­ ative perspective. Thus, our rationally evaluable attitudes express our evalu­ ations of their objects. As such, they are attributable to us in a way that non-­rational items are not. An itch, perspiration, or a rash, do not reveal a person’s evaluative perspective nor anything else about the person’s mind or self. In this respect, non-­rational sensations and bodily episodes are like any other event or condition that is not attributable to an agent: the rotation of the earth, the sunlight entering the room through the window, the room temperature, the breaking of the glass, the stain on the carpet. In some of these cases, agents can be held responsible for bringing about or not preventing these events and conditions, but these events and conditions are not of an agent in the way that the intention to bring them about or prevent them is.

What Is Evaluable for Fit?  85 f. Only attitudes are fundamentally rationally evaluable. The recent focus on attitudes in philosophy of normativity is sometimes accompanied by the proposal that all rationally evaluable items are explained by rationally evaluable attitudes (and only rationally evaluable attitudes are not explained by other kinds of rationally evaluable items). This is the thesis that only attitudes are fundamentally rationally evaluable. Actions, for example, are said to be rationally evaluable only when and because they are expressions of rationally evaluable intentions, which are, of course, rationally evaluable attitudes.4 So the distinction between what is rationally evaluable and what is non-­rational is often described as a distinction between rationally evaluable attitudes and non-­rational attitudes, feelings, moods, sensations, and bodily episodes. It is important to note here that according to this view, nothing physical—­ no movement or occurrence in the body—­ is itself rationally evaluable; physical movements and occurrences are only rationally evaluable as guises of rationally evaluable attitudes. Moreover, many mental states are also non-­rational. So there is a question about which mental states fall on either side of the distinction.5 g. Rationally evaluable items are items that are fit-­evaluable. Rationally evaluable attitudes can be fitting or unfitting to what they are about—­they can be evaluable for fit. Fit is a normative relation between an attitude and what it is about. Different types of attitudes are individuated by the kinds of things that merit them or that they are fitting to. Fear is fitting to (or is merited by) what is fearsome, admiration is fitting to (or merited by) what is admirable, belief is fitting to (or merited by) what is credible, and desire is fitting to (or merited by) what is desirable. It is generally assumed that all rationally evaluable items are fit-­evaluable and, given (f), it is further assumed that most if not all fit-­evaluable items are attitudes.6 However, sensations, pains, and conditions such as fatigue or thirst are neither fit-­ evaluable nor rationally evaluable. These are the seven commonly invoked characteristics of rationally evaluable items. As we saw, it is commonly thought that, when considered in light our evaluative practices, these characteristics apply primarily to attitudes and not to bodily episodes and sensations, which are, therefore, non-­rational.

4  Gibbard (1990: 38–9), Scanlon (1998: 21), Portmore (2011: 63), Smith (2013: 60), and Hedden (2015: ch. 6) endorse this view. Recently, McHugh and Way (forthcoming) have argued for this view at length. 5 Nolfi (2015) offers an answer to the question which mental states are rationally evaluable. However, Nolfi shares the assumption I will later question, namely, that rational evaluability is a property of types of items. 6  Howard (2018) suggests actions can be evaluable for fit.

86 Fittingness

4.3  The initial case against the common view The correct description of our evaluative practices—­of our normal expectations and dispositions—­does not, by itself, settle the question of which items are rationally evaluable. However, if we normally treat a certain item as rationally evaluable this is strong prima facie reason to hold that it is in fact rationally evaluable. To argue that, contrary to our practices, the item in question is non-­rational, one must provide some strong reason against the apparent force of our practices. Alternatively, one can dispute the accepted interpretation of our evaluative practices and argue that we do not treat the item in question as rationally evaluable after all. Consider the example of grief. Stephen Wilkinson argues that grief is non-­ rational by appealing to our evaluative practices (Wilkinson 2000). He claims that lack of normal grief in response to loss does not strike us as a rational failure as long as one’s beliefs and desires are rational and justified. By contrast, Donald Gustafson argues that, contrary to our evaluative practices, grief is always ir­ration­al. Gustafson appeals to a theory of rationality according to which a rational state must enable the agent to realize states of affairs she desires; he then argues that grief fails to do so (Gustafson 1989). Responses to these and similar arguments contest Wilkinson’s interpretation of our evaluative practices regarding grief (Jollimore 2004) as well as Gustafson’s theoretical assumptions about rationality (Cholbi 2017; Marušić 2018; Moller forthcoming). Despite these debates—­and perhaps as a result of them—­the common view today is that grief is rationally evaluable and that this is supported both by our evaluative practices and by commonly accepted theoretical assumptions about rationality and justification.7 I, too, share the view that grief is rationally evaluable, but I think we should go even further and maintain that, for example, the headaches, sleeplessness, and loss of appetite that normally accompany grief (or that are elements of grief) are also rationally evaluable. To argue for this de­part­ ure from the common view, I first dispute the common depiction of our evalu­ ative practices and then question the theoretical assumptions that underlie it. Start with the following example: your heartrate. According to the common view of rationally evaluable items, both the bodily episode and the feeling of your heart racing are non-­rational. Consider characteristics (a) to (g) with respect to the following case: RUNNING TO THE BUS Your heart races after running to catch the bus. 7  But the idea that grief is rationally evaluable has also given rise to theoretical questions about the expiration of reasons and the temporality of fit. See Moller (2007, forthcoming); Marušić (2018, forthcoming); Na’aman (2021); Schönherr (forthcoming).

What Is Evaluable for Fit?  87 (a) It makes little sense to evaluate the rationality or justification of your accelerated heartrate after running to catch the bus; (b) while there is a clear cause there is no reason in light of which your heart is accelerating in this case; (c) we should not expect your heartrate to decrease in response to changes in your evaluative judgements, so your racing heart is not judgement-­sensitive; (d) we also do not suppose that you must be able, if challenged, to reason to your accelerated heartrate; (e) your racing heart is not an expression of your evaluative point of view nor is it attributable to you in the way your intentions or regrets are; (f) your heartrate and your experience of it are not attitudes and they are not explained as rationally evaluable by rationally evaluable attitudes; (g) your heartrate and your experience of it are not evaluable for fit. So far, the common view seems correct when assessed in light of the characteristics of rationally evaluable items outlined about. In RUNNING TO THE BUS, your racing heart as well as your sensation of your racing heart are non-­rational. Now consider a different example: BEAR ENCOUNTER It is a dark night and you are camping alone in the woods. As you fall asleep near the campfire, you hear a noise, turn around, and see a bear standing over you. The bear is examining you, looking you up and down. In any moment, he might strike you and it will all be over. Your heart is racing. In fact, your heart is beating at the exact same pace as it was beating after you ran to catch the bus: you are in a very different predicament but have the very same heartrate. I begin my push against the common view by making an initial case that your accelerated heartrate upon facing the bear is rationally evaluable: it has crucial characteristics of a rationally evaluable item and gives us reason to revise other purported characteristics of rationally evaluable items. Let us consider each characteristic in turn: (a)  It should be uncontroversial that the fear you experience as you face the bear is rationally evaluable. Indeed, your fear might very well be both rational and justified since you are in fact in great danger. Your racing heart seems rational and justified in response to the bear for the same ­reasons that your fear is rational and justified. Had you responded in the same way to a mouse, both your fear and your racing heart would seem rationally criticizable and unjustified. That said, fear is a kind of item to which we normally apply norms of rationality and justification, whereas the same cannot be said of an accelerated heartrate, as demonstrated by RUNNING TO THE BUS. Despite this difference, in BEAR ENCOUNTER the presence of the bear seems to justify and rationalize your ra­cing heart as well as your fear. I return to this issue below.

88 Fittingness (b)  There seems to be a reason in light of which your heart is racing: you are in great danger. That you are in danger is not merely a causal explanation of your racing heart, but a reason-­for-­which explanation. To see this, consider the statement: ‘my heart is racing because there’s a bear approaching me.’ Now contrast BEAR ENCOUNTER with a deviant-­ causal-­ chain variant in which a bear approaches you, unseen, and steps on a button that triggers a mechanism that causes you to see a hologram of a bear, which in turn causes your heart to race. In this variant, the claim ‘my heart is racing because there’s a bear approaching me’ is only true when we switch to using a ‘because’ of causal explanation. The fact that the relevant sense of ‘because’ seems to change between the cases supports the idea that in the original BEAR ENCOUNTER case ‘because’ denotes a reason-­ for-­which explanation.8 Another indication that there is a reason for which your heart is racing is the fact that it is possible to ignore the reason for which your heart is racing. Start, again, with fear. Suppose I’m scheduled to give a talk later today and I fear being exposed as a sham. One friend tries to reason with me. She points out, for ex­ample, that I’ve proven myself as a worthy philosopher in the past, or that this talk was received well on other occasions. I am not a sham and will not be exposed as one and therefore my fear is unjustified. Perhaps she makes good points, perhaps not. But in any case, I am not persuaded and my fear persists. A second friend tries a different approach: she offers me a tranquillizer. The tranquillizer mitigates my fear, but it ignores the reason in light of which I am afraid, namely, that I deem it likely that I be exposed as a sham. The tranquillizer has a non-­ rational impact on my fear. Now, as I take the tranquillizer and my fear dissipates my heartrate drops. But just as in taking the tranquillizer I ignore the reasons-­for-­ which I fear, so I ignore the reason-­for-­which my heart races. The tranquillizer has a non-­rational impact on my heartrate. So there is a reason-­for-­which ex­plan­ ation of my accelerated heartrate; it is the same as the reason-­for-­which ex­plan­ ation of my fear. Had my first friend been successful in persuading me that I will do well in my talk, both my fear and my heartrate would rationally dissipate due to a change in the reasons I take myself to have. (c)  As in the case of fear, we should expect your heartrate to decrease upon judging that you are out of danger or that the bear is not dangerous after all (if, e.g. the bear turns out to be your friend in a very realistic bear costume). Your racing heart is as much a reflection of your judgement as your fear is. So your racing heart in facing the bear is judgement-­sensitive. This might seem too quick. When Scanlon introduces the notion of ‘judgement-­ sensitive attitudes’ he writes: ‘These are attitudes that an ideally rational person would come to have whenever that person judged there to be sufficient reasons

8  I thank Selim Berker for this point and example.

What Is Evaluable for Fit?  89 for them and that would, in an ideally rational person, “extinguish” when that person judged them not to be supported by reasons of the appropriate kind’ (Scanlon 1998: 20). To judge that one is in danger is not to judge that one’s ­heartrate is supported by sufficient reason. So the fact that one’s heartrate would decrease upon judging one is out of danger does not yet show that one’s heartrate is judgement-­sensitive. And if it never makes sense to ask whether one’s heartrate is supported by sufficient reason then an ideally rational agent would not make judgements about the matter and the notion of judgement-­sensitivity would not apply to heartrates. So to accept the claim that one’s heartrate is judgement-­ sensitive in BEAR ENCOUNTER one must already accept the ­conclusion that this claim is meant to support, namely, that one’s heartrate is rationally evaluable. However, Scanlon’s characterization of judgement-­ sensitive attitudes is am­bigu­ous with regard to the content of the relevant judgements. On the first, narrow reading of Scanlon’s statement, a judgement-­sensitive attitude is sensitive only to the judgement . This is the reading the objection presupposes. On the second, broad reading of the statement, a judgement-­sensitive attitude is also sensitive to the judgement where r constitutes sufficient reason for this attitude, or where the absence of r implies that this attitude is not supported by ­reasons of the appropriate kind. To see that the broad reading is more plausible, consider the following. An ideal­ly rational person who judges

would believe p even if she does not make the further judgement . Similarly, an ideally rational person who judges would have the intention to φ even if she does not make the further judgement . And, finally, an ideally rational person who judges would fear the bear even if she does not make the further judgement . It is therefore more plausible that an item’s judgement-­sensitivity includes an item’s sensitivity to a person’s judgements about facts that constitute reasons for the item; the item is not merely sensitive to judgements about whether the item is supported by ­reasons. So the fact that your heartrate drops when you judge that you are no longer in danger can be an indication that your heartrate is judgement-­sensitive in this case. (d)  If your fear is challenged, you should be able to reconstruct the reasoning that leads to it—­e.g. the bear is only a few feet away and might kill me in one stroke, so I’m in great danger. Adopting Hieronymi’s terminology, we can say that just as an intention to φ embodies one’s answer to the question of whether to φ, fear of the bear embodies one’s answer to the question of whether the bear poses a danger to oneself. We form a rationally evaluable attitude by answering a question about its content. It is arguable that one’s racing heart in response to the bear is

90 Fittingness also an embodiment of one’s answer to the question of whether the bear poses a danger. If that is so, then you can reason to your racing heart in the very same way that you reason to your fear. (e) We are inclined to view your racing heart in this case as an expression of your evaluative point of view and to attribute it to you in the same way we at­tri­ bute your fear to you. Just like your fear, so your perspiration, shivers, and racing heart express your understanding and appreciation of your predicament. Another indication that we attribute these responses to you is the fact that if your shivers, perspiration, and racing heart continue after the bear has gone away, we might appropriately reason with you to persuade you that you are no longer in danger. Since we take your physical symptoms to be justified in response to perceived danger, and we take you to be rational, we expect these symptoms to go away once you judge that you are out of danger. We would not have the same ex­pect­ ation if your heartrate were drug-­induced. (f)  It is true that your racing heart and your sensation of it are not themselves attitudes. However, the fact that they bear crucial characteristics of rationally evaluable items and that in the bear encounter case they are closely associated with the attitude of fear, suggests that there is a connection—­perhaps an explanatory connection—­between the rational evaluability of bodily episodes and sensations and the rational evaluability of attitudes. I consider this connection in section 4.6. Perhaps, it might be suggested, the connection is merely causal—­your rationally evaluable fear causes your heart to race—­and the appearance of rational evaluability is illusory? However, at this stage in the argument, I am only concerned with how things appear to be. I consider this proposal and offer reasons against it in the next section (section 4.4). (g)  If only attitudes are fit-­evaluable then your racing heart, which is not an attitude, is not fit-­evaluable. However, in light of the fact that your racing heart bears many characteristics of fit-­evaluable items, the following possibilities should be considered. First, it is possible that things other than attitudes are fit-­evaluable. Second, perhaps bodily episodes and sensations are fit-­ evaluable when and because they stand in a certain relation to attitudes that are fit-­evaluable. Again, more on this in section 4.6. This concludes my initial case for the claim that our evaluative practices treat at least some bodily episodes and sensations as sometimes rationally evaluable. It is worth noting that the example we have been considering is by no means an outlier. Our rationally evaluable attitudes are often accompanied by bodily ­episodes, conditions, and sensations that seem to be as rationally evaluable as the attitudes they accompany. In our daily life, we do not normally draw the distinction, commonly drawn by philosophers, between attitudes and other mental and physical states and events. It is therefore striking that the expansive philosophical literature on rationally evaluable attitudes rarely considers such cases and assumes that

What Is Evaluable for Fit?  91 only attitudes are rationally evaluable. By considering objections to my initial case, we will find a possible explanation for this widespread neglect.

4.4  A response on behalf of the common view As far as I can tell, the only discussion in the recent literature of cases such as BEAR ENCOUNTER appears in Angela Smith’s work.9 In formulating her rational relations account of responsibility for attitudes, Smith argues that ‘nonintentional mentals states, such as physical pains, sensations, and physiological conditions such as hunger or thirst . . . are not rationally sensitive to our evaluative judgements or our wider cognitive and evaluative commitments’ (Smith 2005: 257). Smith goes on to consider a possible objection: One might object here that many of these physical states do, in fact, seem to be directly connected to our evaluative judgments. The nausea that I feel before having to speak in public, for example, seems to be a direct result of my evalu­ ative judgment that such public performances are both important and also fraught with opportunities for failure. The butterflies that I feel in my stomach before boarding a roller-­coaster also seem to be a direct result of my evaluative judgment that such a ride is scary and somewhat dangerous. Does it follow on my account, then, that these physical states are also attributable to me for purposes of moral assessment?  (257–258)

In response, Smith insists that ‘The relation between a person’s physical states and her evaluative judgments is purely causal’ and therefore our responsibility for these states ‘flows from the responsibility we have for the evaluative judgments which constitute their causal triggers’ (258). Smith proposal is an example of characteristic (f) above: we can be responsible for our physical states only in­dir­ ect­ly, in the way that we can be responsible for other states of affairs, that is, via our direct responsibility for our evaluative judgements. This explanation of the phenomenon makes it compatible with the view that only attitudes are fundamentally rationally evaluable but it does not constitute a reason to endorse this view. Why explain away the apparent rational evaluability of physical episodes and sensations? More specifically, given that certain physical episodes and sensations seem to have the characteristics of rationally evaluable items, what reason do we have to insist that they are merely caused?

9  But see fn. 11, where I mention Jennifer Corns’s related discussion of the rational evaluability of pleasantness and unpleasantness (Corns 2019).

92 Fittingness Smith specifies characteristics of rationally evaluable items that, she claims, physical states lack. She points out that when the causal connection between our evaluative judgements and our physical states fails, we are not thereby open to rational criticism. However, where regret, guilt, or remorse are fitting, distress—­ with its mental and physical components, such as accelerated heartrate—­is fitting as well. To feel fitting fear of the bear without feeling distress is to be vulnerable to rational criticism; to feel fitting remorse about one’s crime without feeling distress is to be vulnerable to rational criticism. In response, Smith might argue that, due to the strong causal connection between fear and remorse, on the one hand, and distress, on the other, it is hard to believe or even to imagine that one experiences fear or remorse without experiencing distress. The absence of distress is therefore not rationally criticizable itself but is a strong indicator of the rationally criticizable absence of a rationally evaluable attitude. But, again, we must ask whence the insistence that the relation between distress and our evaluative judgement is merely causal? Whether distress is a component of fitting emotions or a distinct phenomenon (a point which I will come back to in section 4.6), our evaluative practices indicate that we sometimes view it as fitting in the same way and for the same reasons we view the relevant emotions as fitting.10 Thus, judging that I’ve done you wrong, it is not enough that I apologize and change my ways, I should also regret the wrong and, in regretting, be distressed by it. In another appeal to our evaluative practices, Smith claims that it makes sense to ask a person to defend or justify ‘her shame, jealousy, fear, or admiration’ but it does not make sense to ask her to defend or justify her nausea (258). But consider the kind of nausea a guilt-­ridden person might feel. The person might say, ‘I’m sick to my stomach thinking about what I’ve done,’ meaning it quite literally. ‘What justifies this reaction?’ a friend might respond, ‘as far as I can tell you’ve done nothing wrong’. Or consider a case where the guilt-­ridden person describes her sensations and feelings without realizing she is feeling guilty. Her friend might help her identify these sensations and feelings as guilt. The friend might say: ‘what you’re feeling is guilt, but what do you have to feel guilty for?’ and this might be a first step towards helping the guilt-­ridden person see that her nausea, irritation, and sleeplessness, are not non-­rational but rationally evaluable, and that there is no justification for them. Maybe once she recognizes these facts, the feelings and sensations she’s been suffering from will rationally dissipate. Alternatively, they might be reinforced by the realization that they are justified indeed. Of course, we do not expect people to defend claims about the fit of their bodily episodes and sensations. But the same is true of guilt. In justifying her guilt, the guilt-­ridden person would explain what she had done and why it was wrong 10  Corns (2019) argues in detail for a similar claim—­specifically, that the pleasantness or unpleasantness associated with a wide range of mental phenomena is rationally evaluable. She calls this ‘hedonic rationality’.

What Is Evaluable for Fit?  93 to do so. Such explanation would be sufficient to justify her guilt given that guilt is a fitting reaction to the fact that she committed the wrong in question. She need not, in defending her guilt, also argue that guilt is a fitting response to the fact that she committed a wrong. It is easy to see this in the case of belief. In defending her belief that p, a person need not argue that belief is a fitting response to the fact that p is credible; she needs only to defend her judgement that p is credible. Similarly, in defending her nausea in response to the wrong she committed, a person need not defend the claim that her nausea is fitting in response to the wrong. Rather, she is expected to offer the same reasoning she would offer in defence of her guilt. The fact that she committed a horrible wrong—­e.g. that she had ruined someone’s life—­is the reason for her guilt as well as her nausea. In light of the above, I think it is difficult to make a case only on the basis of a survey of our actual evaluative practices that, as the common view holds, our evaluative practices treat all bodily episodes and sensations as non-­rational. However, there are theoretical assumptions that seem to provide strong reason to accept that sensations and bodily episodes must be non-­rational. These assumptions might also explain why philosophers have generally interpreted our evalu­ ative practices as drawing a clear line between attitudes, on the one hand, and sensations, feelings, and bodily episodes, on the other.

4.5  A theoretical presupposition of the common view The idea that sensations, feelings, and bodily episodes can be rationally evaluable might seem implausible to anyone who assumes that such items lack intentionality and that intentionality is a necessary condition for rational evaluability. Thus, Smith writes: [P]art of the reason that it would make no sense to demand justification in the case of sensations and other nonintentional mental states is precisely because they are not directed upon an object or state of affairs, and hence the idea of “getting it wrong” or “being justified” in the experiencing of the state does not really have application. Directedness upon an object, or intentionality, then, seems to be a necessary condition of direct responsibility in the sense I am trying to capture.  (Smith 2005: 258)

Plausibly, for an item to be rationally evaluable it must have intentionality—­it must be about something.11 Your fear is about the danger the bear poses and can 11  However, actions seem to pose a counterexample. While an intention to eat dinner is about eating dinner, eating dinner does not seem to be about anything and therefore lacks intentionality. But since eating dinner is rationally evaluable, intentionality cannot be a necessary condition for rational evaluability. However, if the reason for eating dinner is the reason to intend to eat dinner—­as McHugh

94 Fittingness therefore be rationally evaluable; since your racing heart is not about anything it cannot be rationally evaluable. I believe this thought underlies much of the discussion of rationally evaluable attitudes and I wish to consider it more carefully. To begin, if intentionality is indeed a necessary feature of rationally evaluable items, then what is distinctive of rationally evaluable items is not that we can demand and provide reasons for them, but that they are fit-­evaluable. Let me explain. Fit-­related reasons are reasons that count in favour of an item being fitting to its object: fit-­related reasons for a belief in a proposition count in favour of the proposition being credible, fit-­related reasons for an intention to φ count in favour of φ -ing being worth doing. Therefore, fit-­related reasons presuppose that the item they support or oppose has intentionality: the item is about something with respect to which it can succeed or fail to be fitting. Many believe we can have value-­related reasons (or ‘wrong-­kind reasons’) for fit-­evaluable attitudes such as belief, intention, amusement, shame, regret, etc.12 These value-­related reasons do not bear on the fittingness of an attitude but on the value of having it. If believing that I will do well on the exam improves my chances of doing well, then I have reason to intend so; the value of having the belief explains my value-­related reason. If intending to drink poison will win me a great sum of money, then I have reason to intend so; the value of intending to drink the poison explains my value-­related reason. Now note that it can also be valuable to digest, to experience a pleasant sensation on one’s skin, or to have fever in response to infection, although such states are not fit-­evaluable. But if we can have reasons to be in valuable conditions independently of fit, it would seem that we can have reasons to be in valuable conditions that are not fit-­ evaluable. So alongside value-­ related reasons to have fit-­evaluable items (e.g. belief), we can have value-­related reasons to have items that are not fit-­evaluable (e.g. fever). To block this implication, we would need an explanation of why the arguments that support value-related reasons for fit-­ evaluable items do not work in the case of items that are not fit-­evaluable. If, as Smith and many others maintain, rationally evaluable items are necessarily evaluable as justified with respect to their objects, then rationally evaluable items are necessarily fit-­evaluable, because for an item to be fit-­evaluable is for it to be evaluable as justified with respect to what it is about. It follows that the fact that an item is such that we can demand or offer reasons for it does not determine whether the item is rationally evaluable or non-­rational. Since we can have value-­related reasons for items that lack intentionality, we can have reasons for and Way (forthcoming) argue and Smith seems to hold—­then actions are not themselves rationally evaluable except as expressions of intentions. Nevertheless, the theory of intentionality and fit-­ evaluability that I propose in the next section can account, I believe, for the intentionality of actions. 12  Examples of defenders of value-related reasons: D’Arms and Jacobson (2000); Rabinowicz and Rønnow-­Rasmussen (2004); Danielsson and Olson (2007); Rosen (2015); Howard (2016, 2019); Leary (2017).

What Is Evaluable for Fit?  95 non-­rational items. Therefore, what is crucial for rationally evaluable items is not that we can demand and offer reasons for them, but that they are fit-­evaluable. To be sure, there are those who are wrong-­kind-­reasons sceptics: they deny that there are normative value-­related reasons.13 For them, all reasons are fit-­ related reasons, so there is no problem in appealing to reasons in order to distinguish rationally evaluable items. I do not wish to take a stand on whether there are value-­related reasons. For my purposes, it is only important to remember that rationally evaluable items are necessarily fit-­evaluable and therefore responsive to fit-­related reasons, whether fit-­related reasons exhaust the space of reasons or not. In light of the claim that rationally evaluable items are fit-­evaluable, Smith’s objection to the apparent rational evaluability of sensations and bodily episodes might be reconstructed thus: The argument from fit-­evaluability: 1. Rationally evaluable items are necessarily fit-­evaluable. In other words, if an item is not fit-­evaluable, then it is non-­rational. 2. For an item to be fit-­evaluable it must have intentionality, i.e. it must be directed to, or be about an object, broadly construed. 3. But ‘sensations and other non-­intentional mental states’ lack intentionality, they are not directed to and are not about anything. Conclusion: Sensations and other non-­intentional mental states are non-­rational. I do not take issue with (1) and (2).14 My issue is with (3). Smith’s argument targets sensations—­which she takes to be non-­intentional—­and non-­intentional mental states. Insofar as she is only concerned with non-­intentional phenomena, I have no disagreement with the argument. But the question is whether all bodily episodes and sensations are non-­intentional and whether the mental states she considers to be non-­intentional are in fact so. Smith believes that the nausea she feels before having to speak in public and the butterflies she feels in her stomach before boarding a roller-­coaster are non-­ intentional—­that is, they are not about anything, they lack representational content. In particular, her nausea is unlike her fear and hope, which are directed towards the danger that the public performance poses and the promise it holds.

13  Examples of sceptics about value-­related reasons: Hieronymi (2005); Skorupski (2010); Parfit (2011); Way (2012). 14  Döring (2008) argues for (2). In defending the claim that hedonic tone (i.e. pleasantness or unpleasantness) is rationally evaluable, Jenifer Corns argues that either hedonic tone is reducible to something representational or some rationally evaluable mental phenomena are non-­representational (Corns 2019: 244–245). The latter disjunct is a rejection of (2) above. The suggestion I go on to make is along the lines of Corns’s former disjunct, but there is a crucial difference. I do not think that the intentionality of a mental or physical phenomenon must be explained by the kind of phenomenon it is or by the kind of phenomena it is reducible to.

96 Fittingness Fear and hope are therefore fit-­evaluable while nausea and butterflies in one’s stomach are not. But why does Smith assume this? Why does she take for granted that her nausea or the butterflies in her stomach are not about the objects of her fear and hope? One possible thought is that nausea is often non-­rational; it often occurs without being about anything at all. It might therefore seem that nausea is not the kind of state that has intentionality. Similarly, since headaches can occur without being about anything, one might conclude that headaches that are associated with grief do not share the intentionality of grief. And, finally, since your heartrate can accelerate merely due to the fact that you were running to catch the bus, your accelerated heartrate during fear might seem to lack fear’s intentionality. The underlying assumption is that if a given item has intentionality, it must have it due to the type of item it is. Belief and intention are essentially intentional, they are types of attitudes that have representational content. Since nausea (like headaches and racing hearts) can lack intentionality, it is not the type of item that has intentionality, so any instance of nausea must also lack intentionality. The idea that intentionality is determined by the type of mental or physical phenomenon under consideration leads to the following inference: Intentionality by type If there are tokens of phenomenon of type m that lack intentionality, then there are no tokens of m that have intentionality. I propose that our evaluative practices give us at least prima facie reason to doubt Intentionality by type and the assumptions that underlie it. One and the same type of item can be fit-­evaluable on one occasion but not on another, where in the first instance it has intentionality and in the second it does not. In section  4.6 I propose a possible explanation for this phenomenon.

4.6  A narrative account of fit-­evaluability I have argued that rationally evaluable items are fit-­evaluable and that to be fit-­ evaluable they must have intentionality. I have also argued that we treat sensations and bodily episodes as sometimes fit-­evaluable, sometimes not. The question I wish to consider in this section is how can the same type of sensation or bodily episode be fit-­evaluable (and therefore have intentionality) on one occasion and not fit-­evaluable (because lacking intentionality) on another? My aim is to find a theory that can answer this question. The theory I will adopt and elab­or­ate on cannot be fully worked out within the confines of this chapter and the task of defending it will also have to wait for another occasion. However, the fact that our evaluative practices suggest it is a reason in its favour.

What Is Evaluable for Fit?  97 Let us return to the examples of heartrates. What might explain the fact that your accelerated heartrate after running to catch the bus is not fit-­evaluable but the same heartrate while facing the bear is? And how can the same heartrate be about nothing at all in one case and about the danger you are facing in the other? The clue to answering these questions should be clear by now: when you are fa­cing the bear your heartrate is associated with your fear and both seem to share the same intentional object. It is implausible that both your heartrate and your fear just happen to be about the same object; it is more likely that they have the same object because they are systematically related. A theory that answers our question would (1) describe a single element that explains, at once, the object of your heartrate and the object of your fear in BEAR ENCOUNTER, and (2) claim that a similar kind of explanation is lacking in RUNNING TO THE BUS. A natural view of the systematic relation between accelerated heartrate and fear is that the former is an element of the latter. Consider, for example, Peter Goldie’s description of the complexity of emotions: An emotion is complex in that it will typically involve many different elements: it involves episodes of emotional experience, including perceptions, thoughts, and feelings of various kinds, and bodily changes of various kinds; and it involves dispositions, including dispositions to experience further emotional episodes, to have further thoughts and feelings, and to behave in certain ways. (Goldie 2000: 12–13)

To be sure, Goldie is here describing a richer and more enduring phenomenon than a short-­lived emotional episode, such as the fear in BEAR ENCOUNTER. However, as Darwin’s depiction of fear makes clear, short lived emotional episodes are often quite complex and involve various bodily occurrences and sensations. Thus, whether as a short-­lived reaction or as a more enduring state that involves patterns of sensations, imaginings, thoughts, and motivations, fear is a complex phenomenon that, arguably, includes a person’s accelerated heartrate as one element. This does not yet answer the question of intentionality and fit-­evaluability. While many theorists of emotion allow that emotions are complex, most deny that all the ingredients of an emotion share its intentional object. Rather, many assume that emotions must have some essential ingredient that is itself intentional and thus explains the intentionality and fit-­evaluability of the emotion. To name two leading families of view, judgemental theories of emotions assimilate emotions to evaluative or normative beliefs or judgements,15 while perceptual theories construe emotions as perceptual experiences of evaluative properties.16

15  For example: Solomon (1976); Greenspan (1988); Nussbaum (2001). 16  For example: Johnston (2001); Döring (2007); Tye (2008); Tappolet (2016).

98 Fittingness On many of these views, the bodily episodes that are elements of fear lack intentionality, but fear has intentionality because it is, at its core, a judgement or a perceptual experience of danger. These prominent views would deny that an accelerated heartrate is fit-­evaluable when it is a component of fear. Other views, however, take a more holistic approach. Deonna and Teroni, for example, argue that emotions are ‘distinctive types of bodily awareness, where the subject experiences her body holistically as taking an attitude towards a certain object’ and that we should ‘move away from the curiously atomistic approach to bodily sensations implicit in many accounts of their role in emotions and recognize that, in emotions, these sensations are typically aspects of a whole pattern that constitutes a world-­directed attitude’ (Deonna and Teroni 2012: 79). Deonna and Teroni elucidate the relevant patterns of bodily sensations in terms of action-­ readiness. Thus, ‘fear of the dog is an experience of the dog as dangerous, precisely because it consists in feeling the body’s readiness to act so as to diminish the dog’s likely impact on it (flight, preemptive attack, etc.), and this felt attitude is correct if and only if the dog is dangerous’ (81). On this view, one’s racing heart is fit-­evaluable when it is part of fear’s pattern of action-­readiness. Deonna and Teroni’s view is compatible with a view of emotions that appeals to narrative (83). According to Goldie, the different elements of a given emotion are structured as a recognizable emotion-­type by a narrative in which they are embedded (Goldie 2000: 13). This suggestion follows Ronald de Sousa’s idea that narratives, and specifically ‘paradigm scenarios’, define the character of our emotions: We are made familiar with the vocabulary of emotion by association with paradigm scenarios. These are drawn first from our daily life as small children and later reinforced by the stories, art, and culture to which we are exposed. Later still, in literate cultures, they are supplemented and refined by literature. Paradigm scenarios involve two aspects: first, a situation type providing the characteristic objects of the specific emotion-­type . . . and second, a set of characteristic or “normal” responses to the situation, where normality is first a bio­ logic­al matter and then very quickly becomes a cultural one. (de Sousa 1987: 182, emphasis in original)

According to de Sousa, emotion types are patterns that involve bodily episodes, affects, sensations, thoughts, and motivations, and are determined by paradigm scenarios or narratives. We draw on these familiar narratives to interpret ­situ­ations we face and how we interpret these situations explains our emotional reactions to them. Drawing on this picture, my proposal is that emotion-­patterns inherit their fittingness from the fittingness of the narratives in which they are embedded. To

What Is Evaluable for Fit?  99 unpack this idea, I will first explain what I take fitting narratives to be and then explain how they determine the fittingness of emotions. Very roughly, and without getting into various controversies about the nature of narrative, we can characterize narrative as a representation of a series of events and of the people involved in them, delivered from a certain perspective or perspectives. Moreover, narratives attribute to the events they depict a certain coherence and meaning, as well as evaluative and emotional import. Employing this brief characterization, we can identify ways in which narratives can succeed or fail to be fitting. To begin, note that narratives are representations of sequences or processes. In  general, the events and things depicted in a narrative are not themselves a narrative. So a life-­narrative is a narrative of a life, not a life that is a narrative; a self-­narrative is a narrative of a self not a self that is a narrative (Goldie 2012: 153–154). Since narratives are representations they might also misrepresent. Narratives can misrepresent in various ways: they might distort facts and causal connections, fail to note relevant information, etc. I will call such misrepresentations factual misrepresentations. The first way in which narratives can fail to be fitting is by including factual misrepresentations. There is, however, an important caveat with regard to factual misrepresentation in narrative. Some narratives do not purport (and are not expected) to faithfully represent things as they are, e.g. fictional narratives. The representation of facts in such narratives is not true, but it is not a misrepresentation either. Therefore, fi ­ ctional narratives cannot fail to be fitting due to factual misrepresentation. Another way in which narratives can misrepresent concerns the reactions they elicit or invite. Narratives are typically engaging—­they engage the emotions and evaluative judgements of the audience. Thus, a narrative can misrepresent by elicit­ing or calling upon emotional and judgemental reactions that are not fitting to the events it depicts. For example, a narrative can falsely present an action as shameful, thereby eliciting the unfitting judgement that it is shameful, or it might falsely present an action as contemptible, thereby eliciting unfitting contempt for it. I will call such misrepresentations: emotional and evaluative misrepresentations. The second way in which narratives can fail to be fitting is by including emotional and evaluative misrepresentations. Emotional and evaluative misrepresentations are possible even in fictional narratives, which do not purport to depict factual truth. For example, a fictional narrative about a serial killer might elicit an unfitting reaction to the violence it depicts if it elicits, e.g. admiration for the killer. Of course, whether such reaction is indeed unfitting is debatable, but the fact it might be unfitting is sufficient to show that fictional narratives can, in principle, misrepresent in this way.

100 Fittingness A narrative that includes no factual, emotional, or evaluative misrepresentations is a completely fitting narrative.17 Narratives can be more or less fitting given the degree of accuracy and the quality of their representation.18 Drawing on the narrative view of emotion described above, I propose that bodily episodes and sensations can have intentional objects and be fit-­evaluable when and because they are explained as elements of fit-­evaluable emotional reactions by the agent’s fit-­evaluable narrative representation of the situation. This dense formulation is meant to invoke three explanatory connections. First, the fact that the agent represents the situation by a fit-­evaluable narrative explains the occurrence of the agent’s bodily episodes and sensations. Second, these bodily episodes and sensations have intentionality and are fit-­evaluable because the narrative that explains their occurrence portrays the reaction pattern they are part of as fit-­evaluable. Finally, the fittingness of the agent’s narrative explains the fittingness of the agent’s bodily episodes and sensations.19 Thus, the proposal is that my fit-­evaluable narrative in BEAR ENCOUNTER explains the occurrence of my fast heartrate, my fast heartrate is about something because the narrative that explains it portrays the fear of which it is part as being about something, and my heartrate is fitting when the narrative that explains its occurrence as part of fear is fitting. Suppose you take a pill that makes your heart accelerate and then you encounter a bear. As you face the bear your heart races but this bodily episode is independent of your understanding of the situation. In such a case, your accelerated heartrate is non-­rational even if you actually fear the bear. This is explained by the fact that your racing heart is not explained by your narrative representation of the situation. In BEAR ENCOUNTER, all that is mentioned is that you encounter a bear and that your heartrate accelerates, so we assume that your heart races because of how you represent the situation to yourself—­that is, as having a certain meaning and significance, relating your recent past (setting up camp in the forest) to the present moment (facing the enormous bear) and your immediate future (dying or surviving the bear encounter). The narrative understanding we at­tri­ bute to you leads us to think that your heartrate is an element of your ­fit-­evaluable fear. To illustrate the role of fit according to my proposal, let me introduce another example. Stuck in traffic on my way home from work, I scratch my chin. A fellow driver mistakes my hand movement for an insult and responds with anger. I cannot hear him but I see his facial expressions and hand movements through the

17  As noted above, a fictional narrative does not include any factual misrepresentations because it does not purport to depict facts at all. So a fictional narrative can be perfectly fitting. 18  Although I acknowledge that whether fittingness is gradable is a controversial issue. See Berker in this volume. 19  My proposal is in line with Neta’s (2018): ‘A series of events or states in the agent can amount to the agent’s being committed to something only by virtue of the agent’s representing those very same events or states as appropriately responsive to, or expressive of, that commitment’ (298).

What Is Evaluable for Fit?  101 windshield. If I didn’t know better, I might think he is having some kind of ­seiz­ure. However, given my understanding of the story he must be telling himself, his frantic movements strike me as elements of anger. Thus, the fact that he represents the situation as he does explains his sensations and bodily episodes as elem­ ents of anger. Moreover, his bodily episodes and sensations have intentionality and are fit-­evaluable because they are explained as such by his fit-­evaluable narrative, according to which I intentionally insulted him. However, his fit-­evaluable narrative is not fitting: it misrepresents my intention and the evaluative significance of my hand movement. Since his narrative is not fitting, the anger explained by his narrative—­which includes his racing heart, his facial expressions, his hand movements, and all the rest—­is not fitting either. There is, however, also the possibility that I am mistaken in my interpretation of the situation. Maybe the driver is not angry at all but is, in fact, having a seiz­ ure. In this case, the driver’s bodily episodes are non-­rational because they are not explained by his fit-­evaluable narrative representation of the situation. In fact, my own narrative representation of the situation is unfitting. Let me briefly conclude. I claimed that sometimes bodily episodes and sensations seem to have the characteristics of fit-­evaluable items. I also suggested that suspicion of this initial impression might be explained by the assumption that intentionality, which is necessary for fit-­ evaluability, is type dependent. For instance, fear, belief, and desire are types of items that have intentional objects and therefore they satisfy a necessary condition for fit-­evaluability. Since bodily episodes and sensations are not types of items that have intentionality, no bodily episode or sensation has intentionality and therefore none is fit-­ evaluable. I responded to this line of reasoning by rejecting the assumption that intentionality and fit-­evaluability are type dependent and developing a preliminary theory that explains why bodily episodes and sensations are sometimes fit-­evaluable. The theory is that bodily episodes and sensations can have intentional objects and be fit-­evaluable when and because they are explained as elements of fit-­evaluable emotional reactions by the agent’s fit-­evaluable narrative representation. Thus, I  believe that narratives should play a more central role in our philosophical ­the­or­iz­ing about fit and practical reason more generally.

References Boyle, M. (2011) ‘Making up your mind’ and the activity of reason. Philosophers’ Imprint 11/17. Brady, M. (2018) The rationality of emotional and physical suffering. In Philosophy of Pain: Unpleasantness, Emotion, and Deviance, ed. D. Bain, M. Brady, and J. Corns. London: Routledge. 81–94. Broad, C. D. (1930) Five Types of Ethical Theory. London: Routledge.

102 Fittingness Cholbi, M. (2017) Grief ’s rationality, backward and forward. Philosophy and Phenomenological Research 94, 255–272. Corns, J. (2019) Hedonic rationality. In Philosophy of Suffering ed. D. Bain, M. Brady, and J. Corns. London: Routledge. 229–250. Danielsson, S. and J.  Olson. (2007) Brentano and the buck-passers. Mind  116, 511–522. D’Arms, J. and D. Jacobson. (2000) Sentiment and value. Ethics 110, 722–748. Darwin, C. (1872/2009) The Expression of the Emotions in Man and Animals. Intro and commentary by P. Ekman. London: Harper Perennial. Deonna, J. and F. Teroni. (2012) The Emotions: A Philosophical Introduction. London: Routledge. Döring, S. (2007) Seeing what to do: Affective perception and rational motivation. Dialectica 61, 363–394. Döring, S. (2008) Conflict without contradiction. In Epistemology and Emotions, ed. G. Brun, U. Kuenzle, and D. Dogouglu. Farnham: Ashgate. 83–104. Gibbard, A. (1990) Wise Choices, Apt Feelings. Cambridge, MA: Harvard University Press. Goldie, P. (2000) The Emotions: A Philosophical Exploration. Oxford: Oxford University Press. Goldie, P. (2012) The Mess Inside: Narrative, Emotion, and the Mind. Oxford: Oxford University Press. Greenspan, P. (1988) Emotions and Reasons. London: Routledge. Gustafson, D. (1989) Grief. Noûs 23, 457–479. Hedden, B. (2015) Reasons without Persons. Oxford: Oxford University Press. Hieronymi, P. (2005) The wrong kind of reason. The Journal of Philosophy 102, 437–457. Hieronymi, P. (2009) Two kinds of agency. In Mental Actions, ed. L.  O’Brien and M. Soteriou. Oxford: Oxford University Press. 138–161. Hieronymi, P. (2014) Reflection and responsibility. Philosophy and Public Affairs 42, 3–41. Howard, C. (2016) In defense of the wrong kind of reason. Thought 5, 53–62. Howard, C. (2018) Fittingness. Philosophy Compass 13, e12542. Howard, C. (2019) The fundamentality of fit. Oxford Studies in Metaethics 14, 216–236. Johnston, M. (2001) The authority of affect. Philosophy and Phenomenological Research 63, 181–214. Jollimore, T. (2004) Meaningless happiness and meaningful suffering. The Southern Journal of Philosophy 42, 333–347. Leary, S. (2017) In defense of practical reasons for belief. Australasian Journal of Philosophy 95, 529–542.

What Is Evaluable for Fit?  103 Marušić, B. (2018) Do reasons expire? An essay on grief. Philosophers’ Imprint 18, 1–22. Marušić, B. (forthcoming) On the Temporality of Emotions: An Essay on Grief, Anger, and Love. Oxford: Oxford University Press. McHugh, C. (2017) Attitudinal control. Synthese 194, 2745–2762. McHugh, C. and J.  Way. (forthcoming) All reasons are fundamentally for attitudes. Journal of Ethics and Social Philosophy. Moller, D. (2007) Love and death. Journal of Philosophy 104, 301–316. Moller, D. (forthcoming) Love and the rationality of grief. In The Oxford Handbook of Philosophy of Love, ed. C. Grau and A. Smuts. Oxford: Oxford University Press. DOI: 10.1093/oxfordhb/9780199395729.013.35 Moran, R. (2001) Authority and Estrangement. Princeton, NJ: Princeton University Press. Moran, R. (2017) Frankfurt on identification. In his The Philosophical Imagination: Selected Essays. Oxford: Oxford University Press. 136–157. Na’aman, O. (2021) The rationality of emotional change: Towards a process view. Noûs 55, 245–269. Neta, R. (2018) Rationally determinable conditions. Philosophical Issues 28, 289–299. Nolfi, K. (2015) Which mental states are rationally evaluable, and why? Philosophical Issues 25, 41–63. Nussbaum, M. (2001) Upheavals of Thought. Cambridge: Cambridge University Press. Parfit, D. (2011). On What Matters, vol. 1. Oxford: Oxford University Press. Portmore, D. (2011) Commonsense Consequentialism. Oxford: Oxford University Press. Rabinowicz, W. and T. Rønnow-Rasmussen. (2004) The strike of the demon: On fitting pro-attitudes and value. Ethics 114, 391–423. Rosen, G. (2015). Real definition. Analytic Philosophy 56, 189–209. Ross, W. D. (1939) Foundations of Ethics. Oxford: Clarendon Press. Scanlon, T.  M. (1998) What We Owe to Each Other. Cambridge, MA: Harvard University Press. Schönherr, J. (forthcoming) Two problems of fitting grief. Analysis. Skorupski, J. (2010). The Domain of Reasons. Oxford: Oxford University Press. Smith, A. M. (2005) Responsibility for attitudes: Activity and passivity in mental life. Ethics 155, 236–271. Smith, M. (2013) The ideal of orthonomous agency, or the how and why of buckpassing. In Thinking about Reasons, ed. D. Bakhurst et al. Oxford: Oxford University Press. 50–75. Solomon, R. C. (1976) The Passions. Indianapolis, IN: Hackett Publishing. de Sousa, R. (1987) The Rationality of Emotions. Cambridge, MA: MIT Press.

104 Fittingness Tappolet, C. (2016) Emotions, Values, and Agency. Oxford: Oxford University Press. Tye, M. (2008) The experience of emotion: An intentionalist theory. Revue Internationale de Philosophie 243, 25–50. Way, J. (2012). Transmission and the wrong kind of reason. Ethics 122, 489–515. Wilkinson, S. (2000) Is ‘normal grief ’ a mental disorder? Philosophical Quarterly 50, 289–304.

5 Fitting Emotions Justin D’Arms

5.1  Introduction Emotional responses to the world are assessed on various different grounds. Prudential assessments concern whether a response is beneficial or harmful to have. Moral or ethical assessments concern whether emotional responses are morally good or bad to have, or whether they are something an admirable or vicious person would feel. Another familiar kind of assessment, on which I focus here, asks whether an emotion is appropriate in one specific sense: whether it is fitting to its object, or, equivalently, whether the object merits this response. While those terms may be abstruse, the sort of assessment I am talking about is an everyday matter. When we think that the boss’s anger at the messenger is misplaced because he is not responsible for the bad news he brings, or decide that Woodrow Wilson was not worthy of the admiration we formerly had for him, or believe that a student is overly embarrassed by what was really a quite minor error, we assess the reactions as unfitting. This assessment of fittingness (or ‘fit’) is different from assessments of prudence and virtue. Indeed, they sometimes conflict. One can sensibly think an emotion is fitting while thinking it harmful or less than virtuous. For instance, a sick child might have disgusting symptoms—­symptoms that merit disgust—­even if it would be bad for the parent to feel disgusted and even if a good parent wouldn’t, perhaps because he would be so focused on caring for his child. In such a case disgust is appropriate in one sense—­it gets something right—­even if it would be better not to be disgusted. The possibility of competition between fittingness, prudence, and virtue helps to illustrate their differences. But even when an emotion is fitting, beneficial, and in accordance with virtue, these remain distinct normative assessments.1 Assessments of an emotion as fitting or not presuppose that there is something the emotional episode is about, sometimes called its particular object, and then 1 These distinctions are drawn and defended in D’Arms and Jacobson (2000). In D’Arms and Jacobson (forthcoming) we argue that accepting the possibility of conflict between fitting and virtuous emotions is necessary in order to embrace a robust and plausible kind of pluralism about value alongside a realistic moral psychology of virtue. Justin D’Arms, Fitting Emotions In: Fittingness: Essays in the Philosophy of Normativity. Edited by: Christopher Howard and R. A. Rowland, Oxford University Press. © Justin D’Arms 2022. DOI: 10.1093/oso/9780192895882.003.0005

106 Fittingness ask whether the emotion is appropriate to that object. The particular object of emotion might literally be an object (a cherished vase, an intruder), but it need not be. It might be something much broader, such as an event or a state of affairs. Here I will ignore difficulties about how to identify the particular object of an emotion, and treat ordinary emotion ascriptions (anger at the messenger, embarrassment over the mistake) as tolerably accurate.2 Fittingness is a relation between a response and its particular object, in a context. I introduced the idea of emotional fittingness through examples rather than a definition for a reason. It is not easy to explain fittingness in other terms, and it may not be possible (Howard 2018). Several English terms sometimes express this assessment through different constructions: an emotion can be ‘fitting’, ‘appropriate’, or ‘correct’ to its object, and an object can ‘merit’, ‘call for’, or be ‘worthy of ’ the relevant emotion. But while these phrases have slightly different connotations, none seems more basic or fundamental; they don’t explain, analyse or provide a reduction of one another. Nevertheless, I will try to explicate emotional fittingness in ways that clarify the uses above and help to supply ground rules for assessments of fit. I start from the widely accepted supposition that there is a connection between the fittingness of various emotions and specific (dis)values—­that is, specific ways of being valuable or disvaluable. It is fitting to admire something if and only if that thing is admirable, fitting to be amused if and only if it is funny, fitting to be ashamed of it if and only if it is shameful, fitting to be afraid if and only if it is fearsome (or dangerous), and so on for a variety of emotions and corresponding values.3 (I’ll now call all of these ‘values’, eliding the ‘or disvalues’ for economy.) In brief, unfitting emotions fail to match the evaluative significance of their objects, whereas fitting ones involve such a match. I assume that the concept of fitting emotions is at least constrained by claims like these, linking various fitting emotions to various sentimental values.4 For now let’s leave it open precisely which values are co-­extensive with fitting emotions.

2  Though not unproblematic, the assumption that episodes of discrete emotions have intentional objects is very standard in philosophy. It is often said to be an or the essential difference between emotions and moods. See Scarantino and de Sousa (2018) for a survey. 3  In the philosophy of emotion, these values are sometimes referred to as the formal objects of the paired emotion kinds, a medieval way of saying that they are standards of correctness for those emotions (Deonna and Teroni 2012; Rossi and Tappolet 2018). Thus emotions are treated as having two kinds of intentionality: each emotional episode has a particular object, and each emotion type has a formal object against which the particular object of any given episode of that type must be assessed in order to determine its correctness. 4  This is roughly what Chris Howard calls the ‘Crucial Desideratum’ on any view about fittingness, in a discussion to which I am indebted. He says that it must come out true that ‘something’s being valuable is extensionally equivalent to its being fitting to value, and that parallel equivalences hold between more specific (dis)value properties and the fittingness of certain, correspondingly specific ways of (dis)valuing’ (2018: 2). I am not sure whether valuing in general, or every specific way of valuing, are always assessable for fit. My claims are specific to fitting emotions. But Howard’s examples of

Fitting Emotions  107 Another uncontroversial claim about emotional fittingness that is less-­often discussed will be useful as well, though it complicates the simple biconditionals above. Recall the student’s unfitting embarrassment over a minor mistake. In the normative sense of ‘embarrassing’, something is embarrassing if and only if it merits embarrassment. Does this mean that we must either say that minor mistakes are not embarrassing, or else reject the intuitive thought that a student’s great embarrassment was unfitting? It does not. Both emotions and sentimental values come in strengths or degrees, and the fittingness of any given emotional response requires some kind of match or proportionality between the strength of the response and the strength or degree of value possessed by its object. This is why the student’s great embarrassment is unfitting, even if the mistake was indeed a bit embarrassing. Minor mistakes can merit a little embarrassment, rather than a lot. In general, how ashamed, embarrassed, afraid (and so on) it is fitting to be towards something covaries with how shameful, embarrassing, or fearsome (and so on) the object is. An emotion must have the right ‘size’ for its object as well as the right ‘shape’, in order to be a fitting response.5 Let me briefly situate my topic in relation to some recent literature. One issue has been whether fittingness is basic, or whether it can be understood in terms of other normative notions such as reason, or ought (Chappell 2012; McHugh and Way 2016). I take no stand on that. Related questions concern the normative strength or significance of fittingness: are there (ever or always) reasons against unfitting emotions, or in favour of fitting ones (Maguire 2018; Howard 2018)? Is not feeling a response that would have been fitting had you felt it a violation of standards of fit (ever or always)—or are the only such violations errors of commission, like those with which I began (Fritz forthcoming; Whiting 2021)? I will help myself here to talk of fittingness being normative, of standards of fit, and of certain considerations being reasons of fit, without addressing all those interesting meta-­normative questions. Rather than explicating emotional fittingness by locating fit among other normative relations, I try to explicate it on the basis of the distinctive considerations that ground judgements of fit for emotional responses. Some recent literature suggests that there are or may be moral and prudential varieties of fittingness—­so that assessments of fittingness do not per se fix which kinds of normative grounds are relevant to assessing whether their target is fitting or not.6 But the assessments of fitting emotions that I am discussing are not about specific ways of valuing in the relevant passage are all emotions (as is common in this literature), so my claims apply to them. 5  See D’Arms and Jacobson (2000: 73). The idea that evaluative attitudes must be proportional to the value of the object in order for the attitude to be fitting is not much discussed in the recent normativity literature (but see Maguire 2018 and Berker, this volume). It was common in earlier discussions (e.g. Broad 1954: 209); and it will be important in a later argument. 6  Berker (this volume), Kauppinen (2014). Berker makes some compelling points about the fittingness relation in service of a rather radical conclusion: that fittingness is a distinct family of normative

108 Fittingness the prudential or moral fittingness of emotions—­whatever those might be. I don’t believe that this limits the significance of my topic unduly. The assessments that I am considering are the ones that have currency in ordinary thought, and that connect fitting emotions with values in the ways discussed above. I treat fittingness as an assessment of emotions on the basis of a particular sort of grounds, which (I will suggest) arise from the nature of the emotions themselves. What that means, and how to identify these grounds, will be discussed as we progress. Finally, most of the extant literature about fittingness theorizes it in ways that are supposed to apply to beliefs, intentions, and sometimes even to actions. My focus here is on the fittingness of emotions, and my claims and arguments will not do much to explicate fittingness in other domains. But what I will say has affinities with a family of proposals about fittingness in other contexts, according to which fittingness involves standards of correctness that are internal to the ­attitude or activity in question. The view I will defend allows that emotional ­fittingness is a matter of correctness, while denying that it is a matter of representational accuracy.

5.2  Sentimentalism and Response Dependence Although the biconditional link between fitting emotions and values is uncontroversial, the explanation for it is not. I will consider two alternatives, beginning with the sentimentalist explanation that I favour. Most basically, it holds that emotional responses are prior to and explain various evaluative concepts and the values that they are concepts of. Values explained in this way are response dependent: things count as shameful and funny because they are fitting objects of shame and amusement, not the other way around. The version of this view that Daniel Jacobson and I develop and defend elsewhere, rational sentimentalism, is not a fully general theory of value.7 It applies to a specific range of emotions (which I call natural emotions) and of values (sentimental values). The natural emotions are pan-­cultural psychological kinds whose nature and concerns are not settled by cultural construction or evaluative reflection but, to a considerable extent, given by human nature. Each natural emotion is a distinctive kind of motivational state involving a syndrome that includes action tendencies, relation, on a par with the evaluative and deontic families. Just as judgements using evaluative or deontic concepts can have what he calls different ‘flavours’ (moral, prudential, epistemic, aesthetic . . . ), so too, he suggests, might judgements from the fit family. That raises the question of what is the ‘flavour’ of fittingness associated with the canonical examples of fittingness in our biconditional. I think most of the literature about fittingness assumes that it has only that one flavour, as do I. If it doesn’t, then I don’t see how we could home in on fittingness by contrasting it with prudential and moral assessments of emotion, in the way that I tried to do at the beginning of this chapter. 7  This view is set forth most completely in our forthcoming book Rational Sentimentalism, where many of the claims advanced in the next few sections are explained and defended in greater detail.

Fitting Emotions  109 goals, feelings, and in some cases bodily changes and expressive behaviour. It is an empirical claim that there are any natural emotions, and a partly empirical question which of the emotion terms in common usage refer to such kinds.8 Though not every emotion term names a natural emotion, we argue that there are quite a few such kinds, including many of the paradigmatic emotion types. These include not only so-­called basic emotions such as anger, fear, and disgust, but cognitively complex emotions that few if any other animals experience, such as shame, guilt, envy, jealousy, pride, regret, and amusement. Sentimental values are values that are linked in particular to the natural emotions in fitting emotion-­ value biconditionals like those above. (Thus if there were no natural emotions, there would be no sentimental values in my sense.) The sentimental values with which the natural emotions are paired arise from and depend upon those emotions. But the nature of this dependence is complex. Something’s being shameful does not depend on anyone actually being ashamed of it. Rather, it depends on whether the emotions are fitting. That is because sentimental value concepts are not dispositional but normative; their function is to regulate emotional responses on the basis of reasons and their content reflects that function. Judgements about when they apply expresses standards for what things merit the particular response. What it takes for something to merit a response depends on the nature of that response. As I see it, emotions involve an engagement with their particular objects that treats the object as mattering in various distinctive, sui generis ways. And for it to be fitting to feel some emotion F towards some object (x) is for that response to be correct: that is, for x to matter in the distinctive way that one would treat it as mattering by having F towards it. To think that x has the corresponding sentimental value property F is to think that x does indeed matter in that very way. This is what I mean by claiming that sentimental values are response dependent. There is nevertheless a different sense in which one can say that whether a trait is shameful (etc.) depends not on shame but on facts about the world—­about the trait and its possessor and various relational truths about the social environment. For instance, if it is shameful how little Brentano you have read, that depends on various facts about you and your social role. It depends on things like how much Brentano you’ve read, and to what extent you hold yourself out as an expert on  ­fittingness and emotions, seeking and accepting various benefits on those grounds.9 Those are the kinds of facts that make it shameful (or not)—the 8  These empirical claims are quite controversial at the moment, thanks especially to the influential work of Lisa Feldman Barrett, much of which is summarized in Barrett (2017). See D’Arms (2022) for some relevant discussion. If I am right that there are such kinds, the claim that English terms like ‘fear’ and ‘shame’ refer to them is defensible partly on the basis of common usage and partly on the grounds that natural kinds are the most eligible referents for terms whose intension lies in their vicinity (Lewis 1984). 9  This is a place where some of the issues we are ignoring about the particular object of emotion make a difference to what is the most perspicuous way to frame claims about fit. Treating fittingness

110 Fittingness ­fit-­making facts. But the evaluative feature that these facts do or don’t make your acquaintance with Brentano have, shamefulness, depends on shame in the sense explained above. Being shameful just is being such as to merit the specific kind of disvaluing that is involved in being ashamed of something. This is a kind of value whose nature and significance depend upon a relationship to an emotion, which is a particular affective and motivational way in which things matter to people. The fact that most people reading this chapter will readily understand the relevance of the sorts of facts adduced above to determining shamefulness in this situ­ation (whatever they conclude about your case) is due not only to their being academic philosophers but also to their sharing this part of the human emotional repertoire with us. So far I have stated these claims about sentimental values primarily in conceptual terms, but according to rational sentimentalism the response dependence of sentimental values is not only conceptual; sentimental value properties are also response dependent. If instead sentimental value concepts referred to a property whose normative significance could be adequately accounted for without any appeal to emotions and the various contingent psychological concerns that they support and manifest, then the appeal to emotions would be otiose in giving an account of the values to which sentimental value concepts refer.10 So I agree with Selim Berker that ‘a metaethical view whose appeal to sentiments is itself explained in terms of something deeper is not one that is aptly described as sentimentalist’ (this volume, p. 10) But I deny that the sentimental values can be fully explained by ‘deeper’ facts having nothing to do with the sentiments—­facts that could be fully understood and the significance of which could be vindicated in ways that make no appeal (even tacitly) to contingencies of human emotional psychology. Jacobson and I argue against attempts to give such explanations in a number of cases. Perhaps some values can be fully explained in non-­sentimental terms, but one of the central claims of rational sentimentalism is that this is not so for the values tied to natural emotions.

simply as a relation between an emotion and its object requires building a lot into the object in order to get the right verdicts. For instance, suppose that your child, overhearing your confessions of shame but not fully understanding them, became ashamed of how little Brentano she had read. That would be unfitting—­she is five and has no reason to have read Brentano. So either the particular object of your shame includes a great deal about you and your circumstances that would not normally be cited in an emotion ascription, or the fittingness of shame (and other emotions) depends not just on its object but on various features of the context that are not part of the object but help to determine a response’s fittingness to it. I find the latter way of talking more perspicuous in some cases. This is why I will sometimes say in the text above that the fittingness of an emotion is a relation between the emotion and its object in a context. But I think one can reach the same verdicts about the fittingness of responses on either way of talking. 10 This is the view that Christine Tappolet defends, which she calls Sentimental Realism (Tappolet 2016).

Fitting Emotions  111

5.3  Cognitivism, Response Independence, and the Alethic View A sharp contrast to the sentimentalist explanation of the relation between ­emotions and value is found in the cognitivist tradition in the philosophy of emotion. According to cognitivism, emotions are or involve evaluative thoughts.11 While the nature and significance of the values appealed to in these thoughts is seldom discussed in detail, cognitivists characteristically assume that they are independent of the emotions in which the thoughts occur. For instance, Philippa Foot argued that Humean sentimentalism mistakes the order of explanation between emotional responses and value concepts. Contrary to Hume, ‘the explanation of the thought comes into the description of the feeling, not the other way round’ (Foot 1963: 76). Her example was pride. According to Foot, it is a conceptual truth that one could not feel pride ‘towards something one did not believe to be in some way splendid and in some way one’s own’ (1963: 76). And this is not simply a necessary condition on feeling pride, but one that is supposed to fully capture the value with which pride is concerned. Foot holds that emotions are at least partly constituted by evaluative beliefs, and that the content of these beliefs—­the values that they are about—­is entirely independent of these emotions. Otherwise, her cognitivist claim would not ground the objection to sentimentalism that Foot intends. While Foot is more explicit than other cognitivists in asserting that the evalu­ ative thoughts that are necessary for emotions concern response independent values, it is fair to say that this is a basic assumption within the cognitivist tradition. Many considerations support that attribution, perhaps the clearest of which is that cognitivists take distinct emotion types to have different ‘defining propositions’ or ‘constitutive thoughts’ which, they argue, are essential to distinguishing one emotion from another.12 Feelings, motives, patterns of attention, and other symptoms of emotions do not suffice to differentiate them (Nussbaum 2001; Roberts 2003). Let us call these other symptoms ‘noncognitive’, to make explicit their contrast with the putatively necessary thoughts, even though all of them are cognitive in a more capacious sense of that term. The crucial point is this: if the content of the constitutive thoughts that cognitivists attribute to emotions were 11  Cognitivist theories differ both on whether one must believe the defining proposition, or if some less demanding propositional attitude suffices; and on whether having the thought is just a necessary condition, or if it is also sufficient for being in the emotional state. Neo-­stoics Solomon 2003; Nussbaum 2001) identify emotions with evaluative judgements. Judgementalists (Foot 1963) hold that the putatively defining proposition must be believed, while quasi-­judgementalists (Greenspan 1988; Roberts 2003) allow for a weaker propositional attitude. While significant for some purposes, these differences will not concern us here. 12  For example: ‘It seems necessary to put the thought into the definition of the emotion itself. Otherwise, we seem to have no good way of making the requisite discriminations among emotion types’ (Nussbaum 2001: 30).

112 Fittingness granted to depend on noncognitive aspects of emotions, then the individuation the thoughts allegedly enable would actually depend on something already present in the responses. So even when it is not made explicit, this central argument for cognitivism depends on there being different response independent evaluative thoughts for each emotion type. Cognitivists don’t tend to talk about ‘fittingness’ or what ‘merits’ different emotions. Perhaps this is because, given cognitivist assumptions, there is an easy way to explain the sorts of intuitive assessments of emotional appropriateness that we are discussing in more mainstream philosophical vocabulary. The n ­ atural cognitivist proposal is that for an emotion to be appropriate is for the thoughts that are (partly) constitutive of it to be true. That is Gideon Rosen’s suggestion (2015: 71–2): Fear involves the thought that one is in danger, misery the thought that things are going badly, and so on. These emotions are appropriate in the intended sense only when the relevant thoughts are true, and this suggests an analysis: For an emotion to be appropriate just is for its ingredient thoughts to be true.

Rosen offers the above account as part of what he calls the Alethic View of blameworthiness. On this view, for someone to be blameworthy for something just is for it to be appropriate to resent them for it. And for it to be appropriate to resent them for it just is for the thoughts implicit in resentment to be true. Following this pattern, cognitivists can provide alethic views of all sorts of evaluative concepts linked to emotions. For x to have some evaluative property F is for the thoughts in F to be true of x. So, for instance, Foot can say that for something to be prideworthy just is for it to be ‘in some way splendid and in some way mine’. The combination of the alethic view with the traditional cognitivist idea that emotions involve evaluative thoughts whose content is response independent offers a stark alternative to sentimentalism. If to be prideworthy just is to be splendid and mine, and these are values whose nature and importance is completely independent of pride, then pride is otiose in an account of this value. It might be allowed that the concept ‘prideworthy’ is response dependent. But that concept would be just an optional and rather superficial way of thinking about things that are splendid and mine. Beings who didn’t share our emotional repertoire could ignore pride in their thinking and understand the significance of this value in all its detail by thinking about what was splendid and theirs.13

13  Rosen’s view seems to differ from conventional cognitivism in the respect that is crucial to the latter’s conflict with sentimentalism. He advertises his account as a response dependent theory of blameworthiness; and he is prepared to attribute his ‘implicit thoughts’ to resentment on the basis of facts about its motivational role. So he may not aspire to capture an emotion’s evaluative content in response independent terms. I will return to these points later.

Fitting Emotions  113 Fortunately for sentimentalism, there are many reasons to doubt the cognitivists’ claims. Various damning criticisms have called different aspects of the theory into doubt.14 In Rational Sentimentalism, Jacobson and I argue in addition that response independent cognitivist accounts of natural emotions inevitably fail to capture their true evaluative content because that content depends crucially on their affective and motivational nature. We consider, for instance, cognitivist attempts to explain the prideworthy and the fearsome in response independent terms at some length. Our conclusion is that either thoughts like ‘dangerous’ and ‘splendid and mine’ must be understood as tacitly response dependent, or else they fail to capture the important values associated with pride and fear. These are complex issues, and I won’t try to rehearse our arguments here. But I see them as crucial motivations for pursuing the sentimentalist alternative. Nevertheless, the alethic approach offers an appealing promise. It seems to provide a clear way of thinking about the fittingness of an emotion. Earlier I accepted ‘correct’ as one paraphrase of ‘fitting’, allowing that emotional fittingness is a matter of whether the emotion gets it right about the value of its object. The alethic view offers a familiar model of this correctness. Emotional correctness, on this view, is a matter of the truth of the beliefs that are constitutive of the emotion.15 Since sentimentalism rejects the cognitivist theory of emotion, it cannot avail itself of this simple story. This leaves some important questions for the sentimentalist. What could emotional correctness be, if not truth of some thought that is constitutive of the response? And, which considerations are relevant to the fittingness of different emotions? How are we to discriminate between the conditions under which, say, shame and guilt are fitting? If all that the sentimentalist can say is that for something to be shameful is for shame at it to be fitting, whereas for something to be guilt-­worthy is for it to merit guilt, the theory offers little help in thinking about sensible standards for the values it purports to explain.

5.4  Interpreting Emotions as Appraisals The sentimentalist should say more about how to assess emotions for fittingness. This will help to clarify what things are shameful, guilt-­worthy, and such, and in what sense emotions can be correct. And sentimentalists must do this without appealing to the cognitivist’s proposal that emotions involve constitutive thoughts about response independent evaluative matters. I want to address this challenge. But I also want to temper expectations. Particular questions about what is and is

14 See for instance Griffiths (1997), Deigh (1994), D’Arms and Jacobson (2003), and Scarantino (2010). 15  Rosen contrasts the alethic view with talk of fittingness, but I will treat it instead as a proposal about what fittingness amounts to, as Howard (2018) does.

114 Fittingness not shameful turn on a host of factual and evaluative issues that we should not expect a general account of fitting shame to settle for us. We should ask rather for a way of circumscribing considerations of fit for different emotions that helps explain differences between them and clarifies what is at issue between people who disagree over which attributes merit shame. I said earlier that natural emotions involve an engagement with their objects that treats those objects as mattering in various distinctive, sui generis ways, and that for it to be fitting to have an emotion towards an object (and thus for the object to have the relevant sentimental value property) is for that object to matter in that very way. To make progress on the challenge, I think one must flesh out this claim by looking at the different ways that natural emotions engage with their objects, and use that investigation as a basis for interpreting the different ways in which they treat those objects as mattering. Whatever can be said at the most general level about the conditions something must meet to be shameful, enviable, or even funny will derive from insights about the emotions. On the motivational theory of natural emotions that I favour, these emotions are distinct psychological kinds of motivational state.16 Natural emotions can therefore be investigated empirically, through attention to the circumstances in which they arise and resolve, and the experiences and behaviours of people having emotions. Such investigation yields various kinds of data. First, there are observable patterns of elicitation and satisfaction: things that tend to produce emotional responses and events that tend to resolve them, often with some feeling of relief or gratification. While merely distracting events tend to reduce or elim­in­ ate all sorts of emotional responses, most emotions also have more specific satisfaction conditions (for fear, the avoidance of the threat, for guilt, a reconciliation with an offended party). Secondly, most of these emotions have discrete and specific motivational roles. They have characteristic action tendencies: motivational energy that arises irruptively in bouts motivating one urgently and directly towards certain sorts of characteristic behaviours—­in ways that can conflict with evaluative judgement and practical reason. This is true not only of the most obvious examples, such as mo­tiv­ations towards fleeing or hiding in fear, or lashing out in anger. It applies as well to states like envy and guilt, which urgently motivate competitive and con­ cili­atory behaviours (respectively) in ways that may not reflect considered judgements about what to do. These action tendencies, combined with satisfaction conditions, often enable a plausible inference about a broader goal that is characteristic of the emotion type.

16  This theory is developed and defended in Rational Sentimentalism, Chapter 6. Frijda (1986) offers the seminal version of many of the central ideas. Scarantino (2014) shares many of the same commitments.

Fitting Emotions  115 Finally, the natural emotions have characteristic bodily, affective, and cognitive phenomenologies. Bodily feelings include subjective registration of expressive behaviours like blushing, as well as circulatory and muscular preparations for action. Affects include positive and negative feelings, varying arousal levels, and felt attraction or aversion. Cognitive phenomenology includes patterns of attention and experience of oneself as motivated in the aforementioned ways. Equipped with a suitably detailed account of what a given emotion is like, both functionally and as an experience, the next step is to try to interpret the way in which that kind of emotional engagement involves treating its object as mattering. Jacobson and I call this the emotion’s generic appraisal. An appraisal in this sense is not an element in the emotion, nor is it a causal precursor to emotion (as  it is in the Appraisal Theories of many psychologists).17 It is the emotional syndrome as a whole that amounts to an appraisal of its object. The interpretive question is how best to characterize the common way in which anyone in the throes of a given emotion takes its object to be good or bad. A good gloss will ring true to those acquainted with the emotion—­which in the case of natural emotions is almost everyone. In light of the sorts of things that elicit the emotion and that satisfy it, the ways that it motivates, and what it is like as an experience, the gloss will describe a recognizable way that anyone who is angry or ashamed is taking the object of their emotion to matter. The gloss thereby circumscribes the conditions under which that emotion is correct, by describing the way something needs to be in order for that particular kind of emotional engagement to fit its object. A good gloss will provide some insight into the nature of a given sentimental value. It will therefore not be expressed in terms like ‘shameful’ or ‘fearsome’, but in other terms that help to clarify these concepts. Unlike the cognitivist’s constitutive thoughts, however, the terms of a gloss are not intended to describe response independent concepts that one must deploy in order to count as being in this emotional state. Rather, they are rough and ready attempts to describe in language a distinctively emotional way of taking something to matter. Giving a gloss of an emotion’s appraisal is an interpretation of the emotional syndrome as a whole, in light of its nature, rather than a claim about some thought that is a conceptually necessary constituent part of it. To make all this less abstract, let me briefly apply the approach to the example of shame. I begin with a characterization of shame as a natural emotion which aims to capture platitudes rather than provide novel insights, and which is con­ son­ant with the philosophical and psychological literature as well as with my own experience and (I hope) yours.18 The paradigmatic elicitors of shame include weaknesses, whether physical or moral; deformities and signs of disease, including things as commonplace as conspicuous blemishes; the inability or failure to 17  See, for example, Scherer (2001), Roseman (1984), but compare Lazarus (1991). 18  See, for instance, Williams (1993), Tangney and Dearing (2002).

116 Fittingness follow social norms; acts of cowardice and other vices; and other manifestations of incompetence, such as the stupid things one said. Bouts of shame at any of these things are especially likely when they are being perceived by or attended to by others (especially people the subject respects) or when the prospect of such social attention is vivid. Shame’s action tendency—­what it motivates most directly and urgently—­ involves concealment of the object of shame. When ashamed of a visible blemish or scar, one attempts to hide it, even if that requires hiding oneself. When the feature deemed shameful is imperceptible, one takes steps to avoid having it come up in conversation. When concealment is impossible, the motivation can manifest itself in withdrawal and, when even that is impossible, in a wish to disappear. While the elimination of dispositions to shame at some object might best be achieved by ridding oneself of the offending trait, episodes are often diminished or resolved by successfully preventing others from seeing or attending to it. Phenomenologically shame is unpleasant, and focuses attention on the trait of which one is ashamed and on an audience before whom it is or might become exposed. (While shame in private is certainly possible, it often involves an im­agin­ ing an audience.) The forgoing are intended as purely descriptive claims about the most obvious and familiar objects and symptoms of shame, not yet as normative claims about when it is fitting. They are the data from which to build an interpretation of how shame appraises its object, and thus when it is fitting.19 In light of these data, including first-­personal experience that I expect you as a human reader to share, it seems obvious that shame takes something as bad in some way. More spe­cif­ic­ al­ly, shame appraises its object as something that reflects badly on me. While somewhat vague and avowedly rough and ready, this seems to capture a dis­tinct­ ive combination of personal assessment and concern with self-­presentation that accords with the data. Interpretive claims take empirical ones as their starting points, but their status and the grounds on which to debate them are different. If this gloss is good, it should ring true to those in the throes of shame—­even those who are ashamed of things that you or I think are not really shameful. People who disagree, for instance, over whether the behaviour of our dead ancestors can give us reasons for shame today should be able to accept this as a way of clarifying the question under dispute: does such behaviour reflect badly on subsequent generations? And people who feel shame at characteristics that they do not judge to be shameful can accept that this is an explanation of the tension in their attitudes: they are appraising this trait emotionally as something that reflects badly on them, while 19  Though one can hold that even some paradigmatic elicitors of an emotion do not merit it, the rejection of any widespread propensity as unfitting requires a compelling theory of error. And the more widespread and paradigmatic a given elicitor is, the less plausible it will be to interpret the emotion’s appraisal in a way that renders this elicitor mysterious.

Fitting Emotions  117 judging that in fact it does not. If these claims do not ring true, I am open to hearing a better way of characterizing the way in which shame takes something as bad. But hereafter I will presume that this gloss is a plausible interpretation of shame’s appraisal. The approach I have briefly sketched allows the sentimentalist to describe the conditions under which shame is fitting in a way that is substantive without begging questions over which people with different sensibilities disagree. Jacobson and I have used this approach to interpret the appraisals of various emotions, including guilt (D’Arms and Jacobson 2022). To see the difference between the fittingness of shame and guilt, consider the relevance of whether you have agential control over something to the question of whether these emotions are fitting. With respect to guilt, this question is obviously relevant. Guilt over some conspicuous inability that is congenital is clearly unfitting, because guilt focuses on one’s behaviour and takes that behaviour as something for which apology and perhaps restitution make sense. This warrants interpreting guilt as concerned with one’s own motives and actions: roughly, guilt appraises its object as a betrayal or wrongdoing. In contrast, shame does not generically present its object as something you have done. If the gloss of shame above is good, then in order to argue that your lack of control over your inability suffices to make your shame unfitting, one would need to argue that nothing over which you lack control can reflect badly on you. That is a coherent normative position to adopt, but it does not emerge as a norm from the nature of shame. Indeed, it would render unfitting many of the paradigms of shame, including shame over physical characteristics, displays of incompetence, and non-­culpable social failures of self-­presentation. An argument for such a control condition on shame might rest on something like the idea that you are just your will, so that only things that reflect badly on your will can reflect badly on you. Such ideas are familiar, but not very plausible. In short, a close look at what shame and guilt are like helps us see why an agential control restriction on the shameful requires a heroic argument, whereas such a restriction on the guilt-­worthy plausibly emerges from the nature of that response. And these points are secured not by the highly dubious claim that certain beliefs deploying response independent evaluative concepts are necessary parts of feeling ashamed or guilty, but by looking at what these emotions are like as motivational and affective systems, and then interpreting them as appraisals in light of those facts.

5.5  Interpretivism and Alethism The approach I’ve described offers a sentimentalism-­friendly way to understand generic standards of fittingness for a given emotion type. If emotional responses

118 Fittingness are well interpreted as ways of appraising their object, then we can think about whether they are fitting by considering the accuracy of those appraisals. Because the appraisal is avowedly an interpretation of the nature of an emotion as a whole, what is thereby assessed as accurate is the emotion, not some thought that is alleged to be a part of it. To better appreciate the strengths of this approach to emotional fittingness, it will be helpful to compare it with cognitivism and the alethic view. The most natural development of alethism conjoins it with cognitivism. I’ll call this combination ‘cognitivist alethism’. I see three significant problems with this view. There is a way of relaxing the alethic approach that borrows some of the ideas I have discussed and thereby avoids two of those problems. I’ll consider those first two problems and the modifications to alethism that they motivate here and save the last problem for the final section of this chapter. The first problem has already been introduced. That is cognitivist alethism’s commitment to the response independence of emotional evaluation. If the arguments against cognitivism in Rational Sentimentalism succeed, then the evalu­ ative content of emotional responses is response dependent, and efforts to describe them in other terms are either subject to counter-­example or tacitly response dependent. Relatedly, once we focus on natural emotions that are identifiable as psychological kinds, the cognitivist needs to provide evidence that the putatively defining thoughts really do occur in every episode of the relevant emotion type. But although people sometimes have specific evaluative thoughts in the throes of emotion, in other cases the entire evidence for the presence of these thoughts is the occurrence of other emotional symptoms. That is often true in cases of recalcitrant emotion, for instance: the self-­aware phobic tells us that he knows this kind of spider is not dangerous, but he can’t help being afraid of it anyway. The cognitivist must claim that the phobic’s fear involves a thought that is independent of his other emotional symptoms and conflicts with his sincere judgement, to the effect that the spider really is dangerous. But in many cases the only basis for attributing the thought of danger is the fact that he is afraid. The response I recommend to the alethist is to embrace what I’ll call interpretivism rather than cognitivism. Interpretivism grants that the claim that fear involves ‘implicitly’ thinking its object dangerous does not actually identify an independent element in fear, distinct from its noncognitive symptoms. Rather, what having this thought amounts to is simply being interpretable in this way, in light of the way you feel, the ways you are motivated, the things that elicit and satisfy this response, its distinctive patterns of attention, and so on. Since the basis for attributing the thought derives from these other aspects of the emotional response, the interpretivist alethist would grant as well that the content of this thought is at least tacitly response dependent. The emotion is a distinctive way of taking something to matter, which we then try to describe in terms of a thought so as to specify the conditions under which this way of taking things is correct or fitting. At this point, the alethist would be saying what I said about appraisals,

Fitting Emotions  119 except that he calls them implicit ‘thoughts’. I have no substantive objection to this view, of course. And some aspects of Rosen’s discussion indicate that he might accept it.20 But I think it is clearer to use a different term than ‘thought’ to describe emotional evaluation, in light of the long history of cognitivism in the theory of emotion which has a very different understanding of this claim. The second problem for cognitivist alethism concerns what to say about responses that are too great or too little for the value of their objects. Until now, we have been considering how to characterize the generic fittingness conditions of an emotion—­that is, the conditions that are common to all its instances. If we set aside worries about response dependence, the alethic cognitivist at least seems to have a good account of these generic fittingness conditions. Identify a general evaluative thought that all instances of the emotion share, and then say that instances of that emotion are fitting just in case that thought is true. Let’s suppose, for instance, that the thought constitutive of anger is that this person has committed an offence (again, bracketing all the worries above). The trouble is that even if this were true about anger, it would not be true that every episode of anger is fitting so long as it responds to a genuine offence. That’s because of the point noted earlier: how embarrassed, angry, or proud it is fitting to be towards something depends on how good or bad it is in the relevant way. Just as the student was too embarrassed at something that merited some embarrassment, you can be too angry, or not angry enough, at an offence. The mistake in such anger is that it doesn’t match the severity of the offence; this too is a way of being unfitting. Exactly what emotional strength amounts to is a complex issue that I will not try to analyse here. For now it must suffice to note that degrees of felt intensity, strength and direction of motivation, extent of attentional focus, as well as the duration of emotional symptoms, all intuitively contribute to how angry or ashamed someone is, and are commonly regarded that way. 20  Rosen (2015) characterizes his theory of blameworthiness in terms of appropriate resentment as response dependent, and defines appropriateness in terms of the truth of thoughts that he says are ‘implicit’ in resentment. His evidence for the attribution of these thoughts is partly non-­cognitive. For instance, he attributes to resentment the thought that its object deserves to suffer, primarily because the emotion involves a desire for a retributive response (2015: 82–3). That looks like interpretivism. However, he also has a tendency to refer to the thoughts that he says are implicit in emotions as ‘ingredients’, which suggests a different view. There are puzzles about what the alethist’s attributions of thought amount to. For instance, what should Rosen say about someone who resents somebody for a wrong, but explicitly and sincerely denies that he thinks they deserve to suffer? Plenty of people are committed to the theoretical position that no one deserves to suffer, and this does not immunize them to resentment. By accepting interpretivism, Rosen could say that these people count as thinking the person deserves to suffer in his sense, all the same: they are having an emotional thought at odds with their judgement. In part because this insistence on a conflicting thought that the person expressly denies sometimes sounds like fist-­pounding, I prefer giving up the word ‘thought’ here in favour of ‘appraisal’. But I would settle for the acknowledgement that there is nothing more to an implicit thought than an appraisal as I have explained that notion. What about the content of ‘deserves to suffer’? Does it invoke a sophisticated response independent evaluative concept, DESERT, that anyone capable of resentment must possess and deploy? Not on the interpretivist view. Instead, people who are angry at things done to them can be counted as ‘thinking’ their target deserves to suffer, simply because they take what he did as their reason to retaliate. Desert in this sense is response dependent.

120 Fittingness How do alethic cognitivists differentiate fitting anger from excessive anger, over something that is granted to be an offence? Since they think fittingness is a matter of truth of thoughts necessary to be in the emotional state, they must attribute different response independent thoughts corresponding to different amounts of severity. Perhaps the person who is too angry will be said to think that this is an outrage, whereas the one who is fittingly angry merely thinks it rude. But it is not plausible that what makes an emotion unfittingly strong or weak (too big or too small) can be located in response independent thoughts about the extent of offence, danger, and so on. Two people might agree entirely in the response independent thoughts that (according to cognitivist alethism) determine the fittingness of their emotions, and yet they might differ in the size of their emotional responses, in virtue of how strongly they feel and what they are motivated to do. For instance it seems pos­ sible for people to have the same thoughts about the seriousness of some transgression, but differ in how angry they are about it. One might be quite upset and motivated towards significant retaliation, while the other is not. Clarke and Rawlings (2022) make a similar argument against accounts of blameworthiness like Rosen’s. In some cases, people seem to blame to different extents without believing different things about the culprit or his action.21 Intuitively, such differences in the size of their emotional responses are assessable for fittingness. If ­people are angry or amused to different extents while sharing the same thoughts, then, absent some further explanation, at most one of their responses can be proportionate to how much anger or amusement is fitting. Cognitivist alethism can’t make good sense of this. Moreover, there is a lot of variation in the intensity of emotions. Yet we draw far fewer conceptual distinctions among degrees of sentimental value than there are degrees of emotional strength. For instance, our conceptual tools for cat­egor­ iz­ing different levels of disgustingness and funniness are pretty meager. Perhaps ‘gross’, ‘foul’, ‘revolting’, and ‘amusing’ and ‘hilarious’ are ascending differences of degree. But other words in these families don’t seem to distinguish other amounts

21  Consider this example: ‘Sue and Tim are neighbors. Both are awakened early Saturday by their mutual neighbor Al, who is clearing his driveway with his leaf blower. Sue is angry. She marches next door and gives Al what for. Tim is annoyed. He decides not to invite Al to the upcoming poker game he is hosting. Sue and Tim need not disagree, in the representations of Al that are involved in their respective states of blame, about the seriousness of the offense, the disrespect shown by Al’s action, his ability to have done better, or what he deserves. Though they need not represent Al differently, they blame him differently, perhaps because they have different temperaments’ (Clarke and Rawlings 2022: 9–10). But they allow that these differences in response might not be differences in amount of blame. So they pursue another argument: that the same person can blame someone less as time passes, without a change in any of the thoughts that (according to various alethic theories) determine the fittingness of blame. I think that some plausible versions of the Sue and Tim case are counterexamples to cognitivist alethism about blame, so the case against that view from size of blame need not rest on duration of responses. But whatever one says about blame, many emotions seem to differ in size without difference in response dependent thought.

Fitting Emotions  121 of (dis)value. Anger may be the cognitivist’s best case in this regard, because there are more distinguishable categories of offence. But it is not credible that every variation in strength of anger corresponds to independently identifiable and ­subtly different ingredient thoughts about the severity of wrongdoing or the degree of culpability for it, or anything like that.22 The cognitivist might suggest we throw in modifiers to deal with the paucity of distinct evaluative concepts: this person thinks the behaviour was very, very rude, whereas that one thinks it was almost outrageous. But it seems to me fanciful to insist that such cognitive differences are distinguishable as distinct elements in experiences of anger of varying intensity. A cognitivist alethist will eventually have to acknowledge some differences in strength of anger that do not involve differences in evaluative thought. And presumably he must then hold that there is no difference in fittingness in such cases. That seems to me an unwelcome result. If you and your friend have the same thoughts about how badly someone has treated you both, but you are angrier than she is, then at least one of your reactions is not entirely fitting—­it does not match the severity of the offence. Again, the best response is to abandon cognitivism for the interpretive approach. Allow that the severity of a given emotional appraisal (‘thought’ in the alethist’s language) is constituted by the strength of the emotional response—­as evidenced by the sorts of symptoms mentioned above. How serious you take an offence to be is not found in an additional representational element over and above how angry you are. It is simply because you are more angry than your friend that you count as taking the offence to be worse than she does. Assigning evaluative strength to the emotion as a whole allows us to discriminate between appraisals on grounds of size without supposing that these distinctions cor­res­ pond to independently determined differences in cognitive content. I think the alethist had better grant all that, and say that you count as ‘thinking’ that this infraction is worse than your friend thinks it is, simply in virtue of your emotional differences. This sort of alethist would embrace rational sentimentalism’s interpretive methodology.

5.6  Fittingness, Correctness, and Representation If the alethic view can accept the modifications suggested above, then the issues raised in the previous section are not fatal objections to the approach. But the third and final objection that I wish to consider targets the central claim of the

22  Even if you can convince yourself that there are as many articulable variations of this sort as there are degrees of anger (which I cannot), the attribution of these subtle differences is surely not something for which we have evidence independent of the phenomenological intensity, motivation, and duration of the anger.

122 Fittingness alethic approach to fittingness: that the fittingness of emotions is simply a matter of their being true—­that is, of their representational accuracy. Alethism equates the mistake involved in unfitting anger or shame with that of falsely thinking that something is an offence, or that it reflects badly on you. The objection is that these are different mistakes. Even if the same facts that make the thought false also render the emotion unfitting, what makes the emotion unfitting is that it involves a mistaken way of engaging affectively and motivationally with the object. Unfitting anger, for instance, involves being upset and agitated about the offence, and mo­tiv­ated in retaliatory ways against its perpetrator. Its unfittingness depends on these facts, not (or at least not merely) on its being a misrepresentation of what was done as an offence. Sigrún Svavarsdóttir (2014) levels this criticism against the alethic view. She argues that when we use ascriptions of value (or of lack of value) as grounds for criticizing evaluative attitudes, the substance of the criticism is that this object is not worth responding to in this way. The mistake, she says, ‘is not that of misrepresentation but, rather, that . . . [it] is some kind of misplacement or waste of emotional and motivational resources to train them on an object of little or no value’ (2014: 89). I agree that there is a difference between unfitting emotion and false thought. And I agree that an unfitting emotion involves a kind of misplacement or mis­dir­ ec­tion of the emotions characteristic motives and affect—­that is, a misplacement of what I called ‘emotional engagement’ with the object. Nevertheless, the question of whether it is worth investing emotional and motivational resources in something strikes me as different from the question of whether the object merits a particular sort of emotional engagement. This difference is easily missed, because it is natural to treat ‘x merits F’ as equivalent to ‘x is worthy of F’; and it is also natural to treat ‘x is worthy of F’ as equivalent to ‘it is worth having F towards x’. But these two equivalences rely on subtly different senses of ‘worthy of ’. Whether it is worth having an emotional response, in the sense that contrasts with it being a waste of emotional resources, seems to depend in part on the opportunity costs of response. These costs include missing out on assigning one’s attention and emotional energies elsewhere to things that matter (even) more. But whether something is admirable, funny, or shameful does not seem to me to depend on the opportunity costs of those responses. The presence of these individual evaluative properties hinges on specific aspects of the target of assessment, not on the comparative issue of where to invest one’s emotional resources.23 Because emotional fittingness is a relation between the response and its object, it 23  One reason for thinking this comes from the idea that there may be psychological limits on one’s capacity for a given emotional response which do not reflect limits on the number of worldly circumstances that can merit it. Another reason comes from the plausibility of a conflicted pluralism about sentimental values, according to which different and conflicting responses can be merited at the same time, even if providing some of those responses makes it difficult or impossible to provide others.

Fitting Emotions  123 does not take into account everything that bears on broad questions about how to apportion one’s emotional energies among the many things that merit them.24 I accept that whether a response is fitting admits of various contextual qualifications—­it is only fitting to be afraid of a dangerous thing when you are in a context in which it endangers you, for instance. But it does not depend on all the factors that bear on  whether to engage emotionally with the relevant value. Another ­difference between fit and worth is revealed by considering that assessments of emotional fittingness apply equally to positive and negative emotions. Something is shameful or guilt-­worthy only if it merits those responses. But surely the ­mistake involved in being ashamed of something that is not shameful is not that one has wasted one’s shame on it.25 I conclude that the question of what merits emotional responses is different from the question of what things it is worth having them towards.26 Although I don’t think Svavarsdóttir’s notion of worth can be recruited to explain emotional fittingness, her criticisms do seem to me to identify a problem with the alethic approach to emotional fittingness. An emotion’s fittingness seems different from the correctness of a thought. Moreover, defenders of sentimentalism and fitting attitude theories have an independent reason to accept this point. If an emotion or evaluative attitude being fitting were simply a matter of it being a true belief about value, then those theories would have less to recommend them. I  think that the biconditionals from which we started are only able to provide informative explications of various values if emotional fittingness is something more than correct thoughts about evaluative properties.27 And this would be true whether the thoughts in question had response dependent or response inde24  Svavarsdóttir’s proposals are not advanced as a theory of fitting emotions. Though she sometimes seems to intend them to capture a broad range of evaluative attitudes, she is primarily concerned with standards for an attitude she calls valuing that is distinct from and more dispositional than episodes of emotion. So it may be that what it is worth ‘valuing’ in her sense does indeed depend on available alternatives in ways that (I think) what merits an emotion does not. 25  Svavarsdóttir focuses on value and valuing, not on disvalue and disvaluing. She proposes that for something to be disvaluable is for it to be worth valuing its absence (2014: 108). She does not explore differences between different species of disvalue, such as being shameful and being disgusting. Perhaps it is worth valuing the absence of both, but presumably a full account of disvalue would need to differentiate them. There are also questions to consider about when and why it is a waste to spend one’s emotional resources valuing the absence of disvaluable things. 26  Svavarsdóttir might accept that there is a sense of ‘worth’ that is too broad to capture fittingness or merit and intend her use of the term to be understood differently (2014: 103, fn 30). But it is central to the conception of worth she does want to recruit that it grounds criticisms of evaluative attitudes as a matter of wasting ones resources. I don’t think that can be the general problem with unfitting emotions. 27  In earlier work Jacobson and I were less attuned to these issues, and tended to speak of fittingness as a matter of whether the emotion ‘presents’ its object correctly (see for instance D’Arms and Jacobson 2000). That at least sounds like a matter of mere representational accuracy and has rea­son­ ably enough been interpreted that way by some readers. Our presentation of fittingness in Rational Sentimentalism avoids equating fittingness with representational accuracy. What I am saying in this section is a first attempt to state the relations between fittingness, correctness, and representation more explicitly than we have in our collaborative work.

124 Fittingness pendent contents. On one hand, we would learn too little about what it is for something to be admirable or funny by learning that it has one of these properties if and only if the belief that it is admirable or funny is true. That would not be an informative account of funniness or admirability. And on the other hand, if the emotions were instead (partly) constituted by response independent thoughts that captured their evaluative content, then we would learn too much about what it is to be admirable or funny by learning what it is for those thoughts to be true. The appeal to emotions would be otiose in explaining value. Some might conclude from the forgoing considerations that fittingness is a different assessment from correctness.28 If correctness had to be understood as representational accuracy, then I would agree. In that case, I take it that the only way for an attitude to be correct would be the way in which beliefs and perceptions can be correct. But there are other senses of ‘correct’. Fittingness can be understood as equivalent to correctness in the sense in which the correctness of an attitude is a matter of its meeting a standard that is set by the nature of the ­attitude.29 The account sketched in section 5.4 illustrated how standards of fit for shame and guilt arise from their respective natures. The correctness of a belief is also a matter of its satisfying a standard that is set by the nature of beliefs. But because beliefs and emotions are different kinds of attitudes, their respective standards of correctness are different. Beliefs are correct when the proposition believed is true. Emotions are correct (that is, fitting) when the object of the emotion matters (is valuable) in the particular way one treats it as being valuable by having that emotion towards it. Sometimes the very same set of worldly facts make it the case that more than one attitude is correct. Rational sentimentalism has the consequence that the very same considerations can fix the correctness of an emotion towards an object and the correctness of a belief that the object has the corresponding sentimental value property. For instance, rational sentimentalism says that to be shameful is to 28  Though most philosophers seem to treat these as equivalent, some do not. See Berker (this volume), Howard (2018), Howard and Leary (this volume). 29 This broad suggestion about correctness (or fittingness, or right-­kindedness of reasons) is embraced by a number of philosophers, and developed in different ways. Schroeder (2012) says that reasons of the right kind for intention and belief arise from the nature of these attitudes. Deonna and Teroni (2012) argue that the standard of correctness for emotions emerges from the fact that emotions are states of felt action readiness. McHugh and Way (2016) suggest that an attitude’s correctness depends upon its functional role. Some philosophers defend the further claim that standards of correctness or fittingness are constitutive of the attitudes in question, either metaphysically or conceptually (Schroeder 2010; Sharadin 2015). Howard and Leary (this volume) argue against that claim constitutiveness. They also deny that standards of fittingness are ‘internal’ to the attitudes in question. But that is because they equate being internal with being constitutive of. I reject constitutiveness as well, at least for emotional fittingness, but I think it is helpful to think of the standards as internal to the emotions in a different sense: norms of fittingness depend on the nature of the emotions. But the emotions are not (even) partly constituted by being subject to such norms. The view of natural emotions developed in Rational Sentimentalism holds that something’s being an episode of a natural emotion is not constituted by its being subject to a particular norm, but by its being an instance of a natural psychological kind.

Fitting Emotions  125 merit shame. So facts that would make shame towards some object fitting would also make the belief that this object is shameful true. But that doesn’t mean that what makes shame fitting is that it represents the object as shameful. Considering an example may help to clarify these points. Suppose that someone is taunting our friend Sam in a bar, trying to pick a fight, and that Sam retreats rather than fight him. Stipulate whatever facts you think are needed to ensure that his retreat is not shameful: this was a case of sens­ ibly refusing to get drawn into a stupid altercation, not an act of cowardice. The facts you have stipulated make two things true: shame is unfitting, and a belief that what Sam did was shameful would be false. Nevertheless, if Sam was socialized in a way that many people are, he might be ashamed of having retreated in this way. His shame involves being bothered by his own behaviour, motivated to conceal it or deflect attention from the event, and perhaps even wanting to get away from his friends who saw the encounter—­even though we are treating him well. These facts about what shame is like are what make shame an incorrect response to the situation. Given the stipulations, Sam has no reason to be bothered in this particular way by what he did, or by our continued discussions of the encounter. He can hold his head high among us. To feel shame is to treat the episode as reflecting badly on Sam, but it doesn’t. Sam might also believe that his retreat was shameful. What makes his belief incorrect has to do with the nature of belief as a representational state. It is belief ’s nature to play certain roles in instrumental and theoretical reasoning and to be linked to spontaneous avowal under good conditions. (Beliefs about the fittingness of emotions in particular figure in reasoning that aims to regulate those emotions and the motivations that they characteristically involve.) These sorts of facts about the nature of belief make it the case that belief is a kind of representational state that is correct only when the content believed is true. Because it is not true that what Sam did was shameful, his belief is incorrect. Its incorrectness as a belief arises from the fact that inferences from that content will produce false conclusions, that avowals of it will mislead, that reliance on it will lead to actions one lacks reasons to perform, and so on. In this case, the stipulated facts that make Sam’s shame unfitting also make his belief false. But the full explanations of the incorrectness of the belief and the shame are different, because of the different ways in which these facts show those two different kinds of mental states to be mistakes. In this section I have distinguished the unfittingness of an emotion from the sort of mistaken representation that would render a corresponding evaluative belief false. In that respect I side with Svavarsdóttir and against the alethic view. I have noted a sentimentalism-­specific reason for adopting this position: we need to distinguish the unfittingness of emotions from the falseness of their putatively constitutive beliefs in order for emotional fittingness to explain value in the way that rational sentimentalism requires. But I have also offered some less

126 Fittingness theory-­driven reasons to draw the distinction, which arise from the different natures of emotion and belief. Denying that emotional correctness is equivalent to representational accuracy probably also aligns me against what I will call ‘representationalism’ in a current controversy in the philosophy of emotion.30 Representationalists argue that the relationship that an emotion stands in to the value its object must possess in order for the emotion to be correct should be understood as that of representation (Tappolet 2011; Rossi and Tappolet 2018). Their opponents deny this (Deonna and Teroni 2012; 2015). Though I am happy to allow that emotions present (or represent, if that doesn’t add much) their objects as having evaluative features, I  think of such talk as a useful shorthand rather than an explanation of what ­emotions are or what fittingness is. Earlier I sketched an argument that, although natural emotions are not even partly constituted by beliefs, they can helpfully be interpreted as appraisals of their objects. Such interpretations help us think about whether an emotion is fitting by articulating what its object would have to be like in order for this appraisal to be accurate. We can then say that Sam’s shame appraises his retreat from the bully as reflecting badly on him. In that limited sense, I interpret emotions as having evaluative contents, and as presenting their objects as having certain evalu­ative features. This sort of representational language affords a useful shortcut, allowing us to say that considerations of fittingness are the ones that bear on whether the emotion gets it right about its object, or whether the object is as the emotion appraises it. This is only a shortcut however, not an explanation of what emotions are or what fittingness consists in. I think it could be replaced without loss by saying that because of what Sam’s shame is like, it involves treating his retreat as something that reflects badly on him. The forgoing account of emotions as natural kinds of affective and motivational states that are interpretable as appraisals allows us to say several plausible things at once. Assessments of emotional fittingness concern the correctness of emotions. This correctness isn’t a matter of accurate representation, but a match between the emotional engagement and the object. Nevertheless someone who is in the throes of emotional engagement with an object appraises that object as having certain (response dependent) evaluative features, and having them to a degree that depends on the strength of emotional response. They thereby take the object as having a kind and amount of evaluative significance, and their response can be said to misrepresent the object if it lacks that value. This relaxed stance about representation preserves various intuitive idioms about emotions while

30  I say that my stance ‘probably’ aligns me in this way, because philosophers use talk of ‘representation’ in many ways, and I am not always sure I understand the different senses of the term. Rossi and Tappolet (2018) use ‘cognitivism’ rather than ‘representationalism’ to label the position about emotional correctness that I am disputing here.

Fitting Emotions  127 avoiding the errors of cognitivism. And it is compatible with an attractive sentimentalist theory of value.31

References Barrett, Lisa Feldman (2017). How Emotions Are Made. New York: Houghton Mifflin Harcourt. Broad, C.  C. (1954). ‘Emotion and Sentiment.’ The Journal of Aesthetics and Art Criticism 13(2): 203–14. Chappell, Richard (2012). ‘Fittingness: The Sole Normative Primitive.’ The Philosophical Quarterly 62: 684–704. Clarke, Randolph and Piers Rawlings (2022). ‘True Blame.’ The Australasian Journal of Philosophy. DOI: 10.1080/00048402.2022.2033285. D’Arms, Justin (2022). ‘Emotion Kinds, Motivation, and Irrational Explanation.’ In John Doris and Manuel Vargas, eds. The Oxford Handbook of Moral Psychology. Oxford: Oxford University Press. D’Arms, Justin and Daniel Jacobson (2000). ‘The Moralistic Fallacy: On the “Appropriateness” of Emotions.’ Philosophy and Phenomenological Research 61: 65–90. D’Arms, Justin and Daniel Jacobson (2003). ‘The Significance of Recalcitrant Emotions (or, Anti-QuasiJudgmentalism).’In Anthony Hatzimoysis, ed. Philosophy and the Emotions. Cambridge: Cambridge University Press. D’Arms, Justin and Daniel Jacobson (2022). ‘The Motivational Theory of Guilt (and its Implications for Responsibility).’ In Andreas Brekke Carlsson, ed. Self-Blame and Moral Responsibility. Cambridge: Cambridge University Press. D’Arms, Justin and Daniel Jacobson (forthcoming). Rational Sentimentalism. Oxford: Oxford University Press. Deigh, John (1994). ‘Cognitivism in the Theory of Emotions.’ Ethics 104: 824–54. Deonna, Julien and Fabrice Teroni (2012). The Emotions: A Philosophical Introduction. New York: Routledge. Deonna, Julien and Fabrice Teroni (2015). ‘Emotions as Attitudes.’ Dialectica 69(3): 293–311. Foot, Philippa (1963). ‘Hume on Moral Judgement.’ Reprinted in Foot (1978). Virtues and Vices. Oxford: Oxford University Press. Frijda, Nico (1986). The Emotions. Cambridge: Cambridge University Press. Fritz, Jamie (forthcoming). ‘Unfitting Absent Emotions.’ Oxford Studies in Metaethics.

31  I am grateful to Chris Howard for very helpful comments on a previous draft. Work on this chapter was supported by a sabbatical leave from the Ohio State University.

128 Fittingness Greenspan, Patricia (1988). Emotions and Reasons: An Inquiry into Emotional Justification. London: Routledge. Griffiths, Paul (1997). What Emotions Really Are. Chicago: University of Chicago Press. Howard, Chris (2018). ‘Fittingness.’ Philosophers Compass. https//doi.org/10.1111/ phc3.12542. Kauppinen, Antti (2014). ‘Fittingness and Idealization.’ Ethics 124(3): 572–88. Lazarus, Richard (1991). Emotion and Adaptation. New York: Oxford University Press. Lewis, David (1984). ‘Putnam’s Paradox.’ Australasian Journal of Philosophy 61: 343–77. Maguire, Barry (2018). ‘There Are No Reasons for Affective Attitudes.’ Mind 127(507): 779–805. McHugh, Connor and Jonathan Way (2016). ‘Fittingness First.’ Ethics 126: 575–606. Nussbaum, Martha (2001). Upheavals of Thought: The Intelligence of Emotion. Cambridge: Cambridge University Press. Roberts, Robert (2003). Emotions: An Essay in Aid of Moral Psychology. Cambridge: Cambridge University Press. Roseman, I. J. (1984). ‘Cognitive Determinants of Emotion: A Structural Theory.’ In P.  Shaver, ed. Review of Personality and Social Psychology, Vol. 5: Emotions, Relationships, and Health. Beverly Hills, Calif.: Sage. Rosen, Gideon (2015). ‘The Alethic Conception of Moral Responsibility.’ In Randolph Clarke, Michael McKenna and Angela Smith, eds. The Nature of Moral Responsibility: New Essays. New York: Oxford University Press. Rossi, Mauro and Christine Tappolet (2018). ‘What Kind of Evaluative States Are Emotions? The Attitudinal Theory vs. the Perceptual Theory of Emotions.’ Canadian Journal of Philosophy, DOI: 10.1080/00455091.2018.1472516. Scarantino, Andrea (2010). ‘Insights and Blindspots of the Cognitivist Theory of Emotions.’ British Journal for the Philosophy of Science 61: 729–68. Scarantino, Andrea (2014). ‘The Motivational Theory of Emotions.’ In Justin D’Arms and Daniel Jacobson, eds. Moral Psychology and Human Agency: Philosophical Essays on the Science of Ethics. Oxford: Oxford University Press. Scarantino, Andrea and Ronald de Sousa (2018). ‘Emotion.’ Stanford Encyclopedia of Philosophy. Scherer, Klaus (2001). ‘Appraisal Considered as a Process of Multilevel Sequential Checking.’ In K.  Scherer, A.  Schorr, and T.  Johnstone, eds. Appraisal Processes in Emotion: Theory, Methods and Research. Oxford: Oxford University Press. Schroeder, Mark (2010). ‘Value and the Right Kind of Reason.’ Oxford Studies in Metaethics 5: 25–55. Schroeder, Mark (2012). ‘The Ubiquity of State-Given Reasons.’ Ethics 122(3): 457–88. Sharadin, Nathaniel (2015). ‘Reasons Wrong and Right.’ Pacific Philosophical Quarterly 96:1–29. Solomon, Robert (2003). Not Passion’s Slave: Emotions and Choice. Oxford: Oxford University Press.

Fitting Emotions  129 Svavarsdóttir, Sigrún (2014). ‘Having Value and Being Worth Valuing.’ The Journal of Philosophy 111(2): 84–109. Tangney, J.  P., and R.  L.  Dearing (2002). Emotions and Social Behavior: Shame and Guilt. New York: Guilford Press. Tappolet, Christine (2011). ‘Values and Emotions: Neo-Sentimentalism’s Prospects.’ In C. Bagnoli, ed. Morality and the Emotions. Oxford: Oxford University Press. Tappolet, Christine (2016). Emotions, Values, and Agency. Oxford: Oxford University Press. Whiting, Daniel. (2021) ‘Aesthetic Reasons and the Demands They (Do Not) Make.’ The Philosophical Quarterly 71(2): 407–27. Williams, Bernard (1993). Shame and Necessity. Berkeley, Calif.: University of California Press.

6 Intuitions of Fittingness Philip Stratton-­Lake

6.1 Introduction In The Right and the Good Ross claims that basic propositions of the form ‘keeping your promise is prima facie right’ and ‘harming someone is prima facie wrong’ are self-­evident (2002: 29, 33, 40).1 In the Foundations of Ethics he retained his view about the epistemic status of fundamental moral propositions (1939: 283, 315, 320), but understood them as claims about what is fitting.2 So he came to believe that claims about which acts are fitting to what features of the situation are self-­evident (1939: 315).3 A dominant view is that only attitudes are fit-­apt. As Na’aman notes in his contribution to this volume: ‘Fit is a normative relation between an attitude and what it is about . . . It is generally assumed that all rationally evaluable attitudes are fit-­ evaluable and that most if not all things fit-­evaluable are attitudes’ (p. 5). Since actions do not have intentional objects, and are not attitudes, it may be claimed, they are not fit-­apt. But I see no reason other than theoretical neatness to suppose that it is impossible for an action to be fitting to some aspect of the situation, as Ross claimed. So I will assume that Ross is correct to assume that acts are fit-­apt, as many intuitionists did before him. Ross also accepted that it is self-­evident that admiration is fitting towards what is intrinsically good (1939: 279), though he rejected a fitting attitude analysis of goodness.4 He thought such knowledge is no different from our knowledge of other a priori truths, such as those of mathematics (2002: 29, 32). So, according to

1  Many thanks to Selim Berker, Jonathan Dancy, and the participants of a workshop on this volume for incredibly helpful comments on an earlier draft of this chapter. 2  He also came to endorse a certain sort of subjectivism about duty under the influence of Prichard. But this change is not relevant to my topic here, so I set it to one side. 3  In their contribution to this volume McHugh and Way claim that fittingness is an overall rather than a contributory relation (p. 4) By contributory they mean pro tanto. But I see no reason why we should accept this view. 4  This is because he thought that what makes admiration fitting is that its object is intrinsically good in an indefinable sense. But someone like Ewing, who endorses a fitting attitude account of goodness could reply that it is not the goodness of the good thing that makes admiration fitting, but the features that make it good. Hurka rejects this response (Hurka 2014: 61). Philip Stratton-­Lake, Intuitions of Fittingness In: Fittingness: Essays in the Philosophy of Normativity. Edited by: Christopher Howard and R. A. Rowland, Oxford University Press. © Philip Stratton-­Lake 2022. DOI: 10.1093/oso/9780192895882.003.0006

Intuitions of Fittingness  131 Ross, our knowledge of what is fitting to what gets subsumed under a rationalist approach to our knowledge of the a priori in general. I want to explore this approach to the epistemology of fittingness. I consider some ways in which one might think that our knowledge of fittingness may be problematic in ways that our knowledge of what is good or right is not. I maintain these worries are not serious. I finish by considering whether we need a different sort of intuition in relation to the all in fittingness of actions and to fitting attitudes. Some have argued that intuitions of all in fitting acts cannot be purely intellectual since these intuitions must be intrinsically motivational. Intuitions of  fit emotion, on the other hand, may be purely intellectual. I consider how a rationalist epistemology might accommodate such action-­guiding intuitive prop­ os­itions. I reject a view of practical intuitions as emotions, and propose an alternative simpler, and more unified account based on the nature of rational agency. On this alternative view there are not different types of intuitions in relation to fitting actions and attitudes: rather there is a single type of intuition that can play different roles. I begin by outlining the specific rationalist epistemology of the a priori that I favour. This is not the place to argue for this view or defend it from objections. I just plan to lay out the bare bones so that we can ask whether there are any special problems about the claim that we can know what is fitting to what in this way.

6.2  Outline of a rationalist epistemology of fittingness All self-­evident propositions are intuitive in the sense that we have non-­inferential justification for believing them,5 but not all intuitive propositions are self-­evident. This is in part because self-­evidence is factive, whereas intuitiveness is not, and partly because not all intuitive propositions can be known, even if true. They may be too vague or complex for that. I think this is true of intuitions about trolley cases. The intuitive proposition that I may divert the trolley to save the five, even though one will die if I do, seems straightforward enough. But all sorts of things are lurking in the background, that cast doubt on our confidence in the truth of this proposition. The fact that it seems wrong to kill one person to save five in the bridge version, may weaken our confidence in the original switch case. One might get round this with reference to the doctrine of double effect, but that is a controversial doctrine. Furthermore, the fact that it is a greater wrong to kill than to let die, may make us doubt whether this greater wrongness is sufficient to make pulling the lever to save five wrong. Finally our knowledge of the sort of distorting effects that empirical psychology 5  Some self-­evident propositions may not seem intuitive at first sight. These are what Audi calls mediate self-­evident propositions (Audi 2019: 362).

132 Fittingness seems to discover6 may cast doubt on all trolley case intuitions. So although the proposition that it is permissible to pull the lever in straightforward switch cases is intuitive, I do not think it is self-­evident. That proposition may well be false, and even if it is true I do not think we know it. If, however, we change the proposition to make it simpler and to exclude the complicating factors that cast doubt on the truth of intuitions about overall rightness, I think the answer may well be self-­evident. Consider the following two propositions: 1. The fact that pulling the lever would save five people counts in favour of pulling it 2. The fact that pulling the lever would kill one person counts against pulling it. I think 1 and 2 are not only intuitive but are self-­evident. They do not get embroiled in issues about double effect, or whether killing is morally worse than letting die. One can assent to these two propositions regardless of what one’s intuition is about whether we ought to pull the lever. Finally, because these prop­ os­itions are quite abstract, it is hard to see how our intuition of them could be distorted by irrelevant facts, like being ‘up close and personal’.7 But I do not want to get embroiled in a debate about which intuitive prop­os­ itions are self-­evident. My discussion of trolley cases was just to illustrate the distinction between the intuitive and the self-­evident. But since all self-­evident propositions are intuitive propositions, I shall just focus on the latter, and what sort of direct justification we can have for believing these propositions. On a very influential view, what plays the basic justificatory role in intuitive (and self-­evident) propositions is an adequate understanding of them (see, for example, Audi 1997: 40–1, Sosa 2007: 60; Shafer-­Landau 2003: 247). What counts as an adequate understanding is quite involved. It involves more than a bare semantic grasp, or the ability to translate the proposition into some other language one might know. It also involves the ability to see certain entailments, to distinguish the proposition from near neighbours, to see the relevant relations between the proposition’s constituent concepts, being able to recognize a suitable range of instances of the general truth, and so on.8 But regardless of how adequate understanding is understood, I reject the idea that it is a basic justifier for synthetic a priori truths.9 This is not to say that it does not have an important epistemic role. It does. It just doesn’t have the role that Audi and others think it does—­namely, the role of being a basic justifier. What 6  See Sinnott Armstrong (2006). 7  James Andow casts doubt on some of these claims (2018). 8  In a recent paper Audi lists nine conditions (2019: 364–72).

9  See Stratton-­Lake (2016).

Intuitions of Fittingness  133 has that role is a clear intuition of the intuitive proposition: it is this that justifies believing intuitive propositions. Not all accounts of intuition make this possible. But a very plausible account of intuition that is widely accepted does—­that is the view that intuitions are intellectual seemings (see Bealer 1996; Bedke 2008; Bengson 2015; Chudnoff 2013; Dancy 2014; Huemer 2005; and Kauppinen 2013). Seeming here is not a hedging term used to express doubt, but a quasi-­experiential state, analogous to perceptual seemings. On this view intuitions are distinct from beliefs, hunches, guesses, or gut reactions. Just as when I look at the wall it seems to be white, so when I get a clear grasp of certain propositions they seem to be true. Just as the wall’s seeming white provides defeasible justification for believing that it is white (absent undercutting defeaters), so p’s seeming true provides default justification for believing that it is true.10 For intuitions, so understood, to be basic justifiers, the seeming must be explained in the right way. It may well be that the seeming is explained by evolutionary forces,11 or more extravagantly, by a machine that has made me such that the proposition seems true whenever I consider it, even though my understanding is no better than it was before the machine fiddled with my brain, and when the proposition did not seem true (Markie: 2013). In such cases the default justification provided by the seeming will be undercut because the seeming is explained in the wrong way. If however, the seeming is explained by an adequate understanding of the proposition, then it will be explained in the right way, and so the seeming will justify a belief with the same content. This is the important epistemic role of adequate understanding mentioned earl­ier. Understanding doesn’t play the role of justifier—­that role is played by the seeming. Adequate understanding is, rather, a necessary condition of a seeming having the sort of justificatory role it has, by providing the right sort of ex­plan­ation of why that proposition seems true. So adequate understanding is not a basic justifier, but as a condition of something else (intuition) being a basic justifier. Intellectual seemings seem to me to provide an adequate justification of our beliefs of what is fundamentally good, pro tanto right, and fitting. This is not to say that justification cannot be provided by means of argument. It is just to say that such inferential justification is not needed.12 We may now ask whether there

10  Kauppinen describes these as ‘non-­sensory, non-­doxastic, spontaneous, primitively compelling, and putatively non-­inferentially justify’ (Kauppinen 2013: 364), and Dancy describes them as ‘conscious, contentful, nonfactive, representational, and presentational’ (2014: 791–2). 11  For a compelling argument that evolution cannot shape our intuitions or moral beliefs in a way that casts doubt on their truth or justifiability, see Hanson (2016) and Morgenson (2015). 12  Sinnot-­Armstrong argues that the need to check that our intuitions are not distorted in various ways means that when they justify, they do not do so non-­inferentially. But that is to confuse a justification for believing the proposition, and for believing that there is non-­inferential justification for believing the proposition. See Ballantyne and Thurow (2013).

134 Fittingness are any special problems with the view that our intuition of propositions about fittingness can justify our belief about fittingness.

6.3  Different fittingness relations If an intuition of a proposition of the form ‘p is fitting to q’ is to justify believing that p is fitting to q then that seeming must be based on an adequate understanding of that proposition. So we would need an adequate understanding of the key terms in this proposition, including that of fittingness. One worry about in­tu­ itions of fittingness might be that the very concept of fittingness is too vague to satisfy Audi’s conditions of adequate understanding. Although he does not endorse this view, Howard notes how common it is for philosophers to claim that the concept of fittingness is vague, and imprecise at best.13 Tom Hurka expresses such scepticism about fittingness. Sometimes he says that it either collapses into goodness or rightness when adequately understood,14 whilst at other times he says it is simply the concept of ‘ought’ with a distinctive ground—­that of matching.15 A further way in which the concept of fittingness might be vague relates to whether it entails ability. Ross and Ewing thought that in relation to actions it does, and in relation to emotions it doesn’t (Ross 1939: 55; Ewing 1949: 133). This is because they thought that the relevant notion of ability is being under the control of the will. Since emotions are not under the control of the will, yet can be fitting, the fittingness of emotions does not entail ability. If the entailment of ability is internal to fittingness as it applies to actions as Hurka maintains (2014: 77),16 this would make the term ‘fitting’ not only vague but ambiguous. It could mean either fittingness in the ability dependent sense or in the ability independent sense. All of this would cast doubt on whether there is some unified relation about which we could have clear intuitions. But I agree with Howard that there is nothing muddled or vague about the concept of fittingness, and I think he shows that it cannot be reduced to other more basic normative notions.17 Furthermore, The Ross/Ewing understanding of the relevant notion of ability is questionable at best. It is true that we cannot choose to have some emotion by an act of will, but that does not mean that there is not a relevant sense of ability that applies to coming to have emotions. As Connor McHugh argues (2017), emotions may well be under our control in the sense that they are sensitive to our judgements about whether 13 ‘Reports that one has prodigious difficulty “latching onto” the notion of fittingness, or that “fittingness” is a cooked-­up concept, entirely unfamiliar in ordinary thought, tend to ring loudly in Q&As for talks in which the notion of fittingness centrally figures. This current tendency is, I think, unfortunate’ (2018: 2). 14  In its practical sense Hurka claims, fittingness is simply an ought with a distinctive ground (2014: 63, 77). 15  See his contribution to this volume. 16  See also his contribution to this volume, p. 9. 17  See also Berker’s contribution to this volume.

Intuitions of Fittingness  135 we have good reason to have them, whether we ought to have them, or whether they are all in fitting. If he is right there is no vagueness about whether ‘fitting entails can’, and ‘fitting’ is not ambiguous between an ability dependent sense and an ability independent sense. There is, nonetheless, some reason to think that there are different types of fittingness relations, and if that is right ‘fitting’ would seem to pick out different relations. Take Berker’s examples of the shameful, the lovable, and the credible. These are plausibly analysed as being the fitting object of shame, love, and belief. But the fittingness relations in each seem different. The fittingness that relates to love in a lovable kitten seems closer to what Jonathan Dancy calls an enticing reason,18 and lacks any peremptory character. There is no sense in which you ought to love any or all of lovable kittens, even if there are no opposing reasons. It is just that they have lovable features that make loving any of them an in­tel­li­gible option. But it is not plausible to suppose that the sort of fittingness involved in the credible is anything like an enticing reason. For a start the credible is nothing to do with the fun, attractive, exciting, pleasant, etc. There is nothing enticing about a proposition. Furthermore, the reasons we have to believe some credible prop­os­ ition can take us to an ought. If there is sufficient reason to believe that p, and no reason not to believe p, then at least sometimes we ought to believe p. So the fact that some proposition is fitting to believe seems closer to a straightforward ­normative reason to believe that proposition, and is thus different in style from the fittingness involved in the lovable.19 Finally, the fittingness involved in the shameful seems different from both the lovable and the credible. Shame for my disgraceful act is not merely an intelligible option, or something I have some reason to feel, but seems very close to saying that I ought to feel shame. So we seem to have three different types of fittingness—­ one with no peremptory character at all, one with pro tanto peremptory character, and one with all in peremptory character. Does the fact that the generic notion of fittingness includes very different relations raise a problem for our knowledge of what is fitting? I do not think it does. All it means is that we have to get clear on which fittingness relation is relevant to the intuitive proposition we are considering, and that will involve seeing which type of fittingness is relevant to the case. Exactly the same point could be made 18  ‘Enticing reasons are to do with what would be fun, amusing, attractive, exciting, pleasant, and so on. They can be stronger and weaker, and they are often strong enough for action. But (as I understand the matter) they never take us to an ought; it is not true of an enticing reason that if one has one of them and no reason of any other sort, one ought to do what the reason entices one to do’ (Dancy 2003: 99). 19  I do not claim that the fitting relation and the reasons relation are the same, or that there is nothing more to F’s being a fitting object of some attitude or act, than F having properties that give us reason to have some attitude or do some act. For that to be true it would have to be the case that if one has some reason to believe P then it is fitting to believe p. But wrong kind of reasons cases cast doubt on this implication. (See Howard 2018: 3–4.)

136 Fittingness about our knowledge of what we have reason to do. The term ‘reason’ could mean either normative or motivating reason. But that ambiguity hardly casts doubt on our knowledge of the proposition that we have reason to avoid agony. It just means that we have to be clear what type of ‘reason’ is at play in our understanding of this proposition, and that this is a normative claim rather than a psychological one.

6.4  Intuition internalism Broad, Ross, Ewing, and all previous intuitionists thought that acts as well as emotions can be fitting. Ross applied his distinction between the pro tanto and the all things considered to fittingness as well, and went on to define pro tanto duty as fittingness to an aspect of the situation, and all things considered duty as the act that is most fitting to the total situation (1939: 53). Since propositions about what is fitting, or called for, are in the action-­guiding business, this raises the question about whether propositions about fitting actions require a different type of intuition to propositions about fitting emotions. One reason to suppose that they do is that intuitions of action-­ guiding ­propositions must be intrinsically motivational. The idea here is the intuitional correlate of moral judgement internalism. According to judgement internalism, if I judge that I ought to Φ, then either I am motivated to Φ, or I am irrational (Smith 1994).20 According to what I call intuition internalism,[do we want to split the word ‘internalism in this way?’] if it seems to me that I ought to Φ (which for Ross meant that Φing is most fitting to, or called for by, the total situation) then I  am motivated to Φ. Dancy and Kauppinen argue that to capture the intrinsic motivational nature of action-­guiding intuitions, these seemings cannot be purely intellectual. They must also involve some conative mental state, and this conative mental state is, they claim, emotion.21 So on their view we have two types of seemings. Purely intellectual seemings and practical seemings, where the latter are essentially emotion involving. Practical seemings must, they think, involve emotion to capture what I have called intuition internalism. 20  Moral belief internalism requires a rationality constraint in order to accommodate cases such as weakness of will, or accidie. Intuition internalism, however, does not need this constraint. Of the listless Dancy claims that such a person ‘may know that he has reasons, but these reasons are not presented to him as such; they are only represented to him. He has the right belief, but belief is not enough, since it is merely representational’ (2014: 802). Dancy follows Bengson in distinguishing presentational states from representational states, where the former have an experiential, passive quality to them that the latter do not. Seemings are presentational whereas beliefs are representational, lacking as they do presentational phenomenology. So for Dancy, if someone is in a state of accidie they could not have an intuition that they ought to Φ, where Φing is some action, though they could have the belief that they ought to Φ. 21  As we will see, they choose emotion as opposed to some other conative mental state, such as desire, as they want this motivational state to be one that is at the same time a representational state and emotions are better suited to this than desires.

Intuitions of Fittingness  137 I will say more about their view in the next section. Before I get onto that I want to clarify intuition internalism. The first thing we need to note is the distinction between an intuition of all in fittingness being intrinsically motivational and its being essentially motivational. Such an intuition is essentially motivating if it motivates whenever one has that state. An intuition is intrinsically motivating if it can motivate without any help from some other mental state, such as a desire. There is nothing in the idea of an intrinsically motivating state that implies that it must motivate in every case, and so must be essentially motivating. The idea is that when it does motivate, it does it without the help of anything else. Given this distinction we may ask whether intuition internalism claims that practical in­tu­ itions are essentially motivational or just intrinsically motivational? This is not clear. At times Dancy seems to think that practical seemings are essentially motivational (2014: 802), so if you are not motivated, you don’t have them.22 But elsewhere he seems to think that practical intuitions lead to action only if assented to (2014: 795, 796).23 But if practical intuitions must be believed if they are to be motivational they could neither be essentially nor intrinsically motivational. This is because their motivational force would depend upon having a certain belief. So if intuition internalism is to capture the idea that practical intuitions are at least intrinsically motivational, it should abandon the assent condition. Is it plausible to suppose that a practical intuition could motivate even if it is not believed? Kauppinen makes a good case for the view that it is. He interprets Huck Finn as having, and being motivated to act by, an intuition that he ought not to turn Jim in to the slave hunters whilst believing that he should turn him in. This scenario is plausibly interpreted as a case in which Huck’s intuition motivates him to act, even though he does not believe it is true, and so does not assent to it (2013: 371–2). Kauppinen also gives the example of someone who believes it is wrong to give money to beggars (because it will probably make their situation worse), but in the face-­to-­face situation the beggar’s desperate situation may present giving as called for, and one might give for that reason (2013: 366). Chudnoff also argues that intuitions can motivate by themselves. Beliefs, he states, have a kind of permanence that intuitions do not, and it is implausible to suppose that for every time we are motivated by some intuition, that we have to acquire a corresponding belief: If intuitions about reasons for action play action-­guiding roles, their ability to do so does not depend on their leading to beliefs. You can act in light of the 22  It is unclear what Kauppinen thinks. He writes: ‘moral beliefs themselves are not essentially motivating, while moral intuitions do intrinsically motivate’ (2013: 366). He here switches from the essential to the intrinsic, so does not seem to distinguish the intrinsic from the essential. 23  He seems to claim that action is assent. But that does not fit with Kauppinen’s cases.

138 Fittingness justification you have for believing that you have a reason to act without forming the belief that you have that reason to act. In fact, this is likely the norm. Beliefs are mental states with a certain degree of permanency. Most of our actions have little significance beyond the moment of their occurrence. It would be silly to form standing beliefs about what reasons you have for all the actions you perform in life. So it seems that if intuitions play action-­guiding roles, their doing so does not depend on their leading to beliefs. There is reason to think, then, that beliefs are more dispensable than intuitions as guides to action.  (2013: 169)

This seems right to me. Think of all the times when one responds to one’s situ­ ation because one can see something, such as the brake lights coming on in the car in front on the motorway. I see that and brake. It is, I think, implausible to suppose that I collect all the corresponding beliefs. So both Kauppinen and Chudnoff give us good reason to suppose that intuitions can be intrinsically mo­tiv­ational. I leave aside the question of whether they are essentially mo­tiv­ation­al, as intrinsic motivation is what motivates the view that practical intuitions cannot be purely intellectual, as Kauppinen and Dancy claim. It is to capture the intrinsically practical nature of practical intuitions that Dancy and Kauppinen claim that practical intuitions must either be (Dancy) or be constituted by (Kauppinen) an emotion. If they are right then my intuition that it is all in fitting for me to do some act would always be emotion-­involving, whereas my intuition that some emotion is fitting to something would (­para­dox­ic­al­ly) be purely intellectual. It is this idea that I wish to explore in the rest of this chapter. I will propose an alternative unified account of our ­knowledge of the fitting. Rather than having two types of intuition of the fitting, ­I propose that there is one type of intuition that can play two different roles in rational agents.

6.5  Intuitions of fittingness and emotion The view we are considering is that to capture the intrinsic motivational nature of practical intuitions we must assume they are, or are constituted by, emotion. But the link between emotion and motivation is complicated in Dancy. He claims that intuitions must be emotional to be motivational, but he rejects the idea that it is intuition that motivates. When discussing the emotion of anger he claims that it can present its object, or more precisely, its ground, as a reason. This is the presentational aspect of anger. It presents the world as being a certain way. This seeming is also motivational, but that does not mean that it motivates. What motivates is not the anger, but its ground (2014: 801–2). So although emotion is needed for motivation, Dancy does not think that emotions, or intuitions understood as emotions, motivate.

Intuitions of Fittingness  139 How are we to make sense of the view that such intuitions are intrinsically motivational but do not motivate? I think a plausible reading is that this claim is a variant of his view that desire is not what motivates, but is the state of being mo­tiv­ated. What motivates, he claims, is the reason one desires to do some act, and the desire is the state of being motivated (2014: 803). I think his view is that in the case of practical intuitions the emotion-­intuition does not motivate: the emotion is the state of being motivated and what motivates is the reason the emotional state presents as calling for action. That would make sense of the need for emotion, and his claim that it is not the emotion that motivates. If that is right, then on Dancy’s view, the mental state of being motivated can be realized in different ways, either as having some desire or as having some emotion. The claim is that intuitions of fittingness as they relate to action involve emotion whereas intuitions of fittingness as they apply to attitudes is purely intellectual. If there is a need for practical intuitions so understood, they would have to be limit­ed to all in propositions, i.e. to propositions about what we are obligated to do, or about which action is most fitting to the total situation.24 Typically an obligatory act will be in some respect pro tanto wrong—­that is, an act that is most fitting to the total situation will be unfitting to some aspect of that situation. But it is, I think, very implausible to suppose that for every intuited consideration that is unfitting to the act I ought to do, we must feel some negative emotion or mo­tiv­ ation. I can see no reason to suppose that we must have any motivation at all to do some act other than the one we ought to do.25 So intuitions of fittingness would only be practical at the verdictive level—­that is, when one has an intuition that some act is most fitting to the total situation. If, then, emotion is needed to capture the practical nature of such propositions, it is only needed for intuitions about what is most fitting for me to do. Ross would deny that all in fittingness is the sort of thing that we could know directly (2002: 30–1), but that seems overly pessimistic. I see no reason to suppose that we couldn’t know directly that some act is all in wrong, or unfit to the total situation in act types, such as Harman’s cat-­burning scenario, or in relation to act tokens, such as the treatment of Jesse Owens when he arrived for the post-­Olympic reception at the Waldorf-­Astoria Hotel.26 Second, even if he is right that we could not directly know such propositions, we may, nonetheless, have non-­inferential justification for believing them, and that is all that is needed here.

24  Judgement internalism is not about judgements at the pro tanto level, but about all in judgements about what we ought, all things considered, to do. The same should apply, I think, to intuition internalism. 25  Perhaps we intuit that the fact that our act is unfitting to some aspect of the situation makes compunction a fitting response. But we do not need to have an emotion to motivate having some emotion. So such intuitions would on Dancy’s view be intellectual. 26  Because of his ethnicity he was forced to use the freight elevator.

140 Fittingness

6.6  Rational agency Must we suppose that our intuition of all in fittingness in relation to acts must be, or be constituted by, emotion to account for the motivational role of such in­tu­ itions? In this section I outline an alternative account based on the idea of a rational agent that seems to me at least as plausible as the alternative, and does not force us to suppose that there are different types of intuition for intuitions about the fittingnesss of some emotion versus the overall fittingness of some action, or of fittingness at the pro tanto level and at the overall level. On the view I propose there are not two types of moral intuition: the intellectual type and the motivational-­emotional type, but one type that can play two roles—­the justifying role and the action-­guiding role. So I will finish by exploring this more unified account of moral intuitions. The idea is that we can give an adequate explanation of how certain intuitions can be intrinsically motivational from the very idea of a rational agent. The rele­vant notion of a rational agent is one whose attitudes and actions tend to track their judgements about whether those attitudes and actions are justified, or whether we have most/conclusive reason to have those attitudes or to do the rele­vant act. As Scanlon puts it: A rational agent is, first, one that is capable of thinking about the reasons for certain actions or attitudes, and for reaching conclusions about which of these are good reasons. Second, a being is a rational agent only if the judgments that it makes about reasons make a difference to the actions and attitudes that it proceeds to have. A perfectly rational agent would always have attitudes and perform the actions that are appropriate according to the judgments about reasons that he or she accepts. A rational agent will, for example, generally intend to do those actions that he or she judges him or herself to have conclusive reason to do, and believe a proposition if he or she takes him or herself to have good evidence for its truth.  (2014: 54)27

The relevant attitudes are what Scanlon calls judgement sensitive attitudes: These are attitudes an ideally rational person would come to have whenever that person judged there to be sufficient reasons for them and that would, in an

27  It is worth noting that the view that emotions are presentational is not well equipped to capture the judgement sensitivity of emotions. This is because, on this view, the seeing that someone is in dire need is part of the emotion of sympathy, or empathy, rather than something prior to sympathy to which our sympathy is a response. So I am not sure how they can accommodate the ordinary thought that we feel sympathy for someone’s situation because we see that their situation gives us sufficient reason to feel sympathy for them.

Intuitions of Fittingness  141 ideal­ly rational person, ‘extinguish’ when that person judged them not to be supported by reasons of the appropriate kind.  (1998: 20)

Scanlon makes reference to an ideally rational agent. None of us meet this standard, but we are rational to the extent that we match up to this ideal, and our particular attitudes are rational if they cohere with our normative judgements about whether we have conclusive reason to have them. The idea here is that when rational agents have the attitudes, or do the actions, they judge they should have, this judgement provides a sufficient explanation of why they have those attitudes, or do those actions, and also makes sense of this in normative terms.28 Having such attitudes is just what one would expect of a rational agent who accepted such judgements (Scanlon 2014: 55).29 To be rational one does not have to have every attitude one judges one has sufficient reason to have. These are not attitudes that we must have, but are attitudes it makes sense for us to have given our normative beliefs. But one’s attitude would be irrational if one judged it to be groundless, or one judged one ought not to have it. Furthermore, one would be pro tanto irrational if one didn’t do the act one judged one ought to do (unless one were restrained, or incapacitated in some other way). This conception of a rational agent means that there is no need to posit some desire or emotion to explain how some normative judgement is motivational. There is no mystery if a rational agent’s attitudes and actions line up with her judgements about whether she should have those attitudes or do some act. All that is needed is that those states are both able to explain her coming to have some attitude or to do some action, and can rationalize or make sense of this (Scanlon 2014: 53). A belief that I have conclusive reason to do some act can just as well rationalize my doing it as a desire, or some relevant emotion, can. I would add that an intuition that I have conclusive reason can do the same.30 There is 28  The normative judgement does not have to be true for complying with it to be rational. This is where what we have most reason to do and what it is rational to do can come apart. They come apart when the normative proposition believed is false. See also McHugh and Way in this volume, p. 7 for other objections. 29  To make this account of rationality more plausible we need to add a qualification for judgement to do the rationalizing work—­one that Parfit adds (2011: 111ff). That is, that the normative prop­os­ ition believed has to be such that, if it were true it would be a reason to do that act. Unless we do this certain beliefs might explain action but fail to rationalize it. 30 Some philosophers clam that judgement and emotion cannot be separated in this way. For instance, Sabine Roeser claims that: With such emotions, we cannot separate the cognitive from the affective aspect. They are two sides of the same coin. In the same vein, it is futile to ask whether the affective or the cognitive response comes first. Experiencing indignation means having a judgement and a feeling. Experiencing indignation is not the response to an initial, purely cognitive moral judgement. Forming the judgement and having the feeling go hand in hand. Based on factual information of a situation, we form a moral judgement that is cognitive and affective at the same time.  (Roeser: 2010: 180)

142 Fittingness no need to suppose that this intuition is a desire or an emotion for it to do this mo­tiv­ation­al work.31 This is true not only of judgements and intuitions of conclusive reasons, but also of fittingness. If we have an intuition that the total situation makes Φing the most fitting thing to do, or as called for by the situation, then there is no mystery—­ nothing in need of further explanation—­if we Φ for that reason. Our taking the situation to justify some response is all that is needed to explain and make sense of our action. Chudnoff illustrates this point with his account of the mental act of inferring. Suppose I believe that p. On entertaining this thought I have the intuition that p  entails q. But this is not like the intuition that 2 + 2 = 4. For this intuition (or more precisely its content) not only justifies the belief that p entails q, but in and of itself can lead me to infer that q, because it justifies this inference. Inferring is an act, albeit a mental act, so this intuition is not merely intellectual but also practical. How can it play that dual role? To suppose such mental acts necessarily involve emotion is not only unnecessary but is deeply implausible. As stated above, all we need is the idea that rational agents are the sorts of beings that tend to line up their attitudes and actions with what they take to justify those attitudes and actions. Part of what it is to be a rational agent is for me to come to believe q if I see that p  entails q, for the intuition that p entails q justifies inferring q, and a rational agent requires nothing more to infer q. If an intuition can explain and justify a mental act of inferring without any emotion being required, I see no reason why an intuition that some act is most fitting to, i.e. called for by, the total situation could not explain and justify the mental act of intending, or as Ross and Prichard would say, ‘setting oneself ’, to do that act. It would explain (and rationalize) why I intended to do that act because I took the situation to justify doing that act. What else would a rational agent do if she came to believe that the situation required her so to act? She might not act as she thinks she should, but that would be a case of local irrationality. So I don’t think we need to regard intuitions as emotions to capture their explanatory role. Their explanatory role can be fully explained by their justificatory But no argument is given for this supposed inseparability and the idea that such emotions are judgement sensitive attitudes gives us reason to suppose that they are separable, with the emotion tracking the judgement (or in our case, the seeming). If the emotion is in part the judgement or seeming, then we cannot say that it is a response to apparent reasons, as it would be identical with apparent value. But then it is hard to see what it is a response to. There is a tendency among such authors to suppose that if emotions are separated from such judgements then they will be understood as a sort of blind impulse (see Nussbaum 2001: 25). But on the view I am defending this assumption is a mistake. 31  This is all consistent with the idea that it must give rise to a desire to motivate, if the state of desiring to Φ is the state of being motivated to Φ. But the intuition, or its content would be what does the motivating here, and the desire is wholly explained by the intuition.

Intuitions of Fittingness  143 role in rational agents, and this lends support to the idea that intuitions of fitting action, on the one hand, and those of fitting emotion/attitudes, on the other, are of the same kind. But the motivational role is not the only argument for the view that practical intuitions must be emotions or are emotional. Kauppinnen argues that emotion is needed in order to capture the phenomenology. So I now turn to the issue of whether the idea that intellectual intuitions have a dual role (justificatory and a practical) can account for the phenomenology.

6.7 Phenomenology Consider Kauppinen’s beggar example again. He writes: Consider walking past a dirty, pathetic beggar on the street on a rainy day. I am fairly convinced by arguments that we should not give money to street beggars, because it provides people with wrong kinds of incentive and it will not in fact benefit them. In fact there is a high risk of making the beggar worse off. Yet when I look at the weather-­worn face and hear the polite request, it sometimes seems to me that I really ought to give this time. I am not only inclined to believe that I should give this time, but also feel the motivational pull. Sometimes such seeming can result in action even if I do not change my belief. (2013: 366, my emphasis)

The phenomenology in such cases is the feeling the motivational pull of the action called for by the beggar’s situation in this face-­to-­face encounter. Kauppinen claims that this felt pull is best explained by the intuition being constituted by an emotion which both presents the beggar’s situation as calling for a certain response—­ giving—­and motivates my responding in that way (Dancy would say that it is the beggar’s situation that motivates, rather than the emotional presentation of it, but he can make the same point about the phenomenology). If this intuition were merely intellectual, it is claimed, it would lack this phe­nom­en­ology. It would, so the argument goes, be more like the cold, detached, intuition that two plus two equals four. So the challenge for the dual role view is whether an intellectual intuition can account for the felt pull even when it is contrary to our belief. I think this challenge can be met by drawing on what Sosa says about in­tu­ itions. Sosa understands intuitions as felt attractions to assent (2007: 44ff). I do not think this account of intuition is correct, as I do not think that if I have a felt attraction to assent to p, then I have an intuition that p. Whether I have an in­tu­ ition that p depends on why I feel attracted to assent. If it is because p fits with some theory I already hold, or p is flattering, then I may well not have an intuition

144 Fittingness at all.32 If on the other hand I feel attracted to assent to p because p seems true, I do have an intuition. But then the intuition is not the felt attraction, but what explains it—­the seeming true.33 So as I understand things the felt attraction to assent is distinct from, yet explained by, an intuition understood as an intellectual seeming. Here the phenomenology is explained by the force of the seeming true. I think the same would be true of our intuition that the beggar’s situation calls for us to give. One would feel the pull of giving, just as one could feel attracted to assent to some proposition, but that felt pull is not itself the intuition. It is rather distinct from and explained by the intellectual intuition that giving is called for, or more precisely, explained by the strong intuition that the beggar’s dire circumstances call for me to give. Would one feel that pull without the emotion of sympathy? I don’t think we should be too quick to assume that we would not. I can feel the pull towards believing some intuitive proposition even if I do not believe it. But this phe­nom­ en­ology does not require some sort of emotional engagement with the prop­os­ition, or of propositions of the relevant type. I have a strong mathematical intuition that there are more natural numbers than even numbers, but I do not believe this. Whenever I try get this proposition clearly in view, I cannot see how things could be otherwise, and feel a strong attraction or pull to assent. (This pull is even stronger given I do believe there are more natural numbers than prime numbers.) I can resist this pull by reminding myself of the proof that there are just as many even numbers as natural numbers. But the point is that there is a distinctive phenomenology here—­a felt pull that needs resisting. In this respect the mathematical case is just like the beggar case, except that I manage to resist the pull. But if one can feel the pull to assent to some math­em­ at­ic­al proposition that seems true without engaging any emotion, I see no reason to suppose that this cannot be true in the practical case. Just as the phenomenology in the mathematical case is explained by the force of the seeming rather than by some emotion, so the phenomenology of the beggar example may be explained by the strength of the presentation of his situation as calling for me to give, rather than by some emotion. Just as in the mathematical case, the felt attraction to assent is the result of the intuition, and thus distinct from it, so in the beggar case the felt pull to give is the result of the intuition, and is thus distinct from it. It might be argued that it is not the felt pull of the seeming that the emotion explains, but the seeming itself. One might claim that the beggar’s situation would not seem to call for me to give, or at least would not seem so forcefully without 32  Sosa thinks that intuitions are felt attractions to assent based solely on our understanding of the proposition (2007: 60). I gave reason to reject this view earlier in this chapter. 33  I should add here that Sosa understands his account as giving positive content to Bealer’s seeming account. But once we abandon the view that what grounds the attraction to assent is our understanding of the proposition and replace it with that proposition’s seeming true, his view cannot be regarded as telling us what seemings are.

Intuitions of Fittingness  145 the emotion of sympathy or empathy. To assess whether this is always true, we need to consider different examples, as the plausibility of this proposal may relate to the specifics of this situation. What about reverse cases, where I think I ought to do some act, but I don’t want (in the directed attention sense of desire)34 to do that act, perhaps because it would involve a significant sacrifice? Could it seem then that I ought to act? Consider a case where Bob’s elderly parents can no longer care for themselves, but have a strong aversion to going into a care home. Bob has a large enough house that they could move in with him, but that would impose significant constraints and burdens on him, and because of this he does not want them to move in. His aversion does not reflect well on him. He might wish that he were the sort of generous person who would not think twice about doing this for his parents, and focused more on the benefit to them rather than on the cost to him. But he is who he is, and it’s too late to change now. The question is, could it seem to Bob that he ought to care for his parents? I think it could. When he thinks about all of the things his parents have done for him, and the sacrifices they made, it may well seem quite forcefully that he has a duty of gratitude to look after them. Does this seeming presuppose an emotion? If it does it is not sympathy, as we are imagining Bob as focused on the costs to himself rather than the benefits to his parents. Must he have a feeling of gratitude to have this seeming? I don’t think so. As I am imagining the case the experience is more like the acquisition of an unwanted debt, rather than a genuine feeling of gratitude. When Bob considers the situation clearly, it presents itself as in­escap­ able, but this presentation does not flow from a feeling of gratitude. It is more like the feeling of having to pay a big tax bill, and gratitude doesn’t get in here. Nor could it flow from a feeling of guilt, as the guilt could only occur after he fails to look after them. Is this seeming explained by a sense of duty? That’s what it looks like. But the sense of duty is not a feeling or emotion. To act from duty is to do some act because you think it is the right thing to do, or as I would say, because of the ­reasons that you believe make it the right thing to do.35 So I do not think we need to assume that some emotion is engaged to explain Bob’s intuition that he ought to look after his parents. This picture of a cold Kantian agent is not attractive, and is certainly not a moral ideal.36 But the point is that it could seem to such a cold, but dutiful person that he ought to act in ways he would rather not. If that is correct it suggests that having certain moral in­tu­ itions need not presuppose emotion.

34  See Scanlon (1998: 39). 35  Stratton-­Lake (2000). 36  See for instance Michael Stocker’s famous discussion of the Kantian good-­willed individual visiting his friend in hospital solely out of duty (Stocker 1976: 462).

146 Fittingness

6.8 Conclusion I have argued that our knowledge of what is fitting to what can be accommodated within a rationalist epistemology of the kind that intuitionists endorse. Worries about whether propositions about fittingness can be adequately understood, whether they reduce down to intuitions about what is good or right, or whether the term ‘fitting’ is ambiguous, do not present any serious problems for our knowledge of what is fitting. I finished by considering whether we have to suppose that our intuitions of the all in fittingness of action must be different than our intuitions of the fittingness of attitudes, and argued that they need not be. Rather than supposing that there are two types of intuition here, a practical-­emotional one and a purely intellectual one, we need only assume that there is a single type of intuition that plays two roles—­justifying and explaining. In some cases the seeming justifies and explains belief: in others it justifies and explains action. And the phenomenology—­the felt pull do to the act that seems most fitting—­can be accommodated by the seeming tending to give rise to a felt attraction to assent or to act. We need not suppose that our intuitions of an act’s being all in fitting are emotional in a way that our intuitions about the fittingness of some attitude are not. It may be that unless we have certain feelings and emotional dispositions, certain moral propositions would not seem true to us, or at least their seeming would not have the force it does. But the case of the cold Kantian agent suggests that such seemings need not presuppose certain emotions.

References Andow, J. (2018) ‘Are Intuitions about Moral Relevance Susceptible to Framing Effects?’ Review of Philosophy and Psychology 9: 115–41. Audi, R. (1997) Moral Knowledge and Ethical Character. Oxford: Oxford University Press. Audi, R. (2019) ‘Understanding, Self-Evidence, and Justification’, Philosophy and Phenomenological Research, 358–81. Ballantyne, N., and Thurow, J. (2013) ‘Moral Intuitionism Deflated?’ American Philosophical Quarterly 50: 411–21. Bealer, G. (1996) ‘On the Possibility of Philosophical Knowledge’. Noûs 30 (Supplement: Philosophical Perspectives, 10, Metaphysics, 1–34. Bedke, M. (2008) ‘Ethical Intuitions: What They Are, What They Are Not, and How They Justify’. American Philosophical Quarterly 43 (3): 253–70. Bengson, J. (2015) ‘The Intellectual Given’. Mind 124(495): 707–60. Berker, S., (2022) ‘The Deontic, the Evaluative, and the Fitting’. In Fittingness. Edited by Howard, C. and Rowland, R. Oxford: Oxford University Press. Chappell, R. (2012). ‘Fittingness’. Philosophical Quarterly 62: 684–704.

Intuitions of Fittingness  147 Chudnoff, E. (2013) Intuition. Oxford: Oxford University Press. Dancy, J. (2003) ‘What Do Reasons Do?’ The Southern Journal of Philosophy XLI: 95–113. Dancy, J. (2014) ‘Intuition and Emotion’. Ethics 124(4): 787–812. Ewing, A. C. (1949) The Definition of Good. Abingdon: Routledge. Hanson, L. (2016) ‘The Real Problem with Evolutionary Debunking Arguments’. Philosophical Quarterly 67 (268): 508–33. Howard, C. (2018) ‘Fittingness’. Philosophy Compass 13: e12542. https://doi. org/10.1111/phc3.12542. Huemer, M. (2005) Ethical Intuitionism. Basingstoke: Palgrave Macmillan. Hurka, T. (2014) British Ethical Theorists from Sidgwick to Ewing. Oxford: Oxford University Press. Kauppinen, A. (2013) ‘A Humean Theory of Moral Intuition’. Canadian Journal of Philosophy 43 (3): 360–81. Markie, P. (2013) ‘Rational Intuition and Understanding’. Philosophical Studies 163(1): 271–90. McHugh, C., and Way, J. (2016) ‘Fittingness First’. Ethics 126: 575–606. Mogensen  A. (2015) ‘Do Evolutionary Debunking Arguments Rest on a Mistake about Evolutionary Explanations?’ Philosophical Studies 173: 1799–817. Nussbaum, N. (2001) Upheavals of Thought. Cambridge: Cambridge University Press. Parfit, D. (2011) On What Matters, Volume1. Oxford: Oxford University Press. Roeser, S. (2010) ‘Intuitions, Emotions and Gut Reactions in Decisions about Risks: Towards a Different Interpretation of “Neuroethics”’. Journal of Risk Research 13(2): 175–90. Ross, W. D. (1939) The Foundations of Ethics. Oxford: Clarendon Press. Ross, W.  D. (2002) The Right and the Good. Edited by P.  Stratton-Lake. Oxford: Clarendon Press. Scanlon, T.  M. (1998) What We Owe to Each Other. Cambridge, MA: Harvard University Press. Scanlon, T. M. (2014) Being Realistic about Reasons. Oxford: Oxford University Press. Shafer Landau, R. (2003) Moral Realism: A Defence. Oxford: Oxford University Press. Sinnott-Armstrong, W. (2006) ‘Moral Intuitionism Meets Empirical Psychology’ in Terry Horgan and Mark Timmons (eds), Metaethics after Moore, Oxford: Oxford University Press, 339–66. Smith, M. (1994) The Moral Problem. Oxford: Blackwell. Sosa, E. (2007) A Virtue Epistemology. Oxford: Oxford University Press. Stocker, M. (1976) ‘The Schizophrenia of Modern Ethical Theories’. Journal of Philosophy 73: 462. Stratton-Lake, P. (2000) Kant, Duty, and Moral Worth. London: Routledge. Stratton-Lake, P. (2016) ‘Intuition, Self-Evidence, and Understanding’. In Oxford Studies in Metaethics. Edited by Shafer-Landau, R. Oxford: Oxford University Press, 28–44.

SECTION T WO

FIT T INGN E SS, R E AS ONS , NOR MAT I V I T Y

7 Reasons and Fit Garrett Cullity

7.1  Introduction When we evaluate something by using words that contain suffixes like ‘-worthy’, ‘-able’, ‘-ful’, ‘-some’, ‘-ive’, or ‘-ing’ (words such as ‘praiseworthy’, ‘desirable’, ‘shameful’, ‘awesome’, ‘repulsive’, or ‘boring’) we assert a relationship between the object we are evaluating and the response-­type that precedes the suffix: between the praiseworthy and praise, the desirable and desire. ‘Fitness’ is one name that has been given to the relation.1 Adopting this usage, good things are fit objects of favour-­responses; bad things of disfavour-­responses.2 The relation of ‘fit’, so understood, seems to be central to our evaluative thought. We also talk of reasons for responses of action, thought or feeling—­reasons that are objective in the sense that whether you have such reasons is independent of whether you think so, and normative in the sense that what you ought to do, think, or feel is determined by what such reasons require of you overall. If we are looking to understand the relationship between value and normative reasons, then, a good question to ask is: What is the relationship between between reasons and fit? Until recently, a widespread and influential view has been that we should treat normative reasons as primitive, and appeal to them to explain the relation of fitness. Fitness-­relationships are a special class of reason-­relationships: a fit response to an object is one for which there is sufficient reason of the ‘right kind’. But recently, some philosophers have started to explore the opposite explanatory project, treating fitness as more basic and invoking it to explain what it is for something to be a normative reason.3 1  When A bears relation R to B, I shall say that R is the relation, and A’s bearing R to B is the relationship. 2  Versions of this thought belong to a tradition that travels from Aristotle (1999: 1139a24–30) through Aquinas (1964–1975: 1a2ae.24,1, 1a2ae.59,2) to Brentano (1969: Sect. 22, 23, 27, p. 74, and Appendix IX, Sect. 13)—see also Chisholm (1986: Ch.5); and Chisholm (1976); and thence to Moore (1903: pp. 204–50; Broad (1930: pp. 233–5); Ross (1939: pp. 271–89); Ewing (1947: 163–4); Lemos 1994: p. 15); Gaus (1990: Sect. 6); Zimmerman (2001: p. 199); and Svavarsdóttir (2014: p. 91). 3  See McHugh and Way (2016); Howard (2019); Danielsson and Olson (2007); Sharadin (2015: Sect. 3); LeBar (2013: Ch. 7); and Chappell (2012). Also relevant are Setiya (2013); and Thomson (2008: p. 121). Garrett Cullity, Reasons and Fit In: Fittingness: Essays in the Philosophy of Normativity. Edited by: Christopher Howard and R. A. Rowland, Oxford University Press. © Garrett Cullity 2022. DOI: 10.1093/oso/9780192895882.003.0007

152 

Fittingness

This chapter is a contribution to the latter, ‘fitness-­priority’ project. It defends the view that a normative reason for an action, thought, or feeling is a con­sid­er­ ation that is fit for determining what one will do, think, or feel. In a rough-­and-­ ready slogan: reasons for action are the facts we are fitly disposed to act on. My aim will be to lay out the attractions of this idea in what follows, showing that it provides an appealing explanation of how our normative reasons can have the different sources they do and, more generally, of the relationship between normative reasons and value. I begin with a brief description of these explanatory attractions, then set out the assumptions I am making about ‘fitness’, before providing a careful formulation of my proposal, explaining in just what sense a disposition to act on a fact can be a fit response to it, describing what I see as the proposal’s strengths more fully, and responding to some of the main objections it invites.

7.2  Two Explanatory Questions The most prominent attempts to explain the relationship between the reason- and fitness-­relations have been reasons-­priority views. Using ‘responses’ as a term cover­ing all the attitudes, feelings, and actions for which we have reasons, these have the general form: (R)  For object O to be fit for response R by person S is for there to be sufficient reason of the right kind for S to make R to O —with particular proposals of this form differing in their specifications of what it is for a reason to be of the ‘right kind’.4 4  One kind of reasons-­priority view dispenses with the ‘of the right kind’ qualifier, instead simply analysing fitness in terms of reasons and sufficiency: see Skorupski (2010: p. 83). The ten most prom­ in­ent attempts to explain the qualifier are: (i) object-­given reasons (Parfit 2011: vol. 1, p. 45; Piller 2006)—for objections, see Rabinowicz and Rønnow-­Rasmussen (2004) and Schroeder (2012); (ii) reasons for the correctness of R (Danielsson and Olson 2007)—for objections, see Hieronymi (2005); (iii) reasons appearing in the intentional content of R (this is considered and objected to by Rabinowicz and Rønnow-­Rasmussen 2004: pp. 414–21); (iv) reasons bearing on the truth of the content of R (D’Arms and Jacobson 2000a)—for objections, see Schroeder (2010: Sect. 2); McHugh and Way (2016: p. 597); (v) nonderivative and noninstrumental reasons (Stratton-­ Lake 2005)—for objections, see Schroeder (2012); (vi) reasons bearing on the question the answering of which is the making of R (Hieronymi 2005)—for objections, see Schroeder (2010: Sect. 2); Sharadin (2015: Sect. 2); (vii) reasons that are evidence that R conforms to a constitutive standard of correctness (Sharadin 2015)—the objections of Hieronymi (2005) and Schroeder (2010: Sect. 2) are also relevant here; (viii) reasons that make R appropriate independently of the value of making R (Raz 2009: Sect. 3—the challenge for a reasons-­priority version of this is to interpret ‘appropriateness’ in­de­pend­ent­ly of fitness);

Reasons and Fit  153 Reversing this order of explanation, fitness-­priority views treat fitness as the more fundamental relation and invoke it to explain what it is to be a normative reason. These have taken three main forms. Examples of the first form are the proposal that a reason for a response is an explanation of the fitness of the response (Chappell 2012), or evidence of its fitness (Thomson 2008; Sharadin 2015). These proposals begin with the class of responses that are fit responses to their objects, and then account for reasons as the facts that bear some specified relation to the fitness of those responses—­such as the relation of explaining, or supplying evidence. Their general form is: (F1)  For fact F to be a reason for S to make response R to O is for F to bear relation X to O’s being fit for R by S.5 Views of this kind place a restriction on the responses for which we have reasons: when a response is (evidently) not a fit response to its object, there is no reason to make it. So the fact that admiring an evil demon will prevent him from tormenting someone is not a reason for admiring him: it is at most a reason for causing yourself to admire him, or wanting to admire him.6 A second kind of fitness-­priority view rules differently on such cases: there are two kinds of reasons for a response such as admiration—­reasons that explain why the response is fit, and reasons that explain why wanting to make it is fit (Howard 2019). The latter are the ‘wrong kind of reasons’ to bear on whether the demon is admirable, but they are reasons for admiring him.7 This view has a dis­junct­ ive form:

(ix) reasons that are shared by necessarily anyone who is engaging in R, just because they are engaging in R (Schroeder 2010: Sect. 3—for objections, see Sharadin 2015: Sect. 2 and McHugh and Way 2016: n. 54); and (x) reasons that are neither provided nor enabled by facts about the additional consequences of having R (Rowland 2017: p. 216). 5  These proposals are analogues of the claims that reasons are explanations of ought-­facts (Broome 2004: p. 41) or evidence of ought-­facts (Kearns and Star 2008). 6 For the demon example, see Crisp (2000: p. 459), and Rabinowicz and Rønnow-­Rasmussen (2004). Those who hold that the ‘wrong kind of reasons’ for attitudes are not reasons for the attitudes themselves but instead for wanting to have them or causing oneself to have them include Gibbard (1990: p. 227); Parfit (2011: pp. 50–1 and App. A); Skorupski (2007: p. 10; Kolodny (2005: pp. 550–1); and Way (2012: p. 490). For the opposite view, see D’Arms and Jacobson (2000b), Rabinowicz and Rønnow-­Rasmussen (2004), Danielsson and Olson (2007), Hieronymi (2005), and Raz (2009). 7  The ‘wrong kind of reasons problem’ is a label for two different problems. The first is a problem for the ‘buck-­passing’ view that what it is for O to be good is for there to be sufficient reason to favour O (namely, the problem that cases of the kind just mentioned in the text appear to generate coun­ter­ exam­ples). The second is the problem of giving an account of what exactly distinguishes the con­sid­er­ ations that intuitively are ‘right kind of reasons’ from ‘wrong kind of reasons’ (note 4 lists ten prominent proposals). For the first usage, see D’Arms and Jacobson (2000b: pp. 734–5); Rabinowicz and Rønnow-­Rasmussen (2004: p. 393); Way (2012). For the second, see Hieronymi (2005: n. 2); Schroeder (2012: p. 461); Sharadin (2015: p. 3). For a clear account of the two problems, and further discussion, see Gertken and Kiesewetter (2017).

154 

Fittingness

(F2)  For fact F to be a reason for S to make response R to O is for F either: (i)  to bear relation X to O’s being fit for R by S; or (ii)  to bear relation X to R’s being fit for response R* by S.8 In Howard’s version, response R* is the attitude of wanting; another view of this form would be one in which R* is the action of causing to occur. The third kind of reasons-­priority view is illustrated by McHugh and Way’s (2016) proposal that to be a reason for a response is to be a premise in a fitness-­ preserving pattern of reasoning from fit responses to that response. This has the form: (F3)  For fact F to be a reason for S to make response R to O is for F to be fit to produce R-to-O through process P. In McHugh and Way’s version of (F3), process P is fitness-­preserving reasoning. Like (F1) and unlike (F2), they deny that the ‘wrong kind of reasons’ really are reasons: admiring the demon is not the output of a fitness-­preserving pattern of reasoning (although wanting to admire him could be).9 My own view is also a version of (F3); but it differs from McHugh and Way’s view. It says that a reason for you to make a response is a fact to which you are fitly responsive in determining what response you will make. Process P is not fitness-­preserving reasoning; it is S’s own response determination. Why take this idea seriously? One motivation for the fitness-­priority project in all its forms is that it makes room for a fitting attitude analysis of goodness (according to which an object’s goodness is its fitness for a pro-­attitude), while avoiding problems for the ‘buck-­passing’ view that analyses goodness in terms of reasons.10 While that does seem to me an attraction of fitness-­priority views, my emphasis will be on two other attractions of the particular proposal I am going to develop. It offers promising answers to two explanatory questions. As we have seen, fitness-­priority theorists take different sides on the question whether the instrumental advantages of an attitude such as admiration are ­reasons for admiring. But suppose we turn now from the attitude of admiration to the action of praise. Here, it cannot seriously be doubted that reasons for this action could come from two sources. They could come from facts about the fine qualities that make someone praiseworthy; or they could come from facts about the usefulness of the action of praise in furthering some other worthwhile end. If praising a torturer will prevent him from torturing your children, that fact would clearly be a reason for performing the action of praising him—­and not just for 8  Howard (2019: p. 232). 9  However, they do allow for reasons for some responses that are not fit: see McHugh and Way (2016: p. 589). 10  McHugh and Way (2016: pp. 575–7).

Reasons and Fit  155 wanting to perform that action, or causing yourself to do so. This prompts the following question. Is there an account of what a reason is that explains how ­reasons can have these two sources—­how reasons for praising can come either from the praiseworthy qualities of the object of praise, or from the instrumental advantages of praising? This question invites the following simple answer. What the fine qualities in the first case and the instrumental advantages in the second have in common is this: they are facts it makes sense to be responsive to in deciding whether to praise someone. If ‘making sense’ here refers to the fitness-­relation, we have the outline of an account of the relationship between reasons and fit that answers our question. What a reason for action is is a fact that is fit for determining whether one performs that action. That is how facts of these two different kinds come to qualify as reasons. The second explanatory question is this. A good account of the relationship between the domain of normative reasons and the domain of value should explain an often-­remarked difference between them.11 When we see ourselves as having a reason to react to an object in some way, we go further than just seeing the object as good or bad. You could register an object’s goodness without thinking it has any relevance to you: seeing yourself as having a reason to respond to it is now acknowledging that relevance. A normative reason has a recommending force—­a directiveness—­that goes beyond what you attribute to an object when you evaluate it. What account is it possible to give of this difference? Here, too, the view I am proposing can help us. On this proposal, your reasons are the facts that fitly bear on your own self-­direction—­facts that you are fitly responsive to in determining what you will do. Experiencing reasons as directive is appreciating their relevance not just to your ranking of objects as good or bad, but to your own self-­ determination in responding to them. So the view I am proposing—­the view that reasons are facts that are fit for response determination—­has this attraction: it promises a unified account of what it is for a fact to be a reason that supplies a simple and plausible answer to these two questions. And at least on the face of it, this seems to give it an advantage over the other views just surveyed. In relation to the first question—‘Is there an account of what a reason is that explains the two sources of reasons for actions such as praise?’—some of these views will apparently struggle even to recognize the explanandum; and the others will struggle to explain it. Versions of (F1) and (F3) that deny that there are reasons to admire the unadmirable demon seem equally committed to denying that there are reasons to praise the unpraiseworthy torturer. The disjunctive view (F2) does recognize that reasons have two different sources, but while it asserts that this is so, it doesn’t give an account of what reasons are that explains why this is so. And most proponents of the ­

11  Dewey (1998: pp. 318–19). See also Darwall (2006: p. 280); Svavarsdóttir (2014: p. 99).

156 

Fittingness

reasons-­priority view (R) are reasons-­primitivists, who decline to offer any analysis of what a reason is, depriving them of an answer to the first question. It is true that (R) does not logically require reasons-­primitivism: you could combine it with some further analysis of reasons. But you would not be able to give the simple answer just described: the closest you would be able to get would be to agree that reasons for an action are facts it makes sense to take account of in deciding whether to perform it, while declining to identify ‘making sense’ with fitness, instead introducing a third kind of ‘making sense’ relation whose relationship to both reasons and fitness would then need to be explained.12 Likewise, in relation to the second question—‘Is there an account of what a reason is that explains the directiveness that reasons have but fitness need not have?’—(R) will apparently struggle to recognize the explanandum, since it treats all fitness-­relationships as reason-­relationships; and other fitness-­priority views of forms (F1–3) will struggle to explain it. Although they allow for fitness-­relationships that are not reason-­relationships—­reasons are explanations of fitness, or evidence of fitness, or premises in fitness-­preserving patterns of reasoning—­it is not clear how those ways of being related to fitness explain the directiveness of normative reasons. We will need to look more closely at these two issues: I return to them below (Section 7.5). The rival views (R) and (F1–3) cannot be dismissed so quickly: there are avenues of response that have to be considered. Moreover, even if my view has a better answer to these two questions than its rivals, they might be superior in other respects, making them better overall. However, this does at least suggest that a fitness-­priority view of the form I am proposing is worth closer examination. The rest of the chapter seeks to work out this proposal in satisfactory detail. First, I spell out my assumptions about the relation of fitness, by reference to which I propose to analyse reasons. Then a careful formulation of the proposal is presented. With that in hand, I return to the two explanatory challenges it helps us to meet, and point to a third source of support for it.

7.3  ‘Fit’ The fitness-­priority project explains what it is for something to be a normative reason by invoking the relation of fitness. As I shall understand it, it makes a claim about the relationship between fitness and reasons themselves, and not about what it is to have thoughts about them—­about our concepts of them. The fitness-­priority

12  Can proponents of (R) turn the tables here? Doesn’t (R) have the mirror-­advantage of being able to answer the question, ‘Why are facts of these two kinds facts you are fitly responsive to in deciding whether to praise?’ by saying they are both reasons? I respond to this objection in Section 7.6.

Reasons and Fit  157 project offers a compositional explanation of what it is for something to be a normative reason that invokes the relation of fit: a property-­reduction. What is fitness? My answer to this is stipulative, and ostensive: it is the relation we refer to in evaluative uses of the suffixes I began with.13 This usage makes us all familiar with the fitness-­relation itself, but ordinary language does not give us an obvious name for it; so philosophers have had to co-­opt other terms to serve this purpose: ‘fittingness’,14 ‘fitness’,15 ‘correctness’,16 ‘worthiness’,17 ‘worth’,18 ‘merit’,19 ‘appropriateness’,20 ‘suitability’,21 ‘requiredness’,22 ‘aptness’,23 and so on. None of these terms is ideal, since they all have other uses in ordinary discourse, and this can be misleading. For example, sometimes you should not praise a praiseworthy action: it might be impertinent, or embarrassing.24 It would then be quite natural to express this overall verdict in ordinary speech by saying that praise would not be fitting or appropriate in the circumstances—­but when they are used in that verdictive way, ‘fitting’ and ‘appropriate’ do not denote the ‘-worthiness’ relation in praiseworthiness.25 This makes ‘fitness’, to my ear, a better choice: it sounds natural to say in such a case that the action is fit for praise although I should not actually praise it. However, I should emphasize: this is simply a matter of finding a convenient noun to serve as a name for a familiar but nameless relation.26 In calling this relation ‘fitness’, I am not saying that the relation obtains in virtue of some sort of structural match between the properties of the object and the properties of the response (or between the-­making-­of-­the-­response-­to-­the-­object and human nature).27 Nor am I subscribing to Clarke and Balguy’s view of ‘eternal and necessary fitnesses’ and all of the epistemological worries that provokes.28 I shall use the phrases ‘R is a fit response to O’, ‘it is fit to make R to O’, and ‘O is R-worthy’ as equivalent ways of asserting the same relationship. About the fitness-­relation, I make two unargued assumptions. The first is that the various evaluative terms I began with are used to make claims about the same relation. If theorizing with this assumption runs into difficulties, we might be 13  We also refer to it in some non-­evaluative uses of those suffixes (‘surprising’, ‘noteworthy’), but not in others (‘vulnerable’). 14  Ewing (1947: pp. 132–3); Broad (1930: pp. 164–5, 219). 15  Cudworth (1996); Clarke (2002); Price (1974); Ross (1939: p. 52). 16  Brentano (1969: p. 18). 18  Svavarsdóttir (2014: p. 91). 17  Ross (1939: p. 279). 19  McDowell (1985: pp. 118–20). 20  Moore (1903: pp. 204–5); Gaus (1990: Sect. 6). 21  Ewing (1947: p. 165). 22  Chisholm (1986: Ch. 5); Zimmerman (2001: p. 86). It may be better to read these authors as proposing a deontic analysis of (the relation I call) fitness, rather than an alternative name for it. 23  Gibbard (1990: p. 6). 24  See the discussion of ‘the moralistic fallacy’ in D’Arms and Jacobson (2000a). 25  As I read them, several of the authors cited above—­including Ewing, Broad, and Moore—­do not apply this distinction: often, their use of ‘fitting’ or ‘appropriate’ is the verdictive one. 26  If this relation is so important to evaluative thought, how come our language has no name for it? Because languages are shaped for practical purposes. Our evaluative practice can get by fine without a word for this relation; it is only those of us who theorize about it who need one. 27  Cumberland (2005: p. 484); for discussion, see Irwin (2008: Sect. 532). 28  Clarke (2002); Balguy (1976).

158 

Fittingness

forced to abandon it; but it is implicit in the literature, and I retain it. One ground for doubting this assumption is the following. Javier’s praiseworthiness sounds like a relationship between only two relata: Javier and praise. When I say that Javier is praiseworthy I am not saying that he is praiseworthy by any type of person in particular. But desirability is different. A ticket to a public dissection can be desirable for you but not for me. Clearly desirability involves three relata: the desirable object, desire, and those for whom it is desirable. If praiseworthiness involves a two-­place relation and desirability a three-­place relation, they cannot be the same relation. So in order to retain my first assumption, I add a second: fitness is a three-­place relation, with an object-­place, a response-­place, and a responder-­place.29 In some cases, there is no restriction on occupants of the responder-­place: Javier is praiseworthy-­by-­anyone. There is no need to indicate that in ordinary speech, so only two of the three relata feature in the idioms we actually use. But the relation itself is a three-­place relation. To these assumptions, I add four disclaimers. First, having introduced fitness ostensively, I leave open whether it has a further analysis. Deontic,30 alethic,31 evaluative,32 and rationalist33 analyses have all been proposed; some treat it as irreducible.34 I remain neutral about that: my view is that fitness is prior to normative reasons, but leaves open whether something else is prior to fitness. Secondly, our topic—­the relationship between reasons and fitness—­is separate from the question which substantive fitness-­relationships there are: which objects are desirable, which actions are blameworthy, and so on. And, thirdly, it is also separate from the epistemological question what warrants such judgements. I assume that we do make such judgements—­we do so whenever we think something is desirable—­but the question of what justifies us in doing so is not tackled here. In particular, I do not claim that judgements about reasons can be justified by appealing to epistemically prior judgements about fit. The final disclaimer concerns whether actions are among the responses that can enter fitness-­relationships. Opinions divide on this.35 Accepting it is encouraged by the view that we find relationships of fitness wherever we find the intuitive distinction between the ‘right kind of reasons’ and the ‘wrong kind of reasons’ for a response: actions such as praising, as we saw, are subject to that distinction. 29 One might consider emulating T.  M.  Scanlon’s treatment of reasons, and adding a fourth, circumstance-­place: (Scanlon 2014: pp. 31, 37). But for our purposes, we can treat relevant variations in the circumstances as variations in the object of the fit response. 30  For accounts of fitness in terms of requirement, see Ewing (1947: Ch. 5); Chisholm (1986: pp. 52–3); Rabinowicz and Rønnow-­Rasmussen (2004: pp. 391–2). For an account in terms of permission, see Shah and Velleman (2005). 31  D’Arms and Jacobson (2000b); Tappolet (2011: pp. 119–20). 32  Johnston (1989: p. 158). 33  Anderson (1993: Ch. 1); Gibbard (1990: Ch. 1). 34  McHugh and Way (2016). For criticism of the four analyses just referred to, see Svavarsdóttir (2014: Sect. IV). 35  This is accepted by Schroeder (2010: pp. 32–3), and Sharadin (2015: pp. 22–4); but denied by Raz (2009: p. 14), Hieronymi (2013: pp. 117–18), and Heuer (2011: pp. 170–1).

Reasons and Fit  159 On the other hand, opposition to it is encouraged by the view that fitness is always an attitude-­object relationship in which the attitude correctly represents a property of the object.36 Desire is fit when it correctly represents its object as desirable; admiration, when it correctly represents its object as admirable. But since actions do not represent properties of objects, they are not candidates for assessment as fit or unfit. Although I am more sympathetic to the first view than the second, my fitness-­ priority proposal can be formulated in a way that makes room for both of them, so I shall do that. Proponents of the second view will want to insist that Javier’s praiseworthiness is his fitness not for the action of praising him but instead for the attitude of positive evaluation expressed by that action. To allow for either view, I will therefore say that an action such as praise ‘is or expresses’ a fit response to an object: the first disjunct allows for the first view and the second disjunct the second.

7.4  Fit Response Determination So far, we have a slogan—­reasons for action are the facts we are fitly disposed to act on—­and an indication of why it is worth trying to turn it into a proper proposal. Now we need to do that. We need a formulation that covers more than just reasons for action: an adequate account of normative reasons must cover the ­reasons we have for beliefs, desires and emotions too.37 But at the same time, we need to retain the idea that the responses for which we have reasons are responses that we ourselves determine. There is a difference between a piece of behaviour, or a feeling, or a thought, that just happens to me, and something that I do or feel or believe for a reason. I shall mark this difference by saying that cases of the latter kind are cases of response determination—­where what I mean is that these are cases in which my response is determined by me, rather than merely being produced by a causal process that culminates in something that happens to me. Of course, it is a central question in the philosophy of action, and also of emotion and of belief, just what account should be given of the agential element that must be present in order for a response to be attributable to me—­to be something that I do, or feel, or believe.38 I am not going to offer my own account of this: what I am going to require is simply that the correct account upholds two standard assumptions. One is that a response that I make is guided by my awareness of the situation or other object to which I am responding. (Explaining exactly what kind of ‘guidance’ is necessary 36  D’Arms and Jacobson (2000a: pp. 66, 78). For objections, see Howard (2018: Sect. 4). 37  For one line of resistance to this, see Maguire (2018). 38  For discussion, see the contributions to O’Brien and Soteriou (2009).

160 

Fittingness

for agency is one of the main tasks that the correct account needs to acquit.)39 The other is that this must be interpreted in a way that covers unreflective responses. The actions, emotions, and beliefs that are attributable to me need not be produced by any episode of deliberation or reasoning. Catching a ball can be something you do, guided by your awareness of the trajectory of the ball and the state of the game, even if it remains under the control of the brain’s ‘System 1’.40 Within the broad class of cases of response determination—­those in which I determine my response—­I then want to focus on a subset of cases that are of special interest. These are cases in which, in making a response, I am guided by my awareness of the fitness of the response I am making. I take it that this condition, too, is routinely met in unreflective cases. Admiring Javier can be a response to your awareness of his admirable qualities without being the outcome of some episode of deliberation about whether to admire him. I do not claim that this further condition is always met whenever we determine our own responses: we can akratically do or believe or feel things without thinking our responses are fit responses to their objects. But it is often met. With this in hand, I can now provide a more careful statement of my proposal. Normative reasons are the facts that are fit for determining responses in the way just described—­that is, through my awareness of the fitness of the response I am making. This gives us a two-­condition account of what it is for a fact to be a normative reason for me to make a response, with two references to fitness. The first condition is that the fact makes the response a fit response to an object; the second is that my awareness of that is fit to guide my response determination. To complete the account, we need to make one further refinement to the first condition. Earlier, we noticed the two ways in which reasons for praising can arise. They can come from the qualities that make someone praiseworthy. Then the response (praising) for which one has a reason is or expresses a fit response to its own object (the praiseworthy person). Or, when that is not so, they can come instead from what is achieved by praising—­preventing torture, for example. But now we can add a further point. In the latter case, there is still a fitness-­relationship to which one’s reason for action is connected, albeit less directly. Torture is fit for aversion: fit either for actions of avoidance and prevention (if actions can be fit) or for the aversive attitudes those actions express (if not). And in praising the torturer, you are doing what prevents torture. So there is still a description of your action of praising—­namely, as doing what prevents torture—­under which it is or expresses a fit response to an object (namely, torture). Thus, the two ways in which reasons for praising can arise correspond to two ways in which reasons can 39  See Mele (2003), esp. Ch. 8. 40  On ‘System 1’, see Kahneman (2011: Part I). For the details of the ball-­catching case, see Dienes and McLeod (1993).

Reasons and Fit  161 be connected to fitness. The connection can be direct, when one’s praise is or expresses a fit response to what one is praising; or indirect, when one’s praise instantiates some other fit response. I now generalize this claim. Reasons are always connected to fitness-­relationships, either directly or indirectly. This, too, is part of what it is to be a reason. Reasons are the facts that, because of their connection to the fitness of a response, are fit for determining whether one will make that response. Phrasing these two conditions more carefully, my proposal is this: Fit Response Determination (FRD).  For fact F to be a reason for person S to make a response of type R to object O in circumstances C is for it to be true that: (i) F makes it the case that, were S to make R-to-O in C, S’s making R-to-O would meet a description under which it is or expresses a fit response to an object; and (ii) S’s being disposed towards the determination of R-to-O in C is a fit response to (i). For F to be a reason against R is for it to be a reason for not-R. A more accurate (if less catchy) slogan is this: normative reasons are the fitness-­making facts that are fit for our response determination. Condition (ii) is phrased in terms of dispositions towards determining a response. This allows it to account for the very common (perhaps ubiquitous) cases in which we have reasons that count in different directions. I may have competing reasons to watch a movie this evening or to prepare my lecture instead. The facts about the quality of the movie are facts that make it worth watching, and my awareness of those facts fitly disposes me to determine my action in one direction, by watching the movie; but at the same time the facts about my professional responsibilities make the lecture worth preparing properly, and my awareness of those facts can also fitly dispose me to determine my action in the other direction, by preparing my lecture. One might worry about whether there really are any fitness-­relationships between objects of form (i) and responses of the form mentioned in (ii). Earlier, we noted the controversy over whether actions can enter fitness-­relationships. But (FRD) requires that this can be true of ‘being disposed towards the de­ter­min­ ation of response R’. Why agree to that? However, the answer to this is that both of the rival tests that we identified for whether a response enters fitness-­relationships—­both the one that rules actions in and the one that rules actions out—­are met in this case. According to one test—­the one that rules actions in—­we find fitness-­relationships wherever, intuitively, there is a distinction to be made between the ‘right kind of reasons’ and the ‘wrong kind of reasons’ for a response. We do find that distinction here. You could be given a powerful incentive to be disposed to determine your responses in a

162 

Fittingness

random way: perhaps you are in a ‘Schelling case’ where acquiring irrational dispositions is the best way to render yourself invulnerable to others’ threats.41 Intuitively, there is a distinction between the considerations that are fit to guide your response-­determining dispositions and those that you could have the ‘wrong kind of reason’ to allow to guide them. The other, stricter test—­the one that rules actions out—­was that a response can only enter a fitness relationship when it correctly represents a property of its object. But although actions fail this test, the cases of response determination covered by (FRD)’s condition (ii) will pass it. These are cases in which my disposition to determine a response is guided by my awareness of (i)—that is, my awareness of the fitness of the response or what it expresses (under some description). In response determination of this kind, I am representing my response or what it expresses as fit (under some description). So when the response I am determining is fit, this representation will be correct, and my determination of the response will also pass the second test for participating in relationships of fitness. To appreciate how the two conditions of (FRD) work, we can check some applications. First, we can apply it to the two kinds of reasons for praise. If Javier has done something fine, then that fact can make it the case that praising him meets a description (namely, praising a praiseworthy person) under which this is (or expresses) a fit response to him, so condition (i) is met; and my awareness of (i) can fitly dispose me to praise him, so that condition (ii) is met as well: then this fact qualifies as a reason for me to praise him. But also, the fact that it will prevent my children from being tortured qualifies as a reason to praise the torturer, since (i) that fact makes it the case that praising him meets the description doing what prevents my children from being tortured, which is (or expresses) a fit response to the prospect of their torture, and (ii) my awareness of that fitly disposes me to praise the torturer. So this proposal accommodates reasons of both kinds for the action of praising. It does not require that actions themselves are fit responses to objects; but it does require that every action for which you have a reason has a description under which it is or expresses your fit response to an object.42 This may seem difficult to satisfy for those actions we refer to using intransitive verbs: sitting still, for example.43 That has no obvious object, but we could have a reason to do it. However, once we identify the reason, we will see that condition (i) is satisfied. Suppose you have been taken hostage, and sitting still is the best way not to annoy the hostage-­taker. Then that fact makes it the case that sitting still meets the 41  See Schelling (1960). As before, there will be scope for a debate here over whether facts about threat-­invulnerability are reasons to have the irrational response-­dispositions, or only to cause yourself to have them. 42  If you hold the more relaxed view described in Section 7.3, the disjunct ‘or expresses’ can be dropped from (FRD). 43  Compare Gertken and Kiesewetter (2017: p. 8).

Reasons and Fit  163 description reducing the likelihood of harm, and under that description it is or expresses a fit response to the prospect of harm. (It is a fit response if actions can themselves enter fitness-­relationships; if not, it still expresses the fit response of preferring a lower to a higher likelihood of harm.) So (FRD) applies readily to reasons for action. It also applies readily to the evidence that gives us reasons for belief: (i) facts that are evidence of the truth of a proposition can make it the case that increasing the credence you attach to that proposition meets the description proportioning your credence to the likelihood that the proposition is true, and (ii) your awareness of this can fitly dispose you to increase the credence you attach to the proposition. It applies readily to reasons for evaluative attitudes such as admiration: (i) facts about Olga’s resolute nature can make it the case that your admiration meets the description admiring Olga’s resolve, which is a fit response to her admirable qualities, and (ii) your awareness of this can fitly dispose you to admire her. And it applies readily to reasons for emotions such as fear: (i) facts about danger can make it the case that something is fearsome, and (ii) your awareness of this can fitly dispose you to be afraid. More generally, (FRD) generates normative reasons in relation to any type of response that is subject to one’s own determination. This seems to me a significant strength of the account. For example, it accommodates the reasons we can have for and against forms of response determination themselves. One way of determining responses of action is through deliberation. But when deliberation would be time-­wasting or distracting or dangerous, that fact is a reason not to deliberate. If I am walking over a crocodile-­infested river on a rickety rope-­bridge, deliberating about the danger may make it likelier that I will fall off. (FRD) neatly allows for this, because deliberation itself is one possible substitution-­instance for R. When F is the fact that deliberating about the danger will make it likelier that I  will fall, this does qualify as a reason against deliberating, since (i) this fact makes it the case that deliberating about the danger meets the description doing what increases my chance of getting eaten and (ii) a fit response to this is my being disposed not to deliberate about the danger. So this fact qualifies as a reason against deliberating, as it should.

7.5  Two Explanations We now have a properly formulated proposal that supersedes our initial slogan. It needs to be defended against further objections. But first, let us return to the two explanatory challenges we began with and examine how (FRD) addresses them. The first challenge was to give an account of what a reason is that explains the two different sources of reasons for actions such as praising. We have just seen how (FRD) does that, accommodating not only reasons for response-­types that

164 

Fittingness

bear fitness-­relations to their own objects, but also reasons for those that do not. Facts of either kind can make a response (or the attitude it expresses) fit under some description, and our awareness of this can fitly dispose us to make that response. In Section 7.2, I claimed that some rival views (of forms (F1) and (F3)) ­apparently struggle to allow that there are indeed two sources of reasons for actions of praising, and others (of forms (F2) and (R)) will struggle to explain this. However, we might now wonder how forceful that objection is. If condition (i) of (FRD) succeeds in capturing both sources of reasons, it can be used as a template for constructing versions of the rival views that do so too—­for example, a version of (F1) on which reasons are explanations of why a response is fit under some description; or a version of (F3) on which reasons for a response are premises in fitness-­preserving patterns of reasoning to that response under some description. And while it may not be obvious how views of forms (F2) and (R) can by themselves explain why reasons have the two sources they do, nor is it obvious that they cannot be supplemented in a way that explains this. So on its own, the success of (FRD) in meeting the first explanatory challenge has limited force as a case for preferring it over its rivals. Where the case becomes stronger is when it is combined with the second challenge, which we can now examine more closely. This was to explain an often-­noticed difference between the domain of normative reasons and the domain of value. To evaluate something is to assess its su­per­ ior­ity or inferiority in a dimension. Thought about normative reasons goes further: normative reasons are directive in a way that value need not be. One way to approach this difference is via Dewey’s observation that ‘[t]here is an intrinsic difference, in both origin and mode of operation, between objects which present themselves as satisfactory to desire and hence good, and objects which come to one as making demands upon his conduct which should be recognized.’44 It is true, as we have noted, that a normative reason need not get as far as making a demand that should be recognized: it might be outweighed. But normative ­reasons, as the facts that can count towards such demands, have a force that value-­facts need not have. Perhaps the directions our reasons give us can be of an invitational rather than an imperative kind;45 but even so, they go beyond something’s simply being either good or bad. It is important not to overdraw this contrast. The norms of fitness that govern our responses to the good and the bad set standards that we cannot simply opt out of. If I hate what is lovable or admire what is despicable, then the unfitness of my responses is a defect in me. But the world and its history are full of lovable and despicable things, most of them remote from me. Their positive and negative 44  Dewey (1998: pp. 318–19). See also Darwall (2006: p. 280); Svavarsdóttir (2014: p. 99). 45  See Dancy (2006) on ‘enticing reasons’.

Reasons and Fit  165 value is not enough by itself to establish the kind of connection to me that gives me a normative reason to love or despise them. We see this difference reflected in the common view that reasons are capacity-­ dependent, whereas value is not.46 If I become incapable of fitly responding to something’s goodness, that does not make its goodness disappear, but my in­cap­ acity to make a response can deprive me of a normative reason to make it. So if fitness is the relation that good objects bear to favour-­responses then fitness, unlike normative reasons, is capacity-­independent. Suppose Javier and I are the last two people left alive; when I suffer a head injury, Javier nurses me with great care and humanity. If my injury has deprived me of the capacity to appreciate what he is doing, that would not extinguish the goodness of what he does. His actions are still praiseworthy and admirable, although there is no one able to praise or admire him. Indeed, his actions are especially worthy of gratitude by me, despite my incapacity to make that response. This is one of the respects in which my injury is an impairment: I cannot make the responses that his care fitly elicits. However, my incapacity to make a response does deprive me of normative ­reasons to make it. It becomes false that I ought to respond to Javier’s actions with praise or gratitude, once I lose the capacity to do so. We saw earlier how other views of the relationship between reasons and fit struggle to account for this. (R) implies that all fitness-­relationships are reason-­ relationships; and (F1–3), although they do not carry that implication, relate ­reasons to fit in a way that does not explain this difference between normative reasons and value. (FRD) gives us the following explanation. Reasons are the fitness-­ making facts that fitly bear on your own self-­determination. By itself, the fitness of a response to some object need have no such bearing. But your reasons are the facts that do bear on how your own self-­direction is fitly exercised—­the facts to which you are fitly sensitive when you determine what your own responses to the world are to be. Experiencing reasons as directive is the experience of that. This explains why reasons are capacity-­dependent. If you lack the capacity to make a response, that response is not available for your own determination, so no fact fitly bears on your determination of that response. Since condition (ii) cannot be satisfied for such responses, you cannot have reasons to make them. This is the main respect in which my proposed fitness-­priority view differs from McHugh and Way’s version of (F3)—the view that reasons are premises in fitness-­preserving patterns of reasoning. They rightly want to avoid suggesting that the existence of a normative reason depends on whether it features in some psychological episode of reasoning, and therefore emphasize: ‘we do not mean

46  Compare Streumer (2007) and (2016). Arguably, for F to be a reason for S to make response R, S must have two different capacities: both the capacity to make R and the capacity to see F as a reason to make R. For further discussion of the sort of ‘capacity’ that is relevant to the possession of a reason, see Cullity (2016).

166 

Fittingness

that anyone must actually perform this good reasoning. Patterns of reasoning are abstract.’47 But this makes their version of (F3) markedly different from mine: theirs detaches reasons from particular agents; mine ties them to an agent’s capacities for response determination. Where they analyse reasons in terms of their place within reasoning, conceived of as an abstract structure, I instead want to explain the reasons for you to make a given response as the facts that are fit for your determination of your own response, and as therefore dependent on your having the capacity to determine it. It may be tempting to resist this line of thought by distinguishing between whether there is a reason to make a response, and whether it is a reason for me to make it. Facts about my capacities, one might insist, are relevant to the latter, not the former. When I lose the capacity to praise Javier, that does not affect whether there is a reason to praise him, but only my access to that reason.48 So if (R) and (F1–3) are rewritten as claims about when there is a reason to make a response, the capacity-­dependence I attribute to reasons will not apply. However, there is a two-­part reply to this. First, it has changed the subject. Our question concerns the relationship between normative reasons and fit—­where a normative reason is one that bears on whether I ought to make the response for which it is a reason.49 Javier’s fine qualities cannot be the sorts of reasons that bear on whether I ought to praise him, if I lack the capacity to do so. But secondly, if we insist on distinguishing between whether there is a (non-­normative) reason for a response and whether it is a normative reason for me to make it, we then invite the question what is the relationship between these two kinds of reason—­ and (FRD) readily supplies an answer. The non-­normative reasons there are are the facts that meet condition (i). Of these, the reasons that are normatively dir­ect­ ive for my responses are the facts that also satisfy condition (ii). Again, moves can be made on behalf of the rival views to mirror this treatment of reasons as capacity-­dependent. One could simply add ‘provided S has the capacity to make R’ to each of (F1–3). However, the challenge is not just to assert the capacity-­dependence of normative reasons but to explain it. (FRD) is an account of what reasons are that explains this: as far as I can see, this is a merit that is not matched by the other accounts. Notice that the two explanatory challenges I have described relate to two op­pos­ite ways in which reasons and fitness come apart. The first points to the way in which the class of responses for which we have reasons is broader than the class of fit responses: you can have a reason to praise someone who is not fit for 47  McHugh and Way (2016: p. 586). 48  See e.g. Crisp (2006: p. 43); Heuer (2010); Gardner (2004: pp. 55–6); Millgram (1996); Schroeder (2007: pp. 165–7); and, for discussion, Wedgwood (2015: p. 124). 49  I do not assume that there is a single normative ‘ought’: the claim is that all normative oughts are capacity-­sensitive. For example, it is clear that neither the fact-­relative ‘ought’ that we use in giving advice nor the evidence-­relative ‘ought’ that we use in evaluating a person’s rationality apply to impossible responses. On the different uses (Parfit says ‘senses’) of ‘ought’, see Parfit (2011: Sect. 21); and Broome (2013: Ch. 2).

Reasons and Fit  167 praise. But the second points to the way in which it is narrower: my incapacity can deprive me of a reason to praise praiseworthy Javier. We have seen that ­modifications can be introduced to other views to allow them to accommodate both of these features. However, where (FRD) goes further is in offering a unified fitness-­priority account of what normative reasons are—­the fitness-­making facts that are fit for our response determination—­that generates a straightforward explanation of both features. It is true that (FRD) has a complex, two-­part structure. However, this two-­part structure is not simply an ad hoc attempt to get the extension of the class of ­reasons right. It plausibly reflects the pair of capacities that we exercise in being responsive to normative reasons: first, the capacity to appreciate the fitness (under some description) of the responses we could potentially make; and secondly, the capacity to be guided by that awareness in determining the responses we do make.50 Reasons are the facts to which we are fitly responsive in that pair of ways. (FRD)’s condition (i) specifies the class of facts that potentially bear on my response determination; condition (ii) then restricts my normative reasons to those facts that actually are fit to do this. This is the third significant source of support for my proposal, alluded to at the end of Section 7.2. (FRD) gives an account of what normative reasons are that faithfully reflects the structure of our responsiveness to such reasons. We saw in Section 7.2 that opinions are split over whether the ‘wrong kind of reasons’ are indeed reasons. I do not think that there is an obviously correct view about this that should guide our theorizing about reasons: we need a theory of reasons, defended on other grounds, to tell us what to say. (FRD) carries the following corollary. Suppose you could admire at will: you could perform the mental action of admiring the threatening demon just as readily as you could perform the speech-­act of praising the threatening torturer. Then you would have the same reasons for making the unfit response in both cases. However, most of us lack the capacity to admire at will. So, since normative reasons are capacity-­dependent, most of us would lack a reason to admire the demon; we would only possess ­reasons to want to do so and to cause ourselves to do so. Whether a fact is a reason for us to make a response depends on whether we are able to guide the de­ter­ min­ation of our response through the awareness of that fact, and not on whether it is the ‘wrong kind of reason’.

7.6  Objections (FRD), then, has some significant attractions: it provides a single unified treatment of the two explanatory challenges that faithfully reflects the structure of reasons-­responsiveness. I acknowledge that this leaves open the possibility that 50  Compare Korsgaard (1996: pp. 92–3).

168 

Fittingness

there are other respects in which rival views may be superior. However, rather than investigating that possibility here, I turn now to considering some more direct objections. First, we can ask whether there are any counterexamples showing that (FRD) gets the extension of the class of reasons wrong. But secondly, even if there is no such counterexample, the reductive claim it makes could still be mistaken. Potential counterexamples to (FRD) belong to two possible classes. One comprises non-­reasons that it misclassifies as reasons: false positives.51 Here, one might worry about testimonial facts. The fact that some other knowledgeable person assures me that F is a reason for action A is something that could properly lead me to decide to do A, but the fact about her testimony would not thereby acquire the status of a reason for A—­that role would be played by F itself (if the testimony is accurate). The other class of potential counterexamples comes from reasons that the proposal fails to classify as such: false negatives.52 Are there not some reasons which you cannot fitly act on, because it is impossible to act on them? For example, you can have reasons to rectify your own ignorance. Suppose you do not know that your flight leaves at 2 o’clock. Can’t I be referring to this fact when I advise you that you have a reason to look up the flight schedule?53 But deciding to look up the flight schedule could not be a fit response to your awareness of that fact: your awareness of it would remove the ignorance you have a reason to rectify. We can take these cases in turn. The apparent false positive is a good counterexample to the initial slogan, but not to the more carefully formulated proposal. (FRD) does not claim that whenever a fact fitly bears on my determination of a response it is a reason for the response. To be a normative reason for the response, a fact must (i) make it the case that that response is fit, under some description, and (ii) the determination of the response must itself fitly respond to that. But in the testimonial case, condition (i) need not be met. The fact about the other person’s testimony may be evidence that there is a description under which my action is (or expresses) a fit response to an object, but it need not make it the case that this is so. So the testimonial fact can be evidence that I have a reason for acting; but need not itself be a reason—­just as the objection supposes. This reply draws attention to something I have not attempted to explain. I have offered no analysis of the making-­the-­case relation referred to in condition (i). The question of what analysis (if any) there is of that relation lies at a level of detail below what I am attempting here. Different treatments of that question are 51  For help in thinking about false positives, I am grateful to Daniel Whiting, Conor McHugh, and Alex Gregory. 52  For help in thinking about false negatives, I am grateful to Krister Bykvist and Jonas Olson. 53  Schroeder (2007: p. 165) gives the related example of your enjoyment of surprise parties: can’t I be referring to this when I advise you that you have a reason to go home, even though an awareness of the reason-­giving fact would spoil the surprise?

Reasons and Fit  169 possible: they should be treated as different available variants of the proposal I am advancing.54 I say that a testimonial fact need not meet condition (i); but notice that it might do so. Perhaps if I act on your advice that will be a way of expressing my respect for your good judgement, and under that description doing A will be a fit response to your good judgement. Then that is an additional reason for doing A (a reason additional to F, which you have told me about). When we consider the different ways in which testimonial facts can and cannot provide reasons for action, we find that these tend to support (FRD). Now let us turn to the apparent false negative: the case of ignorance about your flight schedule, where you apparently have a reason you cannot act on. Here, my response is to distinguish these two facts: (1) that you do not know that your flight leaves at 2 o’clock, and (2) that you do not know what time your flight leaves. I agree that (FRD) does not allow (1) to qualify as a reason to look up the flight schedule: an awareness of the fact that you do not know that your flight leaves at 2 o’clock is itself impossible. But (2) is still available as a reason for looking up the schedule. Once we make this distinction, I think it sounds odd to cite (1) as a reason for you to do that—­even as adduced by an omniscient third party. When I advise you that you have a reason to look up the schedule, and you ask me what it is, I could sensibly offer you (2), but not (1). Once we notice that (FRD) has no difficulty admitting (2) as a reason for the action in this case, it is hard to see why its disqualifying (1) should be seen as a serious objection to it.55 Objections of those two kinds challenge the extensional adequacy of (FRD): they question whether it classifies the right facts as reasons. A further kind of objection accepts the extensional adequacy of (FRD) but rejects its reductive claim. The following Euthyphro-­challenge needs an answer. When a fact does fitly bear on your response determination, isn’t that because it is a reason? The fact that the heat will be unpleasant is a reason not to travel through the desert: isn’t it because of this that you are fitly guided by this fact in deciding not to travel? Another fact about travelling through the desert is that you will be breathing while you travel. But that fact lacks the status of a reason either for or against the action; therefore, it is not fit for guiding your decision. This suggests that (FRD)

54  However, one interpretation that is clearly unsatisfactory is to interpret ‘making the case’ causally. An antecedent cause of my making a promise need not be a reason for me to keep it. 55  How about the surprise party case (from n. 53)? Again, we should distinguish two candidate reasons: (1) that there will be a surprise party for you at home and (2) that something pleasant will happen if you go home. As in the flight case, (1) is disqualified from being a reason for going home—­ you could not be fitly responsive to that fact—­but (2) is not. True, we could stipulate that believing (2) will prevent you from getting the pleasure. But that would just give us another case like the rope-­ bridge example: given the rest of your psychology, thinking about a reason will prevent you from achieving what you have a reason to achieve. That is not enough to prevent (2) from being fit for disposing you to act on it, and hence from being a reason. It is another example of a reason you ought not to think about because of the consequences of doing so.

170 

Fittingness

gets things the wrong way around. When a reason-­giving fact fitly bears on your self-­determination, that is because it is a reason; it is not a reason because it fitly bears on your self-­determination. However, the earlier discussion gives us a ready diagnosis of the apparent force of this objection. As we saw in Section 7.5, condition (i) of (FRD) can be read as an account of when ‘there is a reason’ to make a response. Read in this way, (FRD) analyses the normative reasons that bear on our self-­determination in terms of fit responsiveness to the non-­normative ‘reasons there are’ specified by (i), which are themselves analysed in terms of fitness. So interpreted, the proposal does indeed say that it is because ‘there is a reason’ not to travel through the desert—­namely, the fact that the heat will be unpleasant—­that you are fitly guided by this fact in deciding not to travel. It is when your awareness of this fact fitly guides your self-­ determination that you have a normative reason to act on it. The attractiveness of the claim to which the Euthyphro objection appeals, instead of furnishing a case against (FRD), tends to support it.

7.7  The No-­Priority Alternative I have set out a proposal for explaining what it is to be a normative reason in terms of the prior relation of fit. Normative reasons are the fitness-­making facts that are fit for our response determination. There are of course further questions that need to be answered in order to demonstrate that this view of what reasons are can be successfully integrated with the rest of our thought about reasons. For example, an account will be needed of what it is for a reason to have a weight, and of the relationship between reasons and rationality.56 However, I hope at least to have cleared the first hurdles that are required to make a case for this proposal. I have pointed to some advantages I claim (FRD) has over the alternatives. Let me close by emphasizing what is at stake between them. A real, non-­verbal alternative to (FRD) requires more than just that there is some true interpretation of a sentence such as: (R)  For object O to be fit for response R by person S is for there to be sufficient reason of the right kind for S to make R to O. If I rewrote (FRD) by removing its talk of ‘fit’ and using the phrase on the right-­ hand side of (R) to refer to the fitness-­relation instead, it would still say the same thing. For (R) to amount to a real alternative, what is needed is an account of what it is for a reason to be of the right kind, giving us the materials for a 56  Notice that an account of reasons in terms of the fitness of being disposed to make a response allows for three ways in which reasons can differ in weight, since dispositions, responses, and fitness itself all apparently come in degrees.

Reasons and Fit  171 reductive explanation of what it is for R to be a fit response to O that invokes three independent properties: the property of there-­being-­a-­reason-­for, the property of sufficiency, and the property of right-­kindedness. Next, any proposed account of right-­kindedness needs to be defended against its rivals.57 Once that is done, there are five further explanatory challenges. The first is to explain the relationship between there being a reason of the right kind for R and someone’s having the kind of normative reason that bears on whether she ought to make R. The next three require matching the explanatory credentials of the proposal I set out above: accounting for the two different sources of normative reasons, through their direct or indirect connection to fitness-­relationships; explaining the directiveness of normative reasons; and explaining the relationship between what reasons are  and what it is for us to be responsive to them. The last is to explain away counterexamples. Suppose these challenges could all be met. Where would that leave us? We would then have two rival priority-­claims: a compositional explanation of fitness in terms of reasons, proposed by (R); and a compositional explanation of reasons in terms of fitness, proposed by (FRD). We would then need to look for some further tie-­breaking argument to decide between the two competing priority-­ claims. The following possibility would then arise. In the absence of a compelling further argument, we should reject both of them. We would lack a good reason to treat either property as prior to the other. This would actually push us in the direction of a different view, which drops the reductive ambition of relating the two properties compositionally, and instead just makes a generalization about how their instantiation-­conditions are related to each other. For example, one might retreat from the compositional proposal (FRD), advocating instead the corresponding no-­priority claim with the same two conditions: (FRDNP)  Necessarily, fact F is a reason for person S to make a response of type R to object O in circumstances C if and only if: (i) F makes it the case that, were S to make R-to-O in C, S’s making R-to-O would meet a description under which it is or expresses a fit response to an object; and (ii) S’s being disposed towards the determination of R-to-O in C is a fit response to (i). The success of a rival project explaining fitness by reference to reasons could force the abandonment of the claim that the fitness-­relation is more basic than the relation of being-­a-­reason-­for, but not the claim that they are biconditionally related 57  See note 4.

172 

Fittingness

in the way claimed by (FRDNP). It could be true that there is more than one way of explaining the relationship between reasons and fit, and this is one of them.58 The main argument for preferring the stronger, reductive claim (FRD) is that it answers an explanatory question that the weaker, biconditional claim (FRDNP) leaves unanswered: namely, why is the biconditional true? (FRD) offers us the answer: because meeting those two conditions is what it is for a fact to be a reason. Since (FRD) answers this question, we should therefore prefer it until a successfully developed alternative compels its rejection. But even if a successful alternative did compel that, we would still have learnt something from this discussion: (FRDNP)’s account of the relationship between reasons and fit could still be correct.59

References Anderson, Elizabeth. 1993. Value in Ethics and Economics (Harvard University Press: Cambridge, Mass.). Aquinas, Thomas. 1964–1975. Summa Theologiae (Blackfriars: London). Aristotle. 1999. Nicomachean Ethics (Hackett: Indianapolis). Balguy, John. 1976. The Foundation of Moral Goodness (Garland: New York). Brentano, Franz. 1969. The Origin of Our Knowledge of Right and Wrong (Routledge & Kegan Paul: London). Broad, C. D. 1930. Five Types of Ethical Theory (Routledge & Kegan Paul: London). Broome, John. 2004. ‘Reasons’ in R. Jay Wallace, Philip Pettit, Samuel Scheffler, and Michael Smith (eds), Reason and Value: Themes from the Moral Philosophy of Joseph Raz (Clarendon Press: Oxford). Broome, John. 2013. Rationality through Reasoning (Wiley Blackwell: Oxford). Chappell, Richard. 2012. ‘Fittingness: The Sole Normative Primitive’, Philosophical Quarterly, 62: 684–704. Chisholm, Roderick M. 1976. ‘Brentano’s Theory of Correct and Incorrect Emotion’ in Linda L. McAlister (ed.), The Philosophy of Brentano (Duckworth: London). Chisholm, Roderick  M. 1986. Brentano and Intrinsic Value (Cambridge University Press: Cambridge).

58  Another might be (RNP): Necessarily, O is fit for response R by S if and only if there is sufficient reason of the right kind for S to make R to O. 59  I am grateful to Selim Berker, Chris Howard, and Richard Rowland for extremely helpful written comments on a previous draft of this chapter: thanks also to Robert Audi, Krister Bykvist, Ronald de Sousa, Alex Gregory, Tom Hurka, Brian McElwee, Conor McHugh, Jonas Olson, Christine Tappolet, Pekka Vayrynen, Daniel Whiting, and audiences at conferences of the Australasian Association of Philosophy, Joint Session, SLACRR, Australasian Moral Philosophy workshop, and departmental seminars at ACU, ANU, Leeds, Notre Dame, Southampton, and Toronto.

Reasons and Fit  173 Clarke, Samuel. 2002. A Discourse Concerning the Unalterable Obligations of Natural Religion (Garland: New York). Crisp, Roger. 2000. ‘Review of Joel Kupperman, Value  .  .  .  and What Follows’, Philosophy, 75: 458–62. Crisp, Roger. 2006. Reasons and the Good (Oxford University Press: Oxford). Cudworth, Ralph. 1996. A Treatise concerning Eternal and Immutable Morality, with a Treatise of Freewill (Cambridge University Press: Cambridge). Cullity, Garrett. 2016. ‘Demandingness, ‘Ought’, and Self-Shaping’ in Marcel van Ackeren and Michael Kühler (eds), The Limits of Moral Obligation: Moral Demandingness and Ought Implies Can (Routledge: London). Cumberland, Richard. 2005. A Treatise of the Laws of Nature (Liberty Fund: Indianapolis). D’Arms, Justin, and Daniel Jacobson. 2000a. ‘The Moralistic Fallacy: On the ‘Appropriateness’ of Emotions’, Philosophy and Phenomenological Research, 61: 65–90. D’Arms, Justin, and Daniel Jacobson. 2000b. ‘Sentiment and Value’, Ethics, 110: 722–48. Dancy, Jonathan. 2006. ‘Enticing Reasons’ in R. Jay Wallace, Philip Pettit and Samuel Scheffler, and Michael Smith (eds), Reason and Value: Themes from the Moral Philosophy of Joseph Raz (Oxford University Press: Oxford). Danielsson, Sven, and Jonas Olson. 2007. ‘Brentano and the Buck-Passers’, Mind, 116: 511–22. Darwall, Stephen L. 2006. ‘The Value of Autonomy and Autonomy of the Will’, Ethics, 116: 263–84. Dewey, John. 1998. ‘Three Independent Factors in Morals’ in Larry A. Hickman and Thomas M. Alexander (eds), The Essential Dewey, Vol. 2: Ethics, Logic, Psychology (Indiana University Press: Bloomington). Dienes, Zoltan, and Peter McLeod. 1993. ‘How to Catch a Cricket Ball’, Perception, 22: 1427–39. Ewing, A. C. 1947. The Definition of Good (Macmillan: New York). Gardner, John. 2004. ‘The Wrongdoing that Gets Results’, Philosophical Perspectives, 18: 53–88. Gaus, Gerald. 1990. Value and Justification: The Foundations of Liberal Theory (Cambridge University Press: Cambridge). Gertken, Jan, and Benjamin Kiesewetter. 2017. ‘The Right and the Wrong Kind of Reasons’, Philosophy Compass, 12: 1–14. Gibbard, Allan. 1990. Wise Choices, Apt Feelings (Clarendon Press: Oxford). Heuer, Ulrike. 2010. ‘Reasons and Impossibility’, Philosophical Studies, 147: 235–46. Heuer, Ulrike. 2011. ‘Beyond Wrong Reasons: The Buck-Passing Account of Value’ in Michael Brady (ed.), New Waves in Metaethics (Palgrave Macmillan: Basingstoke). Hieronymi, Pamela. 2005. ‘The Wrong Kind of Reason’, The Journal of Philosophy, 102: 437–57.

174 

Fittingness

Hieronymi, Pamela. 2013. ‘The Use of Reasons in Thought (and the Use of Earmarks in Arguments)’, Ethics, 124: 114–27. Howard, Christopher. 2018. ‘Fittingness’, Philosophy Compass, 13: e12542. Howard, Christopher. 2019. ‘The Fundamentality of Fit’, Oxford Studies in Metaethics 14: 216–36. Irwin, Terence. 2008. The Development of Ethics: A Historical and Critical Study (Oxford University Press: Oxford). Johnston, Mark. 1989. ‘Dispositional Theories of Value’, Proceedings of the Aristotelian Society (Supplementary Vol.), 63: 139–74. Kahneman, Daniel. 2011. Thinking, Fast and Slow (Penguin: London). Kearns, Stephen, and Daniel Star. 2008. ‘Reasons: Explanations or Evidence?’, Ethics, 119: 31–56. Kolodny, Niko. 2005. ‘Why Be Rational?’, Mind, 114: 509–63. Korsgaard, Christine  M. 1996. The Sources of Normativity (Cambridge University Press: Cambridge). LeBar, Mark T. 2013. The Value of Living Well (Oxford University Press: Oxford). Lemos, Noah M. 1994. Intrinsic Value: Concept and Warrant (Cambridge University Press: Cambridge). Maguire, Barry. 2018. ‘There Are No Reasons for Affective Attitudes’, Mind, 127: 779–805. McDowell, John. 1985. ‘Values and Secondary Qualities’ in Ted Honderich (ed.), Morality and Objectivity (Routledge: London). McHugh, Conor, and Jonathan Way. 2016. ‘Fittingness First’, Ethics, 126: 575–606. Mele, Alfred R. 2003. Motivation and Agency, (Oxford University Press: Oxford). Millgram, Elijah. 1996. ‘Williams’ Argument against External Reasons’, Noûs, 12: 345–64. Moore, G. E. 1903. Principia Ethica (Cambridge University Press: Cambridge). O’Brien, Lucy, and Matthew Soteriou (eds). 2009. Mental Actions (Oxford University Press: Oxford). Parfit, Derek. 2011. On What Matters (Oxford University Press: Oxford). Piller, Christian. 2006. ‘Content-Related and Attitude-Related Reasons for Preferences’, Royal Institute of Philosophy Supplement, 81: 155–81. Price, Richard. 1974. A Review of the Principal Questions of Morals (Oxford University Press: Oxford). Rabinowicz, Wlodek, and Toni Rønnow-Rasmussen. 2004. ‘Strike of the Demon: On Fitting Pro-Attitudes and Value’, Ethics, 114: 391–423. Raz, Joseph. 2009. ‘Reasons: Practical and Adaptive’ in David Sobel and Steven Wall (eds), Reasons for Action (Cambridge University Press: Cambridge). Ross, W. D. 1939. Foundations of Ethics (Clarendon Press: Oxford).

Reasons and Fit  175 Rowland, Richard. 2017. ‘Reasons or Fittingness First?’, Ethics, 128: 212–29. Scanlon, T. M. 2014. Being Realistic about Reasons (Oxford University Press: Oxford). Schelling, Thomas  C. 1960. The Strategy of Conflict (Harvard University Press: Cambridge, Mass.). Schroeder, Mark. 2007. Slaves of the Passions (Oxford University Press: New York). Schroeder, Mark. 2010. ‘Value and the Right Kind of Reason’, Oxford Studies in Metaethics, 5: 25–55. Schroeder, Mark. 2012. ‘The Ubiquity of State-Given Reasons’, Ethics, 122: 457–88. Setiya, Kieran. 2013. ‘What Is a Reason to Act?’, Philosophical Studies, 167: 221–35. Shah, Nishi, and David Velleman. 2005. ‘Doxastic Deliberation’, Philosophical Review, 114: 497–534. Sharadin, Nathaniel. 2015. ‘Reasons Wrong and Right’, Pacific Philosophical Quarterly, 96: 1–29. Skorupski, John. 2007. ‘Buck-Passing about Goodness.’ in T.  Rønnow-Rasmussen, B. Peterson, J. Josefsson, and D. Egonsson (eds), Hommage á Wlodek: Philosophical Papers Dedicated to Wlodek Rabinowicz (Lund University: Sweden) [online resource]. Available at www.fil.lu.se/hommageawlodek. Skorupski, John. 2010. The Domain of Reasons (Oxford University Press: Oxford). Stratton-Lake, Philip. 2005. ‘How to Deal with Evil Demons: Comment on Rabinowicz and Rønnow-Rasmussen’, Ethics, 115: 788–98. Streumer, Bart. 2007. ‘Reasons and Impossibility’, Philosophical Studies, 136: 351–84. Streumer, Bart. 2016. ‘Reasons and Ability’ in Daniel Star (ed.), The Oxford Handbook of Reasons and Normativity (Oxford University Press: Oxford). Svavarsdóttir, Sigrún. 2014. ‘Having Value and Being Worth Valuing’, The Journal of Philosophy, 111: 84–109. Tappolet, Christine. 2011. ‘Values and Emotions: Neo-Sentimentalism’s Prospects’ in Carla Bagnoli (ed.), Morality and the Emotions (Oxford University Press: Oxford). Thomson, Judith Jarvis. 2008. Normativity (Open Court: Chicago). Way, Jonathan. 2012. ‘Transmission and the Wrong Kind of Reason’, Ethics, 122: 489–515. Wedgwood, Ralph. 2015. ‘The Pitfalls of “Reasons”’, Philosophical Issues, 25: 123–43. Zimmerman, Michael  J. 2001. The Nature of Intrinsic Value (Rowman & Littlefield: Lanham, Md.).

8 Value-­First Accounts of Normativity R. A. Rowland

8.0  Introduction It is tempting to think that all our reasons for action derive from value.1 We might be drawn to this idea in various ways. We might be ‘spellbound’ by the idea that we could never be required to do what’s worse rather than what’s best.2 Or we might think that it is people, places, objects, artefacts, and states of affairs’ valuable features that explain why we ought to, and have reasons to, respond to them in various ways. It is people’s value that explains why we should respect them. Admirable people’s good features—­that is, the features of them that make them admirable—­are what give rise to reasons to admire them and make it fitting to admire them. The value of natural beauty is what gives us reasons to protect it. And the good-­making features of a good state of affairs, a good present, and a good holiday resort are what give us reasons to desire a good state of affairs, give a good present, and visit or recommend a good holiday resort.3 But the view that value explains, grounds, or gives rise to our reasons for actions and attitudes has fallen out of favour. Instead the dominant view is now that for something to be valuable is for it to be a fitting object of pro-­attitudes, or to have the higher-­order property of having other properties that are reasons for pro-­ attitudes. For instance, what makes a holiday resort a good one is that it has other features that are reasons to recommend or visit it (its sandy beaches, pleasant amenities, stunning views, etc.). On these buck-­passing (reasons-­first) and fitting-­attitude (fittingness-­ first) accounts of value, reasons and fittingness are explanatorily prior to value. Those arguing for such buck-­passing and fitting-­attitude accounts have con­ sidered rival value-­first accounts of reasons and fittingness that explain reasons and fit in terms of value. And they have pointed out several quite severe problems with value-­first accounts of reasons and fittingness. But these opponents of value-­ first accounts have only considered direct rather than indirect value-­first accounts

1  I would like to thank Rachel Achs, Selim Berker, Chris Howard, Tom Hurka, Austen McDougal, Oded Na’aman, and Jonathan Way for comments on previous versions of this chapter. 2  See Foot (1985: 196) and Portmore (2011: 32–3). 3 Though these ideas might attract one to a value-­first view, buck-­passing and fitting-­attitude accounts of value may well be compatible with them; see e.g. Rowland (2019: 59–65). R. A. Rowland, Value-­First Accounts of Normativity In: Fittingness: Essays in the Philosophy of Normativity. Edited by: Christopher Howard and R. A. Rowland, Oxford University Press. © R. A. Rowland 2022. DOI: 10.1093/oso/9780192895882.003.0008

Value-First Accounts of Normativity  177 of reasons and fittingness—­largely because it is only direct value-first accounts of reasons and fit that have been clearly proposed.4 According to direct views, the fittingness of ϕ-ing and reasons to ϕ are explained by the value of ϕ-ing, the value of that which it is fitting to ϕ in response to, or the value of the object of the attitude ϕ. This chapter aims to consider the prospects of indirect value-­first accounts of fit and reasons, indirect value-first accounts of normativity. The indirect account I propose explains the fittingness of ϕ-ing and reasons to ϕ in terms of the value of being guided by standards according to which it is fitting to ϕ and there are reasons to ϕ. Fittingness-first and reasons-first attempt to explain all of normativity in terms of reasons and fit. If I’m right, we might ultimately explain all of normativity, indirectly, in terms of value. I have previously argued at length for a buck-­passing account of value, and against value-­first accounts of reasons. The aim of this chapter is to explore whether I and others have been too quick to dismiss value-­first approaches to reasons and fit by exploring the possibility of an indirect value-­first approach. As I will explain, an indirect value-­first approach may seem able to avoid all of the objections that I and others have levelled against value-­first views.5 But there do still seem to be costs to an indirect value-­first approach which I will explain and explore the severity of. And, ultimately, it may still seem that an indirect value-­ first approach, hardened against objections though it is, does not quite do as much explanatory work as fitting-­attitude and buck-­passing accounts of value. My aim here though is primarily exploratory. I hope to explain how an indirect value-­first account of reasons and fit can work and to try to make the strongest case for such a view. After making this case I try to honestly assess whether an indirect view like that which I consider has as many theoretical virtues as fitting-­attitude and buck-­passing accounts of value, whether it leads to other problems that direct value-­first views do not face, and whether an indirect value-­first account is as plausible as the fitting-­ attitude and buck-­passing accounts of value with which it competes. Section 8.1 briefly explains how direct value-­first views work and explains several objections that have been made to them which have not yet been answered by value-­firsters. Section 8.2 proposes a new indirect value-­first account of r­ easons and fittingness. Section 8.3 tries to make the strongest possible case for the view that this indirect value-­first account is as explanatorily powerful as fittingness-­ first and reasons-­first views (or as explanatorily powerful as fitting-­attitude and buck-­passing accounts of value). Section 8.4 explains how this indirect value-­first 4  Raz (2001: 164–6), Heuer (2004), Olson (2006), Wedgwood (2009), Orsi (2013), and Maguire (2016) propose direct value-­first accounts of reasons; Moore (1903: 24–6) held that value explains what we ought to do. Cowie (2019) and Maguire and Woods (2020) do, however, argue for indirect accounts of epistemic normativity in particular; similarly rule consequentialists argue for indirect accounts of moral normativity. 5 See Way (2013), Howard (2018: 4–5), Rowland (2019: chs 2, 3, 5), and Kiesewetter (2022). Scanlon (1998: 95–100), Suikkanen (2005: 516–22), and Stratton-­Lake and Hooker (2006) also argued against value-­first accounts of reasons, but Olson (2006) and Orsi (2013) showed that value-­first accounts can be constructed that avoid their objections.

178 Fittingness account can address the objections discussed in Section 8.1. Section 8.5 tries to give an honest assessment of the prospects of this indirect value-­first view, whether it will create problems elsewhere, whether it is really as explanatorily powerful as fittingness-­first and reasons-­first accounts of normativity, and whether we could really have reason to accept an in­dir­ect value-­first account over such views.

8.1  Problems with Direct Value-­First Accounts There are broadly two versions of a value-­first account of reasons and fittingness that have been discussed. The first explains reasons and fit in terms of the value of the object or thing that provides a reason to act or that it is fitting to have a pro-­ attitude towards. The relevant value properties are normally, or most plausibly, thought to concern final value or non-­instrumental goodness simpliciter.6 So on this view, if it is fitting to admire someone, this is because they have properties that make them (non-­instrumentally) good simpliciter—­for instance, they are kind, creative, or virtuous. Similarly, on this view, there are reasons to desire happiness because it is good simpliciter in itself and there are reasons to desire friendship because friendship is either intrinsically good simpliciter or is a means to other things (such as happiness) that are good simpliciter. And, on this view, our reasons to perform actions derive from, and consist in, the valuable features of outcomes, which we can bring about by acting. Let’s call this view, the value-­of-­ the-­object-­first view, or VO-­first for short. A second value-­first account explains the fittingness of ϕ-ing and reasons to ϕ in terms of the value of ϕ-ing. On this view, if there is a reason to perform an action, this is due to the value of that action. And fitting attitudes are fitting in virtue of the value of having them. Let’s call this view, the value-­of-­the-­attitude-­ first view, or VA-­first for short. On this view, our reasons for action consist in the  value of our performing these actions or the value that we create via our actions.7 And fitting attitudes are explained by the value of having fitting attitudes, that is, the value of desiring the desirable, believing the credible, admiring the admirable, and so on. On VO-­first, the value of the happiness, the relatively large amount of leisure time, and the understanding that an academic career provides are what explains the reasons to desire it and pursue it; on VA-­first the value of (or created by) pursuing a career that involves happiness, a relatively large amount of leisure time, and understanding is what explains why you have reason to pursue an academic career. On VO-­first, the value of the objects of fitting admiration is what makes fitting admiration fitting, the fact that kindness and creativity are valuable 6  For a defence of the idea that there are things that are good simpliciter see Rowland (2016b). 7  There may not be a great difference between these two accounts’ views of reasons for action.

Value-First Accounts of Normativity  179 features of someone makes it fitting to admire them; on VA-­first the value of admiring someone who is kind and creative is what makes it fitting to admire them. The most plausible version of the VA-­first account holds that in order for it to be fitting to have an attitude, it must be non-­instrumentally valuable to have it. Although it is valuable to admire a demon who will punish you if you do not admire it, it is only instrumentally valuable to have this attitude and so it is not fitting to have it. But why think that it is non-­instrumentally valuable to have all attitudes that it is fitting to have? Brentano (1969: 22–3), Moore (1903: 204, 208–9, 211, 217), Nozick (1981: 429–33), and Hurka (2001: ch. 1) argue that it is derivatively non-­instrumentally valuable to have pro-­attitudes towards that which is good in itself and to have con-­attitudes towards that which is bad in itself: it is good to love the good and good to hate the bad, good in itself to admire the ad­ mir­ able and desire the desirable, and so on. On this view, it is non-­ instrumentally valuable to desire those things that are desirable in themselves and to admire the admirable. This is because things that are intrinsically desirable are intrinsically valuable and so to desire them for their own sake is to have a pro-­ attitude towards that which is good in itself. Similarly, admirable features of people are features of them that are non-­instrumentally valuable, and so, to admire the admirable is to have pro-­attitudes towards things that are non-­instrumentally valuable. Hurka calls this view a recursive view of the non-instrumental value of pro-attitudes because on this view the non-­instrumental value of some things, pro-­attitudes, is entirely the result of the intrinsic value of other things. And whenever something is finally valuable that thereby creates the additional non-­ instrumental value of valuing it accordingly. So, we can call the non-­instrumental value that pro-­attitudes towards the good and con-­attitudes towards the bad have recursive value. With the idea of recursive value in place it can be argued that fitting admiration is ad­mir­ation that is non-­instrumentally valuable because it is admiration of that which is good in itself. And (right kind) reasons to desire things for their own sake are reasons to have attitudes that are non-­instrumentally good because they are ­reasons to desire that which is good in itself, and it is good in itself to desire that which is good.8 8  The recursive value view relies on the general idea that there are types of non-­instrumental value beyond intrinsic value. This general idea is plausible since there are things that are good for their own sake but not in virtue of their intrinsic properties. For instance, Audi (2005: 125–6) argues that, although the only intrinsically valuable things are mental states, beautiful artworks are non-­ instrumentally derivatively valuable. On this view, the value of beautiful paintings derives from the value of the mental state of appreciating these paintings, but the value of these paintings is not best understood as merely instrumental value since these paintings are ‘not a means (in any ordinary sense) to the value of experiencing them, since [they are] partly constitutive of [the valuable] experience[s]’ that their value derives from. Somewhat similarly, Kagan (1998: 285) and Rabinowicz and Rønnow-­Rasmussen (2000: 41) argue that the pen that Lincoln used to sign the emancipation proclamation and one of Princess Diana’s dresses are finally valuable but not in virtue of their intrinsic features; these things are finally, but not intrinsically, valuable.

180 Fittingness Both VA-­first and VO-­first accounts are direct value-­first accounts of reasons and fittingness: they give accounts of reasons to ϕ and the fittingness of ϕ-ing directly in terms of the value of ϕ-ing or in terms of the value of an object that gives us a reason to ϕ or makes it fitting to ϕ. Direct value-­first accounts face problems with three types of reasons and fit.

8.1.1  Fitting Belief and Reasons for Belief Not everything that we have reason to believe is good or bad, so a VO-­first view of reasons for belief and fitting belief seems like a non-­starter. And there are plaus­ibly cases of things that it is fitting to believe and things that we have reasons to believe which it is of no value for us to believe. There is no value to believing that there are 8,294,400 blades of grass on the field at Camp Nou, or that there are six specks of dust on your desk. But you might have excellent evidence that these things are true; in this case, it seems that it would be fitting to believe these things and that you would have reason to believe them.9 If this is right, then direct VA-­ first views are false.

8.1.2  Fitting Emotions Is it always non-­instrumentally good to envy the enviable, fear the fearsome, be annoyed by the annoying, and to dread the dreadful?10 It seems not. But it is fitting to have these attitudes and there are right kind reasons to have them. In this case, it seems that VA-­first views are false, since we cannot analyse right-­kind reasons for these attitudes and the fittingness of having these attitudes in terms of the non-­ instrumental value of these attitudes. And we cannot hold that an attitude is fitting whenever it is instrumentally valuable to have it, since it is not fitting to admire an evil demon even when it will punish you if you do not admire it. VO-­ first views may also seem to face problems with some of these attitudes. Perhaps people who are enviable and lives that are enviable have good-­making features—­ otherwise why would they and their lives be enviable? But it is less clear that annoying things have bad simpliciter making features or that all fearsome things have bad-­making features.

9  See Rowland (2019: 41–52) and Howard (2018: 4). 10  See Howard (2018: 4). On how it might be instrumentally good to envy the enviable, see Protasi (this volume).

Value-First Accounts of Normativity  181

8.1.3  Deontological Reasons Many deontologists hold that what we have most reason to do can outstrip what it is of most value for us to do. Think about the footbridge trolley case: many de­ontolo­gists hold that we have most reason to refrain from pushing the heavy man even though pushing the heavy man would be best because it would save the most lives and do the most good. But direct value-­first views hold that what we have most reason to do is explained by what it is valuable to do or by the valuable features of the world that we can bring about. So, these views seem to be in conflict with the existence of deontological reasons.11 Similarly, it has been argued that we have reasons to obey legitimate authorities even when doing so is not good simpliciter at all and that there are reasons to respect property rights even when doing so is not good or finally valuable.12 The problem here is best understood, I have argued, as a problem about neutrality. The problem here is not that there are such deontological reasons that outstrip value and so direct value-­first views are false. Rather the problem is that the view that there are deontological reasons that outstrip value is a view that it is coherent to hold and that is somewhat plausible. And, other things equal, our views about the nature of reasons and fit should not imply that coherent and somewhat plausible first-­order normative views are mistaken.13 A lot more could be said about whether value-­first views truly face problems with these three types of reasons and fit. But hopefully it should be clear how there are prima facie problems here; more thorough cases that there are problems for value-­first accounts with these three types of reasons and fit have been made elsewhere. In the rest of this chapter I want to discuss how these problems can be evaded by a different kind of value-­first account.14 11  See Way (2013: 36–7) and Rowland (2019: 24–30), who both discuss and argue against the many possible ways for proponents of value-­first views to respond to this argument such as by appealing to agent-­relative value. 12  See Kiesewetter (2022). 13  See Rowland (2019: ch. 2). Reasons- and fittingness-­first views have been held to face problems making sense of the contrast between deontology and consequentialism. But elsewhere I’ve argued that this seeming problem can be shown to be an illusion; see Rowland (2019: 147–9). 14  Two other kinds of problems for direct value-­first views have been raised. First, in Rowland (2019: 32, 22–4) I argued that VA-­first views problematically depend on a controversial axiology because they depend on the recursive value thesis for their plausibility. However, I am now not so sure that the recursive value thesis is problematically controversial since it only holds that it is derivatively non-­instrumentally valuable to have pro-­attitudes towards the good. And so it is compatible with hedonistic and desire-­satisfaction-­based accounts of final value, for instance; see Rowland (2022b: 128–30). Second, Howard (2018: 4) and Kiesewetter (2022) argue that direct value-­first accounts imply that value explains reasons and fit when intuitively it does not. Kiesewetter (2022) argues that even if it is in some way valuable for us to keep our promises, it is implausible that our reasons to keep our promises are always constituted by or explained by the value of so keeping them: sometimes our reasons to keep our promises consist in the fact that we promise or that we are obligated to do so. However, a VA-­first view should probably be understood in terms of

182 Fittingness

8.2  An Indirect Value-­First Account One way to solve these problems for the value-­first approach is to abandon the attempt to directly explain reasons and fit in terms of value. This involves abandoning the view that we should understand the fittingness of each particular instance of fitting ϕ-ing in terms of the value of ϕ-ing or the fact that makes it fitting to ϕ on that occasion; and similarly abandoning the view that we should understand reasons to ϕ directly in terms of the value of ϕ-ing or the value of the fact that is a reason to ϕ (for instance, we should not analyse the reason to believe that there are 8,294,400 blades of grass on the field at Camp Nou in terms of the value of believing this). Instead we seek to explain reasons and fittingness in terms of value in a more indirect fashion. The strategy here is somewhat similar to how rule consequentialists and other indirect consequentialists ditch the idea that what we ought to do is always what it would be best to do but still retain the idea that we can explain what we ought to do and what’s morally right and wrong in terms of what’s best. To start, let’s think of how rule consequentialism works. We consider the set of all possible moral codes. These different moral codes give different verdicts about what’s right and wrong and what moral reasons there are. For instance, a common-­sense Rossian moral code tells us that we have reasons to benefit ourselves and others, make amends if we have wronged others, keep our promises, tell the truth, do what’s just, and refrain from harming others. According to a (crude) egoistic moral code, there are only moral reasons for us to promote our own well-­being, and so we have no reasons to keep a promise just because we made one or to make amends just because we did wrong. And, according to a certain kind of Nietzschean moral code, there are none of the moral reasons that the common-­sense moral code says there are. Rule consequentialists say that the correct or genuine moral code is the one that it would be best if we internalized. And rule consequentialists like Brad Hooker say that this code is the common-­ sense Rossian code rather than the egoistic or Nietzschean code: things go better

good-­makers: for R to be a right kind of reason to ϕ is for ϕ-ing to have properties that make it good. Now if it is good for us to keep our promises and our reason to keep them is that we are obligated or have so promised, then it may be that what makes our keeping our promises good is that we promised or are obligated to keep them. In this case, the view that reasons are value-­makers may seem to deal with Kiesewetter’s case. Similarly, Howard (2018: 4) says that what explains why it is fitting to admire Sharon, or why she is admirable, is not that it is valuable to admire her—­even if it is in fact valuable to admire her but VA-­ first views seem committed to the view that what explains why it is fitting to admire her is that it is valuable to admire her. This kind of issue no longer seems problematic to me. This is because pro­pon­ ents of VA-­first views can hold that what ultimately explains why it is fitting to admire Sharon is her valuable features, but her valuable features make it the case that it is valuable to admire her, and this makes it the case that it is fitting to admire her. (For how proponents of VA-­first can make this move see the discussion of the recursive value thesis above).

Value-First Accounts of Normativity  183 in our society if we all internalize the Rossian code rather than the Nietzschean or egoist code.15 Now we can be guided by a code or standard either instrumentally or for its own sake: someone can be guided by the norms and requirements of the no-­ snitching code just because they don’t want to get murdered rather than because they think it’s right not to snitch; we can be guided by a legal standard for its own sake, because we respect the law (we have taken up the Hartian internal point of view on it) or just because we don’t want to be locked up. On the value-­first approach that I want to explore: Indirect Value-­First (IV-­First). For R to be a genuinely normative reason to ϕ is for R to be a reason to ϕ according to a standard that it is non-­ instrumentally better simpliciter for us to be guided by for its own sake. For it to be fitting to ϕ is for it to be fitting to ϕ according to a standard that it is non-­instrumentally better sim­plici­ter for us to be guided by for its own sake.16 There are lots of different possible standards for fitting admiration, fitting desire, fitting envy, reasons for pro-­attitudes, and reasons for action. Only some of these standards tell us what it is in fact fitting to do and what we have genuinely normative reasons to do. For instance, there are possible standards for fitting admiration that are misguided, e.g. consider a standard that holds that it is fitting to admire people who are evil, malicious, and powerful.17 According to IV-­first, value explains which of these standards tell us what it is actually fitting to do and what we really have reasons to do, and value explains this via the value of being guided by the standard on which there are these reasons and on which it is fitting to have these attitudes. It is non-­instrumentally better (simpliciter) to be guided by the standard of admiration that holds that it is fitting to admire kind, creative, and smart people than it is to be guided by the standard of admiration that holds that it is fitting to admire evil, malicious, and powerful people. On this view, for there to be the reasons that the Rossian code says there are is for it to be 15  Hooker (2000). 16  One might wonder why IV-­first is framed in terms of betterness rather than value: why does IV-­ first hold that there is a genuinely normative reason to ϕ iff there is a reason to ϕ according to a standard that it is non-­instrumentally better to be guided by rather than holding that there is a genuinely normative reason to ϕ iff there is a reason to ϕ according to a standard that it is non-­ instrumentally good to be guided by? IV-­first is articulated in terms of betterness because it might be to some extent non-­instrumentally good to be guided by an incorrect moral standard rather than to not be guided by any moral standard at all. Since being guided by some close-­to-­correct moral standard might involve respecting others and caring about their well-­being. But nonetheless it is not better to be guided by the incorrect moral standard than to not be so guided, since it is better to be guided by the correct moral standard than the incorrect one—­this involves being virtuous rather than being close to virtuous; and it is better to be guided by the correct moral standard than to not be so guided. It is of course possible that whichever way we frame IV-­first—­in terms of non-­instrumental betterness or non-­instrumental goodness—­it will lead to counter-­examples. 17  Or that there are genuinely normative reasons to admire such people and/or that it is robustly normatively fitting to admire them; more on this idea in Section 8.3.

184 Fittingness non-­instrumentally better simpliciter for us to be guided by the Rossian code than to not be guided by this code. IV-­first has a similar structure to a rule-­consequentialist account of right and wrong. But, in addition to being a view about all kinds of reasons and fit (rather than about right and wrong and only moral reasons), IV-­first also differs from rule-­consequentialism in that it focuses on the non-­instrumental betterness of being guided by a standard. According to rule consequentialism, ϕ-ing is wrong iff the best moral code holds that ϕ-ing is wrong, where the best moral code is the one that it would be best for us to internalize: our internalizing this code would produce the best consequences. The consequences relevant to rule consequentialism involve instrumental consequences, e.g. that if we internalized a particular code less assault would occur and people’s well-­being would be promoted because they could rely on one another better (because, for instance, people would break their promises less). According to IV-­first, it is fitting for us to ϕ iff it is non-­ instrumentally better for us to be guided by a code which holds that it is fitting to ϕ rather than to not be guided by such a code. According to IV-­first, the instrumental consequences of our being guided by such a code are irrelevant to whether it is fitting for us to ϕ. In order for IV-­first to be extensionally adequate it must be non-­instrumentally better to be guided by the correct moral standard and the (correct) standard for admiration. It does seem non-­instrumentally better to be guided by these standards. It seems good in itself to be guided by the correct moral standard for its own sake (de re rather than de dicto at least), since virtue seems non-­instrumentally good, and it seems non-­instrumentally good to care about not wronging people, respecting them, and making them well off rather than hurting or harming them.18 The correct prudential standard holds that there are reasons to have pro-­ attitudes towards things that are in fact intrinsically good for us. So, being guided by the correct prudential standard involves having pro-­attitudes towards things that are in fact intrinsically good for us. In this case, if we accept the recursive value thesis discussed in Section 8.2, on which it is non-­instrumentally valuable to have pro-­attitudes towards things that are intrinsically good, it must be non-­ instrumentally valuable to be guided by the correct prudential standard; being guided by this standard involves having non-­instrumentally valuable attitudes. Similarly, the correct standard for admiration is the standard that holds that it is

18  Furthermore, if we hold the recursive value thesis discussed in Section 8.2, and people are finally valuable, it will be recursively valuable to respect people and to care about not wronging them, since having this attitude towards people will involve valuing the valuable. Since recursive value is non-­ instrumental value, this means that it is non-­instrumentally valuable to be guided by the correct moral code. One might worry that if hedonism about final value is correct, this will imply that it is not non-­ instrumentally valuable to be guided by the correct moral standard, and that this means that the combination of hedonism and IV-­first would imply that there are no genuinely normative moral reasons. I discuss this problem in Rowland (2022b: 128–30).

Value-First Accounts of Normativity  185 fitting to admire people, features, and things that are in fact admirable. So, being guided by this standard involves admiring the admirable. According to the recursive value thesis, it is non-­instrumentally valuable to admire the admirable. So, it is non-­instrumentally valuable to be guided by the correct standard of admiration.19 Let’s consider some particular examples to see how IV-­first works. Suppose that the correct moral code holds that we have most reason to refrain from pushing the heavy man off the bridge in the footbridge trolley case. IV-­first says that what makes it the case that there is in fact most reason to refrain from pushing is that, even though it is in fact better to push, it is non-­instrumentally better to be guided by the moral code that holds that we have most reason to refrain from pushing. If one is guided by this code, one is more virtuous than one otherwise would be, or one is responding to people and the features of people which make us worthy of respect and have dignity—­the features of us which make us have value—­better than one would be if one was not guided by this standard. Similarly, suppose that the correct standard of what makes someone admirable holds that being kind, creative, smart, and insightful are features of someone that make it fitting to admire them. Zoë has these features. So, according to this standard it is fitting to admire Zoë. Suppose these features are valuable features of someone. So, if we are guided by this standard for the admirable—­if we accept this standard and admire these features and those who have them—­then we have pro-­attitudes towards valuable things. In this case our being guided by this standard is non-­ instrumentally valuable; it is better for us to be guided by this standard than to not be, since (other things equal) our not being guided by this standard involves our not having pro-­attitudes towards (or having fewer pro-­attitudes towards) valu­able things. According to IV-­first, it is fitting to admire Zoë; this standard gets things right about who it is fitting to admire. This is because the features of people that this standard claims makes it fitting to admire them are features of people that are valuable. In virtue of this fact, being guided by this standard is non-­instrumentally valuable, since it involves having pro-­attitudes towards things that are valuable in themselves. The value of being guided by this standard, which derives from the value of the admirable properties themselves, makes it the case that this standard gets things right about which attitudes are fitting.

8.3  IV-­First Is Explanatory IV-­first does not seem to give a reductive account of reasons or fittingness, since reasons and fittingness figure in IV-­first’s analysis of genuinely normative reasons and genuine fittingness. In this case one might wonder whether IV-­first is a 19  For a discussion of the extent to which the plausibility of IV-­first depends on the recursive value thesis, see Rowland (2022b: 128–30). For a defence of the recursive value thesis, see Rowland (2022b: 114) and Hurka (2001: ch. 1).

186 Fittingness genuine competitor to reasons- and fittingness-­first accounts of normativity, or to buck-­passing and fitting attitude accounts of value. Since these views give reductive accounts of value but IV-­first does not give a reductive account of reasons and fit. However, I think that we should primarily see the interesting issue being debated between fittingness-, ­reasons-, and value-­first accounts as one of the metaphysical explanatory priority between these properties rather than one of what can be reduced to what. Furthermore, the main reasons to be interested in giving reductions of value in terms of reasons or fit or vice versa are that reductive views of value, fit, and ­reasons have been argued to have theoretical virtues and explanatory pay-­offs that non-­reductive views do not have. But as I’ll argue in the rest of this section, IV-­first in fact seems to have these theoretical virtues and explanatory pay-­offs without providing a reductive account of fittingness or reasons. First, one reason that many have been attracted to reductive views is that they have been argued to be able to explain a necessary connection in a way that non-­ reductive views cannot. Reductive accounts of value and reductive accounts of reasons have been held to explain why: Necessary Connection. Necessarily, if X is of final value, then it is fitting to have pro-­attitudes towards X and there are reasons for us to have a pro-­ attitude in response to X. Necessary Connection is very plausible. For instance, if achievement is finally valu­ able, it seems it is fitting to desire it and if friendship is good in itself, there are reasons to want it for its own sake. Reductive direct value-­first views explain this connection because they hold that for it to be fitting for us to have a pro-­attitude towards X is for X to be finally valuable, or for having a pro-­attitude towards X to be finally valuable. Fitting-­attitude and buck-­passing accounts of value explain this necessary connection because they hold that for X to be finally valuable is for it to be fitting for us to have pro-­attitudes towards X or for us to have reasons to have pro-­attitudes towards X. But no-­priority views of the relationship between ­reasons/ fit and value, which do not reduce either value to fit/reasons or vice versa, cannot explain this necessary connection, or so it has been argued.20 IV-­first seems able to explain this necessary connection without giving a reductive account of fit or reasons. If X is finally valuable and the recursive value thesis holds, then it is non-­instrumentally valuable to have pro-­attitudes in response to X, since according to the recursive value thesis it is non-­instrumentally good to have pro-­ attitudes towards the good. In this case, it will be non-­ instrumentally better to be guided by a standard S that holds that it is fitting to have pro-­attitudes towards X than to not be guided by such a standard, since being guided by S will involve having pro-­attitudes towards X, and so being 20  See Suikkanen (2005: 525–8) and Rowland (2019: 55–9, 65–9).

Value-First Accounts of Normativity  187 guided by S will be non-­instrumentally valuable. So, if IV-­first holds, this necessary connection is explained. Second, existing reductive views can explain why: Value Does Not Provide Reasons. The fact that X is valuable never on its own provides a non-­derivative reason to have a pro-­attitude towards X and is never the reason why it is fitting to have a pro-­attitude towards X. For instance, the fact that a holiday resort is a good one is no (non-­derivative) reason to visit it or recommend it to friends; more specific features of the holiday resort, such as its pleasant amenities, tasty food, and stunning views are (­non-­derivative) reasons to visit it and recommend it to friends. Similarly, that someone is good is no reason to admire her; other features of her such as her strength of character, intellect, resolve, creativity, and kindness are reasons to admire her. Non-­reductive views have been held to fail to explain why value doesn’t itself provide reasons; reductive views, such as buck-­passing and fitting-­ attitude accounts of value have been shown to explain this.21 But it seems that IV-­first does allow us to explain why value doesn’t itself provide (non-­derivative) reasons to have pro-­ attitudes towards something and why the value of something—­rather than its value-­making features—­does not make it fitting to have pro-­attitudes towards it. For, according to IV-­first, value is not that which directly explains why it is fitting to have any attitude or provides reasons to perform actions or have any attitude. Instead, value plays a background role, that of explaining why the facts that make attitudes, such as admiration, fitting make them fitting and explaining why certain facts are reasons to have certain attitudes. On IV-­first, facts that are (genuine) reasons and facts that are fit-­makers are just facts that are reasons and fit-­makers according to standards that it is valuable to be guided by. Value makes facts into normative reasons to admire on this view rather than itself being a reason to admire.22 A third reason for wanting a reductive account is that reductive views are more qualitatively parsimonious than non-­reductive views, and we should prefer more qualitatively parsimonious theories to less qualitatively parsimonious theories: a 21  See Scanlon (1998: 97), Stratton-­Lake and Hooker (2006: 154–6), Crisp (2008: 263–4), Schroeder (2011), Parfit (2011: 39), and Rowland (2019: 59–65). The buck-­passing account of value explains why the fact that a holiday resort is a good one is never a (non-­derivative) reason to visit it or recommend it to friends. This is because, on this view the property of being good is just the higher-­order property of having other reason-­providing properties. And the property of having other properties that provide reasons cannot itself provide non-­derivative reasons to have pro-­attitudes towards something; the non-­derivative reasons to have pro-­attitudes towards something are instead provided by the lower-­ order features of things that themselves provide reasons. The higher-­order property just reflects or consists in a summation of normative facts, it is not itself normatively important on this picture; see Rowland (2019: 59–65). 22  See Schroeder (2007: ch. 2) for discussion of how certain types of facts can play this purely background normative role.

188 Fittingness theory on which there is only fittingness holds that there are fewer kinds of en­tities than one which holds that there is both fittingness and value; and, just as we should prefer theories that do not posit new entities over those which do, we should prefer a theory that implies that there are fewer kinds of things to one that implies that there are more.23 A case can be made that IV-­first is as parsimonious as fittingness-­first and reasons-­first accounts. In order to see this case we need to understand the recently much-­discussed distinction between formal and robust normativity. We have formal normativity as soon as we have any standard of correctness. If we make up a game where everyone is required to wear one red item of clothing every other day, or where whenever someone says ‘cat’ we say ‘hat’, then we have reasons to say ‘hat’ whenever someone says ‘cat’, and we are required to wear red on alternate days of the week. But these reasons and requirements are not like moral and prudential reasons and requirements; if we never wear red on alternate days or do not say ‘hat’ whenever someone says ‘cat’, we don’t seem blameworthy or to be missing something normatively in the way that we seem to be if we never do what we have moral or prudential reason to do and never do what we’re morally or prudentially required to do. Similarly, all social norms and conventions have formal normativity but not all social norms and conventions seem to be robustly or authoritatively normative in the way that the correct moral, prudential, and epi­stem­ic norms seem to be. According to the ‘no-­snitching code’ we have reason to refrain from telling the police if we’re assaulted, according to the norms of mid-­twentieth-­ century US high-­society women shouldn’t wear white after labour day, and according to the norms of masculinity, men shouldn’t cry and are forbidden from wearing mascara. But we’re not going wrong normatively and we’re not criticizable if we fail to live up to these reasons and requirements in the way that we’re going wrong and criticizable if we fail to do what we morally or prudentially ought to do.24 So there is a vast array of codes, some of which give rise to robustly normative reasons and fittingness, most of which do not. Value-­first accounts have standardly been given as accounts of robustly normative reasons. It would be hard to see why there would be any value to an action that there is just some formally normative reason to perform, provided by some hypothetical code of etiquette or some made-­up game; value-­first accounts are not trying to give accounts of these kinds of reasons. Similarly, reasons-­first accounts of value—­that is, buck-­passing accounts of value—­give accounts of value in terms of robustly normative reasons for pro-­attitudes. They do not hold that something is valuable iff there are reasons (merely formal or robustly normative) for everyone to have pro-­attitudes in response to it. That would over-­generate value. Since, according to the Mafia’s 23  See Rowland (2019: 69–72) and Stratton-­Lake and Hooker (2006: 157) for a detailed articulation of this argument and a response to objections to it. 24 See Rowland (2022b). For more on robust and formal normativity see McPherson (2018), Woods (2018), Wodak (2019), Enoch (2019), and Finlay (2019).

Value-First Accounts of Normativity  189 code of omerta, we all have reasons to have pro-­attitudes towards violent ­pay-­back.25 But that there are such (formally normative) reasons does not show that there is any value to such violent pay-­back. Similarly, according to the norms of 1990s high-­school cliques, it is fitting to want bad things to happen to geeks and losers; so it is formally normatively fitting to want bad things to happen to geeks and losers. But this doesn’t mean that fitting attitude accounts of value imply that it is good if bad things happen to geeks and losers. There is no connection between value and such formally normative fittingness. If we accept IV-­first, we can give an account of what makes a standard give rise to robustly normative reasons and robustly normative fittingness. For we can hold that: Value-­First Account of Robust Normativity. For a standard to be robustly ­normative is for it to be non-­instrumentally better (simpliciter) to be guided by that standard than to not be so guided. If we assume that IV-­first is a view about robustly normative reasons and fit and that a standard is robustly normative only if it gives rise to robustly normative reasons and/or fit, then this account of robust normativity—­or something close to it—­is entailed by IV-­first. Elsewhere I’ve argued at length for a value-­based account of robust normativity along these lines.26 So, if we hold IV-­first, we can give an account of robust normativity in terms of value. Reasons-­ first and fittingness-­first views do not seem to have the resources to themselves give accounts of what distinguishes robust from merely formal normativity, since the robustly/merely formally normative distinction is a distinction that arises among reasons and fittingness. It is unclear how these accounts can give an account of why, and what makes, moral reasons authoritatively normative in contrast to ­reasons of etiquette which are not authoritatively normative.27 And there are barely any plausible accounts on the table of what distinguishes merely formally normative reasons from robustly normative reasons. Most accounts of the distinction between robust and formal normativity are accounts of what it is to think that something is that which one robustly (or merely formally) normatively ought to do rather than accounts of what it is for a reason to be robustly rather than merely formally normative.28 And, as I’ve argued elsewhere, there is no similar robustly/formally normative distinction in value.29 25  See McPherson (2018) and Woods (2018). 26  See Rowland (2022b). 27  See Lord and Sylvan (2019). 28 See supra note 24. Could these accounts be built upon to provide accounts of what robust normativity is rather than what it is to think that something is a robustly normative reason? Perhaps. But this work has not been done, and some of these accounts would have very controversial first-­order consequences. For instance, an extended version of Woods’ (2018) account would seem to have subjectivist consequences. And if Lord and Sylvan’s (2019: 66–7) account in terms of agency were extended it would seem to imply that moral reasons are not authoritatively normative because there are agents that are not at all moved by moral standards; see Rowland (2022b: 122–3). 29  See Rowland (2022b: 130–4).

190 Fittingness There were a lot of steps in the argument that I’ve been making. But the general idea is that if we hold IV-­first, we can give a reductive account of what makes a standard or reason robustly normative; if we accept a reasons- or fittingness-­first view we struggle to do this, or at least it is not obvious how we can do this. If this is right, then although IV-­first does not give a reductive account of reasons or fit—­since it does not attempt to give an account of formally normative reasons and fit—­it does give a reductive account of robust normativity. And this means that it is no less qualitatively parsimonious than reasons- and fittingness-­first accounts. The latter accounts reductively analyse value but not robust normativity; IV-­first reductively analyses robust normativity but not reasons or fit. If we assume, with Howard (2019) and Rowland (2019: ch. 11), that we can and should reductively analyse reasons in terms of fit or vice versa, this means that IV-­first is as qualitatively parsimonious as reasons-­first and fittingness-­first accounts.30 So we have a tie in terms of qualitative parsimony. If this is right, then the third reason for wanting a reductive value-­first view to compete with reductive reasonsand fittingness-­first accounts is off the table; IV-­first is no less explanatory, and does not have fewer theoretical virtues, than reasons- and fittingness-­ first accounts in virtue of not offering a reduction of fit or reasons. So, it seems that IV-­first has all the explanatory and theoretical virtues of reductive accounts, and we did not want a reductive account for its own sake, but rather because reductive accounts seem to have these explanatory virtues over non-­reductive views. We’ll explore whether all of this is right in the final section of the chapter when we consider objections to IV-­first and the argument that I’ve been making for it. Before that I want to thoroughly explain how IV-­first can avoid the three problems for direct value-­first views that I outlined in Section 8.2.

8.4  IV-­First Avoids the Problems with Direct Value-­First Accounts In Section 8.2 we saw that direct value-­first accounts face problems with three types of reasons and fit.

8.4.1  Deontological Reasons Unlike direct value-­first accounts, IV-­first does not imply that there are no de­onto­logic­al reasons. It is compatible with IV-­first that we have most reason to ϕ, 30  IV-­first implies that there is value and reasons/fit, but analyses robust and formal normativity in terms of value and reasons/fit, so is committed to only two things. Fittingness- and reasons-­first views imply that there is reasons/fit and robust normativity but analyses value in terms of reasons/fit and robust normativity, and so is also committed to only two things.

Value-First Accounts of Normativity  191 and that it would be fitting to ϕ, even though it would be all things considered better not to ϕ. So, it is compatible with IV-­first that we have most reason to refrain from pushing the heavy man in the footbridge trolley case even though it would be better to push him. This is because it might be non-­instrumentally better to be guided by a standard according to which we should refrain from pushing even though it would be all-­things-­considered better to push; indeed some de­ontolo­gists, such as Kamm (1989: 254), argue for exactly this view. If this is the case, then, according to IV-­First, we have most reason to refrain from pushing even though it would be better not to push.

8.4.2  Fitting Belief and Reasons for Belief Direct value-­first accounts cannot provide a plausible account of all fitting belief and reasons for belief. But IV-­ first can. IV-­ First can hold that it is non-­ instrumentally better to be guided by the correct evidential standard for its own sake, since this involves having positive attitudes to things that are finally valu­ able, such as knowledge and understanding.31 In this case, it is always fitting to believe in line with our evidence and we have reasons to believe things that we have strong evidence for even when there is no value to our believing these things. For instance, it is not valuable to believe that there are 8,294,400 blades of grass on the field at Camp Nou, but if our evidence favours this, then it is fitting for us to believe this and we have (robustly normative) reason to believe this because it is non-­instrumentally better for us to be guided by the correct evidential standard than to not be. And according to this standard, if our evidence favours that p, then it is fitting for us to believe that p and we have reason to believe that p. However, many have argued that actually we shouldn’t think of epistemic ­reasons for belief as robustly normative at all: Cowie (2019) and Maguire and Woods (2020), for instance, argue at length for the view that epistemic reasons are merely formally normative reasons. If knowledge and understanding are not finally valuable—­and considerations such as that nothing is finally valuable are behind Cowie’s (2019) argument that epistemic reasons are formally normative—­ then epistemic reasons are merely formally normative reasons according to IV-­first.32 31  For arguments for and discussion of the view that knowledge and understanding are sometimes finally valuable see Rowland (2022b: 116–17, 2022a) and the references therein. 32  Cowie (2019) and Maguire and Woods (2020) argue that epistemic reasons are merely formally normative even though it is instrumentally better to be guided by the correct epistemic standard. Why not think that there are robustly normative reasons to ϕ if there are reasons to ϕ according to standard S and it is instrumentally better to be guided by S for its own sake than to not be? We shouldn’t say this because it might be instrumentally better for A to be guided by the no-­snitching-­code (or by the norms of masculinity) for its own sake—­because A’s being so guided will stop A from being assaulted or being otherwise seriously harmed. But we shouldn’t think that this makes the no-­snitching-­code or the norms of masculinity robustly normative standards; for they are not robustly normative.

192 Fittingness

8.4.3  Fitting Emotions The other problem for direct value-­first views concerned fitting emotions such as fitting fear and envy. It might seem that IV-­first doesn’t help here because it is not non-­instrumentally or recursively valuable to be guided by standards according to which it is fitting to envy enviable people and according to which there are reasons to fear the fearful. However, I think that with the resources that we’ve used to develop a plausible value-­first view, we can diminish the problem posed by fitting fear and envy for value-­first views. There is only a problem here for value-­first views, such as IV-­first, if we have robustly normative reasons to fear the fearsome and envy the enviable; there is only a problem here if it is fitting in a robustly normative sense to have these attitudes. We might think that reasons to envy the enviable are not robustly normative, loosely speaking because we are not making a normative mistake if we never envy anyone and because the normativity of fear is based in the good instrumental consequences of fearing fearsome things, namely that doing so motivates us to avoid things that are dangerous and harmful to us and those whom we care about. In this case we might understand the normativity of reasons to fear along similar lines to how Cowie and Maguire and Woods understand the normativity of epi­ stem­ic reasons: reasons to fear are formally normative reasons (and the fittingness of fearing the fearful and envying the enviable is not robustly normative). But it is nonetheless for the most part instrumentally beneficial to see there to be reasons to fear fearsome things; the standard on which there are reasons to fear fearsome things, is one that it is instrumentally better to be guided by than to not be guided by. OK, but let’s assume that the fittingness of fitting envy and fear is robustly normative. In this case can value-­first views like IV-­first explain the robust normativity of fitting envy and fear? I think so. If there are robustly normative reasons to fear the fearsome and envy the enviable, then we should hold that we are missing out on the normative world if we never fear the fearsome or envy the enviable. For we are not adequately appreciating the enviable or the fearsome and there are robustly normative reasons to do so; in order to fully appreciate these properties one needs, we might argue, to have the relevant attitudes that they call for, just as we don’t fully appreciate someone’s admirability unless we admire them to some extent. We might then say that to the extent that understanding is finally valuable and having such fear and envy is part of having such understanding, there is something finally valuable about fearing and envying.33 33  IV-­first may seem more plausible here than direct value-­first views. We might say that this value of understanding does not establish that it is always valuable to understand someone as enviable or something as fearsome, so this line of argument does not show that direct value-­first views avoid the problems with fitting envy and fear. But it is valuable to understand the properties of the enviable and the fearsome and this understanding involves sometimes envying enviable things and fearing fearsome

Value-First Accounts of Normativity  193 Furthermore, IV-­first may be able to yield the view that there are robustly normative reasons to envy and fear due to the recursive value thesis. It might be argued that things that are genuinely fearsome are things that have a capacity to do us great bad, and that things that are annoying to us are things that lead to bad experiences for us (think about how annoying music causes a grating experience). More generally, it might be that all right-­kind ­reasons for envy, to be annoyed, and to fear—­and the fittingness of having these emotions—­can be understood in terms of the thick evaluative properties of ­people and things, namely their being enviable, annoying, and fearsome (re­spect­ive­ly). Now there are issues about whether these thick properties can be disentangled into thin (positive and negative) evaluative and descriptive (non-­evaluative) elements. If they can, then, to have one of these thick evaluative properties will be to have a feature that is good or bad in a way. So, suppose that, as sketched above, to be fearsome is to have a capacity to do us great bad. In that case, to fear the fearful would be to have a negatively valenced attitude towards the bad. Given that according to the recursive value thesis discussed in Section 8.1, it is good to hate the bad, it seems that it would then be good to fear the fearful. And it seems somewhat plausible that we can make analogous arguments for the view that it recursively valuable to envy the enviable and to be annoyed by the annoying. But if the property of being enviable and the property of being fearsome cannot be disentangled into thin evaluative and descriptive components and these properties are just irreducible or sui generis evaluative properties, as Jonathan Dancy (1995) and Debbie Roberts (2011) have argued, then we might still hold that it is non-­instrumentally recursively valuable to have evaluative attitudes towards the things that have these properties that match them. The recursive value thesis says that it is good to have positive attitudes towards the good and good to have bad attitudes towards the bad. If there is more to evaluation and evaluative properties than good and bad, as Dancy and Roberts argue, then why not also hold that it is good to have these other evaluative responses to evaluative properties that match them: to envy the enviable and fear the fearsome for instance? To be clear, I’ve not been trying to establish that there are robustly normative reasons to fear or envy or that it is ever robustly normatively fitting to have such attitudes; I am not at all sure that it is robustly normatively fitting to have these attitudes.34 But if you think that it is, there are several ways in which value-­first views, such as IV-­first, can fit with this view of yours.

things. In this case, there is something non-­instrumentally valuable about being at least somewhat guided by standards according to which it is fitting to fear fearsome things and envy enviable people because being somewhat guided by these standards is the only way to fully appreciate, understand, and comprehend the normatively important properties of the fearsome and the enviable. 34  See Rowland (2022b: 117–18).

194 Fittingness

8.5  The Prospects of a Value-­First Account of Reasons and Fit I’ve been exploring whether an indirect value-­first account of fittingness and ­reasons can avoid the problems that direct value-­first accounts face. In particular I’ve been exploring the plausibility of: IV-­First. For R to be a genuinely normative reason to ϕ is for R to be a reason to ϕ according to a standard that it is non-­instrumentally better simpliciter for us to be guided by for its own sake. For it to be fitting to ϕ is for it to be fitting to ϕ according to a standard that it is non-­instrumentally better simpliciter for us to be guided by for its own sake. Even if IV-­first avoids the problems that direct value-­first accounts face, IV-­first may lead to problems of its own. One worry about IV-­first is that it over-­generates fittingness. Suppose that an evil demon will punish everyone unless we are guided by a standard for fitting admiration according to which evil demons are admirable for their own sake; call this the demon’s standard. It might seem that in this case it is non-­instrumentally better to be guided by the demon’s standard for its own sake than to not be guided by it, since, being guided by that standard for its own sake involves a valuable appreciation of the value of those who would be harmed by the demon if we were not guided by this standard. So, it is non-­instrumentally recursively better to be guided by this standard for its own sake than to not be guided by it. And if this is right, IV-­first implies that it is fitting to admire the evil demon. And it is not. So, it over-­generates fittingness. It seems to me that worries about this particular case can be overcome. If one is  truly guided by the demon’s standard, then one does not thereby have pro-­ attitudes to those who would be harmed by the demon if one were not guided by this standard for its own sake. Being truly guided by this standard for its own sake just involves internalizing and endorsing the idea that evil demons have genuinely admirable qualities; it does not involve having pro-­attitudes towards anyone’s value. It is true that it is non-­instrumentally better to have very strong pro-­ attitudes towards the value of people and that if we have these pro-­attitudes we might be motivated to internalize the demon’s standard to avoid p ­ eople being harmed. But having such pro-­attitudes is not part of what it is to be guided by the demon’s standard for admiration for its own sake. So, IV-­first does not over-­ generate fittingness in this evil-­demon case. Another type of worry about IV-­ first concerns its similarity to rule-­ consequentialism. First, it might seem that IV-­first is objectionable because it implies that rule-­consequentialism is the correct moral theory. However, IV-­first definitely does not imply the truth of rule-­ consequentialism. Since rule-­ consequentialism implies that there is moral reason to perform an action only if the moral code the internalization of which has the best consequences overall

Value-First Accounts of Normativity  195 says  that there is moral reason to perform that action. But IV-­first is not concerned with the overall consequences of internalizing or being guided by a particular code or standard; it is only concerned with the non-­instrumental value of being guided by a standard: there could be great instrumental value to being guided by a particular code but no non-­instrumental value to being guided by it. (For example, in certain circumstances there is great instrumental value to being guided by the no-­snitching code or the norms of masculinity because being so guided will stop one from being harmed. But this does not establish that there is any non-­instrumental value to being guided by these standards). Alternatively, it might seem that IV-­first will face analogous objections to those which rule-­consequentialism faces, such as the objection that it involves an objectionable form of rule worship. This might be true. But, of course, analogous replies will also be available. More work needs to be done to see whether the similarity of IV-­first to rule-­consequentialism presents a problem for it. But it is not obvious that it does—­or that it does not. Perhaps the most significant issue with IV-­first is whether it really has as many virtues as the fittingness-­first and reasons-­first views with which it competes. In Section 8.3 I tried to make the strongest possible case that it does. But I’m not sure that this case stands. And more needs to be done to show that this case does stand. First, a proponent of IV-­ first needs to (i) show that proponents of fittingness-­first and reasons-­first accounts cannot give a plausible account of robust normativity, (ii) show that a value-­first account of robust normativity along the lines of that which I suggested in Section 8.3 is plausible, and (iii) show that the formal/robust normativity distinction does not arise within value. I’ve argued for (ii–iii) elsewhere.35 But others may disagree with the arguments that I have made. Second, a proponent of IV-­first may need to show that they can give an account of what unifies all the different types of goodness (attributive goodness, prudential goodness, and goodness simpliciter), that is, what makes them all varieties of goodness. If they cannot do this, then, similarly, IV-­first will be less explanatory than fittingness-­first and reasons-­first views. Since the latter views can give accounts of what unifies all varieties of goodness in terms of reasons and/or fit. They can provide a unified explanation of all different forms of goodness and value in terms of the different sets of agents whom it is fitting to, and who have reasons to, have particular pro-­attitudes. For instance, they can hold that for something to be good simpliciter is for there to be (right-­kind) reasons for everyone, or for it to be fitting for everyone, to have pro-­attitudes towards it; for instance, for pleasure to be good simpliciter is for it to be fitting for everyone to desire it. Attributive and prudential goodness can be understood in terms of the reasons for pro-­attitudes that there are for a more restricted set of agents. For a

35  See Rowland (2022b).

196 Fittingness Lakeland knife to be a good knife—­that is, have the attributive goodness property of being good as a knife—­is for there to be reasons for anyone who has reason to want a knife to want a Lakeland knife; and for exercise to be good for me is for it to be fitting for anyone who has reason to care about me to have pro-­attitudes towards my exercising because they have reason to care about me.36 If proponents of an IV-­first view cannot give an account of what unifies all the varieties of goodness, then they are left with an account on which it cannot be explained what makes these different goodness properties all goodness properties.37 In this case IV-­first would not have an explanatory virtue which reasons- and fittingness-­first have, since these views can explain this.38 Even if proponents of IV-­first can do everything that I’ve been explaining that they need to do in order to show that IV-­first is as plausible and explanatorily powerful as reasons- and fittingness-­first views, it’s not clear to me that we will be left with anything but a tie between IV-­ first and fittingness-­ /reasons-­ first accounts. It’s true that for some things, perhaps most clearly action, value can at least often seem to intuitively come before reasons and fit. But with fitting envy and fear, the fittingness seems to intuitively come before the value. So intuitions of explanatory priority—­if they should be trusted—­don’t seem to tilt towards IV-­ first. And IV-­first doesn’t preserve the spellbinding intuition that we always have most reason to do what’s best. For IV-­first implies that what we have most reason to do can be something that isn’t the best thing to do. So, although the spellbindingness intuition might give us reason to accept certain direct value-­first accounts it does not give us reason to accept IV-­first.39 This is all to say that a more work needs to be done to show that: (a) an indirect value-­first approach can secure all the explanatory advantages that reasons-­first and fittingness-­first views have; (b) an indirect value-­first approach doesn’t face new insuperable objections; and (c) that there are independent reasons to favour an indirect value-­first view that could show that we should accept it rather than a fittingness-­first or reasons-­first view. In some ways this is unsurprising. A lot of work has been done on reasons-­first—­and increasingly on fittingness-­first—­ accounts of value and normativity more generally. In order for an indirect

36  See Rowland (2016a, 2019: ch. 5) and Schroeder (2010). 37  I assume that if we accept IV-­first, we may be able to explain or analyse attributive goodness properties in terms of formally normative reasons and prudential goodness properties in terms of robustly normative prudential reasons. So, I expect that IV-­first will not be less parsimonious than reasons- or fittingness-­first views on account of not being able to analyse attributive or prudential goodness. But that leaves open what the unity is among these varieties of goodness and goodness simpliciter, since according to IV-­first goodness simpliciter is not analysed in terms of reasons; cf. Rowland (2019: ch. 5). 38  See Rowland (2019: ch. 5). 39  Of course, this fact is related to an important virtue of IV-­first, namely that it is compatible with the existence of deontic reasons. My point is just that we do not have reasons to accept IV-­first in virtue of spellbindingness. So, some of the initial attractions of value-­first views are not attractions of IV-­first.

Value-First Accounts of Normativity  197 value-­first approach to be shown to be a serious competitor to these views a similar amount of work would have to be done on all aspects of it. And this amount of work cannot be done in one chapter. But what I do hope to have done is show that an indirect value-­first approach can avoid the problems that have led those of us attracted to reasons-first and fittingness-­first approaches to value and normativity to reject direct value-­first views. And that there may be reasons to be optimistic that if the hard work is done on such an indirect value-­first account of reasons and fit, such an account may be shown to be a strong competitor to fittingness-­first and reasons-­first approaches to value and to normativity more generally.

References Audi, Robert (2005). The Good in the Right. Princeton UP. Brentano, Franz (1969). The Origin of Our Knowledge of Right and Wrong. Translated by Roderick M. Chisholm and Elizabeth Schneewind. Routledge. Cowie, Chris (2019). Morality and Epistemic Judgment. OUP. Crisp, Roger (2008). ‘Goodness and Reasons: Accentuating the Negative’. Mind 117 (466): 257–65. Dancy, Jonathan (1995). ‘In Defense of Thick Concepts’. Midwest Studies in Philosophy 20: 263–79. Enoch, David (2019). ‘Is General Jurisprudence Interesting?’ In David Plunkett, Scott Shapiro, and Kevin Toh (eds), Dimensions of Normativity: New Essays on Metaethics and Jurisprudence. OUP. Finlay, Stephen (2019). ‘Defining Normativity’. In David Plunkett, Scott Shapiro, and Kevin Toh (eds), Dimensions of Normativity: New Essays on Metaethics and Jurisprudence. OUP, 187–220. Foot, Philippa (1985). ‘Utilitarianism and the Virtues’. Mind 94: 196–209. Heuer, Ulrike (2004). ‘Raz on Values and Reasons’. In R.  Jay Wallace, Philip Petit, Samuel Scheffler, and Michael Smith (eds), Reason and Value: Themes from the Moral Philosophy of Joseph Raz. OUP. Hooker, Brad (2000). Ideal Code, Real World: A Rule-Consequentialist Theory of Morality. OUP. Howard, Chris (2018). ‘Fittingness’. Philosophy Compass 13 (11): 1–14. Howard, Chris (2019). ‘The Fundamentality of Fit’. In Russ Shafer-Landau (ed.), Oxford Studies in Metaethics, Volume 14. OUP. Hurka, Thomas (2001). Virtue, Vice, and Value. OUP. Kagan, Shelley. (1998). ‘Rethinking Intrinsic Value’. The Journal of Ethics 2: 277–97. Kamm, Frances (1989). The Limits of Morality. Clarendon Press. Kiesewetter, Benjamin (2022). ‘Are All Practical Reasons Based on Value’. In Russ Shafer-Landau (ed.), Oxford Studies in Metaethics, Volume 17. OUP.

198 Fittingness Lord, Errol, and Kurt Sylvan (2019). ‘Reasons: Wrong, Right, Normative, Fundamental’. Journal of Ethics and Social Philosophy 15 (1): 43–74. Maguire, Barry (2016). ‘The Value-Based Theory of Reasons’. Ergo 3 (9): 233–62. Maguire, Barry, and Jack Woods (2020). ‘The Game of Belief ’. Philosophical Review 129 (2): 211–49. McHugh, Conor, and Jonathan Way (2016). ‘Fittingness First’. Ethics 126: 575–606. McPherson, Tristram (2018). ‘Authoritatively Normative Concepts’. In Russ ShaferLandau (ed.), Oxford Studies in Metaethics, Volume 13. OUP. Moore, G. E. (1903). Principia Ethica. CUP. Nozick, Robert (1981). Philosophical Explanations. Harvard UP. Olson, Jonas (2006). ‘G. E. Moore on Goodness and Reasons’. Australasian Journal of Philosophy 84 (4): 525–34. Orsi, Francesco (2013). ‘What’s Wrong with Moorean Buck-Passing’. Philosophical Studies 164 (3): 727–46. Parfit, Derek (2011). On What Matters, Volume 1. OUP. Plunkett, David, Scott Shapiro, and Kevin Toh (2019). Dimensions of Normativity: New Essays on Metaethics and Jurisprudence. OUP. Portmore, Douglas (2011). Commonsense Consequentialism: Wherein Morality Meets Rationality. OUP. Protasi, Sara (this volume). ‘The Things We Envy: Fitting Envy and Human Goodness’. Rabinowicz, Wlodek, and Toni Rønnow-Rasmussen (2000). ‘A Distinction in Value: Intrinsic and for its Own Sake’. Proceedings of the Aristotelian Society 100 (1): 33–51. Rabinowicz, Wlodek, and Toni Rønnow-Rasmussen (2004). ‘The Strike of the Demon: On Fitting Attitudes and Value’. Ethics 114 (3): 391–423. Raz, Joseph (2001). Value, Respect, Attachment. CUP. Roberts, Debbie (2011). ‘Shapelessness and the Thick’. Ethics 121 (3): 489–520. Rowland, Richard (2016a). ‘Reasons as the Unity among the Varieties of Goodness’. Pacific Philosophical Quarterly 97 (2): 200–27. Rowland, Richard (2016b). ‘In Defence of Good Simpliciter’. Philosophical Studies 173 (5): 1371–91. Rowland, Richard (2019). The Normative and the Evaluative: The Buck-Passing Account of Value. OUP. Rowland, Richard (2022a). ‘Reasons as Reasons for Preferences’. American Philosophical Quarterly 59 (3): 297–311. Rowland, Richard (2022b). ‘The Authoritative Normativity of Fitting Attitudes’. In Russ Shafer-Landau (ed.), Oxford Studies in Metaethics, Volume 17. OUP. Scanlon, T. M. (1998). What We Owe to Each Other. Harvard UP. Schroeder, Mark (2007). Slaves of the Passions. OUP. Schroeder, Mark (2010). ‘Value and the Right Kind of Reason’. In Russ Shafer-Landau (ed.), Oxford Studies in Metaethics, Volume 5. OUP.

Value-First Accounts of Normativity  199 Schroeder, Mark (2011). ‘Buck-Passers ‘Negative Thesis’. Philosophical Explorations 12 (3): 341–7. Stratton-Lake, Philip, and Brad Hooker (2006). ‘Scanlon versus Moore on Goodness’. In Terry Horgan, and Mark Timmons (eds), Metaethics after Moore. OUP. Suikkanen, Jussi (2005). ‘Reasons and Value–In Defence of the Buck-Passing Account’. Ethical Theory and Moral Practice 7(5): 513–35. Way, Jonathan (2013). ‘Value and Reasons to Favour’. In Russ Shafer-Landau (ed.), Oxford Studies in Metaethics, Volume 8. OUP. Wedgwood, Ralph (2009). ‘Intrinsic Values and Reasons for Action’. Philosophical Issues 19 (1): 321–42. Wodak, Daniel (2019). ‘Mere Formalities: Fictional Normativity and Normative Authority’. Canadian Journal of Philosophy 49: 828–50. Woods, Jack (2018). ‘The Authority of Formality’. In Russ Shafer-Landau (ed.), Oxford Studies in Metaethics, Volume 13. OUP.

9 Feasibility and Fitting Deliberation Nicholas Southwood

9.1  Introduction There is a venerable tradition stretching back to Broad (1930) and Ewing (1939; 1947) and well and truly alive today (see e.g. D’Arms and Jacobson 2000; Cullity 2018, ch. 2; cf. Scanlon 1998) that employs the idea of fittingness—­in particular, the idea of fitting attitudes—­to explain truths about value or desirability.1 More recently, a number of scholars have also appealed to fittingness to explain truths involving deontic phenomena such as our reasons for action (see e.g. Way 2017; Cullity 2018: ch. 2; this volume) and even normativity in general (see e.g. Chappell 2012; McHugh and Way 2016; Howard 2019). My question is: might we also look to fittingness, in addition or instead, to explain truths about feasibility? Elsewhere, I have proposed an account of feasibility of just this kind (Southwood 2022). In particular, I have suggested that truths about feasibility should be understood in terms of truths about fitting deliberation. According to this Fitting Deliberation Account of feasibility, whether an action counts as feasible is a matter of whether it has features that are required to be a fitting object of practical deliberation, or deliberation about what to do. I have argued that the Fitting Deliberation Account appears to be unique in being able to do justice to our practices of making and using feasibility claims while providing us with a principled and plausible basis for settling important questions about feasibility—­for ex­ample, whether feasibility has an essential volitional dimension, whether it is at odds with counterfactual flukiness, and so on. If I am right, this suggests an important way in which truths about fittingness might be relevant to certain normative truths, even if it turns out that we cannot fully explain such truths in terms of truths of fittingness (see Rowland 2017). My aim in what follows is to further develop and explore this Fitting Deliberation Account of feasibility—­in two main ways. First, I want to consider two important

1  I am grateful to the editors for helpful comments on an earlier version of the chapter and to many friends and colleagues, especially Garrett Cullity, Tom Dougherty, Sarah Fine, Philip Pettit, Zofia Stemplowska, and David Wiens for valuable discussion about the Fitting Deliberation Account. Research for the chapter was supported by an Australian Research Council Future Fellowship (FT160100409).

Nicholas Southwood, Feasibility and Fitting Deliberation In: Fittingness: Essays in the Philosophy of Normativity. Edited by: Christopher Howard and R. A. Rowland, Oxford University Press. © Nicholas Southwood 2022. DOI: 10.1093/oso/9780192895882.003.0009

Feasibility and Fitting Deliberation  201 challenges to it. The challenges concern two significant features of our talk and thought about feasibility that the Fitting Deliberation Account appears ill placed to vindicate: objectivity and non-­normativity. The objectivity challenge arises given a very natural picture of correct deliberation as depending, not on what properties our actions in fact have, but (exclusively) on our (reasonable) beliefs about whether our actions have those properties. The non-­normativity challenge arises because, whereas feasibility appears to be a non-­normative property, the fittingness of fitting deliberation is a normative property and one that appears to depend on further normative properties. I will suggest that the challenges can be met and that doing so provides us with a valuable opportunity to make important clarifications of the Fitting Deliberation Account. Second, I want to contrast the Fitting Deliberation Account with a rival account that also purports to explicate feasibility in terms of fittingness but that looks to fitting prescription rather than fitting deliberation (cf. McGeer and Pettit 2015). A Fitting Prescription Account might seem to share many of the advantages of the Fitting Deliberation Account. Nonetheless, I will argue that it faces distinctive and serious problems that the Fitting Deliberation Account avoids. Several quick remarks about terminology before we begin. By the ‘fittingness’ of an action or activity or attitude I will mean that it conforms with standards of correctness that are internal to, or constitutive of, the action or activity or attitude in question in the sense that they describe what is needed to fulfil the constitutive aim associated with actions or activities or attitudes of that type (see McHugh and Way 2016, pp. 598–60; cf. Howard 2017, p. 9, n. 4). By ‘feasibility’ I will mean a binary status2 that attaches to the performance of particular actions by particular (individual or collective) agents at particular times (see Southwood 2022; cf.  Gilabert and Lawford-­Smith 2012). By ‘practical deliberation’ I will mean a special kind of activity, namely the exercise of practical reason, which I shall take to be the activity, not necessarily conscious or explicit, of reflective decision-­ making, or undertaking to settle the question of what to do by bringing to bear relevant considerations on the question of whether (or when or how) to perform a particular action (see Southwood 2019).

2  An important challenge that I won’t try to address here concerns the alleged incapacity of the Fitting Deliberation Account to explain feasibility’s apparent gradability: the fact that, rather than merely involving a binary status, feasibility also appears to come in degrees (see Gilabert and Lawford-­ Smith 2012; Lawford-­Smith 2013). Briefly, my view is that this claim about feasibility’s gradability rests on flimsy foundations. Its appeal derives principally from the fact that it appears to be supported by certain linguistic practices. However, we should be wary about assigning much if any significance to such data in adjudicating disputes about feasibility’s conceptual contours (see Southwood 2022). We certainly have a need for various gradable notions—­most obviously the notion of probability that figures in, for example, our judgements regarding expected utility. But it is not clear that a gradable notion of feasibility in particular adds anything of substance to the set of gradable notions that we already have in our conceptual toolkit.

202 Fittingness

9.2  The Fitting Deliberation Account There are many different accounts of feasibility. Some point to some kind or ­flavour of possibility (see e.g. Wiens 2015); others to (counterfactual) probability of success (see e.g. Gilabert and Lawford-­Smith 2012); still others to our dispositions (see e.g. Southwood 2016a); others again to some combination of these elem­ents (see e.g. Stemplowska 2016). While these accounts are importantly different, they are united in embracing a certain metaphysical picture. According to this picture, feasibility is to be identified with some modal status, and the central question is simply to say what this (feasibility-­constituting) modal status is supposed to be; it is a further, downstream question what kind of normatively significant functional status, if any, feasibility might have. But there is an alternative picture. The alternative picture holds instead that feasibility is to be identified with some functional status—­the status of playing some normatively significant functional role—­that may be gauged by attending to the ways in which we use and treat it as appropriate to use claims about feasibility. The relevant (feasibility-­constituting) functional status may be realized by some modal status but is not identical with any such modal status. Rather, for an action to be feasible just is for it to have whatever features are required to play the relevant functional role. To be infeasible is for it to lack such features. If one is tempted by this alternative picture, the obvious question is: how exactly should we characterize this normatively significant functional role? Perhaps the most natural first thought is to look to the role that claims about feasi­bil­ity are supposed to play with regard to (certain) deontic claims about what we ought to do—­in particular, in constraining the truth of such claims (via something like the principle that ‘ought’ implies ‘can’). But, in spite of its initial appeal, I believe we have good reason to reject a view of this kind. Consider the fact that we are disposed to treat judgements of infeasibility as automatically rendering inappropriate deliberation about whether (or when to how) to perform the actions in questions. (For example, it would seem obviously inappropriate to deliberate about when to climb the Matterhorn insofar as we take it to be infeasible to do so (say, because of advanced age, or ill health, or lack of fitness).) The ought-­constraining view cannot satisfactorily explain this. To be sure, it can explain why we are disposed to treat infeasibility as automatically rendering false claims to the effect that we ought to perform the actions in question. But it does not follow from the fact that it is not the case that we ought to perform an action that there is anything inappropriate about deliberating about whether (or when or how) to perform it. (Think of deliberating about whether to perform a supererogatory action.) The Fitting Deliberation Account is based on a different view of feasibility’s functional role, according to which it should be characterized in terms of fittingness (rather than ought)—in particular, in terms of fitting deliberation.

Feasibility and Fitting Deliberation  203 The role of feasibility is to determine which actions are fitting objects of practical de­lib­er­ation. Thus, feasible actions are those that are fitting objects of practical de­lib­er­ation. Infeasible actions are those that are not fitting objects of prac­tical deliberation. What exactly does it mean to say that an action is ‘a fitting object of practical deliberation’? Part of what it means is simply that the action is such that deliberating about whether (or when or how) to perform it accords with certain standards of correct deliberation—­that it is a deliberation-­worthy action, as we might put it. Notice that this is quite different from saying that the action is a decision-­worthy action—­that a decision to perform the action would be a fitting decision (still less the uniquely fitting decision). A decision to do or not to do something is an outcome of deliberation. Hence, the question of whether an action is decision-­worthy is a matter of whether the action has features to which a decision would be a fitting response—­features which in deliberating we are seeking to identify, weigh up, and bring to bear on the question of what to do. By contrast, deliberation involves undertaking to make a decision about whether (or when or how) to do something. Hence, the deliberation-­worthiness of an action is a matter of whether the action has whatever features are needed for undertaking to make a decision about whether (or when or how) to perform it to constitute correct deliberation. To be decision-­worthy, an action must be deliberation-­worthy (or so I shall assume), but not vice versa. Thus, in holding that claims about feasibility are supposed to determine which actions count as fitting objects of practical de­lib­er­ation, part of what we are saying is that they play an option-­fixing role, as opposed to an option-­adjudicating role. That is, they play the role of determining whether actions should be included in, or excluded from, the set of actions that are candidates for consideration as to whether or not to perform them. They do not play the role of determining which actions, among some set of actions, are the actions to perform. Moreover, to say that an action is a ‘fitting object of practical deliberation’ means, not only that deliberating about whether (or when or how) to perform it accords with certain standards of correct deliberation, but that such deliberation accords with certain internal or constitutive standards of deliberation—­standards that are given by what is needed to achieve deliberation’s constitutive aim. This is  quite different from saying that either the action itself or deliberating about whether (or when or how) to perform the action accords with external standards such as standards of morality or prudence. An action may be a perfectly fitting object of practical deliberation in spite of the fact that it would be deeply morally or prudentially inappropriate to perform it or deliberate about it, so long as it possesses whatever features are required to achieve deliberation’s constitutive aim. Thus, according to the Fitting Deliberation Account, claims about feasibility do not merely play an option-­fixing role. Rather, they play a special kind of option-­ fixing role. In particular, they fix our options by determining which actions lie

204 Fittingness within what I have called elsewhere our domain of deliberative jurisdiction—­that is, which actions are such that deliberating about whether (or when or how) to perform them counts as correct or incorrect qua deliberation. This naturally raises the difficult question: what is deliberation’s constitutive aim? I have argued elsewhere that it involves having the aim of manifesting a special kind of self-­determination or self-­authorship whereby we aim to determine how we will act by deciding how to act (see Southwood 2016a). Thus, for an action to be a fitting object of practical deliberation is a matter of its having whatever features are required for deliberating about whether (or when or how) to perform it to achieve this aim. This suggests three different ways in which an action might fail to be a fitting object of practical deliberation and, hence, feasible according to the Fitting Deliberation Account. First, there might be some insurmountable impediment to our bringing about the performance of the action at all. Second, there might be an insurmountable impediment to our deciding to perform the action, Third, there might be some insurmountable impediment to our decision making it the case that we succeed in performing the action—­to our performing the action by deciding to do so—­and, hence, manifesting the requisite self-­ authorship with regard to the action in question. The Fitting Deliberation Account has a number of significant virtues. First, the Fitting Deliberation Account squares well with our practices of making and using feasibility claims. We do as a matter of fact appear to treat feasibility as playing the role of determining the domain of fitting deliberation; and other rival views of feasibility’s functional role such as the ought-­constraining account cannot explain it (since, as we saw, there needn’t be anything deliberatively untoward about deliberating about whether to perform actions that it is not the case that we are required to perform). Second, the Fitting Deliberation Account provides us with a principled basis for settling other difficult and controversial questions about feasibility—­by looking to what is required to do justice to the role of determining which actions are fitting objects of practical deliberation, rather than, say, to our practices of linguistic usage. Consider, for example, the question of whether feasi­ bil­ity requires volitional capacity: i.e. whether we must be able to bring ourselves to perform actions in order for the actions to be feasible for us to perform (see e.g. Estlund 2011 versus Wiens 2016). Here is a reason to think that feasibility does indeed require volitional capacity. Volitional capacity is needed in order for de­lib­er­ation about whether (or when or how) to perform an action to achieve its constitutive aim (i.e. for us to manifest the relevant form of practical self-­ determination). If the action is volitionally inaccessible to us such that there is some insurmountable impediment to our deciding to perform it, this means that there is no way for us to perform the action by deciding to perform the action. Thus, only volitionally accessible actions are fitting objects of practical deliberation and, hence, feasible. Third, unlike its rivals, the Fitting Deliberation Account offers an admirably straightforward explanation as to why claims about feasibility

Feasibility and Fitting Deliberation  205 are to be interpreted in certain fitting deliberation-­compatible ways. Many of the existing accounts of feasibility simply fail to deliver an interpretation of feasibility that is consistent with what it takes to play the role of determining fitting de­lib­er­ ation at all. For example, counterfactual probability-­based accounts imply that volitionally inaccessible actions are feasible so long as we would succeed in performing them if we were somehow to decide to perform them (see Stemplowska 2016; 2021). Others (such as possibility-­based accounts) can only succeed in doing so by making arbitrary or ad hoc stipulations, such as, restricting the domain of actions to which feasibility applies to those that are not counterfactually fluky (see Schwartz 2020; cf. Southwood and Wiens 2016). (The problem arises because counterfactually fluky actions (such as saving the life of an innocent person by hitting the bulls-­eye of a dartboard 100 times in a row) appear to be perfectly possible (irrespective of our preferred notion or in­ter­pret­ation of possibility), but they are not fitting objects of practical deliberation; it would hardly be correct to deliberate about whether (or when or how) to save the life of the innocent person.) According to the Fitting Deliberation Account, by contrast, feasible actions are those that have certain features (that are required to play the role of determining the domain of fitting deliberation) because to be feas­ible just is to have whatever features are needed to play this role.3 In sum, the Fitting Deliberation Account is an account with considerable appeal.

9.3  Two challenges Nonetheless, in spite of its appeal, the Fitting Deliberation Account also faces a number of serious challenges. I want to focus on two challenges in particular.

9.3.1  Objectivity The first challenge is that the Fitting Deliberation Account appears to be incompatible with feasibility’s ostensible objectivity or fact-­dependence. By this I mean that what counts as feasible would appear to be determined by certain facts about, or objective features of, the actions whose feasibility is in question—­as opposed to

3  This naturally suggests an interpretation of the Fitting Deliberation Account according to which it is a purely ‘formal’ account of feasibility that is, in principle, compatible with various other ‘substantive’ accounts such as possibility-­ based accounts, probability-­ based accounts, disposition-­ based accounts, and so on. I think this is fine so long as we understand the alternative accounts as merely providing an account of what realizes the functional role of determining the domain of fitting de­lib­er­ ation. However, it is clearly not right insofar as the alternative accounts purport to be something more: say, accounts of what it is to be feasible. I am grateful to Richard Rowland for raising this issue.

206 Fittingness our attitudes or evidence about whether the actions have such features.4 What counts as fitting deliberation, by contrast, might appear to be a subjective or attitude-­dependent matter inasmuch as it is determined by what we believe (or reasonably believe) to be the case about the actions in question. Suppose that you have inherited a valuable painting and are deliberating about whether to donate it to a museum (as opposed to, say, keeping it, or selling it for a tidy profit). However, unbeknownst to you, there has been a fire in the home of your deceased relative in the course of which the painting has been destroyed. It might seem that there is nothing wrong with such deliberation qua deliberation in spite of the fact that donating the painting is obviously not feasible. To be sure, you are deliberating on the basis of a false belief: that the painting is intact. Yet you are deliberating perfectly correctly given this false but reasonable belief. Or suppose that you wake up in hospital following a surfing accident and are deliberating about whether to get out of bed and go to get a glass of water. However, unbeknownst to you, you have broken your neck and are now a quadriplegic. Going to get a glass of water is obviously not feasible. However, again, it might seem that there is nothing wrong with such deliberation qua deliberation. As we might put it, channelling John Broome, the correctness of correct de­lib­er­ ation appears to supervene on the mind (Broome 2013). Nonetheless, I think we should reject this kind of subjectivism about fitting deliberation.5 What is certainly true is that the rationality of our deliberating about whether to perform particular actions is determined by our (reasonable) beliefs. Thus, it may be perfectly rational for us to deliberate about whether to donate the painting from our deceased relative given that we (reasonably, though falsely) believe it to be intact; and perfectly rational to deliberate about whether to get up out of bed and go to get a glass of water given that we (reasonably, though falsely) believe our spinal cord to be intact. This is hardly surprising. In deliberating about whether to perform an action, we are, in effect, treating the action as a fitting object of practical deliberation. It can surely be rational, if we believe an action to be a fitting object of practical deliberation, or have beliefs whose truth would entail that it is a fitting object of practical deliberation, to treat the action as

4  To be sure, the relevant feasibility-­determining facts include certain psychological facts, including certain facts about our attitudes themselves. Whether it is feasible for us to do certain things is often determined, in part, by whether we know how to do the things in question, which is obviously at least partly a matter of having certain beliefs about what is the case. (For example, it is not feasible for me to perform successfully a complex bit of gastrointestinal surgery since doing so obviously requires having certain knowledge (and, hence, beliefs) about the human digestive system that I lack.) Nonetheless, the point is that what counts as feasible is determined by whether the actions in fact have the relevant objective features, as opposed to our beliefs or evidence about whether the actions have such features. 5 A very different response would be to embrace some kind of subjectivism about feasibility (cf. Hedden 2012). Subjectivism about feasibility potentially squares well with certain (e.g. decision-­ theoretic) ways of thinking about practical reason. However, I won’t say anything more about it here.

Feasibility and Fitting Deliberation  207 such; it does not matter whether or not the action is in fact a fitting object of practical deliberation. The question is whether we should embrace the further claim that the fittingness of our deliberating about whether to perform a particular act is determined by our (reasonable) beliefs. Here is a reason to think not. It would seem to require an implausible view of deliberation’s constitutive aim, as consisting in some merely internal or subjective aim, such as achieving coherence among one’s mental states. But such a view is indefensible. To see why, notice that practical de­lib­er­ation is oriented at the performance of actions, which are partly a matter of what happens in the world. This is not to endorse the view, often associated with Aristotle, that actions are the outcome of deliberation. Rather, actions are part of what deliberation aims at. In deliberating, our aim is, in part, to determine how we will act. There are different views about what exactly this involves. As I noted above, my own view is that the aim of deliberation involves practical self-­determination, or determining whether we will perform certain actions by determining whether to perform them (Southwood 2016a). But, whether or not this is exactly right, so long as the aim of deliberation involves the performance of actions, then it would seem that what is required to achieve this aim cannot be a purely internal or subjective matter. For example, if an action is in fact meta­phys­ic­al­ly or nomologically impossible, then deliberating about whether to perform it obviously cannot achieve deliberation’s constitutive aim since there is no way of determining that we will perform it even insofar as we determine whether to perform it by deciding to do so. The fact that we happen to believe falsely that the action is metaphysically and nomologically possible cannot change the fact. It seems, then, we have compelling reasons to reject subjectivism about fitting deliberation in spite of its considerable prima facie plausibility.6 Should we nonetheless accept a form of semi-subjectivism, as it were, according to which fitting deliberation is determined in part by our (reasonable) beliefs about what is the case? Suppose that you reasonably but mistakenly believe the valuable painting to have been destroyed in a fire in the home of your deceased relative, whereas in fact it was moved to a storage facility the day before the fire and is perfectly intact. Or suppose that, following a serious surfing accident, you reasonably but mistakenly believe yourself to be quadriplegic when you wake up in the spinal unit of a hospital. Under these circumstances, it might seem that deliberating about whether 6  It is worth noting that whereas subjectivism about fitting deliberation has considerable prima facie plausibility, I take it that there is little temptation to embrace subjectivism about fittingness in certain other contexts where it arises or in general. Take fitting attitudes accounts of value. Proponents of such accounts since Broad (1930) and Ewing (1939) have apparently without exception been objectivists about fittingness. This is unsurprising. A fitting attitudes account of value based on subjectivism about fitting attitudes would be a non-­starter. For example, it would imply that drinking the liquid in the glass that one believes to be water but that is in fact petrol would be good or desirable. By contrast, while I have argued that subjectivism about fitting deliberation is also false, it is not obviously false. Indeed, it might seem true.

208 Fittingness to donate the painting and whether to get up out of bed and go to find a glass of water would clearly not count as fitting deliberation in spite of the fact that the painting and your spinal cord are perfectly intact. Moreover, such cases might seem to show that whether an action counts as a fitting object of practical de­lib­er­ ation is determined both by certain facts and by our beliefs about those facts (or perhaps by our knowledge of the facts). If so, this would still be bad news for the Fitting Deliberation Account of feasibility since it would imply that it is infeasible for you to donate the painting to the museum and infeasible for you to go and get a glass of water, whereas these things are surely perfectly feasible. However, it seems to me that even this form of semi-­subjectivism cannot be maintained. Suppose that you do in fact deliberate about whether to donate the painting, or whether to go and get a glass of water. This is, of course, highly unlikely to happen, given your belief that the painting has been destroyed and your belief that you are quadriplegic. But suppose that you have just woken up  and, being still a trifle dopey, have temporarily forgotten about the fire in which you believe the painting to have been destroyed, or the surfing accident in which you believe yourself to have been made a quadriplegic (cf. Broome 2013). Do your false but reasonable beliefs make it the case that donating the painting and going to get a glass of water are not fitting objects of practical deliberation? Let’s grant for the sake of argument that your beliefs suffice to make it irrational for you to deliberate about whether to donate the painting, or whether to go and get a glass of water. But, as we have already seen, what is fitting and what is rational may come apart. So the crucial question is whether engaging in such deliberation would be compatible with achieving deliberation’s constitutive aim. It seems to me that it clearly would be. There is simply no relevant impediment that would prevent your deliberation achieving exactly what it is supposed to achieve. Suppose, for example, that you were to conclude your dopey de­lib­er­ ation about whether to donate the painting by deciding to donate the painting, ringing up the lawyer in charge of the estate to communicate your decision to her, in consequence of which the painting is successfully donated to the museum. Or again suppose that you were to conclude your dopey deliberation about whether to go and get a glass of water by deciding to get a glass of water, in consequence of which you get out of bed and go to get the glass of water. In either case, you would surely count as having manifested the relevant kind of self-­determination: determining how you will act (i.e. donating the painting and getting the glass of water) by determining how to act (i.e. deciding to donate the painting and deciding to get the glass of water). Just to be clear, I am not denying that our beliefs about the modal status of our actions are sometimes relevant to what counts as fitting deliberation. Consider, in particular, cases in which we are disposed to succeed or fail in doing things only insofar as we possess or lack certain beliefs about what it is feasible for us to do. For example, a sprinter might be disposed to succeed in running 100 metres only

Feasibility and Fitting Deliberation  209 insofar as she believes that she can. A mountaineer might be disposed to succeed in scaling a particularly challenging rock-­face so long as but only so long as she doesn’t believe that she can’t. Under these circumstances, in order to reach a verdict about whether the relevant actions are fitting objects of deliberation, we must indeed consider the agents’ beliefs about what is feasible. Notice, however, that this interesting and important phenomenon is perfectly compatible with the truth of wholehearted objectivism about fitting deliberation. For, even in special cases of this kind, it is not the beliefs themselves that are determining whether the actions in question count as fitting objects of practical deliberation. Rather, the fact that agents have or lack certain beliefs is merely causally relevant to whether the actions have certain other properties (say, dispos­ itional or modal properties) that determine whether they count as fitting objects of practical deliberation.

9.3.2  Non-­normativity The brings us to a second important challenge, which is that the Fitting Deliberation Account appears to be incompatible with feasibility’s ostensible non-­ normativity. By this I mean that questions about the feasibility of particular actions appear to be prior to, and independent of, questions about their normative status; and settling questions of the first kind does not appear to require settling questions of the second kind. This claim is plausible in its own right. It is natural to draw a distinction between those properties that are normative (such as being good and being what one ought to do) and those that are non-­normative (such as being circular and being a table), and to think that feasibility intuitively falls on the non-­normative side. Moreover, it would seem to be presupposed by our practices of treating feasibility as a constraint on the truth of normative claims about what we ought to do in circumstances that include potentially intractable uncertainty or disagreement about matters of value. Such practices are only intelligible insofar as truths about feasibility are indeed prior to, and independent of, the relevant normative truths. While most of the main existing accounts of feasibility are perfectly compatible with the claim that feasibility is non-­normative,7 the Fitting Deliberation Account would seem to be straightforwardly incompatible with it given that it purports to explain feasibility in terms of fitting deliberation, clearly a normative notion. It is not clear, however, that this makes feasibility normative in a problematic sense. For one, while the Fitting Deliberation Account characterizes feasibility’s 7  Most but not all. In particular, cost-­based accounts, which hold that feasibility is a matter of what is achievable without undue costs (see e.g. Räikkä 1998; Buchanan 2004; Miller 2013), are clearly incompatible with the claim.

210 Fittingness functional role in normative terms, this is perfectly compatible with thinking that what realizes this role is some purely non-­normative property. For another, even if we focus on the (normative) functional property of being a fitting object of practical deliberation, reaching agreement about whether an action exemplifies it appears to be perfectly possible in circumstances of intractable uncertainty and disagreement. Suppose that Anastasia and Bertrand disagree intractably about the desirability of persuading a mutual friend, Clarissa, to end a romantic relationship. Their disagreement would appear to pose no obstacle, at least in prin­ciple, to their reaching agreement that such a course of action is not a fitting object of practical deliberation and concluding, on that basis, that it is not the case that they ought to persuade Clarissa to end the relationship. It might be objected that that uncertainty and disagreement may presumably encompass, not only questions about reason and value, but also questions about fittingness. But I suggest we have good reason to doubt that such uncertainty and disagreement need be intractable. For there is a principled and normatively light-­weight way of resolving it. Just as a builder may determine the fittingness of a certain material to be used in building the roof of a house simply by consulting her (perhaps implicit) understanding of what it is to be engaged in the practice of building a house (roughly, an activity that has as its aim to produce a habitable dwelling), so too we may determine the fittingness of an action to be a subject of practical deliberation simply by consulting our (perhaps implicit) understanding of what it is to be engaged in the practice of practical deliberation (in my view, roughly an activity that has as its aim to determine how we will act by deciding how to act (cf. Korsgaard 1996)). This is an interpretative exercise; it involves engaging in in­ter­pret­ation of a practice in which we are co-­participants. As such, as it does not require us to take a stand on other controversial normative questions such as questions about the value or desirability of the actions in question. Thus, for example, Anastasia and Bertrand might reach agreement about the feasibility of persuading Clarissa to end her romantic relationship so long as they are in a pos­ition to reach agreement about whether there are impediments of a sort that would be at odds with manifesting the requisite form of practical self-­determination and, hence, about whether there is a decision to make, as it were, about whether to persuade Clarissa to end her romantic relationship. Still, we might worry that the Fitting Deliberation Account remains incompatible with feasibility’s ostensible non-­normativity—­not because it employs the normative notion of fittingness as such, but because what constitutes fitting de­lib­er­ation is dependent on other normative (deontic and/or evaluative) considerations.8 Why 8  If deontic truths partly determine deliberation-­worthiness, this would obviously fatally com­ prom­ise the capacity of feasibility to play the role of a constraint on the truth of deontic claims. If evaluative truths partly determine deliberation-­worthiness, this would fatally compromise the cap­ acity of feasibility to play this role in circumstances of intractable uncertainty or disagreement about matters of value. Either way, it would be very bad news for the Fitting Deliberation Account.

Feasibility and Fitting Deliberation  211 might we think that? One possibility is to point to the role of practical reasons. It is very tempting to suppose that the fittingness of fitting deliberation depends, in part, on the reasons we have to perform and to refrain from performing the actions in question. Take the fact that punching someone in the face would seriously injure them. This obviously constitutes a weighty reason to refrain from punching them in the face. As such, it would seem to have an obvious bearing on  what would constitute fitting deliberation about what to do: by counting (potentially decisively) against punching them in the face. This can be quickly discounted. Whether or not it is true that reasons determine truths about fitting deliberation,9 this does not mean that they determine the relevant truths about fitting deliberation, namely truths about which actions are fitting objects of practical deliberation. As we saw, determining which actions are fitting objects of practical deliberation is a matter of playing an option-­fixing role. By contrast, reasons play an option-­adjudicating role. For example, the fact that punching someone in the face would seriously injure them is a consideration of a kind that has an important, indeed potentially decisive (negative) bearing on the question of whether to punch them in the face. This does not mean that it plays an option-­fixing role; it does not make it the case that punching them in the face is not a candidate for deliberation about what to do (unless there is some sufficiently weighty countervailing consideration).10 Nonetheless, even if reasons as such do not necessarily help to explain truths about which actions are fitting objects of practical deliberation, we might still think that certain reasons do—­in particular, reasons to deliberate. Take morally heinous actions such as putting poison in the coffee of a colleague. I take it that we have decisive reasons, not merely to refrain from putting poison in the coffee of a colleague, but also to refrain from deliberating about whether to do so. Such reasons are clearly the right kinds of considerations to play an option-­fixing role and not merely an option-­adjudicating role. Even so, they are the wrong kinds of considerations to play the right kind of option-­fixing role: namely, fixing our deliberative options by determining our domain of deliberative jurisdiction. Whether an action lies within our domain of our deliberative jurisdiction and whether we have sufficient or decisive reasons to 9  A number of philosophers have argued that it is false since the order of explanation goes in the other direction. That is, our reasons for action are determined by what constitutes fitting deliberation, rather than vice versa (see Way 2017; Cullity 2018, ch. 2; this volume). 10  This is also important in order to answer a potential concern about the Fitting Deliberation Account: namely, that it might appear to be hard to see how we are supposed to distinguish feasibility from reasons for action if we accept an account of the latter that also understands them in terms of fitting deliberation, such as McHugh and Way’s (2016) account of reasons for action as premises in correct reasoning, or Cullity’s (this volume) account of reasons for action as deliberation-­worthy facts. The key is that such accounts understand reasons for action as considerations that are supposed to play a particular kind of option-­adjudicating role (roughly, they are considerations that are appropriate to bring to bear within deliberation), whereas the Fitting Deliberation account understands feasi­ bil­ity as a consideration that is supposed to play an option-­fixing role.

212 Fittingness deliberate about whether (or when or how) to perform it can come apart—­in both directions. On the one hand, there are actions that lie within our domain of deliberative jurisdiction in spite of the fact that we have decisive reasons not to deliberate about whether to perform them, such as putting poison in the coffee of a colleague. While there is clearly something deeply wrong with such de­lib­er­ ation, there is nothing wrong with it qua deliberation. On the other hand, it seems clear that there are actions that lie outside our domain of deliberative jurisdiction in spite of the fact that we have compelling and perhaps even decisive reasons to deliberate about whether (or when or how) to perform them. Consider courses of action where there is no prospect of success, such as eradicating poverty, and yet significant value associated with deliberating about how to do so: say, because of the good consequences of the deliberation; or because such deliberation manifests respect for certain important values. There is clearly something good and right about such deliberation in spite of the fact that there is something wrong with it qua deliberation since it fails to accord with internal standards of de­lib­er­ation, standards that are given by what is needed to achieve deliberation’s constitutive aim. How about if deliberation’s constitutive aim itself encompasses certain normative elements? For example, one possible view holds that the aim of deliberation is to determine what we will do by determining what we ought to do. If this is correct, then it might seem that we simply cannot maintain a strict bifurcation between the internal normative standards of correct deliberation and the external normative standards provided by what we ought and have reason to do since the former will depend on the latter. I myself think that we have good reason to resist characterizing deliberation’s constitutive aim in this fashion given the possibility of what I have called elsewhere consistent normative nihilists who think that there are no truths about what we ought and have reason to do and, hence, who refrain from forming beliefs about, or interrogating the question of, what they ought and have reason to do (see Southwood 2016b). Not only are consistent normative nihilists possible, but it seems possible for them to engage in deliberation: say, deliberation about whether to go for a run after work. But suppose I’m wrong. Suppose that the aim of deliberation is indeed to determine what we will do by determining what we ought to do. If this is right, then presumably it will act as a constraint on what can count as the correct output of practical deliberation, i.e. our decisions. That is, a decision to perform an action will count as deliberatively correct only if we ought in fact to perform the action. But it surely cannot be the case that what counts as a fitting object of practical deliberation is determined by what we ought to do. That would be a very strange view. It would imply that the only actions that are fitting objects of deliberation are actions that we ought to perform. This belies the fact that initiating deliberation appears to presuppose and to derive its point from the fact that we don’t know what we ought to do. If we did already know that we ought to be giving a certain percentage of our income to charity, then why on earth

Feasibility and Fitting Deliberation  213 would we be deliberating about whether to do so? Thus, even if the aim of de­lib­er­ation is to determine what we will do by determining what we ought to do, this aim would hardly be well served by limiting the domain of actions that may enter into deliberation to those that we ought to perform.11 Finally, might truths about which actions are fitting objects of practical de­lib­ er­ation depend on our values or normative belief? I think that this is indeed sometimes the case.12 David Miller gives the nice example of giving up our children to the state (see Miller 2013). As Miller has put it, such a course of action would strike many of us as ‘intolerable’ since it would require us to act in a way that is contrary to our deepest convictions about what is of value and importance. In a case of this sort, it might seem that our values or normative attitudes themselves would constitute an insurmountable impediment to deliberation achieving its constitutive aim by making it the case that we cannot bring ourselves to give up our children to the state; doing so is therefore not a fitting object of practical deliberation. But, of course, this is perfectly compatible with the normativity-­independence thesis. Truths about our values and normative beliefs are not themselves normative truths. Thus, even if truths about what actions are fitting objects of practical deliberation depend on truths about our values or normative beliefs, settling the first does not require settling questions about the normative status of our actions. Even if we are intractably uncertain or intractably disagree about matters of value, this is no impediment to our reaching agreement about what our values are or what we believe to be of value.

9.4  Fitting deliberation versus fitting prescription Appealing as it is, the Fitting Deliberation Account is not the only possible account of feasibility that seeks to explicate feasibility in terms of fittingness; and we might reasonably wonder whether it represents the best incarnation of this idea. There are a number of other possibilities. I want to focus, in particular, on accounts that purport to explicate feasibility in terms of some mode of fitting 11  Might it be well served by allowing those normative considerations that help to determine truths about what we ought to do, such as our reasons for action, to determine the domain of the deliberation-­worthy? Perhaps. But, at the very least, this is not needed to achieve the aim given the possibility of achieving it by instead giving such considerations their due within deliberation: i.e. of playing an option-­adjudicating role. And recall that what counts as deliberation-­worthy is a matter of what is needed to achieve deliberation’s constitutive aim. 12  Though I see no reason to think that it is always the case. The vast majority of us presumably believe that putting poison in a colleague’s coffee would be deeply wrong; doing so would be at odds with our fundamental values and normative convictions. This may well make it the case that it is rationally (as well as morally) objectionable to engage in deliberation about whether to put poison in a colleague’s coffee. Nonetheless, it is at least not obvious that the action of putting poison into a colleague’s coffee lacks any feature that is needed to make it count as correct qua deliberation.

214 Fittingness prescription (advice, recommendation, injunction, exhortation, or some such).13 At least on the face of it, such accounts might appear to share many of the relevant advantages of the Fitting Deliberation Account. It will be helpful, then, to say something about how the Fitting Deliberation Account differs from such accounts and why I take it to be superior.

9.4.1  Fitting advice Consider, first, an account according to which feasibility is to be explicated in terms of fitting advice. Just as actions may be fitting or unfitting objects of prac­ tical deliberation about what to do, so too they may be fitting or unfitting objects of advice about what to do. Moreover, just as deliberation-­worthiness is a matter of conforming with certain internal standards of deliberation that are given by what is needed to fulfil deliberation’s constitutive aim, so too advice-­worthiness is a matter of conforming with certain internal standards of advice that are given by what is needed to fulfil advice’s constitutive aim. Thus, according to what we can call the Fitting Advice Account, feasibility is a matter of possessing whatever features are needed to fulfil the constitutive aim of advice and, hence, play the role of determining whether actions are fitting objects of advice about what to do. The Fitting Advice Account might seem to share many of the virtues of the Fitting Deliberation Account. It will deliver similarly plausible substantive verdicts across a wide range of cases. For example, it can explain the infeasibility of paradigmatically infeasible actions, such as a particular state’s eliminating all social inequalities (at least within a suitably short timeframe) on the grounds that advising the state to eliminate all social inequalities would be inappropriate qua advice. It can explain feasibility’s distinctive importance by pointing to the role it plays within a distinctive and important practice: the second-­personal practice of offering advice. It can explain even more easily than the Fitting Deliberation Account feasibility’s objectivity since I take it that there is no temptation to hold that fitting advice is tailored to the epistemic position of the advisee. Nonetheless, we have good reason to reject the Fitting Advice Account. Consider the fact that truths about the feasibility of our actions appear to be prior to, and independent of, truths about the deontic status of the actions. As we saw, the Fitting Deliberation Account has no difficulty explaining this since truths about

13  I shall assume in what follows that Fitting Prescription Accounts of feasibility are monistic in the sense that they hold that there is some unique (special, privileged) notion of feasibility and that they are aspiring to provide an account of it. An alternative is to embrace some of pluralism about feasi­bil­ ity (see e.g Erman and Moller 2020). If pluralism is true, then the Fitting Deliberation Account and the Fitting Prescription Account needn’t be competitors; rather, they might be aspiring to give accounts of different concepts. I discuss and give some reasons to be sceptical of this kind of feasibility pluralism in Southwood 2022.

Feasibility and Fitting Deliberation  215 the fittingness of our actions to be objects of practical deliberation are clearly prior to, and independent of, truths about their deontic status. It is always possible, at least in principle, to settle the question of whether an action is a fitting object of practical deliberation even insofar as we are intractably uncertain or intractably disagree about whether we ought to perform it. But it does not seem that we are similarly able to divorce the advice-worthiness of our actions from their deontic status in this way. This is due to a crucial difference between advice and deliberation. Deliberation typically—­and arguably necessarily—­ presupposes that we have not yet settled the question of what we ought to do; in deliberation we identify relevant normative considerations, establish their comparative weight, and bring them to bear to adjudicate between actions that we are treating as candidates for what to do. By contrast, advice appears to presuppose that we have already settled the question of what we ought to do and hence undertaken such adjudication; we cannot coherently advise someone to do something if we have not yet formed a view about whether they ought to do it. Advice, unlike deliberation, does not encompass two distinct tasks—­an option-­fixing task that is a precursor to, and precondition for the possibility of, advice and then an option-­ adjudicating task that we undertake in advice itself. Rather, it presupposes that we have already undertaken such adjudication and reached a conclusion about what the agent ought to do. But this appears to have the following consequence: whether an action counts as a fitting object of advice turns, at least in part, on whether the agent ought to perform it; if it is not the case that they ought to perform the action, then it cannot be a fitting object of advice. Consider what this means. It means that the Fitting Advice Account implies that actions are only feasible if we ought to perform them. That is deeply implausible. While many of us are tempted to think that ‘ought’ implies ‘feasible’, it is surely not the case that ‘feasible’ implies ‘ought’. Moreover, it means that there is simply no prospect of treating questions about feasibility as helping us to settle normative questions about what we ought to do in circumstances where we are intractably uncertain or intractably disagree about relevant matters of value. Indeed, it suggests that rather than treating questions about feasibility as helping us to settle normative questions about what we ought to do, we must proceed instead by treating normative questions about what we ought to do as indispensable for settling questions about what is feasible.

9.4.2  Fitting exhortation Might we avoid this problem by looking to some other mode of prescription that does not have this problematic feature? Consider, for example, an account according to which feasibility is understood in terms of fitting exhortation. This is the mode of prescription that is in play when we exhort a child to tidy their room, or

216 Fittingness exhort an athlete to run faster, or exhort a politician to support a piece of legislation. Moreover, just as actions can be advice-­worthy or non-­advice-­worthy (i.e. fitting or non-­fitting objects of advice), so too they can be exhortation-­worthy or non-­ exhortation-­worthy (i.e. fitting or non-­fitting objects of exhortation). Victoria McGeer and Philip Pettit (2015) have proposed an account of responsibility that is based on whether the action is exhortation-­worthy, but it might potentially be extended to feasibility.14 On the face of it, exhortation might seem to be exactly the mode of prescription that we are after. For it appears to be like deliberation and unlike advice in the sense that it does not presuppose that we have already settled the question of what we ought to do. To be sure, it is unlike deliberation in that it does not presuppose that we haven’t settled the question of what they ought to do; it would hardly be atypical or inappropriate to exhort someone to do what we believe they ought to do. Moreover, it is at least arguable that it presupposes that it is not the case that we believe that the exhortee, as it were, ought not to do what we are exhorting them to do. But it seems perfectly appropriate to exhort someone to do something without having settled to our satisfaction the question of whether they ought to do it: say, if we want them to do it, or know that they want to do it, or think that it would be funny for them to do it, or whatever. As a result, unlike the Fitting Advice Account, the Fitting Exhortation Account does not appear to have the problematic implication that whether an action is feasible turns on whether we ought to perform it since it would seem that an action may be a fitting object of exhortation regardless of whether or not we ought to perform it. Nonetheless, the Fitting Exhortation Account faces a different problem. Consider actions that we know agents won’t perform, or are virtually certain not to perform, simply because they are unwilling to perform the actions. (For ex­ample, suppose we know that Professor Procrastinate, a brilliant but utterly unreliable scholar, won’t (or at least is virtually certain not to) compete a particular refereeing assignment (Jackson and Pargetter 1986).) Such actions may count as perfectly deliberationworthy. It might be irrational for Professor Procrastinate to deliberate about, say, when to complete the assignment (assuming that he is also privy to this information). But completing the assignment is a perfectly fitting object of practical deliberation. As Kant (1998, ch. 3) famously observed, prac­tical deliberation takes place under ‘the idea of freedom’. This means, at the very least, that when we deliberate about what to do, we adopt a distinctive standpoint with regard to our future conduct, viewing such conduct as something to be determined by our decisions (whatever they turn out to be), rather than as an impartially construed

14  Notice that I am bracketing a very interesting and important feature of McGeer and Pettit’s account of responsibility according to which claims about someone’s responsibility involve what they call ‘evocative reports’—reports ‘the purpose of which is, at least in part, to evoke or call into being the very possibility reported’ (McGeer and Pettit 2015, p. 178).

Feasibility and Fitting Deliberation  217 event in the world to be predicted. As a result, in deliberating about what to do, it is appropriate to set aside our beliefs (and even our knowledge) about whether we in fact will or are likely to perform the actions in question. But it is at least not obvious that such actions are exhortation-­worthy. On the contrary, exhorting agents to do things that we know they won’t do would seem to be patently unfitting. It is not appropriate for us to set aside our knowledge of agents’ future conduct in exhorting them to do this or that. Exhortation does not operate under the idea of freedom. This means that actions that we won’t perform or that we are virtually certain not to perform cannot be fitting objects of ex­hort­ ation and, hence, feasible by the lights of the Fitting Exhortation Account. This is deeply problematic. Not because it does violence to our linguistic practices (though it surely does). We should not be unduly concerned if an account of feasi­bil­ity departs (even relatively dramatically) from our linguistic practices. Rather, the problem is that if actions that we won’t (or are virtually certain not to) perform cannot be feasible, then, assuming that feasibility is a constraint on the truth of claims about what we ought to do, it follows it cannot be the case that we ought to do things that we won’t (or are virtually certain not to) do. I take it that this would be a clear reductio of the Fitting Exhortation Account. Nonetheless, I don’t want to rest my case against the Fitting Exhortation Account on this objection alone. For a proponent of the Fitting Exhortation Account might respond that actions that we won’t (or are virtually certain not to) perform might nonetheless count as exhortation-­worthy so long as there is some prospect that exhorting agents to perform such actions might fulfil exhortation’s constitutive aim (which is presumably something like inducing the performance of actions on the part of others by getting them to heed the exhortation); and that for this it is enough that we are not robustly disposed to be unresponsive to such exhortation. But now consider agents who won’t (or are virtually not to) perform actions because they are unwilling and who, in addition, are robustly disposed to be  unresponsive to exhortation. I have in mind agents who are (perhaps pathologically) exhortation-­averse; their psychological make-­up is such that exhorting them to perform particular actions is guaranteed to produce non-­ performance in spite of the fact that they would easily succeed in performing the actions if they were to decide to do so off their own bat, as it were. Suppose that Professor Procrastinate is an agent of this kind. Under these circumstances, there is no prospect that exhorting Procrastinate to complete the review will achieve exhortation’s constitutive aim. Thus, his completing the review counts as exhortation-­unworthy and, hence, infeasible according to the Fitting Exhortation Account. Again, this would imply that it is not the case that Procrastinate ought to complete the review (even if he has already accepted it and assuming that feasibility is a constraint on the truth of claims about what we ought to do). This seems plainly mistaken.

218 Fittingness Might the proponent of the Fitting Exhortation Account respond with the charge of ‘tu quoque’? It might seem that the Fitting Deliberation Account is vulnerable to an objection with the same structure. Consider the performance of certain actions on the part of agents who are pathologically deliberation-­averse (rather than exhortation-­averse) in the sense that, while they could easily perform the actions, they are robustly disposed to be unresponsive to their own practical deliberation: say, robustly disposed not to do what they decide. Such actions are presumably not deliberation-­worthy and, hence, will count as in­feas­ible according to the Fitting Deliberation Account. This objection rests on a mistake. Recall that by ‘deliberation’ I don’t mean a conscious, explicit, extended process of reasoning. This is one important kind of deliberation to be sure. And we can certainly imagine agents who are robustly disposed to be unresponsive to deliberation of this kind but for whom we should nonetheless say that the actions in question are perfectly feasible. Rather, by ‘deliberation’ I mean the exercise of practical reason. This may be conscious, explicit, extended. But it may also be unconscious, implicit, and more or less spontaneous. Being pathologically unresponsive to deliberation would therefore involve a fundamental failure of practical reason. It is not obvious that an ‘agent’ of whom this is true should really count as an agent at all, or at least they enjoy agency with regard to the performance of actions where such a failure of practical reason is present; or that the ‘actions’ that such an agent might (easily) bring about should really count as actions of the kind to which the notion of feasibility applies.

9.5  Conclusion My aim in this chapter has been to further develop and explore an account of feasibility—­the Fitting Deliberation Account—­that purports to explain truths about feasibility in terms of truths about fittingness—­in particular, truths about fitting deliberation. I have considered two important challenges to the Fitting Deliberation Account concerning its apparent inability to accommodate feasi­bil­ ity’s objectivity and non-­normativity and argued that both challenges can be met. I then contrasted the Fitting Deliberation Account with two different versions of an interesting alternative—­the Fitting Prescription Account—­and argued that there are significant problems with the latter that do not arise in the case of the former. The Fitting Deliberation Account is of interest for at least two reasons. First, and most obviously, because it constitutes a compelling account of feasibility, a theoretically and practically important idea. But, second, because it represents a concrete illustration of a less familiar way in which the notion of fittingness might be fruitfully deployed to help shed light on the nature of the normative domain. If  the Fitting Deliberation Account is correct, and assuming that feasibility is needed to characterize certain normative truths, then, even if it turns out that

Feasibility and Fitting Deliberation  219 such truths cannot be directly explained in terms of fittingness, fittingness will nonetheless be essential to achieving a proper understanding of the nature of normativity.

References Broad, C. D. 1930. Five Types of Ethical Theory. London: Routledge & Kegan Paul. Broome, John. 2013. Rationality through Reasoning. Oxford: Wiley Blackwell. Buchanan, Alan. 2004. Justice, Legitimacy, and Self-Determination: Moral Foundations for International Law. Oxford: Oxford University Press. Chappell, Richard. 2012. ‘Fittingness: The Sole Normative Primitive,’ Philosophical Quarterly, 62: 684–704. Cullity, Garrett. 2018. Concern, Respect, and Cooperation. Oxford: Oxford University Press. D’Arms, Justin and Daniel Jacobson. 2000. ‘Sentiment and Value,’ Ethics, 110: 722–48. Erman, Eva and Nikolas Moller. 2020. ‘A World of Possibilities: The Place of Feasibility in Political Theory,’ Res Publica, 26: 1–23. Estlund, David. 2011. ‘Human Nature and the Limits (If Any) of Political Philosophy,’ Philosophy & Public Affairs, 39: 207–37. Estlund, David. 2020. Utopophobia: On the Limits (If Any) of Political Philosophy. Princeton, NJ: Princeton University Press. Ewing, A. C. 1939. ‘A Suggested Non-naturalistic Analysis of Good,’ Mind, 39: 1–22. Ewing, A. C. 1947. The Definition of Good. London: Routledge & Kegan Paul. Gilabert, Pablo and Holly Lawford-Smith. 2012. ‘Political Feasibility: A Conceptual Exploration,’ Political Studies, 60: 809–25. Hedden, Brian. 2012. ‘Options and the Subjective Ought,’ Philosophical Studies, 158: 343–60. Howard, Chris. 2017. ‘Fittingness,’ Philosophy Compass, 13: 1–14. Howard, Chris. 2019. ‘The Fundamentality of Fit,’ Oxford Studies in Metaethics, 14: 216–36. Jackson, Frank and Robert Pargetter. 1986. ‘Oughts, Options, and Actualism,’ Philosophical Review, 95: 233–55. Kant, Immanuel. 1998. Groundwork of the Metaphysics of Morals, trans. Mary Gregor. Cambridge; New York: Cambridge University Press. Korsgaard, Christine. 1996. The Sources of Normativity. Cambridge: Cambridge University Press. Lawford-Smith, Holly. 2013. ‘Understanding Political Feasibility,’ Journal of Political Philosophy, 21: 243–59.

220 Fittingness McGeer, Victoria and Philip Pettit. 2015. ‘The Hard Problem of Responsibility,’ in Oxford Studies in Agency and Responsibility, vol. 3, ed. D.  Shoemaker. Oxford: Oxford University Press, 160–88. McHugh, Conor and Jonathan Way. 2016. ‘Fittingess First,’ Ethics, 126: 575–606. Miller, David. 2013. Justice for Earthlings: Essays in Political Philosophy. Cambridge: Cambridge University Press. Räikkä, Juha. 1998. ‘The Feasibility Condition in Political Theory,’ Journal of Political Philosophy, 1: 27–40. Rowland, Richard. 2017. ‘Reasons or Fittingness First?’ Ethics, 128: 212–29. Scanlon, T.  M. 1998. What We Owe to Each Other. Cambridge, MA: Harvard University Press. Schwartz, Wolfgang. 2020. ‘Ability and possibility,’ Philosophers’ Imprint, 20, 1–21. Southwood, Nicholas. 2016a ‘Does “Ought” Imply “Feasible?”’ Philosophy & Public Affairs, 44: 7–45. Southwood, Nicholas. 2016b. ‘ “The Thing to Do” Implies “Can” ,’ Noûs, 50: 61–72. Southwood, Nicholas. 2019. ‘The Question of Practical Reason,’ in Reasoning: New Essays on Theoretical and Practical Thinking, ed. Magdalena Balcerak-Jackson and Brendan Balcerak-Jackson. Oxford: Oxford University Press, 71–90. Southwood, Nicholas and David Wiens. 2016. ‘ “Actual” Does Not Imply “Feasible” ,’ Philosophical Studies, 173: 3037–60. Southwood, Nicholas. 2022. ‘Feasibility as deliberation-worthiness,’ Philosophy & Public Affairs, 50, 121–62. Stemplowska, Zofia. 2016. ‘Feasibility: Individual and Collective,’ Social Philosophy and Policy, 33: 273–91. Stemplowska, Zofia. 2021. ‘The Incentives Account of Feasibility,’ Philosophical Studies, 178: 2385–401. Way, Jonathan. 2017. ‘Reasons as Premises of Good Reasons,’ Pacific Philosophical Quarterly, 98: 251–70. Wiens, David. 2015. ‘Political Ideals and the Feasibility Frontier,’ Economics and Philosophy, 31: 447–77. Wiens, David. 2016. ‘Motivational Limitations on the Demands of Justice,’ European Journal of Political Theory, 15: 333–52.

10 In Defence of the Right Kind of Reason Christopher Howard and Stephanie Leary

10.1  Introduction Why do you love your child? Understood as a request for justification, it’s natural to give two different kinds of answers. On the one hand, you might list some of her lovable features: she’s funny, intellectually curious, kind, and spirited. On the other hand, you might list some ways in which loving her does some good, e.g. by promoting her happiness and by keeping you from ditching her on the side of the road when she’s driving you crazy. The former answer appeals to right-­kind ­reasons (RKRs): considerations to do with whether your child is lovable and thus fitting for you to love. The latter answer appeals to wrong-­kind reasons (WKRs): considerations to do with the goodness of loving your child, whether or not your love for her is fitting. This distinction generalizes for other attitudes like ad­mir­ ation, desire, and belief. Considerations to do with whether something is ad­mir­ able, desirable, or credible constitute RKRs for admiring, desiring, and believing (respectively) because they contribute to the fittingness of these attitudes. Whereas considerations to do with whether it’s good to admire, desire, or believe something constitute WKRs for admiring, desiring, or believing.1 Since the origin of this distinction, there have been WKR sceptics: folks who claim that only RKRs genuinely favour, or justify, having an attitude. For example, these sceptics insist that the only facts relevant to whether you should love some­ one are those relevant to whether love is fitting. The fact that loving someone would do some good at best bears on whether you should want to love them (as an RKR for desiring) or on whether you should try to bring it about that you do.2 Recently, though, many WKR-­enthusiasts have become RKR sceptics: they claim that WKRs are the only facts that genuinely favour, or justify, having an attitude and that putative RKRs for attitudes are merely formally normative ­reasons that arise from attitudes’ constitutive standards of correctness.3 Constitutive standards

1 For a helpful survey of the literature on the RKR/WKR distinction, see Gertken and Kiesewetter (2017). 2  See, e.g. Parfit (2011), Skorupski (2010), and Way (2012). 3  See esp. Côté-­Bouchard and Littlejohn (2018), Mantel (2019), and Maguire and Woods (2020).

Christopher Howard and Stephanie Leary, In Defence of the Right Kind of Reason In: Fittingness: Essays in the Philosophy of Normativity. Edited by: Christopher Howard and R. A. Rowland, Oxford University Press. © Christopher Howard and Stephanie Leary 2022. DOI: 10.1093/oso/9780192895882.003.0010

222 Fittingness of correctness lack normative authority. The reasons arising from such standards don’t bear on what you really ought to or may do, unless you’ve got authoritative reason to be engaged in whatever activity the standards govern. RKR sceptics thus insist that the fact that someone has lovable properties can at best be derivatively relevant to whether you really ought to love them: a person’s lovable properties can be relevant to whether you should love them only if you’ve got authoritative reason to love correctly, i.e. according to the standards internal to love. Your daughter’s lovable properties thereby lack a kind of normative authority: they fail to provide non-­derivatively authoritative normative r­ easons for love, in the sense that they can’t, by themselves, contribute to making it the case that you really ought to, or may, love your daughter.4 We reject both kinds of scepticism: we think both RKRs and WKRs are au­thori­ta­tive­ly normative. Elsewhere, we’ve defended the authoritative normativ­ ity of WKRs.5 Here, we defend the authoritative normativity of RKRs. Our main claim is that RKRs are authoritatively normative in the sense that they play a non-­ derivative role in weighing explanations of an attitude’s overall, authoritative deontic status. This leaves open the question of whether or how RKRs compare with, or weigh against, WKRs.6 It’s also neutral about whether RKRs ever demand, rather than merely justify, the attitudes they favour, i.e. whether RKRs are requir­ ing or only permissive.7 In section  10.2, we argue that the intuitive data about cases favours the view that RKRs are authoritatively normative. We then argue that the kinds of the­or­et­ic­al considerations that RKR sceptics appeal to in order to motivate their scepticism can and should be resisted. In section  10.3, we address the most common theoretical motivation for RKR scepticism, viz. the assumption that RKRs arise from standards of correctness internal to the attitudes. We argue that fittingness is distinct from constitutive correctness and that once we appreciate that RKRs are fit-­makers, rather than constitutive-­correctness-­makers, RKR scepticism is no longer warranted. Then, in section  10.4, we respond to  Barry Maguire’s (2018) recent argument that fit-­ making facts can’t be normative reasons at all since they lack certain essential features of reasons. We argue that fit-­making facts have the features of reasons that Maguire claims they lack.

4  On the distinction between ‘authoritative’ and ‘merely formal’ normativity, see esp. McPherson (2018), Wodak (2019), and Woods (2018). We remain neutral about how exactly this distinction should be drawn, but we take the idea to be intuitive: something’s normativity is authoritative when it’s relevant to settling what you ‘genuinely’ or ‘really’ ought to or may do. Its authoritative normativity is ‘non-­derivative’ if it doesn’t depend on the authoritative normativity of some other factor. 5  See Howard (2016, 2019b) and Leary (2017). 6  On which see Berker (2018), Howard (2019b), and Reisner (2008, ms). 7  For discussion, see Whiting (2021) and (Berker, this volume).

In Defence of the Right Kind of Reason  223

10.2  Intuitive data The main reason to believe that RKRs are authoritatively normative is that this claim best explains intuitions about cases. Consider cases where having an atti­ tude would have no value at all, or even a little disvalue, but there are facts that make the attitude fitting. For example, suppose you meet someone at a party who has many admirable features (they’re intelligent, accomplished, they dedicate themselves to philanthropic pursuits, and they’re sincerely warm and kind), but your privately admiring them would have no good effects on you, them, or any­ one else. Still, it seems you should admire this person.8 Or suppose that while you’re waiting in line to see a movie, an audience member from an earlier show­ ing walks by and blurts out the surprise ending. Believing the fellow movie-­goer’s testimony would be bad for you since it makes the movie less enjoyable. But it still seems like you should believe it.9 The authoritative normativity of RKRs explains these verdicts: if fit-­making facts bear non-­derivatively on whether you should have some attitude, this explains why you should admire someone who’s ad­mir­ able or believe something that’s credible even if doing so would have no benefit or would be slightly harmful. RKR sceptics, on the other hand, must reject these verdicts. We think this is a real cost. These sorts of intuitions are the basic data that any normative theory should explain.10 Can we generalize from cases like the above to the claim that all RKRs are authoritatively normative? We think so, unless there are specific examples to the contrary. Richard Rowland (2022) suggests several examples of attitudes the RKRs for which might seem to lack normative authority: (a) highly specific attitudes like schadenfreude or chrysalism (the tranquil feeling of being indoors during a storm), (b) vicious attitudes like envy or anger, and (c) boredom and depression. But we argue here that none of these examples suggest that some RKRs aren’t authoritatively normative.11 Consider first highly specific attitudes like schadenfreude and chrysalism. Rowland seems to take such attitudes to show that there’s a vast and incredibly

8  We use ‘should’ here in the authoritatively normative sense, but in a way that’s neutral with respect to whether RKRs require, permit, or justify attitudes in a sense that’s stronger than permission, but weaker than requirement (see [Berker, this volume] for discussion of this latter possibility). 9  This case comes from Kelly (2003). 10  One way in which RKR sceptics might try to capture this data is by claiming that whenever there’s an RKR for an attitude, there’s a WKR for the attitude with equal weight because it’s always in some way good to have an attitude for which there are RKRs. We lack the space to discuss this strategy here, but for reasons to reject it; see, e.g. Howard (2019a, 2019b) and Way (2013). It’s worth noting that many RKR sceptics also reject this strategy; see, e.g. Maguire and Woods (2020), Papineau (2013), and Rinard (2019). 11  Lord and Sylvan (2019) also argue that not all RKRs are authoritatively normative by appealing to examples of RKRs for action. We don’t address their examples here because we think their examples rest on the mistaken assumption that RKRs arise from constitutive standards of correctness, which we address in section 10.3.

224 Fittingness diverse array of attitudes, each of which can be fitting with respect to a highly specific object. And Rowland thinks it would be odd to think that there are thereby authoritative reasons to have such attitudes towards those objects. But we think that it’s a mistake to think that attitudes like schadenfreude and chrysalism are distinctive attitudes that have their own fittingness conditions. Rather, they’re ordinary, run-­of-­the-­mill attitudes towards very specific objects: schadenfreude is being pleased at the suffering of others; chrysalism is being relieved at being indoors during a storm. So, when we ask whether schadenfreude or chrysalism can be fitting, we’re asking whether it can be fitting to be pleased by the suffering of others, or to be relieved about being indoors in a storm. These are substantive normative questions that aren’t settled by the mere fact that schaden­ freude and chrysalism, as a matter of conceptual definition, involve having these attitudes towards those objects. The claim that all RKRs are authoritatively nor­ mative thus doesn’t imply that there are authoritative reasons for attitudes like schadenfreude and chrysalism. Whether there are depends on whether these attitudes can be fitting. Next consider vicious attitudes—­envy, in particular.12 Rowland (2022: 117) claims that although envy can be fitting when directed at the enviable, if one fails to envy the enviable one isn’t normatively criticizable. He takes this to show that RKRs for envy aren’t authoritatively normative. We have three responses. First, from the fact that envy is fitting towards the enviable it doesn’t follow that there are RKRs for envy. To see this, consider that for any attitude that can be fitting, it’s an open question what properties comprise the evaluative property to which the fittingness of the attitude corresponds. Similarly, it’s an open question whether that property can ever be instantiated. For example, in the case of envy, it’s an open question what properties comprise enviability and whether this property can be instantiated. Suppose that for someone to be enviable is for them to possess something desirable that you lack, where this difference in possession itself is bad for you.13 It’s highly controversial whether the mere fact that someone has something desirable that you lack could ever be bad for you. Hence, it’s unclear whether anyone can ever be enviable. If not, then it can’t be fitting to envy anyone, and so, there can’t be RKRs for envy. But this is very different from the claim that there are RKRs for envy, but that they lack nor­ mative authority.14 Second, even assuming that someone could be enviable and thus that there could be RKRs for envy, there are possible explanations, compatible with the authoritative normativity of RKRs, for why one needn’t be normatively criticizable for failing to 12  We focus on envy because anger is more contested; see esp. Cherry (2018) and Srinivasan (2018). 13  This account is suggested by D’Arms and Jacobson (2006). 14  Sara Protasi (this volume) defends the claim that envy can be fitting (and hence that the property of being enviable can sometimes be instantiated). By our lights, this defence doubles as a defence of the claim that there are RKRs for envy and that these reasons are authoritative.

In Defence of the Right Kind of Reason  225 envy the enviable. One such explanation is that some (or all) RKRs might not be requiring: if the fact that someone is enviable bears on whether you should envy them in a sense that is weaker than requirement, then you wouldn’t be criticizable if you fail to envy them despite the RKRs to do so.15 Another possible explanation is that there may be strong WKRs against envy in general if envying is always bad for the envier or morally vicious (or both). So, even if someone is truly enviable such that there are authoritatively normative RKRs to envy them, it may nonethe­ less be the case that you ought not to envy them, since there are stronger WKRs against your doing so. Third, and finally, the claim that there are authoritatively normative RKRs for envy seems plausible if we compare envying the enviable against envying the unenviable. Other things being equal, it seems like there’s less (authoritatively normative)16 reason for someone to envy the unenviable than there is for some­ one to envy the enviable, which suggests that there’s some reason to envy the enviable.17 Of course, the claim that there’s less reason to do A than there is to do B doesn’t immediately entail that there’s some reason to do B; there could be less total reason to do A than to do B only because there’s more reason not to do A than there is not to do B. But that doesn’t seem like a viable explanation in the present case: why think that there’s less reason against envying the enviable than there is against envying the unenviable? So instead, what explains the fact that there’s less reason to envy the unenviable than there is to envy the enviable must be that there’s some positive, authoritative RKR for the latter. Rowland’s final examples concern boredom and depression. Again, Rowland (2022: 110-­111) claims that we aren’t criticizable or ‘normatively missing some­ thing’ if we’re never bored or depressed and that this suggests that RKRs for bore­ dom and depression aren’t authoritatively normative. We think it’s important to distinguish between object-­directed attitudes, like belief and admiration, and mental states that aren’t object directed, like mere feel­ ings or moods. We think there are RKRs only for the former. So when Rowland discusses boredom and depression, we think it’s important what kinds of mental states he takes these to be. If he’s talking about boredom and depression as moods, for example, then we think there are no RKRs for these mental states, and that this is what explains why one needn’t be normatively criticizable if one is never in them. But some instances of boredom and depression do seem like object-­directed attitudes—­for example, if you’re bored by a talk or depressed about a pandemic. If Rowland is talking about these kinds of boredom and depression, then what we 15  This explanation is compatible with RKRs being reasons that favour or justify the attitude in a sense that’s stronger than permission, but weaker than requirement (as Berker [this volume] suggests). 16  We omit this clarification from now on, but in what follows it should be understood that we use ‘reason’ to refer to authoritatively normative reason. 17  For arguments of this general form in favour of establishing the existence of a positive pro tanto reason for an attitude, see, for example, Schroeder (2007) and Wodak (2021).

226 Fittingness said above about envy applies here too. First, it’s an open question whether these attitudes can ever be fitting and thus whether there can be RKRs for them at all. Perhaps they can—­perhaps it’s fitting to be bored by the boring and depressed by the depressing and these are both evaluative properties that can be instantiated. But, second, there are alternative explanations for why one needn’t be criticizable if one fails to be bored by a boring talk or depressed by a depressing pandemic: RKRs may not be requiring, or there may be strong WKRs not to be bored or depressed. And, finally, if we compare being bored by the boring or being depressed by the depressing to being bored by the exciting and being depressed by the joyous, there seems to be less authoritative reason to be bored and depressed in the latter cases than there is in the former. In sum, then, we think intuitive verdicts about certain cases can be explained only if RKRs are authoritatively normative and that the intuitive verdicts about Rowland’s cases can be explained in a way that’s consistent with taking all RKRs to be authoritatively normative. Some RKR sceptics like Rinard (2019) and Papineau (2013) argue that we should accept the view that only WKRs are authoritatively normative because it’s simpler and more unified. But we think these virtues provide good enough reason to prefer one theory to another only if the former is equally capable of explaining all the intuitive data. And RKR sceptics can’t explain it. This is a serious cost, which we should accept only if there are further, strong theoretical reasons for thinking that, despite the appearances, RKRs aren’t authoritatively normative after all. So, we turn now to some arguments that purport to establish this.

10.3  Right-­kind reasons as constitutive-­correctness-­makers The main theoretical argument against the authoritative normativity of RKRs starts with the claim that RKRs arise from constitutive standards of correctness. On this view, for a fact R to be an RKR for an attitude A is for R to contribute towards making it the case that A is correct according to a standard that’s consti­ tutive of the kind of attitude that A is. But constitutive standards of correctness aren’t authoritatively normative. From the fact that a chess move would be correct according to the standards constitutive of chess playing, nothing immediately fol­ lows about whether we genuinely ought to, may, or even have reason to, make it. We have reason to make constitutively correct chess moves only if we have reason to be playing chess (and trying to win) in the first place. Indeed, as this example makes clear, constitutive standards of correctness come cheap: we can create them out of thin air by inventing new games, activities, clubs, etc. But authoritative normativity isn’t cheap in this way. So, if RKRs are constitutive-­correctness-­ makers, then they also lack normative authority.18 18  See esp. the references in note 3.

In Defence of the Right Kind of Reason  227 One response would be to argue that at least some constitutive standards of correctness are authoritatively normative.19 Metaethical constitutivists might be sympathetic to this, but we’re not. We tend to agree with our objector that consti­ tutive correctness isn’t authoritatively normative. Our preferred response is to reject that RKRs should be analysed in terms of constitutive correctness. We argue that not only is this analysis not mandatory, but there are good reasons to resist it. Start with some history. The terminology of ‘right-­kind reasons’ has its source in the literature on the ‘wrong kind of reason problem’ for ‘buck-­passing’ analyses of value. Buck-­passing analyses analyse various evaluative properties in terms of reasons for various sorts of response. For example, they might claim that what it is for something to be admirable is for there to be (sufficient) reason to admire it. The WKR problem is the problem that such analyses seem subject to coun­ter­ exam­ples: I might have reasons to admire something that isn’t admirable; perhaps I’ll get the goods if I do. If these reasons really are reasons to admire, then some reasons—­like these—­are of the ‘wrong kind’ to figure in buck-­passing accounts. Right-­kind reasons are of the ‘right kind’ to figure in buck-­passing accounts. What are the right kinds of reasons to figure in a buck-­passing account of admirability? The answer: all and only those reasons to admire that can contribute to making it the case that the object of admiration is admirable. RKRs for ad­mir­ation are thus facts that ensure the existence of the relation to admiration that it takes for its object to be admirable. We have a name for this relation: it’s ‘fittingness’. The fittingness relation is the relation in which a response stands to its object when the object merits—­or is worthy of—­that response. RKRs for an attitude are thus those that are relevant to whether its object is worthy of it. This checks out. For something to be admirable is for it to be worthy of admiration, for something to be desirable is for it to be worthy of desire, and so on. Thus, RKRs for an atti­ tude A are those that can contribute to making it the case that A’s object merits, or is worthy of, the kind of attitude A is. In short: RKRs for an attitude are those that can contribute to explanations of the attitude’s fittingness—­they’re fit-­making facts for the attitude in question.20 The upshot is this: RKRs can be plausibly analysed in terms of constitutive cor­ rectness only if fittingness can. Since RKRs are facts that make responses fitting, they can be correctness-­makers only if fit-­makers are. So, is it true that what it is for an attitude to be fitting, or merited, is for it to be correct according to a

19  This is Lord and Sylvan’s (2019) response. 20  The terminology of ‘fit-­making’ isn’t perfect since ‘making’ is a success verb and, as we’ll explain below (in section  10.4), there can be fit-­making facts for attitudes that aren’t fitting. A better term would be ‘fit-­contributors’, where fit-­contributors stand in a non-­factive ‘contributing-­to-­making-­it-­ the-­case relation’. This clarification is prompted by Kiesewetter (2021), who makes a parallel point regarding reasons as ‘justification-­makers’.

228 Fittingness standard that’s constitutive of the kind of attitude in question? In the last decade this suggestion has become quite popular, but we’re unconvinced. The first thing to note is that, despite its current popularity, the proposed ana­ lysis is revisionary. Although fittingness has only recently come to occupy the limelight in twenty-­first-­century normative philosophy, the relation has a long history. For example, it features prominently in the work of many normative theorists writing in the late nineteenth and mid-­twentieth centuries (Brentano 1889/2009, Brandt 1946, Broad 1930, Ewing 1948) as well as in the work of more recent writers (Feinberg 1970, Gibbard 1990, McDowell 1998, Wiggins 1987).21 But none of these authors holds that the fittingness of an attitude is a matter of its satisfying a constitutive standard of correctness.22 As far as we can tell, this idea has its origins in a relatively recent paper by Mark Schroeder (2010). Since Schroeder’s paper, many others have followed suit.23 Indeed, in a forthcoming paper in which he argues for the authoritative normativity of RKRs, Benjamin Kiesewetter remarks that ‘it seems fairly uncontroversial to say that the right kind of reasons for attitudes are essentially linked to the constitutive correctness stand­ ards for the attitude in question’ (2021: 2).24 But there are reasons to doubt this recently popular view. First, in the case of action, fittingness and constitutive correctness come apart. If you strap on your skates and perform a triple axel in the middle of your department meeting, then you might perform the jump correctly, but the jump (at least in normal circum­ stances) wouldn’t be merited. A response is merited only when it’s in some sense ‘called for’ by the situation, or by certain specific features of it. But there’s no sense in which your correct action—­the triple axel—­is called for in the situation under consideration. Consider also assertion. Some hold that assertion has a constitutive standard such that, by nature, assertions are correct if and only if they’re true.25 On this view, what makes an assertion correct, when it is, is the fact that it’s true. But an assertion isn’t merited in virtue of being true. Many trivial, private, or impolite claims, though true, don’t merit being asserted in most contexts. And some claims that merit assertion—­e.g. ‘I appreciate you helping me move’—may be merited 21  See Howard (2018: 2n4). 22 Brentano uses ‘correctness’ [Richtigkeit] to refer to the relation we’re calling ‘fittingness’, but doesn’t understand an attitude’s correctness in terms of its satisfying a norm that’s constitutive of it. To anticipate: we’re happy to call an attitude ‘correct’ if it’s fitting, as long as the relevant kind of correct­ ness isn’t constitutive correctness, but rather correctness according to a norm that’s external to the attitude. More on this later. 23  See esp. Lord and Sylvan (2019) and Sharadin (2015). 24  Though we disagree with Kiesewetter that RKRs are ‘essentially linked’ to constitutive standards of correctness, we take the present chapter to complement his. While Kiesewetter provides primarily positive arguments for thinking that RKRs, and RKRs for belief in particular, are authoritatively nor­ mative, our main aim here is to defend this thesis by considering and responding to the most common arguments against it. 25  See, e.g. Weiner (2005).

In Defence of the Right Kind of Reason  229 not because they’re true but in virtue of something else, e.g. being an expression of gratitude.26 This suggests that, insofar as fittingness is meritedness or worthiness, fittingness and correctness aren’t the same. The proponent of fittingness-­ as-­ constitutive-­ correctness might respond by conceding that fitting action isn’t necessarily correct action and simply restrict their thesis to the view that what it is for attitudes, specifically, to be fitting is for them to be correct according to their constitutive standards. We find this reply unattractive because it proposes that fittingness isn’t a unified category, that the fittingness of an action is a very different property than the fittingness of an atti­ tude. In our view, this is a mistake. Considerations of theoretical unity, simplicity, etc. suggest that the default view should be that ‘fitting’, as it might be properly predicated of actions, refers to the same property that it does when properly predicated of attitudes. The proponent of fittingness-­as-­constitutive-­correctness who proposes to restrict their thesis to attitudes therefore owes some rationale for thinking that the fittingness of attitudes can be analysed in terms of constitutive correctness, even though the fittingness of action can’t be. What rationale for this might there be? One possible argument for fittingness-­as-­constitutive-­correctness comes from Schroeder (2010). Schroeder observes that both correctness and fittingness are unaffected by WKRs: that WKRs are just as irrelevant to whether a response is constitutively correct as they are to whether a response is fitting. For example, just as the moral or pragmatic benefits of making a chess move are irrelevant to whether the move is correct according to the standard(s) constitutive of chess playing, so too are any benefits of admiration irrelevant to the attitude’s fitting­ ness. As Schroeder summarizes: ‘standards of correctness are unaffected by the kinds of incentives that come into play in Wrong Kind of Reasons scenarios’ (2010: 33). He conjectures that fittingness just is correctness. We have two responses. First, we don’t think Schroeder’s observation suffices to show that fittingness and correctness are equivalent, much less identical. From the fact that A and B are unaffected by C, it doesn’t follow that A and B are the same. My nightstand and clock are equally unaffected by the weather, but they aren’t the same object. Similarly, from the fact that fittingness and correctness are equally unaffected by WKRs, it doesn’t follow that they’re the same relation. Second, Schroeder’s conjecture, if true, would equally well support the view that fitting action is equivalent to constitutively correct action. But, as we’ve just argued, this view seems false. A second argument that the fittingness of an attitude is a matter of its being constitutively correct may seem more promising. This argument starts with the observation that the conditions under which an attitude is fitting strikingly 26  Parallel points hold even if the constitutive correctness condition for assertion turns out not to be truth, but instead knowledge (Williamson 2000, Hawthorne 2004) or justification (Kvanvig 2009, 2011).

230 Fittingness coincide with the conditions under which it’s natural to call the attitude ‘correct’. It’s natural, for example, to call admiration correct when its object is admirable, to call desire correct when its object is desirable, and to call belief correct when its object is credible.27 A simple explanation of this is that fittingness just is correct­ ness. Indeed, if fittingness weren’t correctness, then it’s unclear how the coinci­ dence in question could be plausibly explained. In response, note first the polysemy of ‘correct’. Sometimes ‘correct’ means accurate or true. Some authors identify correctness in this sense with fittingness, claiming that the fittingness of an attitude is a matter of its accurately represent­ ing its object (e.g. Tappolet 2011). This isn’t the sense in which a chess move is ‘correct’ when it satisfies the standards constitutive of chess playing. This latter sense of ‘correct’ amounts to conformity with a norm, whereas the former sense does not. Hence, there’s no sense in which the former sense of ‘correct’ is norma­ tive, whereas there is a clear sense in which the latter is. We reject the view that fittingness is ‘correctness’ in the sense that amounts to accuracy for reasons well-­ rehearsed by others (Schroeder 2010; Svavarsdóttir 2014; Naar 2021; D’Arms this volume).28 So, we’re not happy to say that the fittingness of an attitude amounts to its being ‘correct’ in this sense. However, we’re quite happy to say that fitting attitudes are ‘correct’ in the sec­ ond sense, i.e. in the sense that amounts to their satisfying a norm. Indeed, we accept that attitudes are fitting when they satisfy certain norms that govern them. What we deny is that an attitude’s being ‘correct’ in this sense amounts to its satis­ fying a norm that governs it constitutively. For although constitutive norms of correctness are normative in one sense, viz. formally, they aren’t in another: they lack authority. So it’s the thesis that fittingness amounts to constitutive correctness, specifically, that threatens its authoritative normativity, and hence that of RKRs. So, it’s this thesis we reject. But this is consistent with claiming that when an attitude is fitting, it’s correct according to a norm that’s external to it, i.e. a norm not baked into the attitude’s essence. And we’re happy to call fitting attitudes ‘correct’ in this sense. So, our view easily explains the data that it’s nat­ural to call fitting attitudes ‘correct’. What’s at issue, then, between us and our opponents, is whether the fittingness of an attitude amounts to its being correct according to a norm that governs it 27  It’s also natural to call a belief ‘correct’ when it’s true. As we make clear below, in a certain sense of ‘correct’, this claim is trivial. But as we also make clear, it’s not plausible that the kind of correctness to which this sense of the word refers is identical to fittingness; see also the discussion in note 29. 28  A less discussed reason for resisting this view is that it would make trivial the claim that fitting beliefs are always true beliefs, insofar as it’s assumed (plausibly) that beliefs represent their objects as being true. We take this to be a problem for the view since we think it’s not at all trivial that fitting beliefs are always true. It’s fitting to believe a proposition when it’s credible, or belief-­worthy, and it’s a substantive, normative, and controversial matter whether a proposition can be belief-­worthy, or merit belief, only if true. Indeed, we find it plausible that a proposition could be credible or belief-­worthy even if false—­perhaps if one has excellent evidence for it.

In Defence of the Right Kind of Reason  231 internally or externally. In a neglected discussion, Gideon Rosen (2001) remarks on the difficulty of adjudicating the dispute of whether the norms that govern certain mental states are external to them or instead constitutive of their nature. Borrowing from that discussion, one way to think of the issue is this: the view that correctness (or fittingness) norms are internal to the attitudes they govern entails that were God to suspend the correctness norms, the attitudes they govern would also be eradicated; whereas on the view that correctness norms are exter­ nal to the attitudes, the attitudes would survive. We find the latter idea more attractive, and while we can’t offer an extended argument for it here, we note the following two points in its favour: first, the view that we prefer is compatible with a wider range of views about the natures of fit-­evaluable attitudes and is in this way more ecumenical. For example, the view that correctness norms are constitu­ tive of attitudes may rule out functionalism or other forms of naturalism in the philosophy of mind, whereas our view is clearly consistent with these positions (Rosen 2001). Second, as noted above, our favoured view is compatible with the unity of fittingness as it applies to actions and attitudes whereas our opponent’s view is not. As we’ve already argued, the fittingness of certain actions—­e.g. triple axels and assertions—­can come apart from their constitutive correctness. (Here, too, we’re happy to call these actions ‘correct’ if they’re fitting, but, in light of our arguments, ‘correct’ here can’t mean constitutively correct.) So, again, our op­pon­ ent is left claiming that the fittingness of actions is a different property than the fittingness of attitudes with no clear rationale for doing so. Hence, we think that the burden is on our opponent to explain why we’re compelled to accept that the fittingness of an attitude amounts to its being internally rather than externally correct, particularly given the disunified view of fittingness that this picture entails.29 Our argument here has some important upshots. Because Schroeder and ­others assume that RKRs arise from constitutive standards of correctness, they typically argue that the distinction between RKRs and WKRs is a general distinction that applies to actions as well as attitudes, by pointing to actions or activities that have constitutive standards—­triple axels, tying knots, setting a proper English dining table, etc. This leads Errol Lord and Kurt Sylvan (2019) to claim that there’s also a ‘right kind of reason problem’, which is to explain why RKRs for certain attitudes are authoritatively normative while RKRs for action (and perhaps some attitudes) are not. After all, the fact that tying a knot in a certain way would be a constitutively 29  In principle, what might motivate the view that norms of fittingness or correctness internally rather than externally govern our attitudes? The best example of a possible motivation that we can find comes from Shah (2003), who argues that the claim that belief is constitutively governed by a norm of truth provides the best explanation of a psychological phenomenon he calls ‘transparency’ (roughly, that in deliberative contexts, only what you take to be evidence can be a motivating reason for belief). Shah (2008) also suggests that there may be a similar motivation for thinking that intention is gov­ erned by a constitutive norm. We can’t consider Shah’s position here, but we find plausible existing criticisms of both its psychological and normative claims (for the former kind of criticism, see esp. McHugh 2015; for the latter, see, inter alia, Steglich-­Peterson 2006).

232 Fittingness correct way of tying a slip knot does not seem like a source of authoritatively normative reasons for anyone to tie a knot in the relevant way. But our argument that RKRs are fit-­makers and that fittingness is distinct from constitutive correctness, suggests that actions that have constitutive standards of correctness don’t exemplify the distinction between RKRs and WKRs at all. This doesn’t imply that the distinction isn’t a general one that applies to action too—­we think it does. But it does imply that in order to find actions that exemplify the distinction, we need to look for actions that seem merited as a response, regard­ less of whether performing the action would be of any value.30 Relatedly, our argument suggests that there’s no right kind of reason problem. If constitutive-­ correctness-­makers for action aren’t RKRs, as we claim, then they don’t present a challenge to explain why RKRs for attitudes are authoritatively normative while these merely formally normative reasons for action are not. On our view, all RKRs are authoritatively normative reasons because they’re fit-­makers and not constitutive-­correctness-­makers.

10.4  Formal features of right-­kind reasons Another theoretical argument against the authoritative normativity of RKRs has its roots in a recent paper by Barry Maguire (2018). This argument targets the authoritative normativity of RKRs for affective attitudes in particular. Maguire argues that authoritative normative reasons are essentially gradable and contribu­ tory. They’re contributory in that they constitute ‘incomplete parts of a specific kind of explanation of overall normative facts, such as facts about what you ought to do’ (2018: 780). In particular, reasons contribute to ‘weighing explanations’ of overall normative facts by competing and combining with one another. Reasons are gradable in the sense that they have weights, such that one reason might be weightier than another or provide a greater degree of support for the response it favours. But according to Maguire, fit-­making facts for affective attitudes are nei­ ther gradable nor contributory: they don’t have weights and they never contribute to explanations of overall normative facts by competing and combining. Maguire thus concludes that fit-­making facts, or RKRs, for affective attitudes aren’t really reasons. So, assuming that something can have normative authority only if it is or provides a reason, it follows that RKRs for affective attitudes aren’t authoritatively normative. In principle, there are several ways to reply to this argument. One strategy would be to deny that something can have normative authority only if it is or provides a reason.31 Alternatively, one might reject that reasons are essentially 30  For some discussion, see Howard (2021). 31  This possible response was suggested to us by Jonathan Way.

In Defence of the Right Kind of Reason  233 gradable and contributory, or instead argue that RKRs for affective attitudes do have these features. In this section, we defend the authoritative normativity of RKRs by pursuing this latter option: we grant that something can have normative authority only if it is or provides a reason, and that reasons are essentially grad­ able and contributory, but argue that fit-­making facts for affective attitudes also possess these features essentially. We aim not only to diagnose where Maguire goes wrong in his arguments that these facts lack these features, but to suggest a positive view about what determines the weights of fit-­making facts for affective attitudes and how such facts compete and combine with one another so as to con­ tribute to weighing explanations of overall normative facts.32 Let’s start with Maguire’s argument that fit-­making facts never compete, which proceeds by cases. First, Maguire considers the loss of his grandmother. Upon her passing, he felt both sadness and relief. The fact that his grandmother was suffer­ ing seemed to make the latter response fitting; the fact that she was the family matriarch and ‘had great chat’ up until the end seemed to make the former fitting (2018: 780). Intuitively, these considerations don’t compete with each other. As Maguire puts it: ‘The facts that made the sadness fitting didn’t make the relief unfitting, nor did the facts that made the relief fitting make the sadness unfitting’ (2018: 780). It was fitting to feel both sadness and relief—­sadness that the family matriarch had passed and relief that she was no longer suffering. Second, Maguire considers a case in which your friend Andrew gets a promo­ tion that you’re also up for. According to Maguire, the fact that your friend got the promotion makes it fitting to feel pleased he got it and the fact that you didn’t get the promotion makes it fitting to feel disappointed you didn’t. Again, Maguire claims, ‘the considerations supporting these attitudes do not compete. The fact that you didn’t get the promotion doesn’t make it unfitting to feel pleased and the fact that your friend got it doesn’t make it unfitting to feel disappointed. It is fit­ ting to feel pleased and disappointed in these different respects in this case’ (2018: 787–8). Maguire concludes that fit-­making facts never compete, but instead directly, by themselves, each make a specific and separate attitude fitting. We agree with Maguire that, in the cases he considers, there’s no competition among the fit-­making facts. But we deny that these cases show that fit-­making facts for affective attitudes never compete and that these facts aren’t reasons. As David Faraci (2020) observes, Maguire’s argument seems to presuppose that ­reasons compete per se. But they don’t. Reasons for me to get work done today don’t compete with the reasons for me to relax tomorrow; and reasons for me to believe it’ll rain this afternoon don’t compete with reasons for me to hope it won’t.

32  For other recent responses to Maguire’s arguments, see esp. Faraci (2020) whose responses we largely agree with and build upon here, and Heape (2020), which provides a fruitful exploration of the prospects for reasons-­firsters of adequately responding to Maguire’s arguments; see also McHugh and Way (forthcoming, ch. 7).

234 Fittingness Instead, reasons compete only insofar as they favour incompatible alternatives. Faraci suggests that the explanation for why the fit-­making facts in Maguire’s cases don’t compete is that the attitudes in these cases aren’t incompatible alterna­ tives. We agree. Feeling sad that the family matriarch has passed isn’t in­com­pat­ ible with feeling relief that she’s no longer suffering. Likewise, feeling disappointed that you didn’t get the promotion isn’t incompatible with feeling pleased that your friend did. So, the lack of competition between the fit-­making facts in Maguire’s cases doesn’t suggest that these facts aren’t reasons for the attitudes they make fitting.33 What Maguire needs to show is that fit-­making facts for affective attitudes don’t compete even when the attitudes they make fitting are alternatives. But we think this can’t be shown. We’ll suggest that fit-­making facts do compete in such cases and we’ll explain how this works in practice. First, what does it take for two or more attitudes to be incompatible alterna­ tives? In the case of action, two acts are incompatible alternatives when the per­ formance of one makes impossible the performance of the other. In the case of attitudes, things are different. It’s possible both to believe p and to disbelieve p, but these attitudes are, in some sense, alternatives.34 Our working hypothesis is that affective attitudes are alternatives to one another when they can’t sim­ul­tan­eous­ly be fitting.35 For each type of attitude that can be fitting, there’s some evaluative property such that the fittingness of the attitude is equivalent to its object possessing that property. For example, admiration is fitting just in case its object is admirable, desire is fitting just in case its object is desirable, and awe is fitting just in case its object is awesome. Given this, we have a nice test for whether two attitudes can’t simultaneously be fitting and hence whether they’re in this way alternatives: two attitudes can’t simultaneously be fitting if it’s impossible for their object(s) to bear simultaneously the evaluative properties to which the fittingness of each is equiva­lent. For example, a person can’t simultaneously be both admirable and deplorable, and so it can’t at once be fitting both to admire and to deplore them. Hence, admiring and deploring one and the same thing are alternatives in the relevant sense.

33  McHugh and Way (forthcoming, ch. 7) also make this point. 34  Selim Berker makes this point in an earlier draft of Berker (2018). 35  Faraci (2020: 228) suggests a similar account of when two attitudes are incompatible alternatives. This account looks plausible in the case of affective attitudes and, we think, belief, but less so when it comes to certain conative attitudes such as intention. For example, it can at once be fitting to intend to A and to intend to not-­A (when Aing and not-­Aing are each worth doing) but these attitudes seem clearly to be alternatives to each other. Perhaps a more unified account of when attitudes are alternatives is forthcoming, but since our focus (and Maguire’s) is affective attitudes in particular, the above hypothesis about when such attitudes are alternatives will suffice for present purposes. Thanks to Selim Berker, Jonathan Way, and Alex Worsnip for helpful conversation and correspondence about this.

In Defence of the Right Kind of Reason  235 Of course, a person might have a mix of admirable and deplorable properties. If they do, then it can simultaneously be fitting to admire the admirable proper­ ties and to deplore the deplorable ones: the deplorability of one of a person’s prop­ erties needn’t preclude the admirability of some other property she possesses. So it can be fitting, for example, to admire Angie’s brilliance while deploring her evil intentions. But it can’t at once be fitting both to admire and deplore Angie herself. Could it be fitting both to admire and deplore one and the same property of a person? No. Angie’s brilliance can’t be admirable and deplorable. Plausibly, if Angie is brilliant, her brilliance is admirable; hence, not deplorable. But what if Angie is brilliant, but puts her brilliance to bad use? Then, we think, it’s fitting to admire her brilliance as such, but to deplore the use to which she puts it. We’re now in a position to give a deeper explanation of why there’s no competi­ tion among the fit-­making facts in Maguire’s cases, viz. that the affective attitudes his cases involve can simultaneously be fitting and so don’t count as alternatives. For example, the disappointingness of the fact that you didn’t get the promotion is compatible with the joyousness of the distinct fact that your friend did. Hence, disappointment regarding the former fact and joy regarding the latter fact don’t count as alternatives. Since there can be competition among fit-­making facts for attitudes only if the attitudes are alternatives, this is why there’s no competition among the fit-­making facts in Maguire’s cases. So, how does competition among fit-­making facts work when the relevant atti­ tudes are alternatives? Suppose Angie has a mix of admirable and deplorable properties. What’s the fitting attitude towards Angie herself? In particular, is it fitting to admire Angie, to deplore her, or to feel a kind of ambivalence? We think the answer depends on the outcome of a weighing process that takes as inputs Angie’s admirable and deplorable features. Here’s a brief sketch of how this works. Each of Angie’s admirable features is admirable to a degree. The degree to which an admirable feature is admirable corresponds to its weight. The same goes for Angie’s deplorable features: each is deplorable to a degree, and the degree to which a deplorable feature is deplorable equals its weight. To determine whether Angie herself is admirable or deplorable or neither, we first sum the weights of her admirable features, and then sum the weights of her deplorable features. We then determine the difference between these values. If the difference meets some (possibly vague) threshold, such that the combined weight of Angie’s admirable features is sufficiently greater than the combined weight of her deplorable ones, then Angie herself is admirable, and so fitting to admire.36 If, on the other hand, the combined weight of Angie’s deplorable features sufficiently exceeds the combined weight of her admirable ones, then she is deplorable, and so fitting to 36  Why ‘sufficiently greater’ rather than just ‘greater’? To preclude the possibility that Angie herself could be admirable in a case where her features are just slightly more admirable than they are deplorable.

236 Fittingness deplore. And, finally, if neither of these conditions is met, then Angie is neither admirable nor deplorable, and so it’s not fitting either to admire or to deplore her. Instead, in this last kind of case, it’s fitting to be ambivalent towards Angie, where ambivalence is something like the affective analogue of suspending judgement about a proposition’s truth. Several points of clarification are in order. First, settling whether Angie is admirable (or deplorable) doesn’t settle how admirable she is. There are many views about how to determine how admirable Angie is that are compatible with the above account, but we tend to favour this one: the extent to which Angie is admirable is determined by how much she exceeds the threshold for being ad­mir­ able (rather than deplorable). Roughly, once Angie has met the threshold for being admirable, the greater the degree to which she exceeds that threshold, the more admirable Angie is. Second, one might wonder how our model works in cases where an affective attitude has several alternatives.37 For instance, just as someone can’t be both admirable and deplorable, it’s also plausible that they can’t be both admirable and despicable. So, just as deploring someone is an alternative to admiring them, so too is despising them. We suggest that someone is admirable, rather than deplorable or despicable, if the combined weight of their admirable features is sufficiently greater than each of the combined weights of their deplorable features and their despicable features. And if someone is admirable, they’re fitting to admire, rather than to despise or deplore. In this way, our model can deliver verdicts about which attitudes are fitting, even in cases involving attitudes with multiple alternatives.38 Our model thus explains how facts about features that can contribute to the fittingness of alternative affective attitudes can compete to determine which atti­ tude is fitting. In other words: it explains how fit-­making facts, or RKRs, for alter­ native affective attitudes can compete to determine the (un)fittingness of each. We think this model applies not only in cases like the above, where the common object of a set of alternative attitudes is a person (or some other concrete object), but also in cases where the common object of a set of alternatives is a complex state of affairs, event, process, etc. For example, recall Maguire’s case in which your friend Andrew gets a promotion that you were also considered for. Let the fact that you didn’t get the promotion be a complex state of affairs that includes the fact that you won’t get the raise but also the fact that you won’t work for longer hours. The fact that you won’t get the raise is disappointing, but the fact that you won’t work for longer hours is pleasing. So, on our model, whether you not 37  Thanks to Selim Berker for raising this question. 38  In cases like this, we think it’s most plausible to hold that how admirable someone is is a function both of how much they exceed the threshold for being admirable rather than deplorable and how much they exceed the threshold for being admirable rather than despicable. We’re tempted by the view that in such cases no precise degree of admirability can be specified, and instead the best we can do is to specify an imprecise range.

In Defence of the Right Kind of Reason  237 getting the promotion is disappointing, and thus whether it’s fitting for you to be disappointed by this complex state of affairs, is determined by whether its proper parts are sufficiently more disappointing than they are pleasing. It’s easy to see how this picture might generalize.39 This model of how fit-­making facts can compete also reveals how such facts can combine, such that they sometimes together contribute to explanations of an attitude’s fit. We therefore conclude, contrary to Maguire, that fit-­making facts, or RKRs, for affective attitudes are contributory: they contribute to weighing ex­plan­ ations of overall normative facts by competing and combining. Our model also substantiates the fact that fit-­making facts for affective atti­ tudes are gradable, i.e. that they have weights. The weight of a fit-­making fact for an attitude depends on the extent to which the fact can contribute to making it the case that the object of the attitude has the evaluative property to which the fittingness of the attitude corresponds. For example, loving your daughter is fit­ ting just in case she’s lovable. On our model, your daughter is lovable just in case the combined lovability of her lovable features is sufficiently greater than the combined unlovability of her unlovable ones. So, suppose your daughter’s quirky sense of humour is more lovable than her intellectual curiosity. Then, all else equal, her quirky sense of humour can play a greater role than her intellectual curiosity in making it the case that she’s lovable. Hence, the former is a stronger RKR to love her. Why does Maguire think that fit-­making facts aren’t gradable and where does his argument go wrong? Maguire claims that fit-­making facts aren’t gradable because, strictly speaking, an attitude can’t be more or less fitting. For example, in comparing the deaths of an octogenarian and a twenty-­year old, Maguire claims that while we might say that the tragedy of the youngster’s death makes it more fitting to feel sad about her death, this just amounts to claiming that the tragedy of the youngster’s death makes it fitting to feel more sadness (2018: 790). And although the fittingness conditions for attitudes are themselves gradable—­something can be more or less admirable, fearsome, etc.—Maguire insists that this too only makes it fitting to have certain attitudes to varying degrees: if x is more admirable or fearsome than y, then it’s fitting to admire or fear x more than y. So, Maguire insists that fittingness is an all-­or-­nothing normative property and he takes this to show that RKRs aren’t gradable. The problem with Maguire’s argument is that fittingness needn’t be gradable in order for fit-­making facts to be gradable. To see the point, consider reasons to act. To say that R1 provides a stronger reason to perform act A than R2 is to say that

39  Our model leaves open whether fit-­making facts for affective attitudes with different objects could ever compete. Whether this is possible depends on whether such attitudes can ever count as alternatives. If they can, then our model can also be generalized easily to explain how the fit-­making facts for such attitudes compete.

238 Fittingness R1 (non-­factively) contributes to making it the case that one ought to or may A to a greater extent than R2. But notice that the overall normative statuses of which normative reasons provide contributory explanations—­what you ought to or may do—­are all or nothing. Either you ought to (or may) do A or not.40 The same may hold for fit-­making facts and fittingness: fittingness may be an overall, all or noth­ ing normative status, even though fit-­making facts are gradable in the sense that one fact can contribute to making it fitting to have some attitude more so than some other fact does. For example, as explained above, your daughter’s quirky sense of humour might contribute to making it fitting to love her more so than her intellectual curiosity. So, even if fittingness isn’t gradable, this doesn’t entail that fit-­making facts aren’t gradable. And even if the fitting intensity of an attitude corresponds to the degree of the relevant evaluative property that the attitude’s object possesses, this also doesn’t suggest that fit-­making facts for the attitude aren’t gradable. This is because, as our model explains, this too is compatible with taking the particular features of an attitude’s object to be gradable fit-­makers that determine not only the fitting intensity of the attitude in question, but also whether the attitude is ­fitting in the first place. For example, the particular (un)lovable features of your daughter determine not only how much love it’s fitting for you to feel towards her, but also whether your daughter is lovable, or fitting to love, in the first place.

10.5 Conclusion We’ve argued that the main reason for thinking that RKRs are authoritatively nor­ mative is simply that it’s required to explain the intuitive data. Rejecting this data is a serious cost that should be embraced only if there are strong theoretical ­reasons for thinking that, despite the appearances, RKRs aren’t authoritatively normative after all. And we’ve argued that the main theoretical reasons for think­ ing this that have been offered in the literature so far aren’t compelling: RKRs needn’t and shouldn’t be analysed in terms of constitutive correctness and they have the essential features of normative reasons. We therefore conclude that scep­ ticism about the authoritative normativity of RKRs is unwarranted.

References Berker, S. (2018). A combinatorial argument against practical reasons for belief. Analytic Philosophy, 59: 427–70. Berker, S. (forthcoming). The deontic, the evaluative, and the fitting. In C.  Howard and R. Rowland (eds.), Fittingness. Oxford: Oxford University Press. 40  Faraci (2020: 232) makes this point with respect to permissibility.

In Defence of the Right Kind of Reason  239 Brandt, R. (1946). Moral valuation. Ethics, 56: 106–21. Brentano, F. (1889/2009). The origin of our knowledge of right and wrong. Oxford: Routledge. Broad, C. D. (1930). Five types of ethical theory. London: Routledge. Cherry, M. (2018). The errors and limitations of our ‘anger-evaluating’ ways. In Myisha Cherry and Owen Flanagan (eds), The moral psychology of anger, 49–66. Lanham, MD: Rowman and Littlefield. Côté-Bouchard, C., and Littlejohn, C. (2018). Knowledge, reasons, and errors about error theory. In C. Kyriacou and R. McKenna (eds), Metaepistemology, realism, and anti-realism, 147–71. London: Palgrave Macmillan. D’Arms, J. (forthcoming). Fitting emotions. In C.  Howard and R.  Rowland (eds.), Fittingness. Oxford: Oxford University Press. D’Arms, J., and Jacobsen, D. (2006). Anthropocentric constraints on human value. Oxford Studies in Metaethics, 1: 99–126. Ewing, A. C. (1948). The definition of good. London: Routledge & Kegan Paul. Faraci, D. (2020). We have no reason to think there are no reasons for affective atti­ tudes. Mind, 129: 225–34. Feinberg, J. (1970). Doing and deserving: Essays in the theory of responsibility. Princeton, NJ: Princeton University Press. Gertken, J., and Kiesewetter, B. (2017). The right and wrong kind of reasons. Philosophy Compass, 12: e12412. Gibbard, A. (1990). Wise choices, apt feelings. Cambridge, MA: Harvard University Press. Hawthorne, J. (2004). Knowledge and lotteries. Oxford: Oxford University Press. Heape, A. (2020). How there could be reasons for affective attitudes. Ethical Theory and Moral Practice, 23: 667–80. Howard, C. (2016). In defense of the wrong kind of reason. Thought, 5: 53–62. Howard, C. (2018). Fittingness. Philosophy Compass, 13: e12542. Howard, C. (2019a). The fundamentality of fit. Oxford Studies in Metaethics, 14: 216–36. Howard, C. (2019b). Weighing epistemic and practical reasons for belief. Philosophical Studies, 177: 2227–43. Howard, C. (2021). Consequentialists must kill. Ethics, 131: 727–753. Kelly, T. (2003). Epistemic rationality as instrumental rationality: A critique. Philosophy and Phenomenological Research, 66: 612–40. Kiesewetter, B. (2021). Are epistemic reasons normative? Noûs. (online first). Kvanvig, J. (2009). Assertion, knowledge and lotteries. In P. Greenough and D. Pritchard (eds), Williamson on knowledge, 140–60. Oxford: Oxford University Press. Kvanvig, J. (2011). Norms of assertion. In J. Brown and H. Cappelen (eds), Assertion: New philosophical essays, 233–50. Oxford: Oxford University Press. Leary, S. (2017). In defense of practical reasons for belief. Australasian Journal of Philosophy, 95: 529–42.

240 Fittingness Lord, E., and Sylvan, K. (2019). Reasons: Wrong, right, normative, fundamental. Journal of Ethics and Social Philosophy, 15: 43–74. Maguire, B. (2018). There are no reasons for affective attitudes. Mind, 127: 779–805. Maguire, B., and Woods, J. (2020). The game of belief. The Philosophical Review, 129: 211–49. Mantel, S. (2019). Do epistemic reasons bear on the ought simpliciter? Philosophical Issues, 29: 214–27. McDowell, J. (1998). Values and secondary qualities. Reprinted in his Mind, value and reality. Cambridge, MA: Harvard University Press. McHugh, C. (2015). The illusion of exclusivity. European Journal of Philosophy, 23: 1117–36. McHugh, C., and Way, J. (2016). Fittingness first. Ethics, 126: 575–606. McHugh, C., and Way, J. (forthcoming). Getting things right: Fittingness, value, and reasons. Oxford: Oxford University Press. McPherson, T. (2018). Authoritatively normative concepts. Oxford Studies in Metaethics, vol. 13: 253–77. Naar, H. (2021). The fittingness of emotions. Synthese, 199: 13601–19. Papineau, D. (2013). There are no norms of belief. In T. Chan (ed.), The aim of belief, 64–79. Oxford: Oxford University Press. Parfit, D. (2011). On what matters, vol. 1. Oxford: Oxford University Press. Reisner, A. (2008). Weighing pragmatic and epistemic reasons for belief. Philosophical Studies, 138: 17–27. Reisner, A. (ms). The pragmatic foundations of theoretical reason. Rinard, S. (2019). Equal treatment for belief. Philosophical Studies, 176: 1923–50. Rosen, G. (2001). Brandom on modality, normativity, and intentionality. Philosophy and Phenomenological Research, 63: 611–23. Rowland, R. (2022). The authoritative normativity of fitting attitudes. Oxford Studies in Metaethics. 17: 108–138 Schroeder, M. (2007). Slaves of the passions. Oxford: Oxford University Press. Schroeder, M. (2010). Value and the right kind of reason. Oxford Studies in Metaethics, 5: 25–55. Shah, N. (2003). How truth governs belief. The Philosophical Review, 112: 447–82. Shah, N. (2008). How action governs intention. Philosophers’ Imprint, 8: 1–19. Sharadin, N. (2015). Reasons wrong and right. Pacific Philosophical Quarterly, 96: 1–29. Skorupski, J. (2010). The domain of reasons. Oxford: Oxford University Press. Srinivasan, A. (2018). The aptness of anger. Journal of Political Philosophy, 26: 123–44. Steglich-Peterson, A. (2006). No norm needed: On the aim of belief. Philosophical Quarterly, 56: 499–516. Svavarsdóttir, S. (2014). Having value and being worth valuing. Journal of Philosophy, 111: 84–109.

In Defence of the Right Kind of Reason  241 Tappolet, C. (2011). Values and emotions: Neosentimentalism’s prospects. In Bangoli C. (ed.), Morality and the Emotions. Oxford: Oxford University Press. Way, J. (2012). Transmission and the wrong kind of reason. Ethics, 122: 489–515. Way, J. (2013). Value and reasons to favour. Oxford Studies in Metaethics, 8: 27–49. Weiner, M. (2005). Must we know what we say? The Philosophical Review, 114: 227–51. Whiting, D. (2021). Aesthetic reasons and the demands they (do not) make. Philosophical Quarterly, 71: 407–27. Wiggins, D. (1987). A sensible subjectivism? In his Needs, values, truth: Essays in the philosophy of value. Oxford: Blackwell. Williamson, T. (2000). Knowledge and its limits. Cambridge, MA: Harvard University Press. Wodak, D. (2019). Mere formalities: normative fictions and normative authority. Canadian Journal of Philosophy, 49: 828–50. Wodak, D. (2021). Approving on the basis of normative testimony. Oxford Studies in Metaethics, 16: 183–206 Woods, J. (2018). The authority of formality. Oxford Studies in Metaethics, 13: 217–29.

SECTION THREE

FI T T INGN E SS A ND VA LU E T HE ORY

11 Value and Idiosyncratic Fitting Attitudes Conor McHugh and Jonathan Way

11.1  Introduction Sarah and Stephen are in trouble; only one can emerge unscathed. If Sarah is your friend and Stephen is a stranger then, though you should be concerned about both, it makes sense to be more concerned about Sarah than Stephen. It  makes sense for you to prefer that it is Sarah who emerges unscathed, for example. Cases like this illustrate the partiality problem for norm-­attitude accounts of value. Very roughly, such accounts say that for something to be of value is for there to be norms that support valuing that thing. The problem is that in cases like this, the norms on valuing don’t seem to line up with what’s valuable. It wouldn’t be any better if Sarah escapes—­we can suppose—­but it still makes sense for you to prefer that she does. The aim of this chapter is to explore how norm-­attitude accounts can respond to this problem. More specifically, we argue that our preferred such account—­a fitting-­attitude account—­is just as well p ­ laced to solve the problem as its central rival—­the buck-­passing account. This is important because the buck-­passing account might seem to have an important advantage over the fitting-­attitude account, when it comes to the partiality problem. We argue that this appearance is misleading. After explaining norm-­attitude accounts in more detail (§11.2), we present the partiality problem (§11.3) and explain how the buck-­passer can solve it in a way that is not available to the fitting-­attitude account (§11.4). We argue, however, that the fitting-­attitude account can offer a broadly analogous solution (§11.5). We refine the solution in response to objections (§11.6), and close by discussing a more general family of challenges (§11.7).

11.2  Norm-­Attitude Accounts Norm-­attitude accounts say that for something to be valuable is for there to be norms that support valuing that thing. There are different types of norms and thus different types of norm-­ attitude account. According to fitting-­ attitude Conor McHugh and Jonathan Way, Value and Idiosyncratic Fitting Attitudes In: Fittingness: Essays in the Philosophy of Normativity. Edited by: Christopher Howard and R. A. Rowland, Oxford University Press. © Conor McHugh and Jonathan Way 2022. DOI: 10.1093/oso/9780192895882.003.0011

246 Fittingness accounts, something is of value if it is fitting to value. According to buck-­passing accounts, something is of value if the reasons support valuing it. These are the two forms of norm-­attitude account we will be concerned with.1 These accounts can be specified in several respects. First, while norm-­attitude accounts might apply to all forms of value, we will focus primarily on goodness simpliciter. More precisely, since it is most relevant for the partiality problem, we will focus on the comparative property of betterness simpliciter. Being better simpliciter should be distinguished from being better for (as something might be better for one’s health, or for trees, or for an engine) and from being attributively better (as in better toasters, better holiday destinations, and better assassins), as well as from specific value properties, such as being delightful, awesome, amusing, fearsome, or loathsome. Second, different forms of value can be understood in terms of different forms of valuing. Examples of valuing include desire, taking pleasure in, being glad, certain emotions such as amusement and admiration, and certain actions, such as promoting and preserving. Betterness simpliciter is most naturally understood in terms of norms on preference: for something to be better is for preferring it to be fitting, or supported by reasons; for it to be equally good is for indifference to be fitting, or supported by reasons.2 We assume that preference—­which we understand broadly, to include indifference—­can be directed towards states of affairs that you know are not under your control, such as states of affairs that are in the past. We also assume that preferences need not involve thoughts about betterness.3 Putting these points together, we can state the accounts we will be concerned with as follows: (FA) For X to be better than Y is for it to be fitting to prefer X to Y; for X and Y to be equally good is for it to be fitting to be indifferent between X and Y.

1  There are also prominent accounts which analyse value in terms of what it is rational to value, given full information (Brandt 1979, Smith 1994). However, proponents of these accounts characteristically understand rationality in terms of coherence. So understood, these views count as subjectivist (see later this section). If rationality is instead understood in terms of reasons (as in Parfit 2011), these views would be a form of buck-­passing. 2  For simplicity we treat preference as a sui generis attitude, subject to its own norms. One might instead regard preference as reducible to desire, e.g. to prefer X to Y is to desire X more strongly than one desires Y, to be indifferent between X and Y is to equally desire X and Y. In that case betterness simpliciter might ultimately be understood in terms of norms on these desires, e.g. for X to be better than Y is for the fitting level of desire for X to be greater than the fitting level of desire for Y. This would complicate some of our arguments but not affect them substantively. Finally, note that indifference between outcomes does not imply that you don’t care about these outcomes: indifference is compatible with strongly desiring each. 3  Thus the accounts we will be concerned with avoid the objection that norm-­attitude accounts are circular because they define goodness in terms of thoughts about goodness. (For a version of this objection, against certain norm-­attitude accounts, see Ross 1939: 278–9.)

Value and Idiosyncratic Fitting Attitudes  247 (BP)  For X to be better than Y is for the balance of reasons to support preferring X to Y; for X and Y to be equally good is for the balance of reasons to support indifference between X and Y. Further specifications can be made. For instance, value can be final or derivative. Happiness is a plausible candidate for being finally good. The paradigm derivative goods are instrumental: if money, for example, is a means to happiness, this makes it derivatively good. Value can also be pro tanto or overall. An outcome in which happiness is equally distributed might be better in that respect than one in which happiness is unequally distributed, but worse overall because everyone is still miserable. These distinctions can be captured in terms of the corresponding distinctions in valuing. For example, a fitting-­attitude account can say that something is finally better if it is fitting to prefer it for its own sake and derivatively better if it is fitting to prefer for the sake of something else. Something is pro tanto better—­better in some respect—­if it is fitting to pro tanto prefer it—­prefer it in that respect—­and overall better if fitting to prefer overall. A buck-­passing account can make parallel claims. These distinctions will be important later.4 Why adopt a norm-­attitude of betterness? We mention two attractions. First, an account of betterness should be expected to generalize to other value properties, and to illuminate the similarities and differences between them. Norm-­ attitude accounts are well placed to do this. Specific value properties are evidently tightly connected to certain attitudes—­the delightful to delight, the awesome to awe, and so forth. Norm-­attitude accounts can offer an immediately compelling account of these connections: the delightful is what merits delight; the awesome is what is worthy of awe (Schroeder 2010).5 These properties are thus closely related to goodness simpliciter, which is plausibly identical to the specific value property of being desirable or valuable simpliciter.6 Second, norm-­ attitude accounts can seem an attractive middle-­ ground between two other accounts of value. According to primitivism, values are basic, indefinable properties (Moore 1903). According to subjectivism, to be of value is to be valued under certain non-­normative conditions—­perhaps conditions of full information, coherence, reflection, or the like.7 The problem with primitivism is

4  A further question is how norm-­attitude accounts should be extended to account for incommensurable value (Chang 1997). This will not be crucial here, but we tentatively suggest that X and Y are of incommensurable value when the norms don’t support preferring X to Y, preferring Y to X, or being indifferent between X and Y. 5  In light of such analyses, one might claim that ‘specific value properties’ are better classified as, say, fittingness properties (Berker, this volume). We maintain that being analysed in terms of fittingness does not prevent them from being value properties, but nothing here turns on this classificatory question. 6  Mill 1863, Scanlon 1998, Schroeder 2010, Zimmerman 2007. 7  Influential proponents of sophisticated subjectivism (though not always focusing on goodness simpliciter) include Firth (1952), Brandt (1979), Lewis (1989), Railton (1986), Smith (1994), and Street (2008).

248 Fittingness that it does nothing to illuminate value. Subjectivism offers such illumination, but only at the cost of first-­order implausibility. To take a relatively benign example, subjectivism threatens to imply that money is finally good, since some people value money for its own sake, and might continue to do so under conditions of full information, coherence, and reflection.8 More generally, given the evident possibility of human perversity, it is plausible that for any non-­normative conditions, there can be people who satisfy those conditions but fail to value only what is really valuable.9 Norm-­attitude accounts avoid these problems.10 They join with subjectivism in rejecting the primitivist’s postulation of a basic property of goodness and insisting upon an important connection between value and valuing. But since this connection is normative, it avoids the dubious implications of subjectivism. The cases of perversity which undermine subjectivism are precisely cases of valuing without adequate reason, or of unfitting valuing. Norm-­attitude accounts thus have important attractions. But, as we have seen, norm-­attitude accounts differ. In particular, FA and BP appeal to different normative statuses. FA understands betterness in terms of fitting preference, BP in terms of the reasons to prefer. Reasons are very familiar in normative philosophy: a reason is a consideration that counts in favour of a response, with a certain strength. We will say that the balance of reasons—­or ‘the reasons’—support preferring X to Y when the reasons for preferring X to Y are together stronger than the reasons for preferring Y to X and stronger than the reasons for being indifferent between X and Y. An attitude is fitting when it is merited by, worthy of, or appropriate to, its object. For example, it seems fitting to admire Mandela, fear an onrushing tiger, intend to phone your mother on her birthday, and believe that the Seine flows through Paris. In having these attitudes, you are getting things right. By contrast, it is not fitting to admire Idi Amin, fear an onrushing kitten, intend to ignore your mother on her birthday, or believe that the Thames flows through Paris. These objects are not worthy of, do not merit, and are not appropriate objects of, the corresponding attitudes. In having these attitudes you would be getting things wrong.11

8  Alternatively, if subjectivism is the view that what is finally good is what is valued for its own sake by everyone, then the view seems to imply that nothing is finally good. 9 Of course, some subjectivists have argued that their views can avoid any such revisionist implications—­see e.g. Smith 1994, Korsgaard 1996, Markovits 2014. But their arguments have not been widely accepted. 10  McDowell 1985, Wiggins 1987. For discussion, see D’Arms and Jacobson 2006. 11  For more on fittingness, see e.g. Howard 2018, McHugh and Way 2022: ch.3, as well as the other contributions to this volume.

Value and Idiosyncratic Fitting Attitudes  249 As these examples illustrate, fittingness is an overall rather than contributory status.12 It is thus distinct from the property of being supported by some reason. We will also take it that the fitting-­attitude account takes fittingness to be distinct from being supported by the balance of reasons. This ensures that FA is a genuine rival to BP. One way to see the distinction between fittingness and the balance of reasons is by considering examples of so-­called wrong-­kind reasons. If a terrorist threatens to kill you if you admire Mandela, fear the tiger, or believe that the Seine flows through Paris, there seems a strong reason against these attitudes—­strong enough to tip the balance of reasons. But they remain fitting—­Mandela remains worthy of admiration, the tiger is worthy of fear, and so on. At the least, it seems entirely coherent and intelligible to claim that the threat gives you most reason not to admire Mandela but does not stop him being fit to admire. This suggests that at least the concepts of reasons and fittingness must be distinct. If such claims can be true, then the properties they pick out are also distinct.13 As we have argued elsewhere, such examples reveal a significant advantage of fitting-­attitude accounts over buck-­passing accounts (McHugh and Way 2022: ch.4). When the terrorist threatens to kill you if you admire Mandela, this seems to give you a strong reason not to admire Mandela but he remains fit to admire. A buck-­passing account of the admirable thus seems to imply that Mandela is not admirable. By contrast, a fitting-­attitude account has no such implication. Or suppose the terrorist threatens to kill you unless you prefer the destruction of the whole world to the scratching of your little finger. This seems to give you a strong reason to prefer the destruction of the whole world. But it does not make this preference fitting. The buck-­passing account, but not the fitting-­attitude account, thus seems to imply that it is better if the world is destroyed than if your little finger is scratched. Of course, this is only the first step of an argument. Buck-­passers have various ways of responding to this ‘wrong kind of reason’ problem. Some deny, for ex­ample, that the terrorist’s threat gives you a reason to prefer the destruction of the whole world. Others hold that this is a reason of the ‘wrong kind’ and that the buck-­passing account needs to be restricted to reasons of the ‘right kind’. We discuss these responses elsewhere (McHugh and Way 2022: ch.4.). For p ­ resent purposes, it is enough that it is an advantage of the fitting-­attitude account that such responses are not even needed. Of course, this is just one advantage of fitting-­attitude accounts over buck-­ passing. Perhaps buck-­passing also has important advantages. In the rest of this 12  At any rate, the fittingness we are interested in here is overall—­perhaps there is also contributory or pro tanto fittingness. We also take fittingness to be categorical rather than gradable (see Maguire 2018), but this assumption plays no essential role. 13  For a fuller case for regarding these concepts and the corresponding properties as distinct, see McHugh and Way 2022: ch.3.

250 Fittingness chapter, we consider whether this is so with respect to the partiality problem. Buck-­passers have an attractive response to this problem, one which might seem unavailable to the fitting-­attitude account. That would be an important advantage of buck-­passing.

11.3  The Partiality Problem To understand the partiality problem, a further specification of our norm-­attitude accounts is required. Norms apply to agents. Attitudes are not simply fitting or supported by reasons. Rather, there is a reason for some person, or it is fitting for some person, to have an attitude. So norm-­attitude accounts must specify who the norms they refer to apply to. The natural way to do this is to appeal to norms which apply to all agents, as follows. (For brevity, the clauses about equal goodness are left implicit.) (FA1)  For X to be better than Y is for it to be fitting for all agents to prefer X to Y. (BP1) For X to be better than Y is for preferring X to Y to be supported by the balance of reasons, for all agents. This is the natural way to go because betterness simpliciter seems not to be relativized to individuals. To say that it is better simpliciter if benefits are distributed in proportion to desert is not just to say that this is better for those who are benefitted. Nor is it to say that this is better from your perspective or given what you care about. It is to say, if you like, that it is better from no perspective, or from the perspective of the universe. FA1 and BP1 give a way to capture this.14 However, these formulations run into a simple problem: what it is fitting to value and what the reasons support seem to vary between agents. The most discussed kind of example, and the kind we shall focus on, involves partiality.15 Recall our initial example: Sarah and Stephen are in trouble; only one can emerge unscathed. If Sarah is your friend and Stephen is a stranger, then, although it is fitting to be concerned about both, it also seems fitting for you to be more concerned about Sarah. Sarah’s being your friend makes her an appropriate object of your concern, in a way that goes beyond the concern you might have for a stranger; it thus seems fitting for you to prefer her to emerge unscathed. Likewise, the balance of reasons seems to support this preference.

14  ‘All agents’ means all possible agents in all circumstances—­or at least, all circumstances in which any general conditions on the relevant norms applying (e.g. being able to prefer X to Y) are met. 15  In particular, we focus on positive partiality. There may also be appropriate negative partiality (Brandt 2020). As far as we can see, everything we say could be adapted to apply to negative partiality.

Value and Idiosyncratic Fitting Attitudes  251 But Stephen’s well-­being is no less valuable than Sarah’s, and it may not be ­better that Sarah emerge unscathed.16 The claim that it can be fitting to be partial in this way is sometimes challenged. In a discussion of the partiality problem, Jonas Olson writes: Advocates of FA analyses could maintain that it would be fitting to respond with equal disfavour to the suffering of one’s child . . . and the equal suffering of a stranger . . . [T]he intuition that we should respond with greater disfavour to the . . . suffering of people to whom we stand in personal relations is a moral intuition and it is not clear why moral intuitions should be taken into con­sid­er­ ation in analyses of value.  (2009: 374–5)

Chris Howard takes a similar line, defending the claim that if someone has better qualities than your beloved, it is fitting for you to love that person more: [T]he fittingness relation is the relation in which a response stands to an object when the object merits, or is worthy of, that response. So to say that it would be fitting for you to love someone else more than your beloved is to say only that that person merits or is worthy of more love than your beloved—i.e. that the relevant person is more lovable.  (2019: 131, original italics)

Similarly, one might say: To say that it would be fitting for you to be equally concerned for Sarah and Stephen is to say only that Sarah and Stephen’s interests merit or are worthy of equal concern—neither’s escape would be preferable to the other.

We do not find these arguments convincing. Perhaps Olson is right that partiality is a moral consideration. But moral considerations can bear on what is fitting. For example, outcomes can be bad simpliciter—­and thus fitting to disfavour—­because they are unjust, or because they involve suffering. And we agree that Sarah and Stephen’s interests merit equal concern. The question is how this is to be spelled out. Which agents does it apply to? Should we take it to mean that it is fitting for everyone to be equally concerned with Sarah and Stephen’s interests? Or should we understand it in a more restricted way? The interest of the case is that it suggests the latter. We thus maintain that it may be fitting for you to prefer that Sarah escape—­ and more generally, that partiality can be fitting. Indeed, we take such cases to 16  For an influential statement of this objection, see Blanshard 1961: 287–8. As Blanshard notes, the problem was already discussed by Ewing (1939: 18ff; 1948: 191ff). Note that the partiality problem is not about how to capture the appropriateness of partiality (for recent discussion see Keller 2013, Lord 2016). The problem is rather how to account for the ‘impartial’ value facts, given that some agents are appropriately partial.

252 Fittingness support a stronger claim: that it may be unfitting for you to be indifferent between equally good outcomes. So understood, such cases are counterexamples to FA1. Since they are also cases in which your reasons need not support indifference, they are also counter­examples to BP1. This is the partiality problem. Before moving on, we note two corollaries. First, as noted above, many buck-­ passers hope to solve the wrong kind of reason problem by distinguishing between the right and wrong kind of reasons. Roughly, the right kind of reasons to value something are those which bear on the value of that thing, the wrong kind of reasons are those that don’t. Some hope that this distinction—­however it is to be drawn—­will also serve to solve the partiality problem (Olson 2009, Howard 2019). If what we have just said is right, then this hope is forlorn. One of the central ‘earmarks’ of the distinction between the right and wrong kind of ­reasons is that right-­kind reasons bear on fittingness. It should thus turn out that a response is fitting just in case it is supported by the right-­kind reasons—­indeed, buck-­passers may propose to analyse fittingness in this way (Schroeder 2010). But given that it can be fitting to be partial, it will follow that right-­kind reasons can support partiality. The partiality problem cannot be solved by appealing to right-­kind reasons.17 Second, several philosophers hold that fittingness should be analysed in terms of accurate representation. Just as belief is fitting just if true, other responses are fitting just if they accurately represent their objects (Rosen 2015, Tappolet 2016). On this view, it seems, a response cannot be fitting for you but unfitting for me; the content of that response, whatever it may be, is either accurate or inaccurate. But partiality seems to give rise to just such cases, as when it is fitting for you, but not for me, to prefer that Sarah escape. If preferring that Sarah escape is, say, representing Sarah’s escaping as better, then this representation cannot be ac­cur­ate for you but not for me.18

11.4  The Buck-­Passer’s Solution Several solutions to the partiality problem have been proposed. Some of these are available to both buck-­passing and fitting-­attitudes views.19 But one particularly 17 This point is supported by the other ‘earmarks’ of the right-­/wrong-­kind reason distinction offered by Schroeder (2012). For example, wrong-­kind reasons for an attitude cannot straightforwardly be your (motivating) reason for that attitude. But there is no problem in, e.g. preferring that Sarah escape for the reason that she is your friend. 18  This problem might be avoided by claiming that our preferences have different contents—­for instance, that yours represents Sarah’s escaping as better from your point of view and mine represents it as better from my point of view. But to make such claims, not only about preference but about all attitudes that can be fittingly partial, would be a significant, and not obviously plausible, further commitment. 19  For example, Oddie (2005) argues that value should be understood in terms of what it is fitting to favour when at ‘zero distance’ from an object. Orsi (2013) argues that value should be understood in terms of what it is fitting to favour when behind a ‘veil of ignorance’. While we will eventually agree

Value and Idiosyncratic Fitting Attitudes  253 attractive solution seems only to be available to buck-­passers. This solution appeals to the distinction between agent-­neutral and agent-­relative reasons.20 Agent-­neutral reasons are reasons that are shared by everyone. Agent-­relative reasons are reasons that only some of us have.21 For example, the fact that Sarah is in trouble is a reason to be concerned, and the fact that Stephen is in trouble is a reason to be concerned. Plausibly, these reasons are shared by everyone. The fact that Sarah is your friend is a reason for you to prefer Sarah to escape. This is not a reason that everyone shares. So other things equal, the reasons that are shared by everyone support equal concern. But your reasons support preferring that Sarah escape. The buck-­passer might thus suggest: (BP2) For X to be better than Y is for the balance of agent-­neutral reasons to support preferring X to Y. BP2 avoids the partiality problem: in cases of partiality what is supported by your reasons comes apart from what is supported by the agent-­neutral reasons. This is an elegant response to the problem. The distinction it appeals to, between agent-­neutral and agent-­relative reasons, is one we need to make anyway, and can be drawn clearly and straightforwardly. And it maintains a natural way of capturing the non-­relativity of betterness simpliciter, in terms of the non-­relativity of agent-­neutral reasons. Of course, we might want an explanation of the needed claims about agent-­ relative and agent-­ neutral reasons—­ for instance, of why your agent-­relative reasons to prefer that Sarah escape outweigh the agent-­neutral reasons to be indifferent. But the view has the right structure to solve the problem, and the substantive claims about reasons it appeals to are plausible. Furthermore, this solution seems unavailable to the fitting-­attitude account. The solution turns on the fact that some reasons for a response are shared among all agents while others aren’t, so that the overall status of that response varies between agents. It thus requires that the normative factors in terms of which with these authors that value needs to be understood in terms of what it is fitting for anyone under certain restrictions, we do not think either of these restrictions are acceptable. It is obscure what it is to be at ‘zero distance’ from an object (Bykvist 2009, Orsi 2013). And we take fittingness to be ‘objective’, in the sense that whether a response is fitting need not depend on your epistemic situation (McHugh and Way 2022: ch.3). In that case, being behind a veil of ignorance may make no difference to what it is fitting to favour. We discuss a third solution, due to Zimmerman (2011), in n.24). 20  For a solution along these lines, though not quite in these terms, see Lemos 2011. Schroeder (2010) suggests a solution like this for the parallel problem of temporal partiality (see §11.7). The claim that reasons of partiality are agent-­relative is common in the partiality literature; see, e.g. Keller 2013: 23, Brandt 2020: n.4. 21  Again, ‘everyone’ here means ‘all possible agents in all circumstances in which any general conditions on reasons are met’. For discussion of different uses of the terminology of ‘agent-­neutral’ and ‘agent-­relative’ reasons, see Ridge 2017.

254 Fittingness value is defined are contributory. However, as we explained earlier, fittingness is overall, not contributory. Thus the fitting-­attitude account seems unable to offer a parallel solution. This looks like an important advantage for buck-­passers.

11.5  A Fitting-­Attitude Solution to the Partiality Problem The buck-­passer’s attractive solution to the partiality problem looks like an advantage. We will now argue, however, that the fitting-­attitude account can offer a broadly analogous solution by distinguishing between types of valuing attitudes. As noted, we can value and disvalue things pro tanto—­that is, in certain respects. You might desire to visit Sicily both for the food and the weather; you might prefer to visit Sicily over Devon in these respects but prefer Devon in respect of convenience. The same goes for other valuing attitudes. You might admire someone both in respect of their creativity and their generosity, or resent someone both for what they said and how they said it.22 Return now to our case of partiality. If Sarah is in trouble, there are different respects in which you might desire that she escape. You might desire her escape in that this would involve someone escaping trouble. You might also desire her escape in that it would involve your friend escaping trouble. Both of these responses seem fitting. If Stephen is also in trouble, it is fitting to desire that he escape in that this would also involve someone escaping. But if Stephen is a stranger, then it isn’t fitting to desire that he escape in that he is your friend. So if only one can escape, it is fitting to be indifferent about who escapes in that both would involve someone escaping. But it is fitting to prefer that Sarah escape in that she is your friend. The fittingness of this latter attitude is idiosyncratic: while it is fitting for everyone to be indifferent in that two people’s interests are at stake, it isn’t fitting for everyone to prefer that Sarah escape in that she is your friend. More generally, we can say: Preferring X to Y in respect R is idiosyncratically fitting for an agent A iff it is fitting for A, but not for all agents, to prefer X to Y in respect R. Indifference between X and Y in respect R is idiosyncratically fitting for A iff it is fitting for A, but not for all agents, to be indifferent between X and Y in respect R.

22  What are respects? We take them to be propositions. Thus, when we say, for instance, that you prefer Sicily to Devon in respect of the food, we mean that you prefer it in respect of some proposition about the food (e.g. that it is tastier). One advantage of taking respects to be propositions is that, assuming that reasons are (true) propositions, it allows us to maintain that the respects in which it is fitting to prefer X to Y are also reasons for preferring X to Y (compare McHugh and Way 2022: ch.4). We leave open what else having an attitude in a respect involves, e.g. whether it requires believing the relevant proposition.

Value and Idiosyncratic Fitting Attitudes  255 Our solution to the partiality problem begins by excluding these idiosyncratically fitting attitudes: (FA2) For X to be better than Y in respect R is for it to be fitting for all agents to prefer X to Y in respect R. Since it’s not fitting for all agents to prefer Sarah to escape in the respect that she’s your friend, FA2 implies that Sarah’s being your friend is not a respect in which it is better if she escapes. This seems correct. What about overall betterness? The partiality problem rests on the thought that, other things equal, Sarah’s and Stephen’s escaping would be equally good, overall. Since FA2 says nothing about overall betterness, it does not conflict with this thought. But we also need an account of overall betterness which accommodates it. Our first pass at such an account deploys the notion of a neutral agent: A neutral agent in respect of X and Y is an agent for whom no pro tanto preferences between X and Y are idiosyncratically fitting.

We can then say: (FA3) For X to be better than Y overall is for it to be fitting for a neutral agent to prefer X to Y overall. We take it that it is not fitting for a neutral agent to have an overall preference between Sarah and Stephen escaping. Rather, it is fitting for a neutral agent to be indifferent between these outcomes. Thus FA3 does not imply that it is better that either escape. Why is it fitting for a neutral agent to be indifferent between Sarah and Stephen escaping? Well, the pro tanto preferences it is fitting for a neutral agent to have are the pro tanto preferences it is fitting for everyone to have. And it is plausible that the fittingness of overall attitudes is determined by the fittingness of pro tanto attitudes. Since we are assuming that other things are equal, and thus that any respects in which it is fitting for everyone to prefer Sarah to escape are balanced against respects in which it is fitting for everyone to prefer Stephen to escape, it is plausible that the overall attitude that is fitting here is indifference.23 Note that, since the fittingness of overall preference is determined by the fittingness of pro tanto preferences, this account allows us to capture the plausible idea that overall betterness is determined by pro tanto betterness. Of course, it is difficult to say exactly how pro tanto betterness determines overall betterness. 23 As a reminder, we do not take fittingness to be contributory. Thus the respects here (and throughout) are respects of preference, not of fittingness.

256 Fittingness And similarly, it is hard to say exactly how fitting pro tanto attitudes determine fitting overall attitudes. But we take it that this is a task for first-­order theory. Our proposal is simply that the way in which overall betterness is determined by pro tanto betterness is constituted by the way in which the fittingness of overall attitudes is determined by the fittingness of pro tanto attitudes, for a neutral agent. This solution to the partiality problem is broadly analogous to the buck-­passer’s solution. Whereas the buck-­passer appeals to facts which are reasons for everyone to prefer, we appeal to facts which are respects in which it is fitting for everyone to prefer. Plausibly, these sets of facts are identical. If so, the views take pro tanto betterness to be determined by the same facts. And they can both agree that overall betterness is determined by pro tanto betterness. The two views thus share advantages. Both agree that the non-­relativity of value is to be understood in terms of norms on attitudes that apply to everyone. Both appeal to distinctions which are independently needed and straightforward to draw. And both appeal to substantive claims—­for buck-­passers, the claim that agent-­relative reasons for preference can defeat agent-­neutral reasons, and for us, the claim that fitting idio­ syn­crat­ic pro tanto preferences can make for fitting overall partial preferences—­ which are plausible, even if we would ultimately like a fuller explanation of them.24

11.6  Two Challenges 11.6.1  Which Preferences Are Idiosyncratically Fitting? We said that it is fitting for you, but not for everyone, to prefer Sarah to escape trouble in the respect that she’s your friend. This might be challenged. Perhaps it is fitting for everyone to prefer, for your sake, that Sarah escape in that she’s your friend. After all, suppose that Stephen is Sam’s friend. A stranger to all parties might prefer that Sarah escape in that she’s your friend while also preferring that Stephen escape in that he’s Sam’s friend. These preferences might seem fitting.

24  Our solution bears some resemblance to Zimmerman’s (2011). Zimmerman suggests that we distinguish between the states of affairs that it is fitting to value. To simplify, he suggests that it is fitting for everyone to be indifferent between the states of affairs of Sarah escaping (S1) and Stephen escaping (S2). But it is not fitting for you to be indifferent between the states of affairs of Sarah, your friend, escaping (S1*) and Stephen, a stranger to you, escaping (S2*). He then suggests that the fitting-­ attitude account be restricted only to states of affairs like S1 and S2, and not those like S1* and S2*. However, Zimmerman’s claims about fittingness do not seem plausible. Suppose you are initially indifferent about who escapes, but then remember that Sarah is your friend—­perhaps your temporary amnesia lifts. You may then come to prefer that Sarah escapes. It seems that what has happened here is that you have acquired a preference between states of affairs—­S1 and S2—­between which you were previously indifferent. And this seems appropriate. Furthermore, it is hard to understand what it means to be indifferent between S1 and S2 while preferring S1* to S2*. This combination of preferences seems inconsistent. By contrast, it is clearly consistent and plausibly fitting to prefer Sarah to escape in one respect but not in another.

Value and Idiosyncratic Fitting Attitudes  257 The buck-­passer faces a parallel issue. The buck-­passer says that Sarah’s being your friend is a reason for you, but not for everyone, to prefer that Sarah escape. But Sarah’s being your friend might seem a reason for everyone to prefer, for your sake, that she escape. In response, note, first, that these particular normative claims are not crucial. Our solution requires there to be some respects in which it is fitting for you, but not everyone, to prefer that Sarah escape. The buck-­passer requires there to be some respects which are reasons for you, but not everyone, to prefer that Sarah escapes. The structure of the solutions requires there to be respects that play these roles. But it doesn’t specify which respects play these roles. This point has an interesting upshot. Given natural further assumptions, both views imply that the respects which play these roles are ‘evaluatively empty’: they are not respects in which Sarah’s escaping is better, worse, or just as good as, Stephen’s escaping. To illustrate using our own view: if preferring X to Y in respect R is idiosyncratically fitting, then R is not a respect in which preferring X to Y is fitting for all agents (by the definition of idiosyncratic fittingness). Thus, R is not a respect in which X is better than Y (by FA2). Furthermore, if fitting pro tanto preferences and indifference exclude each other, R cannot be a respect in which it’s fitting for all agents to prefer Y to X or be indifferent between X and Y. Thus, R is not a respect in which Y is better than X or a respect in which X and Y are equally good.25 This may seem to exacerbate the problem. It might seem that respects in which things are better, or of equal value, can nonetheless be respects in which idio­syn­ crat­ic­al­ly fitting attitudes are possible. To stick with our example, Sarah’s being your friend might seem like a respect in which it’s better if she escapes. And if it is not, it might seem a respect in which it’s equally good if she escapes. Yet our solution implies that it is neither. Put most generally then, the problem for our solution is that it requires that in cases of partiality: (i) there is a respect R such that it’s idiosyncratically fitting for the partial agent to prefer the outcome in that respect, and (ii) R is not a respect in which the outcome is better, worse, or equally good. It’s not obvious that there are good candidates to fill this role.

25  Fitting pro tanto preferences and indifference exclude each other if at most one of preference, dispreference, and indifference (in a certain respect) can be fitting for a particular agent. This assumption might be questioned, but we will grant it for the sake of discussion; it plausibly holds in at least some cases of the sort we’re considering. The corresponding assumption for the buck-­passer is that a consideration can be a reason for at most one of preference, dispreference, and indifference. For useful discussion of assumptions of this sort, see Snedegar 2017: ch.2.

258 Fittingness However, we suggest that, properly understood, our illustrative suggestion—­ that Sarah is your friend—­is indeed a good candidate to play this role. ‘That Sarah is your friend’ is ambiguous between what we can call an indexical and a non-­ indexical respect. The indexical respect is that which you would express by saying ‘Sarah is my friend’. The non-­indexical respect is that Sarah is, say, Suzy’s friend (where you are Suzy). We suggest that it is fitting for you to prefer that Sarah escape in the first, indexical respect—­that Sarah is, as you would say, my friend. But it is not fitting for all agents to have this preference (indeed this preference is arguably not even available to agents other than you). By contrast, it may be fitting for you, and for all agents, to have the same attitude—­be that preference or indifference—­in the second, non-­indexical respect—­that Sarah is Suzy’s friend. Furthermore, it is plausible that Sarah’s being your friend, understood indexically, is, as we put it, evaluatively empty. Sarah’s being Suzy’s friend might be a respect in which it is better if Sarah escape, or one in which this is equally good. But Sarah’s being, as you would put it, my friend does not seem a respect in which it is better, worse, or equally good that Sarah escape. As we have emphasized, betterness simpliciter does not seem to be relativized to agents. It is thus hard to see how respects which are so relativized could contribute to betterness simpliciter. These claims commit us to de se propositions and to a role for the de se in some cases of fitting partial concern. That the de se can make this sort of difference is independently plausible—­for instance, that an event will happen to me is widely thought to make it fitting for me to care about it in a distinctive way.26 The idea also fits naturally with the thought that fitting partiality depends on who you are, how you are related to others, or where and when you are located.27 And while de se propositions are not uncontroversial, even their opponents recognize the need to capture the de se at some level; our claims could be reformulated to fit such views.28 There might seem a residual problem. Suppose it is granted that it is fitting for you, but not for everyone, to prefer Sarah to escape trouble in the respect that she’s (indexically) your friend. And suppose it is granted that this respect is evaluatively empty. Nonetheless, one might still be sceptical that all respects in which preference is idiosyncratically fitting are evaluatively empty. Suppose we thought that Sarah’s being Suzy’s friend is in fact a respect in which it’s equally good that 26  Nagel’s (1970) initial characterization of the distinction between (what came to be known as) agent-­relative and agent-­neutral reasons can be understood as marking the significance of the de se. See Pettit 1987 and Darwall 2021 for presentations of Nagel’s distinction along these lines. See Guillot and O’Brien forthcoming for a recent defence of the normative significance of the de se in response to a challenge from Setiya (2015). 27  See, e.g. Brandt 2020: 35. 28  See Ninan 2016 for a helpful discussion and references on the de se. Note that we do not claim—­ though it is perhaps an attractive conjecture—­that all respects in which preferences are idio­syn­crat­ic­ al­ly fitting have a de se or otherwise indexical (e.g. de nunc—­see our discussion of temporal partiality in §11.7) character.

Value and Idiosyncratic Fitting Attitudes  259 Sarah and Stephen escape.29 On our account, it must thus be fitting for all agents—­including you—­to be indifferent between Sarah and Stephen escaping, in this respect. This might seem wrong. After all, since you know you are Suzy, it might seem incoherent for you to both prefer that Sarah escape in that she’s your friend while also being indifferent in that she’s Suzy’s friend. And this might suggest that these attitudes can’t both be fitting.30 However, there is no incoherence here. Compare: it seems fine to prefer that Sarah escape in that she is your friend and be indifferent in that she is the friend of the tallest person in the room, even if you know that you are the tallest person in the room. Why? Because what matters is that she is your friend, not that she is the friend of the tallest person in the room.31 Similarly, what matters is that she is your friend rather than that she is Suzy’s friend, and your preferences can coherently track this. Any further intuition that it is fitting for you to prefer that Sarah escape in that she is Suzy’s friend, is, we suggest, captured by our claim that it is fitting for you to prefer that she escape in that she’s your friend. To sum up: we have argued that our illustrative claim—­that it is idio­syn­crat­ic­ al­ly fitting for you to prefer Sarah to escape in the respect that she’s your friend—­is defensible, so long as it is understood appropriately—­that is, indexically. Furthermore, understanding it this way makes it plausible that this respect is not one in which it is better, worse, or equally good if Sarah escape. More generally, respects of betterness or equal value are always such that the ­corresponding attitude in that respect is fitting for all agents; idiosyncratically fitting preferences are always preferences in other respects.

11.6.2  Overall Betterness and Neutral Agents We said that: (FA3) For X to be better than Y overall is for it to be fitting for a neutral agent to prefer X to Y overall, 29  We have been non-­committal about whether Sarah’s being Suzy’s friend is a respect in which it is better if Sarah escape or one in which it is equally good. Our solution is neutral on this—­as noted, it does not require any particular normative claims. However, one way to reconcile conflicting intuitions is to distinguish different respects: first, that Sarah is Suzy’s friend and Stephen is not Suzy’s friend; second, that Sarah is Suzy’s friend and Stephen is someone else’s (say, Sam’s) friend. Perhaps it is fitting for all agents to prefer that Sarah escape in the first respect and fitting for all agents to be indifferent in the second respect. 30  Compare our criticism of Zimmerman’s (2011) solution (n.24). But note that our claims here are importantly different from Zimmerman’s: he distinguishes objects of overall preference, we distinguish respects of pro tanto preference. We do not claim that it is coherent to prefer overall that Sarah, your friend, escape, and yet be indifferent about whether Sarah or Stephen escapes. 31  These points still apply if we ‘rigidify’ the description to ‘the actual tallest person in the room’; they do not turn on the fact that you might not have been the tallest person in the room.

260 Fittingness where a neutral agent is one for whom no pro tanto preferences between X and Y are idiosyncratically fitting. This appeal to neutral agents might raise concerns. For instance, in many cases there will be no guarantee that any actual neutral agent exists. Thus it seems that the account must appeal to merely possible agents—­but this raises questions about how to understand claims about what is fitting for merely possible agents to prefer (Reisner 2015). It also seems that these neutral agents might need to have some peculiar properties—­for instance, to account for temporal partiality (§11.7) they might have to be atemporal.32 One might doubt whether atemporal agents are possible.33 However, we suggest that the appeal to neutral agents can be understood without raising such difficulties. Roughly it can be understood in terms of the fittingness for all agents of preferences that take only certain respects—­those for which no attitude is idiosyncratically fitting—­into account. The appeal to neutral agents can thus be cashed out as an appeal to the fittingness of these attitudes. To see how this might work, consider that, just as you can prefer something in certain respects but not others, or not overall, you can prefer it in certain (non-­ singleton) sets of respects but not others, or not overall. These sets can include individual respects in which you disprefer it. For example, when choosing a holiday destination you prefer Devon in respect of convenience and in respect of cost, while you prefer Sicily in respect of food and of weather. We can ask which you prefer in respect of the set {convenience, cost, food}—that is, just taking the members of this set into account. It might be that you prefer Devon in respect of this set—­you take the convenience and cost to outweigh the food. But perhaps in respect of the set {convenience, cost, food, weather} you prefer Sicily—­the weather tips the balance in favour of Sicily. Now consider a pair (X, Y). There will be a set S of respects in which it is fitting for all agents to prefer X to Y, or fitting for all agents to prefer Y to X, or fitting for all agents to be indifferent between X and Y. (That is, each member of S satisfies this disjunction; it’s not that there is a particular disjunct they all satisfy.) Thus, S  excludes those respects of preference that are idiosyncratically fitting—­those such that different preferences are fitting for different agents. Then we can say that for X to be overall better than Y is for it to be fitting for all agents to prefer X to Y in respect of S. Thus: (FA4) For X to be better than Y overall is for it to be fitting for all agents to ­prefer X to Y, taking into account only those respects in which no attitude is idio­syn­crat­ic­al­ly fitting. 32  This might not be required if temporal partiality is a first-­person phenomenon, as Parfit (1984: 181ff) suggests. But it is not clear if Parfit is right. See Hare 2008. 33  Such difficulties are standard worries about views which appeal to ‘ideal agents’. This includes some norm-­attitude views (Suikkanen 2008, Kauppinen 2014). If the argument to follow succeeds, this is thus an advantage over such views.

Value and Idiosyncratic Fitting Attitudes  261 The idea is that overall betterness is betterness in the respects that do not make for idiosyncratic fitting preferences. In the case of Sarah and Stephen, S will include respects such as that Sarah’s interests are at stake and that Stephen’s interests are at stake. It will exclude respects such as that Sarah is (indexically) your friend. In this way we will get the result that it is overall equally good that Sarah escape and that Stephen escape. Taking into account only those respects of preference that are fitting for all agents, indifference is fitting. Thus we can cash out talk of neutral agents in terms of respects of preference which are non-­idiosyncratically fitting. Our account of overall betterness does not require any resources over and above what we require to account for betterness in respects. But we can still use talk of neutral agents as a convenient shorthand. We therefore treat FA3 and FA4 as equivalent.

11.7  Related Challenges The partiality problem is an instance of a more general problem. There are a range of cases in which it seems fitting to value things in a way that does not correspond to their value. In this section, we consider whether our solution to the partiality problem helps with these other cases, and if not, whether this gives the buck-­ passer any advantage. A first kind of case involves temporality. Suppose that you know that you either had a very painful operation in the recent past or are due to have a less painful one in the near future. Both operations induce amnesia: if you had the very painful operation, you cannot now remember it, and if you are to have the less painful operation, you will not later be able to remember it. Here it seems fitting to prefer, and the balance of reasons seems to support preferring, that you had the more painful operation in the past—­the one that is over. But the more painful op­er­ ation is worse; things would be better simpliciter if you were instead due to have the less painful operation.34 Our solution to the partiality problem handles this case too. You might prefer the less painful operation in respect of its being less painful while also preferring the more painful operation in respect of its being past. Both of these preferences seem fitting. But the latter preference is idiosyncratic. It is fitting for you now to prefer the more painful operation in that it is past.35 But that attitude was not fitting for you when it was not past. So the fact that the more painful operation is

34  This case is due to Parfit 1984. For discussion of such cases in connection with norm-­attitude accounts, see Oddie 2005, Heathwood 2008, Bykvist 2009, Schroeder 2010, Lemos 2011. 35  And note that indexicality seems essential here—­it is in respect of its being past, rather than of its being on, say, 8 September 2019, that preferring it is fitting.

262 Fittingness past is not a respect in which it is better than the less painful operation. By contrast, it is always fitting to prefer the less painful operation in respect of its being less painful. So its being less painful is a respect in which it is better. And while it seems fitting for you now to prefer overall that you had the more painful op­er­ ation, it seems fitting for a neutral agent to prefer that you have the less painful operation. So FA3 implies that the less painful operation is better. (Of course, it might be fitting for a stranger now to prefer, for your sake, that you had the more painful operation. But this preference is idiosyncratic, depending as it does on the stranger’s temporal location.) Other cases challenge fitting-­attitude accounts of specific values.36 In particular, there seem to be cases in which the fitting response to an object changes even though the object doesn’t. For example, it is fitting to be more amused by a joke when you first hear it than three weeks after you hear it. But the joke doesn’t change, so it doesn’t seem to become less funny. It is fitting to grieve a friend who has just died, but also fitting for grief to subside, although the death does not lose the features that make it sad. Fitting regret and anger seem to follow similar trajectories, at least sometimes, although the mistakes or wrongs they correspond to do not change. In such cases, it’s clear that the fitting responses change, but not so clear that the specific values do.37 These cases raise a host of interesting and difficult issues. Here we can only offer a few brief remarks about their bearing on the choice between fitting-­ attitude and buck-­passing accounts. Our central claim is that these cases differ significantly from the cases involving partiality and temporality, and so we should not expect a uniform account of them. First, the cases involving partiality and temporality involve fitting attitudes which run counter to value: they are cases in which it is fitting to prefer the worse (or the equally valuable). These cases do not. It is not, for instance, that it becomes fitting to disdain the joke or delight in your friend’s death. Rather, it is no longer fitting to be amused by the joke, or saddened by the death, in the way that you were. This means that our response to the partiality problem is inapplicable here. It would not be plausible to say that it is still fitting to be amused by the joke in some respects or to grieve your friend’s death in some respects. The problem is 36 Bykvist (2009) suggests that modal considerations also challenge norm-­attitude accounts of goodness. For example, whether it is fitting to regret an outcome depends not just on how bad it is but how ‘close’ better outcomes were—­thus it is fitting to regret the rain at the picnic more than failing to win the lottery. Such cases are not counterexamples to the norm-­attitude accounts of goodness stated here (Lemos 2011). They do raise issues for norm-­attitude accounts of specific values, of broadly the kind we discuss in this section. 37 On amusement, see D’Arms and Jacobson 2010. On grief, see Marušić 2018. On anger, see Callard 2017, Srinivisan 2018, Marušić 2020. For discussion in connection with norm-­attitude accounts, see Bykvist 2009, D’Arms and Jacobson 2010, and Lemos 2011. For helpful general discussion, see Na’aman 2021.

Value and Idiosyncratic Fitting Attitudes  263 precisely that this seems false. The buck-­passer’s solution also seems inadequate. It’s not that, once you have grieved, you still have a reason to grieve your friend’s death, but one that is outweighed. More plausibly, your reason to grieve has gone—­you have done what it called for. So neither account can explain the values here by appealing to norms which apply in all circumstances. Second, the cases involving partiality and temporality involve a value—­ betterness simpliciter—­which is not relativized to agents. The values in these cases are not so naturally understood in this way. Plausibly, an event can be sad for me but not for you, a joke can be funny in one context but not another, and an act may be initially regrettable but not now, years on. Similarly, tragic events, at least sometimes, get less sad over time. The non-­relativity of betterness simpliciter was an important part of the partiality (and temporality) problem—­it is what motivated the thought that this value should be accounted for by norms on attitudes which apply to all agents. Insofar as the values involved in these cases differ in this way, there is not the same pressure to analyse these values in terms of such norms.38 Plausible norm-­attitude accounts of specific values will thus require care in specifying the circumstances under which the norms on attitudes which constitute these values apply. Exactly how this is to be done will vary between the different specific values. The only way to proceed is by detailed considerations of these different values and the corresponding responses. This is not, in our view, a problem that casts doubt on the plausibility of the norm-­attitude approach. Rather, it’s a reflection of the complexity and subtlety of specific values and the corresponding norms, which in turn reflects the complex role that these values and the associated attitudes play in our lives. Note further that, among norm-­attitude accounts, the buck-­passer again has no clear advantage here. Fitting-­attitude accounts will say that specific values are constituted by the fittingness of valuing something under certain circumstances. Buck-­passers will say that specific values are constituted by reasons to value something under certain circumstances. We see no reason to think that it will be any more straightforward for buck-­passers to specify the relevant circumstances than for proponents of fitting-­attitude accounts.

11.8  Conclusion Norm-­attitude accounts of value face the partiality problem. Buck-­passers can offer an attractive solution to this problem. We have argued that the fitting-­attitude

38  To be clear, we are not claiming that, e.g. events get less sad in a way that directly correlates with it being fitting to be less sad. Perhaps it can cease to be fitting to grieve even when the outcome you grieved for remains sad. Our point is just to illustrate that an account of specific values was always going to be relativized to circumstances or time in a way that betterness simpliciter was not.

264 Fittingness account can offer a broadly analogous and equally attractive solution. And we have argued that other sorts of case where value and norms on valuing diverge pose similar challenges for both sorts of norm-­attitude account.39

References Berker, S. 2022. The Deontic, the Evaluative, and the Fitting. In: C. Howard and R. A. Rowland (eds) Fittingness: Essays in the Philosophy of Normativity. Oxford University Press. Blanshard, B. 1961. Reason and Goodness. Allen and Unwin. Brandt, J. 2020. Negative Partiality. Journal of Moral Philosophy, 17: 33–55. Brandt, R. B. 1979. A Theory of the Good and the Right. Prometheus Books. Bykvist, K. 2009. No Good Fit: Why the Fitting Attitude Analysis of Value Fails. Mind, 118: 1–30. Callard, A. 2017. The Reason to Be Angry Forever. In: M.  Cherry and O.  Flanagan (eds) The Moral Psychology of Anger. Rowman & Littlefield. Chang, R. 1997. (ed.) Incommensurability, Incomparability, and Practical Reason. Harvard University Press. D’arms, J. and Jacobson, D. 2006. Sensibility Theory and Projectivism. In: D.  Copp (ed.) The Oxford Handbook of Ethical Theory. Oxford University Press. D’arms, J. and Jacobson, D. 2010. Demystifying Sensibilities: Sentimental Values and the Instability of Affect. In: P. Goldie (ed.) The Oxford Handbook of Philosophy of Emotion. Oxford University Press. Darwall, S. 2021. Practical Reason and the Second-Person Standpoint. In: R Chang and Sylvan, K. (eds) The Routledge Handbook of Practical Reason. Routledge. Ewing, A. C. 1939. A Suggested Non-Naturalistic Analysis of Good. Mind, 48: 1–22. Ewing, A. C. 1948. The Definition of Good. Routledge & Kegan Paul. Firth, R. 1952. Ethical Absolutism and the Ideal Observer. Philosophy and Phenomenological Research, 12: 317–45. Guillot, M. and O’Brien, L. forthcoming. Self Matters. Ergo. Hare, C. 2008. A Puzzle about Other-Directed Time-Bias. Australasian Journal of Philosophy, 86: 269–77. Heathwood, C. 2008. Fitting Attitudes and Welfare. Oxford Studies in Metaethics, 3: 47–73. Howard, C. 2018. Fittingness. Philosophy Compass, 13: e12542. 39  We received helpful comments on earlier drafts of this material from Rachel Achs, Selim Berker, Chris Howard, Thomas Hurka, Alex King, Oded Na’aman, R.A. Rowland, and Philip Stratton-­Lake. The chapter draws on work from our book Getting Things Right; we are also grateful, therefore, to the many people who helped with the book.

Value and Idiosyncratic Fitting Attitudes  265 Howard, C. 2019. Fitting Love and Reasons for Loving. Oxford Studies in Normative Ethics, 9: 113–39. Kauppinen, A. 2014. Fittingness and Idealization. Ethics, 124: 572–88. Keller, S. 2013. Partiality. Princeton University Press. Korsgaard, C. M. 1996. The Sources of Normativity. Cambridge University Press. Lemos, N. 2011. Intrinsic Value and the Partiality Problem. Philosophy and Phenomenological Research, 82: 697–716. Lewis, D. 1989. Dispositional Theories of Value. Aristotelian Society Supplementary Volume, 63: 113–37. Lord, E. 2016. Justifying Partiality. Ethical Theory and Moral Practice, 19: 569–90. Maguire, B. 2018. There Are No Reasons for Affective Attitudes. Mind, 127: 779–805. Markovits, J. 2014. Moral Reason. Oxford University Press. Marušić, B. 2018. Do Reasons Expire? An Essay on Grief. Philosophers’ Imprint, 18. Marušić, B. 2020. Accommodation to Injustice. Oxford Studies in Metaethics, 15: 263–83. McDowell, J. 1985. Values and Secondary Qualities. In: Honderich, T. (ed.) Morality and Objectivity. Routledge. McHugh, C. and Way, J. 2022. Getting Things Right: Fittingness, Reasons, and Value. Oxford University Press. Mill, J. S. 1863. Utilitarianism. London, Parker, Son, and Bourn. Moore, G. E. 1903. Principia Ethica. Dover Publications. Na’aman, O. 2021. The Rationality of Emotional Change: Toward a Process View. Noûs, 55: 245–69. Nagel, T. 1970. The Possibility of Altruism. Oxford Clarendon Press. Ninan, D. 2016. What Is the Problem of De Se Attitudes? In: M. Garcia-Carpintero and S.  Torre (eds) About Oneself: De Se Attitudes and Communication. Oxford University Press. Oddie, G. 2005. Value, Reality, and Desire. Clarendon Press. Olson, J. 2009. Fitting Attitude Analyses of Value and the Partiality Challenge. Ethical Theory and Moral Practice, 12: 365–78. Orsi, F. 2013. Fitting Attitudes and Solitary Goods. Mind, 122: 687–98. Parfit, D. 1984. Reasons and Persons. Oxford University Press. Parfit, D. 2011. On What Matters. Oxford University Press. Pettit, P. 1987. Universality without Utilitarianism. Mind, 72: 74–82. Railton, P. 1986. Facts and Values. Philosophical Topics, 14: 5–31. Reisner, A. 2015. Fittingness, Value and Trans-World Attitudes. Philosophical Quarterly, 65: 464–85. Ridge, M. 2017. Reasons for Action: Agent-Neutral vs. Agent-Relative. In: E. N. Zalta (ed.) Stanford Encyclopedia of Philosophy. Autumn 2017 edition. Available at: https://plato.stanford.edu/archives/fall2017/entries/reasons-agent/.

266 Fittingness Rosen, G. 2015. The Alethic Conception of Moral Responsibility. In: R.  Clarke, M. McKenna, and A. Smith (eds) The Nature of Moral Responsibility: New Essays. Oxford University Press. Ross, D. 1939. Foundations of Ethics. Oxford University Press. Scanlon, T. 1998. What We Owe to Each Other. Belknap Press of Harvard University Press. Schroeder, M. 2010. Value and the Right Kind of Reason. Oxford Studies in Metaethics, 5: 25–55. Schroeder, M. 2012. The Ubiquity of State-Given Reasons. Ethics, 122: 457–88. Setiya, K. 2015. Selfish Reasons. Ergo, 2: 445–72. Smith, M. 1994. The Moral Problem. Blackwell. Snedegar, J. 2017. Contrastive Reasons. Oxford University Press. Srinivasan, A. 2018. The Aptness of Anger. Journal of Political Philosophy, 26: 123–44. Street, S. 2008. Constructivism about Reasons. Oxford Studies in Metaethics, 3: 207–45. Suikkanen, J. 2008. Consequentialism, Constraints, and Good-Relative-To. Journal of Ethics and Social Philosophy, 3: 1–9. Tappolet, C. 2016. Emotions, Value, and Agency. Oxford University Press. Wiggins, D. 1987. A Sensible Subjectivism? In: Needs, Values, Truth: Essays in the Philosophy of Value. Oxford University Press. Zimmerman, M. J. 2007. The Good and the Right. Utilitas, 19: 326–53. Zimmerman, M. J. 2011. Partiality and Intrinsic Value. Mind, 120: 447–83.

12 Well-­Being as Fitting Happiness Mauro Rossi and Christine Tappolet

12.1 Introduction There is an intuitive connection between well-­being and happiness, so much so that these terms are often taken to be synonymous. This is not how we will use them here. In this chapter, we will take ‘well-­being’ to be an evaluative term, referring to a specific kind of value, i.e. prudential value, while we will take ‘happiness’ to be a psychological term, referring to a favourable mental state or a combination of favourable mental states.1 Accordingly, when we say that an individual has a high well-­being, we mean that that individual’s life is going well for them from a prudential point of view. Instead, when we say that an individual is happy, we mean that they are in a positive mental state, or else a combination of positive mental states. Given this use of the terms, the relation between well-­being and happiness is, if there is any, a substantive one. Even so, many well-­being theorists have argued that well-­being and happiness are intimately connected. There exist different kinds of happiness-­based theories in the literature. These accounts differ with respect to two main dimensions. One concerns the theory of happiness on which they are based. Historically, the most popular theory of happiness is hedonism.2 According to it, happiness consists in a positive balance of pleasures and displeasures. An alternative theory of happiness that has made inroads amongst psychologists and economists over the past forty years, and that has eventually reached philosophers, is life satisfactionism.3 According to it, happiness consists in an attitude of satisfaction towards one’s life.

1  We would like to thank the participants to the SIFA 2018 Conference, at the University of Eastern Piedmont, the participants to the workshop on ‘Well-­being and Affective States’, at the Université Clermont Auvergne, the participants to the workshop on ‘Perspectives on Well-­Being’, at the Centre de Recherche en Éthique in Montreal, and the participants to the ‘Fittingness Conference’ for their helpful comments. We are especially grateful to Selim Berker for his generous and useful comments on a previous draft of this chapter. Many thanks also to an anonymous referee for their very helpful suggestions. 2 Early supporters are Bentham (1789/1961) and Mill (1863/1998). Contemporary supporters include Kahneman (1999); Feldman (2010); Bramble (2016). 3  Supporters of life satisfactionism include, in psychology, Diener and Lukas (2009); in economics, Anand (2009); in philosophy, Sumner (1996) and Suikkanen (2011). Mauro Rossi and Christine Tappolet, Well-­Being as Fitting Happiness In: Fittingness: Essays in the Philosophy of Normativity. Edited by: Christopher Howard and R. A. Rowland, Oxford University Press. © Mauro Rossi and Christine Tappolet 2022. DOI: 10.1093/oso/9780192895882.003.0012

268 Fittingness The second dimension over which happiness-­based theories of well-­being differ concerns how they conceive of the relation between well-­being and happiness. One group of theories holds that well-­being consists in happiness tout court. According to these theories, an individual’s life goes well for them to the extent that, and because, the individual is happy, however happiness is conceived. The second group holds that well-­being consists in properly qualified happiness, e.g. authentic happiness, deserved happiness, and so on. According to these theories, an individual’s life goes well for them to the extent that, and because, the individual is happy, provided their happiness satisfies some additional condition, such as authenticity, desert, and so on. For ease of exposition, we will call the theories in the first group standard happiness-­based theories of well-­being; and the theories in the second group hybrid happiness-­based theories of well-­being. Despite their variety, happiness-­based theories of well-­being have gradually fallen out of favour. Standard theories have been attacked for being either descriptively inadequate, i.e. for not matching our paradigmatic judgements about well-­ being, or normatively inadequate, i.e. for not being normatively useful. These objections have provided the motivation for hybrid happiness-­based theories. Yet, the latter have also been challenged. The main objection raised against them is that they are ad hoc, i.e. they combine different elements (for instance, happiness and authenticity) in a non-­independently justified way. These objections have led many scholars to exploring alternative theories of well-­ being.4 The main competitors include preference satisfactionism,5 perfectionism,6 and objective list theories of well-­being.7 One aspect that these theories have in common, which is relevant for the present chapter, is that they all conceive of the relation between well-­being and happiness as being looser than what happiness-­ based theories assume. According to the typical preference satisfaction theories, for instance, happiness contributes to the individual’s well-­being only if it is the object of the individual’s (actual or rational) preferences. According to standard perfectionist theories, instead, happiness is, at best, a psychological effect of the development and exercise of the distinctively human capacities which an individual’s well-­being consists in. For their part, objective list theories typ­ic­al­ly assign happiness a bigger role, for they normally include happiness within the list of items that constitute well-­being. But even on such theories, happiness is often regarded as merely sufficient and not as necessary for well-­being—­the idea being that an individual’s life can go well for them even if the individual is unhappy, provided that their life abounds of most of the other items in the list.

4  For an overview of contemporary theories of well-­being, see Fletcher (2016b). 5  For an overview of preference satisfactionist theories, see Bykvist (2016). 6  For an overview of perfectionist theories, see Bradford (2016). 7  For an overview of objective list theories, see Fletcher (2016a).

Well-Being as Fitting Happiness  269 This situation raises the question of whether there is any happiness-­based theory that can be successfully defended and that can provide a plausible alternative to the previous theories. In this chapter, we want to offer a positive answer to this question by presenting a new happiness-­based theory of well-­being. For reasons of space, we cannot compare our theory with all the non-­happiness-­based theories of well-­being currently on offer. Our goal in this chapter is more modest. We will argue that our theory is immune from all the main objections raised against ­traditional (standard and hybrid) happiness-­based theories. To the extent that happiness-­based theories can be considered as the intuitive starting point in the elaboration of a theory of well-­being, this is sufficient for raising the status of our theory to that of a serious competitor. Our theory presents two main differences with respect to traditional happiness-­ based theories of well-­being. First, it is based on a theory of happiness, i.e. an affective theory of happiness, which differs from both hedonism and life satisfactionism. Second, while our theory holds that well-­being consists in a qualified form of happiness, it qualifies happiness in a different way than traditional hybrid theories. More specifically, it holds that well-­being consists in fitting happiness. For this reason, we call it the fitting happiness theory of well-­being.8 Our theory can be seen as the combination of the following claims. The first is that happiness consists in a broadly positive balance of affective states such as emotions, moods, and sensory pleasures. The second is that emotions, moods, and sensory pleasures are different kinds of perceptual experiences of evaluative properties. It follows from this that happiness consists in a broadly positive balance of perceptual experiences of evaluative properties. The third claim is that, insofar as happiness is constituted by states that have fittingness conditions, it is possible to assess happiness itself as fitting or unfitting. In particular, happiness is fitting if and only if its constituents are fitting, that is, when their objects possess the evaluative properties they are represented to have. The last claim is a claim about well-­being: it says that well-­being consists in fitting happiness thus defined. Accordingly, our theory holds that fitting happiness is the most basic prudential good. It follows that any other item is prudentially good for the individual if and only if, and because, it contributes to the individual’s fitting happiness or makes the individual fittingly happy. We proceed as follows. In the first part of this chapter, we present our theory of well-­being by discussing in more detail each of the previous claims (Section 12.2). In the second part, we argue that our theory does better than both standard and hybrid happiness-­based theories of well-­being (Section  12.3). As we have seen,

8  It is important to note that, while the notion of fittingness occupies a central place in con­tem­por­ ary meta-­ethical debates, for instance, in debates concerning the analysis or elucidation of evaluative concepts, we use it here to elaborate a first-­order theory of well-­being, that is, a theory of what makes an individual’s life go well for that individual.

270 Fittingness our theory counts as a hybrid theory, since it holds that well-­being consists in a qualified form of happiness, i.e. fitting happiness. We argue that our theory can capture the main insights of the competing hybrid happiness-­based theories in an independently justified way (Section  12.3.1). Next, we argue that our theory is immune from the main objections that plague standard and hybrid happiness-­based theories of well-­being. We consider five objections: the ‘shallowness objection’ (Section 12.3.2), the ‘lack of value objection’ (Section 12.3.3), the ‘inauthenticity objection’ (Section 12.3.4), the ‘passivity objection’ (Section 12.3.5), and the ‘ill-­being objection’ (Section 12.3.6).

12.2  The Fitting Happiness Theory of Well-­Being In this section, we begin by presenting our theory of psychological happiness. For the most part, we will simply illustrate the main tenets of our theory, without providing a detailed defence. This is partly for reasons of space and partly because we have offered such a defence elsewhere.9 Our aim, in this section, is to bring together the different threads so as to present a complete picture in the hope of conveying its attractiveness.

12.2.1  Psychological Happiness Let us start with some preliminaries. When interpreted as a psychological term, ‘happiness’ is sometimes used to refer to a single mental state, such as an emotion or a mood, and sometimes to a combination of mental states. While we do not wish to deny that the term ‘happiness’ can legitimately be used in the former sense, in this chapter we will take ‘happiness’ to refer to a combination of mental states. We think that this is indeed the best way to do justice to the role that happiness plays in deliberation. For example, when we deliberate whether to choose a career as a musician or as a surgeon, and we choose the former because we think it will make us happier, it seems that we are not just thinking of happiness as an emotion or a mood, but as a more encompassing psychological state, one that involves a plurality of mental states. In this sense, providing an account of the nature of happiness requires specifying which mental states happiness consists of, or, equivalently, which mental states are happiness-constituting. Our theory of happiness belongs to a family of theories that has recently emerged as an alternative to hedonism and life satisfactionism, namely, emotional state theories of happiness. The most popular of these is certainly the theory put

9  See (Rossi and Tappolet Manuscript-­b).

Well-Being as Fitting Happiness  271 forward by Daniel Haybron (2008). To motivate our account, it is thus useful to compare it to Haybron’s. According to Haybron, happiness consists in a broadly positive balance of occurrent affective states such as emotions and moods, on the one hand, and purely dispositional states such as mood propensities, on the other. To clarify the terminology: occurrent states are mental episodes that possess a specific phenomenology; by contrast, purely dispositional states are dispositions to experience occurrent states, with no phenomenology over and beyond that of their manifestations. According to Haybron, happiness is determined both by the intensity and by the centrality of these happiness-­constituting states. By intensity, he means phenomenal intensity, that is, the degree to which a state ‘feels’ a certain way. By centrality, Haybron at least primarily means the disposition to cause other affective states or actions. Haybron’s theory aims at offering an account of happiness that, in addition to matching our paradigmatic intuitions, can do justice to the role that happiness plays in the explanation of people’s behaviour and in people’s practical de­lib­er­ation, and that can be useful for normative theorizing. By making centrality a defining feature of happiness, Haybron is able to show how we can predict and explain behaviour by appealing to happiness ascriptions and how happiness can function as an important consideration in normative decision-­making and normative theorizing. We agree with Haybron’s affective approach and with his emphasis on the notion of centrality. However, we think that Haybron’s theory faces some problems, which we have discussed elsewhere (see Rossi  2018, Rossi and Tappolet Manuscript-­a). For present purposes, consider just the following. According to Haybron’s account, both occurrent and dispositional affective states are de­ter­min­ ants of happiness in their own right. One implication is that his account classifies as very happy an individual who has strong positive mood propensities, but who, for some reasons (e.g. because they are asleep, in a coma, or simply in un­favour­able circumstances that prevent their mood propensities to manifest), has no positive occurrent affective states, i.e. no positive emotions or moods. This is counter-­ intuitive. In fact, happiness seems to have an essential phenomenal dimension. The lesson is the following: we need an account that preserves Haybron’s insight about the importance of centrality for happiness, but that can also do justice to happiness’s phenomenology. Here is where our theory comes in. According to our theory, happiness consists in a broadly positive balance of occurrent affective states such as emotions, moods, and sensory pleasures. On our account, positive emotions and moods, as well as sensory pleasures, are happiness-­ constituting states; negative emotions and moods, as well as sensory displeasures, are unhappiness-­constituting states. (Note that, for simplicity, in what follows we will only focus on happiness.) We agree with Haybron that intensity and centrality are two important dimensions of the happiness-­constituting states, which are relevant for determining how happy an individual is. However, we characterize the notion of centrality differently from Haybron.

272 Fittingness Consider this. In general, when we desire happiness, we desire to experience not a series of positive yet disconnected and ephemeral states, but states that have solid bases and some sort of internal coherence. When do the happiness-­ constituting states have these properties? In our view, this is the case when the happiness-­constituting states are based on and reflect the individual’s values.10 For terminological clarity, let us distinguish between what the individual values and the individual’s attitude of valuing these things. What matters for our purpose are the individual’s valuings. Valuings are stable mental states, which play a significant role in practical deliberation. They have robust dispositional properties, in the sense that they dispose the individual to experience a variety of affective states related to the valued objects. In addition, they provide a sort of ‘thematic unity’ to these states. The most important valuings of an individual, i.e. the individuals’ ‘core valuings’, are closely tied to the individual’s own identity.11 Considering this, we can reformulate our previous point by saying that, when we desire happiness, we typically desire to experience states that are based on and reflect our valuings. We are happier, the more our happiness is so based. With this in place, we can distinguish two relevant senses of ‘centrality’. On the one hand, there is what we will call ‘output-­centrality’. This is the extent to which a state is causally productive of other affective states and behaviours. It cor­res­ ponds to Haybron’s understanding of centrality. On the other hand, there is what we will call ‘source-­centrality’. This is the extent to which a state is based on other output-­central states. Valuings can be seen as the most output-­central states in an  individual’s mind. They are indeed the attitudes that most fundamentally guide and motivate our behaviour and define our identity. As for the happiness-­ constituting states, they can be either output-­central or source-­central, or both. That said, if what we have suggested above is correct, what matters for happiness is primarily whether, and to what extent, these states are based on and reflect the individual’s valuings. That is, what matters for happiness is whether its constituent states are source-­central. On our account, then, happiness is a function of the phenomenal intensity and the source-­centrality of the happiness-­constituting states. This idea distinguishes our theory from Haybron’s. We think that our account is perfectly able to do justice to the role that happiness plays in the explanation of  people’s behaviour and in practical deliberation. Insofar as the happiness-­ constituting states are based on and reflect dispositionally robust states like valuings, we can indeed derive a range of predictions about how an individual will act and explanations about their past behaviour from ascriptions of individual happiness. Similarly, we can explain the importance of happiness for practical deliberation 10  See Kauppinen (2013) for the similar suggestion that the degree to which an affective state makes a difference to the happiness of an individual is a function of the degree to which it defines the individual’s perspective. 11  For this characterization of valuings, see, amongst others, Seidman (2009), Scheffler (2010), Svavarsdóttir (2014).

Well-Being as Fitting Happiness  273 by referring to the connection between the happiness-­constituting states and the individual’s valuings. The more happiness is based on the latter states, the more important it will typically be for the deliberating individual. Unlike Haybron’s account, moreover, our theory preserves the intuition that happiness has an inherent phenomenological dimension, since it is constituted by occurrent affective states. By so doing, it avoids the counter-­intuitive implications of Haybron’s theory.

12.2.2  The Nature of the Happiness-­Constituting States Let us move to our next claim, concerning the happiness-­constituting states. This is the most distinctive claim underlying our theory. We can ask the following question to introduce our account. What do emotions, moods, and sensory pleasures have in common, in virtue of which they count as happiness-­constituting states? Our response is that they are all kinds of affective evaluations, that is, evaluations with a phenomenal dimension. Take emotions for a start. Consider an emotional episode such as the fear you experience at the sight of a dog. The dog is the intentional object of your fear, in the sense that your fear is about (i.e. it is directed at) the dog. But your fear is not a merely neutral attitude towards the dog. Rather, your fear is a way of presenting the dog to you as fearsome. Being fearsome is an evaluative property. Thus, by presenting the dog as fearsome, your fear provides you with an evaluation of the dog. Moreover, it provides an evalu­ ation that is distinctively ‘felt’, since fear typically comes with a distinctive phe­ nom­en­ology. The property of being fearsome is often thought to be what identifies all the episodes of fear as belonging to one and the same emotion type. All these episodes are instances of fear because they present their intentional objects as being fearsome. The fearsome is typically called the ‘formal object’ of fear. The notion of a formal object is generally considered to be useful to distinguish between different types of emotions.12 Accordingly, all episodes of joy are instances of the same emotion type, i.e. joy, in virtue of the fact that they present their intentional objects as having the same evaluative property, namely, the joyful. All episodes of anger are instances of the same emotion type, i.e. anger, in virtue of the fact that they present their intentional objects as being offensive. And so on. Generalizing, we can say that emotions are phenomenologically salient ways of presenting their intentional objects as possessing different evaluative properties. As such, different emotion types count as different types of affective evaluations. The claim that emotions are affective evaluations is compatible with different theories of emotions. The theory that we favour, and that one of us has defended at length elsewhere (Tappolet 2000, 2016), is the perceptual theory of emotions.

12  See Teroni (2007).

274 Fittingness According to it, emotions are perceptual experiences of evaluative properties. More specifically, they are perceptual experiences that (non-­conceptually)13 represent their intentional objects as possessing specific evaluative properties. Thus, for instance, an episode of fear consists in a perceptual experience that represents its object as being fearsome, an episode of joy consists in a perceptual experience that represents its object as being joyful, and so on. As stated above, we think that moods and sensory (dis)pleasures are also af­fect­ive evaluations. Like emotions, they are perceptual experiences of evaluative properties. However, they differ from emotions in some important respects. Consider a mood such as apprehension. When you are apprehensive, you have an experience of looming danger. A bit more technically, you have an experience of fearsomeness being instantiated. In this, apprehension is not distinct from fear. We can express this idea by saying that both apprehension and fear are perceptual experiences of the same evaluative property, namely, the fearsome. Unlike fear, however, apprehension does not seem to be directed at anything in particular. This points to a more general feature of moods. Unlike emotions, moods do not seem to have specific intentional objects. This idea has been elaborated in different ways in the literature. For instance, one of us has argued that moods have undetermined objects, that is, objects that the individual is unable to identify. Combined with a perceptual theory, this leads to the view that moods are perceptual experiences that represent undetermined objects as possessing specific evaluative properties (Rossi 2021). Other views are nonetheless possible. For instance, moods can be conceived as having generalized (e.g. Solomon  1976/1993, Prinz 2004), plural (Siemer  2009), probable (Price  2006), or modal (Tappolet  2018) intentional objects. For present purposes, the only point we are committed to is that, while moods are a kind of perceptual experience of evaluative properties, like emotions, they differ from emotions at the level of their intentional object, which is non-­specific. Similar considerations apply to sensory pleasures (and to sensory displeasures). By sensory pleasures, we mean any sensory experience that possesses the property of being pleasant, that is, any pleasant sensory experience. Thus, to understand what sensory pleasures are, one needs to understand, first, what sensory experiences are and, second, what pleasantness and unpleasantness are. As an example, consider the pleasure of tasting coffee in the morning. One of us has defended the following evaluativist account (Rossi Manuscript). First, a sensory experience consists in a perceptual experience that represents an object (e.g. the coffee) as possessing some sensory qualities (e.g. a particular taste). Let us summarize this by saying that a sensory experience consists in a perceptual experience that represents a ‘sensory object’. Next, a sensory experience being pleasant 13  For the notion of non-­conceptual content, see Tappolet (2020). For simplicity, we will omit this qualification in what follows.

Well-Being as Fitting Happiness  275 consists in its additionally representing the sensory object as pleasurable. Note that, on this account, pleasantness and pleasurableness are two distinct properties. Pleasantness is a phenomenological property of the sensory experience, whereas pleasurableness is an evaluative property that the sensory object is represented to  have (and that can genuinely possess if the representation if veridical—­see Section 12.2.3 for more on this). More specifically, pleasurableness is a determinable evaluative property, that is, a property that has different determinate specifications. This means that there are different ways for a sensory object to be pleasurable. For instance, the taste of coffee can be exquisite, delicious, comforting, and so on. One implication is that, while the class of sensory pleasures is unified by the fact that all sensory pleasures have the pleasurable as their formal object, different types of sensory pleasures can be distinguished in terms of the particular way of being pleasurable that their tokens represent. That said, the main point to retain here is that, insofar as sensory pleasures consist in perceptual experiences that represent sensory objects as pleasurable (in a determinate way), and insofar as being pleasurable is an evaluative property, then sensory pleasures too count as perceptual experiences of evaluative properties. In this, they are like emotions and moods. They differ from emotions and moods in two respects. The first is that they always have a sensory object as their intentional object. The second is that, arguably, not all positive emotions and moods represent evaluative properties that are determinate specifications of pleasurableness. Before moving to our next claim, let us say a few words about valuings, as these are relevant for determining the source-­centrality of the happiness-­constituting states. Historically, the most popular accounts have construed valuings either as a form of first-­order or second-­order desires (e.g. to value x is to desire x; or to value x is to desire to desire x) or as a form of value judgement (e.g. to value x is to judge that x is valuable). However, both accounts are subject to serious objections (see e.g. Seidman  2009, Scheffler  2010, Svavarsdóttir  2014). As a result, some scholars have moved to a caring account of valuings, according to which to value an item x is to care about it (e.g. Seidman 2009). We think that this is a step in the right direction, but it is not entirely satisfactory yet. Care is a mental state that belongs to the class of sentiments. The class of sentiments includes, however, other mental states, such as love, hatred, jealousy, and so on. What unifies this class is the fact that all these mental states are multi-­track emotional dispositions, that is, dispositions to experience a variety of emotions related to the intentional object of the sentiment. What matters for present purpose is that all sentiments, not just care, possess the characteristics that are typically associated with valuings. Elsewhere, we have thus proposed to take the notion of valuings to refer to the class of sentiments, rather than simply to a single mental state type such as care, and to consider sentiments as different types of valuings (Rossi and Tappolet Manuscript-­a).

276 Fittingness As dispositional states, sentiments have no phenomenological dimension of their own, over and beyond the phenomenology of the occurrent affective states that they dispose an individual to experience. They have, however, a representational content akin to that of occurrent affective states. In particular, they have an intentional object, which is (non-­conceptually) represented as possessing a particular evaluative property. For instance, when you love your partner, your love has an intentional object, i.e. your partner, which is represented as possessing a particular evaluative property, i.e. the property of being lovable. This type of representational content distinguishes sentiments from simple affective dispositions that do not represent a particular object as having a specific evaluative property. Sentiments are standing affective states that have representational contents of the form ‘x is F’, and that, when activated, generate occurrent states. As a non-­affective example, consider your standing belief that this chapter is entitled ‘Well-­Being as Fitting Happiness’, which we have just activated by writing this line.

12.2.3  Fitting Happiness Let us take stock. We have claimed that happiness consists in a broadly positive balance of affective states such as emotions, moods, and sensory pleasures, where the balance depends, amongst other things, on the relation between these states and the individual’s sentiments. Second, we have claimed that emotions, moods, and sensory pleasures are kinds of perceptual experiences of evaluative properties, and that sentiments are a kind of standing evaluative states. It follows from these claims that happiness consists in a broadly positive balance of perceptual experiences of evaluative properties, which is based on the individual’s standing evaluations. This account has an immediate implication for our understanding of happiness. Happiness can be conceived of as a composite state that affectively informs us about the values (and disvalues) that we encounter in our life. In this sense, happiness counts as a global affective evaluation, which takes into account the individual’s more general evaluative stance. We can now move on to our third claim. As we have seen, emotions, moods, and sensory pleasures are perceptual experiences that represent their intentional objects as possessing specific evaluative properties. As kinds of perceptual experiences, they can be assessed as fitting or unfitting. By fittingness, we mean the same as representational correctness.14 Accordingly, emotions, moods, and sensory pleasures are fitting if and only if the world is as they represent it to be. In other words, emotions, moods, and sensory pleasures are fitting if and only if their objects really possess the evaluative properties that these states represent

14  For criticisms of this understanding of fittingness, see Svavarsdóttir (2014) and Howard (2018).

Well-Being as Fitting Happiness  277 them to possess. Consider, as an example, an episode of fear of a dog that is attacking you. As we have seen, this episode consists in a perceptual experience that represents the dog as fearsome. On our account, such an episode of fear is fitting if and only if there really is a dog attacking you (rather than, e.g. a hologram of a dog) and the dog is really fearsome (rather than, e.g. an innocuous puppy). Similar considerations apply to moods and sensory pleasures. We should add that, on our theory, sentiments have fittingness conditions too. Consider, for instance, a sentiment such as love for a particular person. Arguably, this sentiment can also be assessed as fitting or unfitting. It is fitting provided that such a person genuinely is worthy of love or, equivalently, ‘lovable’. It is unfitting otherwise. Our theory holds that, insofar as happiness is constituted by states that have fittingness conditions, it is possible to assess happiness too as fitting or unfitting. The first condition for happiness to be fitting is that the states that constitute it are fitting, that is, that their intentional objects do possess the evaluative properties that they represent them to possess. Since these states are typically based on sentiments, which play a role in determining the extent to which an individual is happy, another condition must be added for happiness to be fitting: happiness must be based on fitting sentiments. This account has two implications. The first is that happiness can be more or less fitting, depending on the extent to which its constituents and their standing bases are fitting. The second implication is that, insofar as fitting happiness consists in a broadly positive balance of fitting perceptual experiences of evaluative properties, which are based on fitting standing evaluations, then fitting happiness provides the individual, at the same time, with (a) a broadly positive experience of items that are genuinely valuable, and (b) a correct ‘global’ evaluation of their life (at a given time or in an interval of time).

12.2.4  Well-­Being as Fitting Happiness This leads to our final claim. This is simply that fitting happiness is what well-­ being consists in. We can sum up our theory more precisely by reference to three questions that any theory of well-­being must address. First, which items are non-­ instrumentally good for an individual? Second, what makes these items non-­ instrumentally good for the individual? Third, what dimensions of these items are relevant for determining an individual’s degree of well-­being? Our theory holds that the basic prudential good is fitting happiness. This is what well-­being ultimately consists in. Fitting happiness constitutes well-­being in virtue of its being an affective experience of the good. This is what makes it non-­ instrumentally good for an individual. As for the variables that determine the degree of an individual’s well-­being, these are the phenomenal intensity of the fitting happiness-­constituting states and their source-­centrality. In light of what we have said before, it follows that the extent to which an individual’s life goes

278 Fittingness well for the individual living that life is determined by the degree to which that individual affectively experiences genuinely valuable items and by the degree to which these experiences are based on, and reflect, the individual’s standing af­fect­ ive evaluations. One complication should be noted. In some cases, an individual’s happiness-­ constituting states may be fitting at the level of their objects, i.e. insofar as they represent as valuable objects that are really valuable, but not at the level of their intensity, i.e. insofar as they represent as valuable to a given degree objects that are valuable to a different degree. Similarly, in some other cases, an individual’s happiness-­constituting states may be fitting at the level of their objects and intensity, but they may be based on sentiments that are either unfitting or only partially fitting. For instance, this may happen when an individual correctly experiences joy at an event x, but the individual’s joy derives from valuing items of the class X, where not all the members of this class are joyful. In all these cases, the individual’s happiness-­constituting states are only partially fitting. This introduces some difficulties in the estimation of the overall degree of fitting happiness, which we do not have space to explore in this chapter, but which need to be addressed.

12.3  The Case in Favour of the Fitting Happiness Theory of Well-­Being To provide a full defence of our theory of well-­being, we would have to compare it to all the main competitors in the literature. This task clearly exceeds what is possible to do in one chapter. So, in what follows, we adopt another strategy. We start from the intuition that there exists a close connection between well-­being and happiness. We take this intuition to offer a prima facie reason to take happiness-­ based theories of well-­being as the default starting point in the analysis of well-­ being. Next, we argue that our own happiness-­based theory is immune to all the objections raised against competing, standard and hybrid, happiness-­based the­ or­ies of well-­being. In fact, it also presents some important advantages over them. If we are right, then our theory emerges as a plausible account of well-­being, one that deserves close consideration.

12.3.1 Advantages We begin by discussing some advantages that our theory has over alternative happiness-­based theories of well-­being. As we have seen, our theory holds that well-­being goes beyond unqualified happiness and consists in fitting happiness. As such, our theory counts as a hybrid happiness-­based theory of well-­being. In  the introduction, we suggested that hybrid happiness-­based theories face an

Well-Being as Fitting Happiness  279 explanatory challenge. They must provide a convincing, and non-­ad hoc, ex­plan­ ation of why happiness needs to be qualified in a particular way for an individual’s life to go well. As we will now illustrate, we think that our theory does better than its competitors in this respect. Consider the theories according to which happiness must be deserved or authentic to benefit an individual. One may argue that these qualifications are required to make the resulting theory of well-­being descriptively and/or normatively adequate. In other words, these qualifications are required to reach re­flect­ive equilibrium. We do agree that some of the intuitions that motivate these qualifications are very powerful. Below, we will ourselves consider an objection against happiness-­based theories that appeals to the importance of authenticity. That said, it seems to us that the explanatory challenge remains. Qualifying happiness as these theories do can accommodate the data, by better fitting our judgements about particular cases. But can it also explain the data? It seems to us that our theory brings an additional layer to the explanation. Recall that, on our account, happiness consists in a broadly positive affective balance of states that are directed and aim at specific evaluative properties, that is, at determinate specifications of the good. Insofar as happiness is a combination of these states, we can say that happiness is a composite state that aims at the good. As we have seen, happiness can achieve its aim to various degrees. We expressed this idea by saying that happiness can be fitting to various degrees. Considering this, we can say that fitting happiness is happiness that fully achieves its aim. This provides the basis for our account of well-­being. Insofar as we tie well-­being to happiness and we think of happiness as aiming at the good, then it is only natural to think that an individual’s life goes well for them when, and to the extent, that their happiness fully achieves its aim. Thus, the qualification ‘fitting’ that is attached to happiness is not an unrelated element, which is added just to accommodate the data. Rather, this qualification is a way of providing a complete ex­plan­ation of what makes happiness a determinant of well-­being.

12.3.2  The Shallowness Objection In the remainder of this chapter, we explain how our theory avoids the main objections against traditional happiness-­based theories of well-­being. The first objection says that happiness is either too shallow or too ephemeral a psychological state for it to be the ultimate constituent of an individual’s well-­ being. For simplicity, let us call this the ‘shallowness objection’. This objection targets standard and hybrid happiness-­based theories of well-­being alike. The shallowness objection is a challenge both to the descriptive and the normative adequacy of these theories. On the descriptive side, suppose that an individual experiences no displeasure and a number of very superficial and transient pleasures,

280 Fittingness e.g. the pleasure of eating a cracker, the pleasure of a warm shower, etc. Happiness-­ based theories should conclude that the individual’s life goes very well for them. However, the individual’s happiness seems to be too superficial to really benefit them. On the normative side, it appears that, if happiness is shallow or ephemeral, then it cannot play the appropriate normative role. Such an account would, indeed, fail to vindicate the role of happiness in normative theorizing and in ordinary instances of practical deliberation. This is a genuine objection against some happiness-­based theories of well-­ being. For instance, it threatens well-­being theories that adopt a hedonist theory of happiness according to which happiness consists in a positive balance of pleasures and displeasures, and according to which pleasures and displeasures consist in mere qualia or phenomenal properties. If one adopts this theory, then happiness turns out to be a positive balance of mere feelings. It is then justified to attack the resulting theory of well-­being on grounds of descriptive and normative ad­equacy, since mere feelings are both shallow, insofar as they lack depth, and empty, insofar as they lack content. However, the objection does not apply to our own theory of happiness. We conceive the happiness-­constituting states not as mere ‘feelings’, but as states possessing an important cognitive dimension. Through this dimension, they provide the individual experiencing them with information about evaluatively significant circumstances. That is, happiness provides a global affective evaluation of the individual’s situation. This feature of our theory allows us to reject the charge that happiness is empty. In addition, we hold that the degree of an individual’s happiness depends not just on the phenomenal intensity of the happiness-­constituting states, but also on their (source-)centrality. As we have seen, this means that an individual’s happiness is greater, the more its constituents are based on, and reflect, the individual’s valuings. This allows us to reject the charge that happiness is a shallow state.

12.3.3  The Lack of Value Objection The second objection, which we call the ‘lack of value objection’, states that happiness cannot be a constituent of well-­being, because one may derive happiness from items that lack any genuine value. They may lack value either because they are pointless activities, such as when an individual derives happiness from counting the blades of grass in their lawn (Rawls 1971); or because they are genuinely evil activities, such as when an individual derives happiness from torturing a living being; or because, although the activities appear to have value, their value is not actually realized, such as when an individual derives happiness from activities in an ‘experience machine’ (Nozick 1974). In all these cases, the objection goes, happiness does not benefit the individual.

Well-Being as Fitting Happiness  281 This objection targets happiness-­based theories according to which happiness matters for well-­being independently of the value of its sources. According to these theories, if an individual derives happiness from counting the blades of grass or from other pointless activities, it is still the case that their life goes well for them. This is deemed to be a counter-­intuitive implication. There are at least two possible replies open to traditional happiness-­based the­ or­ies. One is to bite the bullet and to say that happiness benefits the individual in virtue of its pleasant character, independently of the value of its sources. A second reply consists in opting for a hybrid happiness-­based theory according to which well-­being consists in value-­based happiness. For instance, a hedonist about happiness may say that what matters for well-­being are only pleasures that are based on valuable items or activities. The problem is that resorting to either option exposes one to the explanatory challenge that we raised in subsection 12.3.1. None of this applies to our theory. As we have seen, our theory holds that happiness contributes to well-­being only when it is fitting. In turn, fitting happiness consists in a broadly positive global affective experience of genuine values. If so, our theory implies that only happiness derived from valuable items matters from well-­being. In this sense, our theory is a form of value-­based happiness theory of well-­being. But unlike other versions, such as the value-­based hedonist theory of well-­being, our theory can be independently motivated, as explained in subsection 12.3.1.

12.3.4  The Inauthenticity Objection The third objection that we consider, which we call the ‘inauthenticity objection’, states that well-­being cannot be based on happiness because happiness may be inauthentic. For example, it may result from adaptation or brainwashing. Consider an individual who, as an effect of perverse conditioning, experiences joy when facing degrading conditions. The individual’s joy does not seem to genuinely benefit them. That is to say that, when happiness is inauthentic, it does not seem to contribute to an individual’s well-­being. This is a serious objection, which has led several authors to adopting hybrid happiness-­based theories according to which well-­being consists in authentic happiness (see Sumner 1996). At first sight, this objection threatens our theory as well. The reason is that happiness may be inauthentic even when it is based on valuable items. So, it seems that fitting happiness is not sufficient for well-­being, since fitting happiness may lack authenticity. One difficulty in discussing this objection has to do with the notion of authenticity. There are many different accounts in the literature and, depending on the account that one adopts, the response to the objection will proceed differently. On a plausible reading, however, happiness is authentic if and only if it is based

282 Fittingness on items that the individual genuinely values. We will assume this account of authentic happiness here. The question for us, then, is whether the notion of fitting happiness is broad enough to capture the relation between an individual’s occurrent happiness and the individual’s most genuine valuings, which is necessary for assessing whether the individual’s happiness is authentic or not. The first thing to note is that, on our view, happiness does depend on the individual’s valuings. As we have seen, happiness depends, amongst other things, on the extent to which emotions, moods, and sensory pleasures are based on, and reflect, the individual’s valuings. The second thing to note is that, on our view, well-­being requires not just that the happiness-­constituting states are fitting, so that the individual derives happiness from genuinely valuable items, but also that the happiness-­constituting states are based on fitting valuings. If valuings are sentiments, this implies that well-­being requires, amongst other things, that the individual’s happiness is based on fitting sentiments. At this stage, one may object that our reply does not really solve the problem of inauthenticity. True, our theory holds that well-­being depends, amongst other things, on the individual’s fitting valuings. But valuings can be fitting, yet in­authen­tic. For example, one may love something that is really good such as music as a result of perverse conditioning. So, if authenticity is necessary for well-­being, then fitting happiness falls short from providing a satisfactory account of well-­being. Once we have reached this stage, however, we are willing to bite the bullet. That is, we are willing to recognize that fitting happiness may be inauthentic, while insisting that it is nevertheless sufficient for well-­being. Our idea is that if an individual affectively experiences genuinely valuable items, if their experiences are based on what the individual values, and if the individual’s valuings are fitting, then the individual’s life goes well for them, even if the individual has acquired their fitting valuings through an improper process. Two considerations may help bite this bullet. The first is that, insofar as the process of value acquisition is improper, this is likely to have an impact on the individual’s affective states. This means that inauthenticity is likely to involve some costs for the individual’s happiness and, thus, for their well-­being, at some point in time. The second con­sid­er­ ation is that, while authenticity is not necessary for well-­being, it does of course matter for morality and justice. This means that, although (e.g.) brainwashing may not reduce an individual’s well-­being, it is nevertheless an impermissible and unjust act.

12.3.5  The Passivity Objection The fourth objection states that happiness-­based theories are too ‘passive’, since they allow for the possibility that an individual may derive well-­being from happiness without exercising their capacity for agency. At the extreme, an individual’s

Well-Being as Fitting Happiness  283 life can go well for them from the merely passive enjoyment of pleasurable ex­peri­ ences, without any form of active engagement from the individual’s part. This is deemed to be counter-­intuitive. We call this ‘the passivity objection’. Our theory acknowledges the important role that agency plays for well-­being. In our theory, agency intervenes at three different levels. First, exercises of agency, e.g. activities, achievements, project realizations, are often the sources of happiness-­constituting states, such as emotions, moods, and sensory pleasures, which contribute to well-­being when fitting. There is more. Some forms of agency are necessary for certain evaluative properties to be genuinely instantiated or, equivalently as we use the terms, for some values to be realized. Think about flow (Csikszentmihalyi 1975). Flow is a state one enters in while performing an activity that requires the exercise of skills. The activity in question is perceived as one that represents a challenge that matches the skills of the individual. Because flows shares all the characteristics of emotions, there is good reason to hold that flow is in fact an emotion (see Tappolet 2022). Indeed, flow appears to be a kind of enjoyment, one that we take when engaged in activities that we perceive as challenges that match our skills. On this account, flow has an activity as its intentional object. Thus, flow requires agency. As a positive affective state, flow can be fitting or unfitting. Granted the assumption that flow is a kind of enjoyment, it follows that flow is fitting when the activity is worthy of enjoyment. More precisely, flow is fitting when the activity is enjoyable in a specific way, that is, as a challenge that matches one’s skills. When this is the case, flow enhances the individual’s well-­being. Combining all this, we get the conclusion that engagement in certain activities is necessary for some values to be realized—­in this case, the value associated with, and represented by, flow—­and, thereby, for the individual’s affective experiences to enhance their well-­being. The third level at which agency intervenes is at the level of valuings. Activities, projects, achievements are, of course, possible objects of an individual’s valuings. It is likely that the individual will experience various occurrent affective states that are based on these valuings. Thus, agency-­based goods will also play a role in determining the source-­centrality of these happiness-­constituting states. Like traditional happiness-­based theories, our theory does not exclude the extreme case where an individual derives well-­being from happiness without exercising their capacity for agency. But, we think, this is how it should be. It seems implausible to hold that well-­being always requires activities or exercises of agency. We think that a theory that recognizes the importance of agency, while allowing for the possibility that well-­being does not consist solely in activities or exercises of agency is more descriptively adequate. In fact, this position has some advantages over the alternative. To give just one example, our theory can do just­ ice to the claim, often made in disability studies, that even individuals with significantly impaired capacities for agency can have a good quality of life.

284 Fittingness

12.3.6  The Ill-­Being Objection Let us consider one last objection. According to it, our theory implies that if unhappiness is wholly unfitting, then an individual’s life does not go bad for them at all. Surely, however, this is incorrect, for unhappiness has a negative phe­nom­ en­ology and its negative phenomenology makes unhappiness bad for the individual experiencing it, even when their unhappiness is otherwise unfitting. Thus, our theory seems to deliver the wrong verdict about the individual’s ill-­being. This objection, which we call ‘the ill-­being objection’, targets all theories according to which ill-­being is determined by qualified unhappiness, such that the qualification functions as a discount factor, which reduces—­and potentially even annuls—­unhappiness’s impact on ill-­being.15 For ease of the discussion, we reformulate this objection by reference to the unhappiness-­constituting states. The idea underlying the objection is that negative affective states, whether they are emotions, moods or sensory displeasures, have a negative phenomenology, which makes them bad for the individual, even when they are unfitting. Our strategy consists in pointing out the similarity between the present objection and an objection that has been put forward against evaluativist theories of unpleasant pain in philosophy of mind, an objection known, alternatively, as the ‘shooting-­the-­messenger objection’ or as the ‘normativity objection’. We will focus on what we take to be the most promising reply that has been offered in this context and see how it applies to the ill-­being objection. Evaluativist theories of unpleasant pain conceive of unpleasant pain as an interoceptive experience that represents a bodily condition as possessing a negative evaluative property. For instance, according to David Bain, ‘[a] subject’s being in unpleasant pain consists in his (i) undergoing an experience (the pain) that represents a disturbance of a certain sort, and (ii) that same experience additionally representing the disturbance as bad for him in the bodily sense’ (Bain 2013: 82). In other words, Bain thinks that unpleasant pain consists in an interoceptive experience that represents a bodily damage as bad for the individual. Thus, according to Bain, the evaluative property that is represented in instances of unpleasant pain is the property of being bad for the individual.16 15  Note that a similar objection could be formulated with respect to well-­being. The idea is that happiness seems to be good for an individual even when unfitting, in virtue of its pleasant character. This contrasts with what our theory of well-­being seems to imply. There are two possible strategies to address this objection. One consists in arguing that well-­being and ill-­being are asymmetric, and that the present objection applies only with respect to ill-­being. For instance, this may be because unpleasantness has a different prudential status than pleasantness. The second strategy consists in holding that the solutions offered to the ill-­being objection can be applied, mutatis mutandis, to the symmetric well-­being objection. We think the latter strategy to be preferable. In what follows, however, we will focus only on the ill-­being objection. 16 Note that, earlier on, we have offered as slightly different theory of sensory displeasures, according to which the relevant evaluative property that is represented in such instances is the

Well-Being as Fitting Happiness  285 Here is the objection that evaluativist theories of unpleasant pain such as Bain’s face. On these theories, unpleasant pain has the function of providing information about the badness of a bodily event for the individual. However, the information provided by an episode of unpleasant pain might be non-­veridical. For instance, in cases of phantom limb experiences, where an individual feels pain in an amputated limb as if it were still attached to their body, unpleasant pain represents a bodily damage as bad for the individual, yet there is really no bodily damage there, since the limb where the individual feels pain has been amputated. So, a fortiori, there is no bodily damage that is bad for the individual. In our ter­min­ ology, such an unpleasant pain is unfitting, because it has an incorrect representational content. This creates the following problem. On this view, unpleasant pain is like a messenger of bad news. But in some cases, the bad news turns out to be false. Now, if unpleasant pain is just a messenger, and if messengers are simply neutral vehicles of information, it follows that when the bad news is false, there is nothing bad for the individual that takes place. However, this conclusion seems to ignore the fact that the individual still experiences an unpleasant pain. Even when the information provided by an unpleasant pain is non-­veridical, the individual’s pain remains unpleasant. Surely, this is still bad for the individual. If not, why would one take a painkiller? The problem is that evaluativist theories of unpleasant pain seem unable to account for this feature. On their accounts, taking a painkiller is the equivalent of shooting the messenger. The similarity between the ill-­being objection and the shooting-­the-­messenger objection is evident. The ill-­being objection states that unhappiness is bad for the individual even when unfitting, because it feels bad. The shooting-­the-­messenger objection states that an unpleasant pain is bad for the individual even when unfitting, because it feels unpleasant (which is a way of feeling bad). Considering this, we can hope that some of the strategies explored to address the latter objection can help address the ill-­being objection as well. There are three main strategies that have been developed to address this objection.17 We consider these strategies in detail in Rossi and Tappolet (Manuscript-­c). For reasons of space, here we will only focus on the strategy that we consider most plausible. Before doing that, we will briefly mention the two other strategies. The first is instrumentalism (Martínez 2015). This strategy accepts the implication that, if an unpleasant pain is unfitting, then it is not bad for the individual in itself. However, it adds that an unpleasant pain may be instrumentally bad for the individual, in the sense that it may have other non-­instrumentally bad consequences for the individual. The second strategy—­perceptualism (Bain 2017)—holds that, while some evaluative representations of a bodily damage are not bad for the property of being unpleasurable. This difference aside, our theory and Bain’s belong to the same family of evaluativist theories. 17  For other strategies, see Bain (2017).

286 Fittingness individual in themselves, such evaluative representations are non-­instrumentally bad for the individual when they appear in a perceptual mode, that is, when they are perceptual evaluative representations of a bodily damage. This means, for instance, that, although the judgement that a bodily damage is bad for you is not itself non-­instrumentally bad for you, the interoceptive experience of the bodily damage as bad for you, which unpleasant pain consists in, is itself non-­ instrumentally bad for you. Crucially, this is true even when the unpleasant pain is unfitting. Both solutions can be generalized so as to apply to our theory. However, we think that they fail to offer a complete story. For this reason, we will move to considering a third possible reply to the shooting-­the-­messenger objection, put forward by Paul Boswell (2016). In our opinion, this offers, when generalized, the most plausible reply to the ill-­being objection. The key notion is that of ‘secondary unpleasantness’. As we have seen, according to evaluativism, a pain being unpleasant consists in a representation of a bodily damage as bad for the individual. Such a representation determines what may be called the pain’s primary unpleasantness, or unpleasantness1. Based on the empirical literature on pain, Boswell claims that some pains also possess a kind of secondary unpleasantness, or unpleasantness2. Secondary unpleasantness is intentionally attached to the unpleasant1 pain itself, in the sense that it consists in a representation of the unpleasant1 pain as bad for the individual. It can be identified as the unpleasantness characteristic of suffering pains. Three things must be noticed. First, according to this account, secondary unpleasantness is part of the pain experience itself. It is not the unpleasantness of a separate, second-­order affective state caused by the pain experience. This distinguishes Boswell’s strategy from instrumentalism. Second, not all unpleasant pains have this kind of secondary unpleasantness, since not all unpleasant pains are suffering pains. Third, the representation of the unpleasant1 pain as bad for the individual, which secondary unpleasantness consists in, is typically a fitting representation, since an unpleasant1 pain is typically bad for the individual in virtue of its unpleasantness1. The combination of the second and third points implies that some pains (i.e. suffering pains) can be non-­instrumentally bad for the individual even if the bodily damage they represent is not bad for the individual. This existential qualifier distinguishes Boswell’s strategy from perceptualism. This strategy can be generalized to our theory. As we have seen, negative emotions, moods, and sensory displeasures are unfitting when they involve incorrect representations of their objects as possessing specific evaluative properties. However, some of these negative emotions, moods, and sensory displeasures may possess the kind of secondary unpleasantness characteristic of suffering pains. Note that, within our theory, unpleasantness2 consists in a representation of the unpleasant state to which it is attached as unpleasurable, rather than as bad for the individual. This representation can be correct, since negative emotions, moods,

Well-Being as Fitting Happiness  287 and sensory displeasures can indeed be unpleasurable, even when they are unfitting. If this is true, then there is a sense in which unfitting negative emotions, moods, and sensory displeasures can legitimately contribute to fitting unhappiness and, thereby, to ill-­being, namely, when these states have a fitting suffering character. In these circumstances, unfitting negative emotions, moods, and sensory ­displeasures contribute to fitting unhappiness and to ill-­being in virtue of their fitting suffering character. The upshot is the following. If we generalize Boswell’s solution to the shooting-­ the-­messenger objection, then we can hold that some unfitting negative emotions, moods, and sensory displeasures are non-­instrumentally bad for the individual in virtue of their phenomenology. They are thus ill-­being-increasing states. This is not true of all unfitting negative emotions, moods, and sensory displeasures, but only of suffering emotions, moods, and sensory displeasures. It seems to us that this is consistent with our considered intuitions. Thus, although the third strategy vindicates the badness for the individual of only some unfitting unhappiness-­ constituting states, this is a plausible feature of, rather than a problem for, our theory.18

References Anand, Paul. 2009. Happiness Explained: What Human Flourishing Is and What We Can Do to Promote It. Oxford: Oxford University Press. Bain, David. 2013. ‘What Makes Pains Unpleasant?’ Philosophical Studies 166 (1): 69–89. Bain, David. 2017. ‘Evaluativist Accounts of Pain’s Unpleasantness.’ In The Routledge Handbook of the Philosophy of Pain, edited by Jennifer Corns, 40–50. New York: Routledge.

18  In footnote 15, we claimed, on the one hand, that an objection similar to the ill-­being objection can be formulated in terms of well-­being, and, on the other hand, that the solutions offered to the ill-­ being objection can be applied, mutatis mutandis, to the symmetric well-­being objection. We can now briefly explain how. Suppose, as it is plausible, that there exists a kind of pleasantness2, which is the equivalent for positive affective states of what unpleasantness2 is for negative affective states. Pleasantness2 is the kind of pleasantness characteristic of enjoyed positive emotions, moods, and sensory pleasures. On our theory, pleasantness2 consists in a representation of the pleasant1 affective state to which it is attached as pleasurable. As with unpleasantness2, this representation can be correct even when positive emotions, moods, and sensory pleasures are otherwise unfitting, since these states can nevertheless be genuinely pleasurable. If this is true, then there is one respect in which unfitting positive emotions, moods, and sensory pleasures can legitimately contribute to well-­being, namely, in virtue of their fitting enjoyed character. This shows that, under certain conditions, a life lived in an  ‘experience machine’ (Nozick 1974) can be well-­being-­enhancing. To the extent that life in the machine contains fittingly enjoyed affective states, then it will contain some prudential goodness. However, insofar as this enjoyed affective states are otherwise unfitting, life in the machine will not be as prudentially good as an experientially identical life outside the machine. This is consistent with, and indeed explains, the intuitions underlying the experience machine objection.

288 Fittingness Bentham, Jeremy. 1789/1961. An Introduction to the Principles of Morals and Legislation. Garden City: Doubleday. Boswell, Paul. 2016. ‘Making Sense of Unpleasantness: Evaluationism and Shooting the Messenger.’ Philosophical Studies 173 (11): 2969–92. Bradford, Glen. 2016. ‘Perfectionism.’ In The Routledge Handbook of Philosophy of Well-Being, edited by Guy Fletcher, 124–34. London: Routledge. Bramble, Ben. 2016. ‘A New Defense of Hedonism about Well-Being.’ Ergo 3 (4): 85–112. Bykvist, Krister. 2016. ‘Preference-Based View of Well-Being.’ In The Oxford Handbook of Well-Being and Public Policy, edited by Matthew D. Adler and Marc Fleurbaey, 321–46. Oxford: Oxford University Press. Csikszentmihalyi, Mihaly. 1975. Beyond Boredom and Anxiety. San Francisco: JosseyBass Publishers. Diener, Ed, and Richard  E.  Lukas. 2009. ‘Subjective Well-Being: The Science of Happiness and Life Satisfaction.’ In Oxford Handbook of Positive Psychology, edited by S. J. Lopez and C. R. Snyder, 63–73. New York: Oxford University Press. Feldman, Fred. 2010. What Is This Thing Called Happiness? Oxford: Oxford University Press. Fletcher, Guy. 2016a. ‘Objective List Theories.’ In The Routledge Handbook of Philosophy of Well-Being, edited by Guy Fletcher, 148–60. London: Routledge. Fletcher, Guy. 2016b. The Philosophy of Well-Being: An Introduction. New York: Routledge. Haybron, Daniel M. 2008. The Pursuit of Unhappiness: The Elusive Psychology of WellBeing. New York: Oxford University Press. Howard, Chris. 2018. ‘Fittingness.’ Philosophy Compass 12: e12542. Kahneman, Daniel. 1999. ‘Objective Happiness.’ In Well-Being: Foundations of Hedonic Psychology, edited by Daniel Kahneman, Ed Diener, and Norbert Schwarz. New York: Russell Sage Foundation. Kauppinen, Antti. 2013. ‘Meaning and Happiness.’ Philosophical Topics, 41 (1): 161–85. Martínez, Manolo. 2015. ‘Pains as Reasons.’ Philosophical Studies 172: 2261–74. Mill, John Stuart. 1863/1998. Utilitarianism. New York: Oxford University Press. Nozick, Robert. 1974. Anarchy, State, and Utopia. New York: Basic Books. Price, Carolyn. 2006. ‘Affect without Object: Moods and Objectless Emotions.’ European Journal of Analytic Philosophy 2 (1): 49–68. Prinz, Jesse  J. 2004. Gut Reactions: A Perceptual Theory of the Emotions. New York: Oxford University Press. Rawls, John. 1971. A Theory of Justice: Justice within a Liberal Society. Cambridge, Massachussetts, Cambridge: Belknap Press. Rossi, Mauro. 2018. ‘Happiness, Pleasures, and Emotions.’ Philosophical Psychology 31 (6): 898–919. Rossi, Mauro. 2021. ‘A Perceptual Theory of Moods.’ Synthese 198 (8): 7119–47.

Well-Being as Fitting Happiness  289 Rossi, Mauro. Manuscript. ‘An Evaluativist Theory of Pleasure.’ Rossi, Mauro, and Christine Tappolet. Manuscript-a. An Affective Theory of Happiness. Rossi, Mauro, and Christine Tappolet. Manuscript-b. ‘Happiness as an Affective Evaluation.’ Rossi, Mauro, and Christine Tappolet. Manuscript-c. ‘Ill-Being and Fitting Unhappiness.’ Scheffler, Samuel. 2010. Equality and Tradition: Questions of Value in Moral and Political Theory. New York: Oxford University Press. Seidman, Jeffrey. 2009. ‘Valuing and Caring.’ Theoria 75 (4): 272–303. Siemer, M. 2009. ‘Mood Experience: Implications of a Dispositional Theory of Moods.’ Emotion Review 1 (3): 256–63. Solomon, Robert  C. 1976/1993. The Passions: Emotions and the Meaning of Life. Indianapolis, Indiana: Hackett Publishing Company. Suikkanen, Jussi. 2011. ‘An Improved Whole Life Satisfaction Theory of Happiness.’ International Journal of Wellbeing 1 (1): 1–18. Sumner, L. W. 1996. Welfare, Happiness, and Ethics. Oxford: Oxford University Press. Svavarsdóttir, Sigrún. 2014. ‘Having Value and Being Worth Valuing.’ Journal of Philosophy 111 (2): 84–109. Tappolet, Christine. 2000. Émotions et valeurs. Paris: Presses Universitaires de France. Tappolet, Christine. 2016. Emotions, Values, and Agency. Oxford: Oxford University Press. Tappolet, Christine. 2018. ‘The Metaphysics of Moods.’ In The Ontology of Emotions, edited by Hichem Naar and Fabrice Teroni, 169–86. Cambridge: Cambridge University Press. Tappolet, Christine. 2020. ‘Emotions Inside Out: The Nonconceptual Content of Emotions.’ In Concepts in Thought, Action, and Emotion: New Essays, edited by Christoph Demmerling and Dirk Schroeder, 257–76. New York: Routledge. Tappolet, Christine. 2022. ‘Sailing, Flow and Happiness.’ In The Philosophy of Sailing: The Sailing Mind, edited by Robert Casati, Cham: Springer, 17–29. Cham: Springer. Teroni, Fabrice. 2007. ‘Emotions and Formal Objects.’ Dialectica 61 (3): 395–415.

13 The Things We Envy Fitting Envy and Human Goodness Sara Protasi

13.1 Introduction Thick and abundant hair; unblemished skin; physical strength; being quick-­ witted; a fast sports car; a lush garden; a doting and supportive partner; well-­ behaved and academically successful children; a reputation on social media; a remunerative job.1 These are only a few of the things we envy. A comprehensive list would be extremely long, perhaps infinite. For something to be an object of envy, the envier has to think of it as both valuable and as something they lack: we envy others for things that we care about but that we don’t have. There are all sorts of objects that fit such a subjective description. Thus, it seems that, in principle, anything could be envied. But not everything is objectively worthy of envy. In fact, one may argue that nothing is truly enviable, or, in other words, that envy is never an apt response to the perceived lack of a good. The idea that envy is systematically misguided or necessarily inappropriate is fairly popular in the history of ideas, both within and without academia, and takes various forms. In popular sayings, envy is presented as irrational (‘envy ­slayeth the silly ones’), maliciously counterproductive (‘envy spoils the good it ­covets’), self-­defeating even, since it ‘slays itself by its own arrows’ and ‘shoots at others, but hits itself ’. In this chapter I focus on a particular version of this critique: the idea that envy is never fitting. Envy’s complex nature makes the analysis of its fittingness conditions complicated, but here I’m particularly interested in one component: the 1  For incisive and constructive criticism on a previous version of this chapter I’m grateful to all the participants to ‘Fit Fest’, and in particular to Rachel Achs, Selim Berker, Louise Hanson, Tom Hurka, Stephanie Leary, Brian McElwee, Oded Na’aman, Mauro Rossi, Philip Stratton-­Lake, and Christine Tappolet. I’m also thankful to Stephen Campbell and Oded Na’aman for in-­depth conversations and precious advice about how to reframe my argument; to Justin D’Arms for clarifications about his and Daniel Jacobson’s view; to Tyler Doggett for helpful feedback on a prior version of this chapter (cf. ‘Conclusion’ in Protasi 2021) and Kariah Phillips for sharing her professional counselling insights on this topic. Finally, thanks to Christopher Howard and Richard Rowland for inviting me to contribute to this volume and for their encouraging and detailed feedback on my contribution. Sara Protasi, The Things We Envy: Fitting Envy and Human Goodness In: Fittingness: Essays in the Philosophy of Normativity. Edited by: Christopher Howard and R. A. Rowland, Oxford University Press. © Sara Protasi 2022. DOI: 10.1093/oso/9780192895882.003.0013

The Things We Envy: Fitting Envy and Human Goodness  291 envied object. If the envied object is never authentically good, then envy is necessarily unfitting. I’ll clarify and expand on this concept momentarily, but let me first emphasize that it differs from another kind of popular critique: the idea that envy is always immoral and imprudent. Many people think that envy proper always motivates the envier to act badly. The sayings above incorporate some of that critique: envy ‘shoots at others’ and ‘spoils the good’. The critique I intend to challenge precedes worries about envy’s alleged moral or prudential badness, and is, rather, about envy’s core concern: relative positioning. The gist of such critique is that authentic goodness doesn’t depend on any sort of  comparison. This is a popular view in everyday conversations about envy (‘you  should focus on what you have, not what others have!’), but can also be found in the philosophical tradition. I defend the opposite view: authentic goodness very often depends on some sort of comparison, and thus, not only can envy be fitting, but furthermore its core concern is systematically connected with our flourishing as humans. Here is a preview of the chapter. I start by presenting my original taxonomy of envy and show how it expands our understanding of the many ways in which envy can be fitting or unfitting (section 13.2). Then, I introduce the position that envy is always unfitting insofar as it’s concerned with the inauthentic good of relative ­positioning, a position exemplified by Immanuel Kant (section 13.3), and contrast it with Justin D’Arms and Daniel Jacobson’s view which is that envy can be fitting insofar as it is an evolutionarily shaped response to a deep and wide human concern, viz. relative positioning (section 13.4). However, I argue that D’Arms and Jacobson don’t go far enough in their defence of fitting envy. First, I expand on their analysis of positional goodness by distinguishing between two claims: an epistemic one, according to which we use implicit or explicit comparison to know what position we occupy in this continuum of goodness, and thus to form judgements of goodness (section 13.5), and a metaphysical one, according to which much human goodness depends on implicit or explicit rankings and positionality (section  13.6). Second, I show that we can be a lot more sanguine about the role of envy in our ethical lives than D’Arms and Jacobson appear to be. They are hesitant to attribute a positive practical role to fitting envy and suggest instead that it might be appropriate to repress it. Against this view, I argue that fitting envy is not only intrinsically valuable qua fitting response to authentic goodness, but can be epistemically, morally, and prudentially valuable (section 13.7). Fitting envy plays a crucial role in safeguarding our well-­being and, in some varieties, even our flourishing as moral agents.

13.2  The Multifarious Nature of Envy, and its Fittingness Conditions In this section I provide a sketch of an account of envy I defend elsewhere (Protasi 2016, 2021), which is supported by extensive empirical evidence and also

292 Fittingness foreshadowed in Aristotle’s Rhetoric and Hume’s Treatise of Human Nature, among other philosophical works. Envy is usually understood as a three-­place relation, involving the subject who feels the emotion (envier); the person towards whom the emotion is directed (envied); and the object that is envied (the good). It’s important to keep in mind the relational and comparative nature of envy (that it’s always targeted at and vis-­ à-­vis another person) so as to distinguish it from a mere desire or wish to have something one doesn’t have (whereby there is no relation to another person) and from coveting something possessed by another (whereby the other person doesn’t constitute a standard of comparison, and which lacks the painful self-­reflective element typical of envy). So, here is a more substantive definition that helps distinguish envy from similar attitudes: envy is an aversive response to a perceived inferiority or disadvantage vis-­à-­vis a similar other, with regard to a good that is relevant to the sense of identity of the envier.2 Let me unpack this definition by using a concrete example. Suppose I confess feeling envy for my sister’s unblemished skin. (That I don’t actually have a sister makes my confession a lot easier.) By confessing this envy, I thereby communicate to you that I’m feeling some sort of pain (my response is affectively aversive), which is brought about by the perception that someone whom I see as similar to me has a superior skin texture; my envy also tells you that I see my own skin as comparatively tarnished; finally, you can also infer that having unblemished skin, or perhaps more generally an unblemished exterior, is something that I care about, or, as some psychologists put it, a ‘self-­relevant’ goal or domain (Miceli and Castelfranchi 2007: 454–6). As mentioned earlier, there’s robust empirical evidence supporting each component of this definition: social psychologists agree that any kind of envy is painful or at least unpleasant to feel; that we only feel it towards someone whom we perceive as similar to us in some respect, someone who’s in the same comparison class as we, and whom we perceive as being in a superior position of some sort; and, finally, that we envy people only for things that we perceive to be good under certain descriptions and that we see as relevant and important to us (Miceli and Castelfranchi 2007; Smith and Kim 2007; Lange and Crusius 2015 provide detailed reviews). 2  I can’t address the many objections that usually arise when presenting such a definition to a philo­soph­ic­al audience; so many different elements whetting the appetite for an endless search for counterexamples! Plenty of scenarios can be concocted to show that this account doesn’t capture all possible cases of envy—­some quite plausible! However, I talk of envy as a concrete culturally specific (think: contemporary ‘Western’ industrialized societies) emotional phenomenon that affects the average person. People vary in their emotional repertoire. When anthropologists say that envy is a universal emotion, that claim isn’t falsified by the fact that a few people never feel envy. Similarly, when psychologists talk about the factors that I’m about to explain, that is compatible with idiosyncratic experiences or the existence of cases that aren’t fully captured by these conditions.

The Things We Envy: Fitting Envy and Human Goodness  293 Naturally, the empirical evidence concerning these conditions isn’t always nuanced. When it comes to similarity between the envied and the envier, for instance, psychologists usually refer to simple demographic factors like age and gender. But philosophers (Aristotle, Francis Bacon, Hume) also discuss this aspect of envy in their analyses, and include similarity based on kinship, spatiotemporal closeness, occupation, and caring about the same things. Sometimes we envy people with regard to general goods (happiness, health, luck in love), or specific but almost universally desired ones (such as a vaccine for COVID-­19)— in these cases, the envied are similar to us qua fellow human beings, and the domain of self-­relevance is very broad (most of us care about being happy and healthy). Even when similarity is broadly construed, however, envy is likely to be more intense if the envied other is similar to us in some respect, such as being ‘near [ . . . ] in time and place and age and reputation’ (Aristotle 2007: 145), as Aristotle remarks in the Rhetoric (Rhet. II.10). In sum, the majority of those who have thought and written about envy, whether philosophers or psychologists, in ancient or contemporary times, agree on what this emotion is. The disagreement begins when it comes to what it mo­tiv­ ates us to do and whether it’s necessarily immoral. The traditional and popular understanding of envy is that of a malicious (thus immoral) emotion, which pushes the agent to bring down the envied and spoil the envied good. This notion, however, has come under severe empirical critique in the last fifteen years: many studies have shown the existence of a genuine kind of envy that is benign, insofar as it motivates to ‘level up’, to try and emulate the envied or to attain what the envied has (van de Ven 2016). The consensus is that, in order for benign envy to arise, the envier has to feel pretty confident and hopeful that they can overcome their disadvantage. If they feel ‘hopeless’ and ‘helpless’ (Miceli and Castelfranchi 2007: 456–9), then they are likely to develop feelings of malicious envy and be motivated to ‘level down’ and engage in aggressive and hostile behaviours, such as sabotaging the envied or spoiling the good among others. Some contemporary philosophers have also developed more positive accounts of envy (La Caze 2001; Thomason 2015). I, too, defend the view that envy need not be bad and argue that a certain species of envy, what I call emulative envy, may even be virtuous in some conditions (Protasi 2021). Emulative envy is characterized by being more concerned with the good than with the envied’s possession of it,3 and by a perception that levelling up to the envied is a likely possibility. In other words, the emulative envier is more preoccupied with the lack of a 3  Focus of concern is a continuous variable. As mentioned earlier, envy always involves three parties, so the envier is always concerned with the envied and the good. But sometimes they are more concerned with lacking the good, whereas other times they are more concerned with the fact that the envied has it. Aristotle makes the same point (but in dichotomous terms) when distinguishing between zēlos (usually translated as ‘emulation’) and phthonos (usually translated as ‘envy’) in Rhetoric II, 10–11.

294 Fittingness c­ ertain good than with the fact that another person has it. The envied’s having the good is both a reminder of the envier’s lack and proof that the envier could also have it (insofar as the envier and the envied are similar). Furthermore, the emulative envier believes that they can improve their condition and achieve the good on their own terms, without taking it from the envied. Thus, they are mo­tiv­ated to ‘level up’ to the envied. Recall my envy for my sister’s unblemished skin. If I feel emulative envy, that means that I think that I might be able to improve the state of my skin, perhaps through an appropriate skin regimen or nutritional plan, and that I care about this good more than the fact that it’s my sister who has better skin. It’s not that I have a rivalry with my sister, but rather that I really care about good skin and my sister’s skin shows me the possibility of achieving it. You may object that this doesn’t sound like envy. This is a reasonable concern which I address extensively elsewhere (Protasi 2016 and 2021, esp. 51–4). For the purposes of the present discussion, and given that the notion of benign envy is becoming mainstream in the empirical literature, I’ll just assume it. I won’t, however, assume my more controversial view that benign envy can be virtuous (note that my example above doesn’t qualify as virtuous, since unblemished skin is not clearly a component of eudaimonia and it actually sounds a bit frivolous and ableist—­fortunately, it’s a purely hypothetical case!). There are three other kinds of envy that I discuss in my work. Inert envy is an ineffective version of emulative envy, which one feels when one is more concerned with the good than with the envied, but doesn’t feel capable of improving one’s situation. It’s a self-­defeating emotion, which aims for something it cannot achieve, and doesn’t motivate anything but sulking and resignation. It’s not malicious, but is prudentially bad, and counterproductive. A good example would be ‘baby envy’: when someone cannot conceive a child and they envy people who can. These enviers often don’t wish the envied lose their good (i.e. lose their fertility or the baby), but their envy cannot in typical cases lead to self-­improvement. Aggressive envy is one species of what is usually called malicious envy. It’s more focused on the envied than the good, and it motivates bringing the envied down to one’s level. It’s the kind of envy that motivates stealing the good and sabotaging the envied; its paradigmatic example is tripping someone during a race in order to get ahead of them. It doesn’t achieve one’s self-­improvement, but it helps to reduce the gap with the envied. This envy motivates morally bad behaviour, but it can bring genuine advantages. Finally, another species of malicious envy is spiteful envy, which is the perverse cousin of aggressive envy, the one we feel when we lack all hope of even subtracting the good, and so it motivates us to spoil it. My usual example here is Othello’s villain: Iago cannot get Cassio’s promotion, or Othello’s beautiful wife. So, he proceeds to destroy their happiness. In a way, spiteful envy is slightly more effective than inert envy, since the envied does lose their advantage. But Othello aims to teach us that this advantage is short-­lived

The Things We Envy: Fitting Envy and Human Goodness  295 (Iago will languish in prison for his misdeeds). Spiteful envy is thus both morally and prudentially bad. Importantly, all of these types can be fitting, but envy is so complex that evaluations of its fittingness are complex too. Many of envy’s facets consist in subjective perceptions; thus, whether one’s envy is fitting depends on whether those ­perceptions are apt. Am I right in thinking that my sister’s skin is smoother than mine? Perhaps I magnify my blemishes or underestimate hers. Is it the case that my sister is similar to me in relevant respects? Even though we’re genetically related (assuming she is my biological sibling), perhaps she’s much younger, in which case a comparison to her may not make much sense when it comes to skin texture.4 Now, one may ask: is it really appropriate given who I am to care so much about this? I may be opaque to myself, lacking self-­knowledge or even deceiving myself. Upon reflection, I may realize that what really bothers me is something else. Perhaps I’m pained by how my parents praise my sister’s appearance over mine, which in turn manifests to me a disparity in love and care. If that’s the case, then my envy for her skin is unfitting, not because unblemished skin isn’t good—­it’s unfitting because an unblemished skin is not relevant to my self-­identity, and it’s a proxy, a trigger, or a symptom of another sort of perceived inferiority.5 It might also be that resentment, not envy, is the fitting response here, if my parents are wronging me. Finally, if we do accept the distinction between different kinds of envy, we might wonder whether one’s perception of control over the outcome (i.e. my ability to obtain the good, or to steal or spoil it) is correct. Suppose I am more focused on the good (my sister’s beautiful skin) than the envied (my sister’s having such beautiful skin). Perhaps I’m overly confident in my capacity to improve an organ as capricious and voluble as the skin; or, vice versa, I’m too insecure or ignorant about the great improvements in skin cosmetics. Either way, whether I’m correct in my perception of the likelihood that I can level up to my sister will determine whether the particular kind of envy I’m feeling is fitting or unfitting. That is, if I’m too optimistic about my capacity to improve my skin, my emulative envy will be unfitting; vice versa, if I’m too insecure, my inert envy will be unfitting. Imagine, instead, that I’m more focused on the envied than on the good. Then perhaps I feel spiteful envy and I’m confident I can switch her beauty products and damage her skin. But in fact beautiful skin is just in her DNA (somehow I don’t share 4  This isn’t to say an older person cannot envy a younger person, but only that their envy might not be fitting. I’m happy to grant this is up for debate—­much depends on what the salient similarity in the eyes of the envier is; but we can easily imagine someone telling the envier: ‘Come on, it doesn’t make sense for you to envy her! She’s so much younger than you! You have great skin for your age!’ Thanks to Steve Campbell for pressing me on this point. 5  Introspection can go wrong in many other ways: I might not even be aware I’m envious, or I might mistake my envy for resentment (a well-­known phenomenon; see e.g. Parrott and Smith 1993). I set these complications aside.

296 Fittingness those particular genes with her!), and so my spiteful envy is unfitting. (Given the envied object, aggressive envy just doesn’t seem possible. This isn’t a flaw in the model, but a reflection of the multifarious nature of envy and envied goods.) So far, we have looked at the fittingness of the envied’s assessment of their pos­ ition vis-­à-­vis the envied (whether they are actually disadvantaged; how similar they actually are); of their perception of self-­relevance (given the envier’s overall system of values, is this particular concern consistent with it?); and of their perception of control over the situation (is it really the case that I can/cannot level up or down with the envied?). But there’s another factor that determines whether my envy is fitting, one that has received more attention in the philosophical tradition than the ones highlighted above: is the object of envy really good for me and my well-­being more generally, and thus truly worth pursuing? The agent sees the lacked object as good, but are they mistaken? If they are, their envy is unfitting. For the purposes of this discussion, I assume both objectivism and pluralism about value; I’m not committed to a specific list, but here are some popular and not too controversial examples of goods: happiness (in the psychological sense, i.e. a positive balance of affective states including sensory pleasures), excellence, beauty, health, knowledge, (rewarding) interpersonal relationships, meaning, and achievement. Thus, I’m leaving out, for now, objects like wealth and the latest tech gadget, or a combination of the two, such as the Black Diamond iPhone 5—­a status symbol for our times. Even though status symbols are often envied, many philosophers doubt that they are actually good for us, and I agree. While happiness, excellence, beauty, health, knowledge, love, meaning, and achievement may be considered relatively uncontroversial examples of goods, suggesting that they are fitting objects of envy may be met with unease, as we are going to see in the next section.

13.3  Against Fitting Envy: Immanuel Kant To reiterate, there are many distinct, if often related, critiques of envy across cultures and philosophical traditions. Some of them derive from distrust about emotions in general (e.g. Stoicism), while some others are specifically concerned with envy and its alleged immorality (e.g. Bacon and Spinoza). Others don’t explicitly or vehemently condemn envy, and yet don’t rush to defend its possible fittingness either (e.g. Descartes and Hume). Among those concerned with envy’s viciousness, Kant deplores both the immoral actions it motivates and its inappropriate focus on relative positioning. In The Metaphysics of Morals, envy is defined as ‘a propensity to view the well-­ being of others with distress, even though it doesn’t detract from one’s own; when

The Things We Envy: Fitting Envy and Human Goodness  297 it breaks forth into action (to diminish their well-­being) it is called envy proper; otherwise it is merely jealousy (invidentia)’ (4:458). Envy ‘proper’ is denounced as a ‘vice of hatred’ (4:458), ‘contrary to a man’s duty to himself as well as to others’ (4:459). Insofar as it goes against the demands of practical reason to respect humanity and make other people’s happiness our end, it’s practically irrational.6 Even mere jealousy is arguably worrisome, given that it’s a distress at the good fortune of others even when it doesn’t affect us directly, and that it motivates immoral actions. Since he talks about a ‘propensity’, Kant probably isn’t thinking about occurrent emotions, but rather dispositions or character traits. Nevertheless, I doubt he’d feel too sanguine about occurrent bouts of envy, either. Both here and elsewhere Kant is concerned with envy’s essentially comparative and competitive nature. I think that for him that is the fundamental problem with envy; therein originates its vicious nature. He discusses social comparison and the ‘malignant inclination’ of envy (6:93) also in Religion within the Boundaries of Mere Reason. In that text, Kant is worried by a typically human form of self-­love, shaped by the concern for what others think of us (6:27). After an initial anxiety for being of equal worth to others, we gradually develop ‘an unnatural desire to acquire superiority for oneself over ­others’ (6:27). Envy is thus both irrational and immoral, and the source of its ir­ration­al­ity and immorality ultimately lies in the process of social comparison, which wreaks havoc in man’s originally good nature. For this reason, Kant rejects both envy and the tendency to compare oneself to others even with regard to pedagogy: while encouraging children to emulate their well-­behaved peers might seem like a good practice, it ends up stirring resentment and envy. Thus, Kant recommends educators to cultivate in children the aspiration to duty itself (Lectures on Pedagogy 9:491–2). Even if one disagrees with my interpretation of Kant, I hope one would still find this type of view not only intelligible, but familiar: authentic goodness is often said to be non-­comparative and non-­positional in everyday conversations. When children lament that their siblings or friends get to have or do something they don’t, parents and educators usually invite them to focus on their own situation, to think of their lot as independent from that of others. (My personal experience as a parent is particularly paradoxical, as I defend envy and the value of competition in theory, and am constantly frustrated by its effects on my children in practice.) When adults express their envy, they often find that self-­help books, meditation apps, psychology blogs, religious sermons, and philosophy op-­eds encourage them to feel gratitude for the good fortune they have, on the one hand, and to 6  Thanks to E. Sonny E. Elizondo for help with this formulation, and for feedback on my understanding of these texts.

298 Fittingness avoid comparisons to others, on the other. According to this perspective, envy is systematically unfitting, because how we stack up compared to others is not truly relevant to our well-­being. Being inferior to the envied is not actually bad, and improving one’s relative position is not actually good. While it’s usually acknowledged that envy, as a natural human emotion, may not be fully eliminated, the general advice is to avoid feeling it as much as pos­ sible, and the most effective way to do that is attacking it at its roots, refraining from comparison and from attributing value to one’s relative standing. The problem with this advice is two-­fold: it’s not psychologically realistic, and it’s not prudentially desirable. It’s not realistic because interpersonal comparison is a very powerful and basic mechanism that cannot be easily extirpated from our minds, and that undergirds many important psychological functions (I say more about this in section 13.5). Even if it were possible to get rid of interpersonal comparison, however, it wouldn’t be desirable, because our conception of goodness wouldn’t be recognizable without it.

13.4  In Favour of Fitting Envy: Justin D’Arms and Daniel Jacobson When it comes to defending envy’s fittingness, the obligatory reference is D’Arms and Jacobson’s seminal work on the fittingness of human sentiments (2000, 2003 2006). I won’t review this work in detail, but only summarize their most persuasive argument in defence of envy’s fittingness. D’Arms and Jacobson (2006) consider envy the ‘most intuitively compelling case’ (p. 106) for the ‘zealot’s’ view that some emotions are systematically unfitting. They happily concede that envy is both ugly and painful, and claim that attempts to embellish it and defend its potentially benign or positive nature are either misleading or confused (ouch!). So, they argue with the ‘zealot’ on their own grounds: if they can prove that even envy can be fitting, then a fortiori their view holds for any other human sentiment. Their defence of envy rests on their well-­known notion of fittingness: ‘to think an emotion a fitting response to some object is to think there is (pro tanto) reason, of a distinctive sort, for feeling the emotion toward it’ (2006: 108). Reasons of fit are distinctive, and differ from moral and prudential reasons, in that they ‘speak directly to what one takes the emotion to be concerned with’ (2006: 108). Emotions—­ they argue—­ are evolutionarily shaped responses to many cross-­ cultural concerns and values. Their pluralistic approach accounts for the fact that human values are often in conflict with one another, both within and across domains: not only is it possible to be faced with genuine ethical dilemmas, but also to experience tension and strife between aesthetic, interpersonal, ethical, and epistemic considerations.

The Things We Envy: Fitting Envy and Human Goodness  299 The human concern individuated by envy is relative positioning: when we feel envy, we, at least sometimes, appropriately react to being disadvantaged vis-­à-­vis a (similar) other (with regard to a domain of self-­importance). This concern is not only deep, that is, ingrained and hard to eradicate in an average human being, but also wide, that is, ‘firmly enmeshed in our web of psychological responses’ (2006: 116). Envy is concerned with positional goods, which are both ubiquitous and, contra Kant, important. Positional goods can be defined as ‘goods the absolute value of which, to their possessors, depends on those possessors’ place in the distribution of the good—­on their relative standing with respect to the good in question’ (Brighouse and Swift 2006: 474). If it can be shown that positional goods are a central and appropriate human concern, then envy is at least sometimes fitting. D’Arms and Jacobson go on to list many positional goods, starting from the most obvious examples, such as sports achievements, job promotions, and being at the top of social hierarchy, all the way to excellence itself: they argue that the very idea of excellence is inherently comparative and relative: ‘which accomplishments count as excellent, or sufficiently good to be worthy of pride, is largely a function of the performance of others (especially those who are nearby)’ (2006: 123). Since relative positioning is a central and appropriate human concern, envy, as a response to this concern, is thus fitting. D’Arms and Jacobson, however, go on to caution against inferring that envy’s fittingness makes it prudentially or morally good: ‘perhaps morality forbids acting from envy or even feeling it’ (2006: 124). D’Arms and Jacobson’s argument doesn’t rest on equating every excellence or accomplishment to a zero-­sum game. Some of the contexts they describe are zero-­sum (i.e. sport races), but many others are not and yet they rely on comparison (whether one is a ‘strong’ runner depends on the comparison class). Thus, I think that D’Arms and Jacobson don’t carefully distinguish between associated but distinct concepts: zero-­sum situations, positional goods, and relative or comparative standards of goodness. The remainder of the chapter is devoted to developing their argument further. First, I consider the role comparison plays in judgements about goodness; then, the role it plays in shaping goodness itself.

13.5  The Role of Comparison in Judgements of Goodness Let’s go back to sibling envy. When I think about the kind of envy I feel towards my imaginary saintly sister, I don’t have to reach too far in my imagination: I can think of sibling rivalries I witnessed from up close, including between my own kids. Envy in children is ubiquitous and often unfitting, and it’s ubiquitous for the same reason it’s often unfitting, namely because their social comparison skills are still developing and they don’t have a good sense of relevant comparison classes

300 Fittingness (so, for instance, they may get upset if a professional artist draws better than them). Furthermore, they have a fluid and ever-­changing sense of their own identity and what matters to them: they haven’t yet ‘specialized’ in one domain, like adults do when they think of themselves as ‘a dancer’ or ‘a philosopher’. Thus, since almost everyone who outperforms them is perceived as similar, and because so many things matter to them, their envy is often all-­encompassing and indiscriminating (Bers and Rodin 1984). But sometimes children’s comparisons are spot on, sensible, and appropriate. And this behaviour—­social psychologists tell us—­is healthy and normal. Susan Fiske calls humans ‘comparison machines’ (Fiske 2011: 13) We habitually and automatically (thus often unconsciously) compare ourselves to others, and this mechanism is essential for a variety of psychological functions, including gathering information for self-­assessment and self-­improvement, self-­esteem protection, and fitting into social groups (Mussweiler 2003; Fiske 2011; Corcoran et al. 2011). One might object that people can be assessed non-­comparatively as well. But there are reasons to doubt this. D’Arms and Jacobson (2006) already present a persuasive discussion of various situations in which being considered a good X soon becomes a matter of being considered better than or worse than another person (along the relevant dimension). After reviewing domains such as scholarship, arts, industry, and athletics (all domains central to our flourishing) they talk of being the ‘Number One Dad’—the dad who mentors and coaches the kids in the neighbourhood. Obviously, this is definitionally a comparative property, since it includes a ranking. But one might want to eschew comparisons and aim to be a ‘Good-­Enough Dad’. On the one hand, the property of good-­enough-­daddyness may be said to be non-­comparative: for instance, a good-­enough daddy does not abuse his children and provides for their essential physical and emotional needs. On the other hand, this is a socially constructed notion, a type of goodness that is unavoidably shaped by human judgements and perceptions, which are in turn subject to historically situated social and cultural norms: good-­enough dads fifty years ago were those who did not beat their children and showed up at dinner time. The criteria for crossing the ‘good enough’ threshold, our judgements about what a good-­ enough dad is like, are much more demanding nowadays. The same holds for many contemporary standards of goodness. Take the arts, in the ancient Greek sense of technai: people run faster and longer, create more complex artworks, invent more impressive technology, achieve ever-­greater feats of virtuosity in all domains. So, whether or not we are considered good children, parents, or spouses, whether we are decreed as competent philosophers, computer scientists, dancers, engineers, or potters—­ any threshold that is used to evaluate success or

The Things We Envy: Fitting Envy and Human Goodness  301 achievement or even just mere adequacy is going to be determined by looking at what other individuals on average accomplish.7 A similar point applies to assessments of physical properties such as good growth or good health, which all rely on some form of comparison or other. When children are evaluated for growth, paediatricians consult charts that are inherently, by design comparative: being a tall child simply means being taller than the average child. Counting as ‘healthy’ at seventy years of age relies on very different indicators than at three or twenty-­five, and counting as healthy in a contemporary industrialized society looks quite different from what it did in preindustrial societies of the past, or in today’s hunter-­gatherer societies. How to understand the notions of health and healthy systems is a controversial topic, but here I set aside the conceptual complications and suggest that, in practice, our judgements of health are unavoidably comparative.8 This point generalizes to most (all?) bodily and mental properties that people deem valuable, whether it’s IQ or endurance and speed, and applies even to judgements of properties like vertical symmetry in a face, which is well-­known to be a crucial feature in judgements of attractiveness (Fink et al. 2006; Young et al. 2011) perhaps because it’s a marker of genetic health and resistance to disease (Jones et al. 2001; Young et al. 2011). Perceptions of facial attractiveness are shown to be biased by assimilative and contrastive biases, both of which depend on comparison; as Pegors et al. (2015) put it: ‘[t]o navigate the social world, it is important to be able to evaluate face attractiveness, but these judgements are always made in relation to a larger social and environmental context.’ Both empirical and anecdotal evidence suggest that our concern for relative positioning is impossible to eradicate. Our deeming something good is often ul­tim­ate­ly a matter of deeming it to be better than. I’ll come back to the implications of these judgements for fitting envy, but I want first to examine a more direct way in which comparison is relevant to human goodness.

13.6  The Role of Comparison in Goodness Some things cannot be good outside of a competition or ranking. The relativity here is metaphysically and logically necessary. This is the case of positional goods mentioned earlier. Positional goods are often discussed in economics, where they are sometimes framed in terms of ‘positional concerns’ (Solnick and Hemenway 1998). 7  Gwen Bradford, the philosophical authority on accomplishments, also argues that difficulty and achievements are attributed on the basis of relevant comparison classes (Bradford 2015: 61–2). 8  See Murphy (2020) for the complex debate on the concept of health. Elizabeth Barnes discusses both comparative health judgements and theories of health in a book in progress. Bloomfield 2001 defends a relativistic understanding of the property of healthiness.

302 Fittingness A standard example of a non-­material positional good is honour, as a social status notion: to be honoured in a certain society means to be anointed, to be given a social status that is higher than other people’s. It’s an ‘essentially hierarchical’ notion as Stephen Darwall puts it (2013: 17). To be in any honour association is valuable because it’s selective and ‘rank-­defining’ (2013: 18). Other examples of positional goods are connected to social status, and in a consumeristic society they are often material goods, objects, or experiences that cost a lot of money and which many people cannot afford, and thus function as status symbols. That paradigmatic examples of conspicuous consumption (defined as ‘expenditure on or consumption of luxuries on a lavish scale in an attempt to enhance one’s prestige’)9 are also paradigmatic examples of envy is not incidental to envy’s bad fame. These tend to be dismissed as inauthentic goods not only by moral philosophers, but also by sociologists and psychologists, for good reasons: evidence shows that they don’t actually bring happiness, and one of the reasons why they don’t is that agents are caught up in an arms race, which is never-­ending and collectively damaging (Frank 1999). But some positional goods aren’t as easily rejected as false idols: educational pedigrees are one such case. Ivy League degrees have value partly (some might say primarily!) in virtue of the social currency they carry, as shown by the success of many an underachieving student who can nevertheless tout having graduated from a prestigious institution (the list of US presidents provides some examples). These degrees are prized because it’s very difficult to be admitted to one of those schools. Similarly with publications: the higher the rejection rate of a journal, the better the journal is taken to be. (We can quibble about the causal chain here: it’s possible that initially a journal is taken to be very good because of its editors—­editors who often come from prestigious institutions, but never mind that—­but soon enough high rejection rate becomes a feature of a good journal.) Variations of this mechanism can be found everywhere: a robust feature of human psychology is that we value being part of exclusive clubs. Groucho Marx’s joke that he doesn’t want to be a member of a club that admits him relies in part on this truth. Furthermore, many goods, whether they are themselves positional or not, have latent positional aspects, as argued by Brighouse and Swift (2006: 478–9). Even if healthiness is not positional in itself, it may be instrumental to achieving pos­ ition­al goods: the healthier a person is, the greater access they have to jobs, education, and other positional or otherwise scarce goods. Similar considerations apply for beauty or wits, insofar as they allow people to get ahead in the world. Indeed, physical attractiveness and intelligence are among the goods for which positional preferences are strongest: in surveys, people consistently claim to prefer lower absolute levels of attractiveness and intelligence for oneself and one’s

9 https://www.lexico.com/en/definition/conspicuous_consumption.

The Things We Envy: Fitting Envy and Human Goodness  303 kids, if that allows them to have them in comparatively higher amounts (Solnick and Hemenway 2005; Hillesheim and Mechtel 2013).10 Note the difference between the claim I’m making here and the claim I made in the previous section. It’s not just that when we judge someone to be attractive or intelligent we are, consciously or unconsciously, comparing them to others in the relevant comparison class; it’s also that attractiveness and intelligence themselves (even if they could be measured according to absolute non-­relational standards) are positional and have indirect positional aspects. This is not to say that all goods are positional, directly or latently. Take such sensorial pleasures and simple joys as hearing birds tweet and children laugh on a sunny winter morning; the warmth and sounds and colour of an autumnal fire; an unexpected caress by a usually reserved lover, or the smile of a stranger. These experiences are arguably good, and even components of our flourishing, and yet don’t depend on comparison, hierarchy, or competition. And there are other goods and components of our well-­being like this, including perhaps knowledge and meaning. My aim in this and the previous section was to show that positional goods and goods with latent positional aspects are numerous and important, and that, furthermore, most of our judgements of goodness are comparative. Thus, not only envy can be fitting, but it can be fitting in many more situations than previously highlighted, even by advocates of envy’s fittingness such as D’Arms and Jacobson. In response, one might argue that our judgements of goodness are sys­tem­at­ic­ al­ly disconnected from actual goodness—­that we’re just wrong when we deem something to be good based on its being better than. Relatedly, one might argue that positional goods are inauthentic goods, false idols: even if humans do care about relative positioning, they shouldn’t. If that were true, then Kant would be right: envy is never fitting. We should not be pained by inferiority to a similar other with regard to a self-­important domain. In the end, this debate boils down to conflicting intuitions about value and the relation it bears to human psychological propensities. I agree with the view that Martha Nussbaum attributes to Aristotle: ‘What we find valuable depends essentially on what we need and how we are limited. The goodness and beauty of human value cannot be comprehended or seen apart from that context’ (Nussbaum 1986/2001: 342). I hope to have shed light on the extent to which comparison and relative positioning affect both goodness and judgements of goodness, which in turn suggests that a conception of human goodness that is entirely noncomparative and nonpositional is unrecognizable as such. 10  Interestingly, in these surveys people rush to note that ‘their positional choices were not mo­tiv­ ated primarily by envy. Many seemed to see life as an ongoing competition, in which not being ahead means falling behind’ (Solnick and Hemenway 2005: 379). Setting aside the fact that people aren’t always aware of their envy, or eager to admit it, not being primarily motivated by envy is compatible with their envy being fitting.

304 Fittingness

13.7  The Multifaceted Value of Fitting Envy To sum up: I have defended a metaphysical claim—­that much human goodness is shaped by positionality—­and an epistemic one—­that we use implicit or explicit comparison to know what position we occupy in this continuum of goodness. Envy’s preoccupation with one’s relative positioning is fitting, insofar as it recognizes fundamental components of our subjective well-­being and flourishing. So far, I have expanded D’Arms and Jacobson’s view on the centrality of relative positioning in our conception of goodness and value. Now, I argue that fitting envy can be also epistemically, prudentially, and morally valuable, which is something that Justin D’Arms, especially, is sceptical about (see D’Arms 2017, and my critique in Protasi 2021: 52–4). Psychologists often talk about the importance of envy’s signalling value. Envy tells us, on the one hand, what we perceive we lack, and, on the other, what we desire and care about.11 However, unfitting envy sends a distorted signal, while fitting envy sends an accurate one: it informs a (self-­aware) envier about im­port­ ant things that concern them. Recall my initial example but tweak it a bit: imagine that I envy my sibling for their ability to forgive a narcissistic parent. If my envy is fitting, it tells me that I perceive myself as less generous and compassionate than someone who is similar to me in a self-­relevant domain, and that this ability I lack is authentically good. Thus, fitting envy has epistemic value for the individual. (It also has epistemic value for the species: after all, we have just spent some time thinking about what envy tells us about human goodness.) But note also something else: my envy—­let’s assume—­is emulative. I really care about being the type of child who can forgive parental narcissism, and I’m pained by the comparison with my sibling only because they and I are similar. If they can forgive our parent, why can’t I? That is, I’m more focused on the good than the envied, who is a benchmark, showing me what I could achieve. Secondly, I happen to feel confident that I can become like them: I saw them go through years of therapy. It was not just a matter of genetic luck: they work hard on their relationship with our parent. I realize I can do that too. Thus, I’m motivated to improve myself, to emulate my sibling. My envy here is neither counterproductive nor misguided—­on the contrary, it focuses on relative positioning with regard to valuable moral goods and it motivates morally appropriate behaviour. If I were to ‘level up’ with my sibling, becoming capable of forgiveness and understanding, my well-­being would increase, and I would be a better person. Fitting envy can thus be morally and prudentially valuable.

11 Exline and Zell (2008) claim also, more controversially, that envy signals alienation or disconnection.

The Things We Envy: Fitting Envy and Human Goodness  305 So, when envy is fitting and emulative, it motivates us to achieve objectively good things within our reach; it pushes us to improve ourselves and aspire to excellence. Fitting emulative envy, far from being a vice that ought to be repressed, plays a central role in our flourishing. But even other varieties of envy have partial value: they all provide information about ourselves, our perception of how we compare to others, of what we value, and so forth. Even the most spiteful envy can be diagnostic, if we’re willing to look inside ourselves: how have we got to the point of wishing ill will to others? What are we disposed to do when in the grip of this ugly emotion? Obviously, any emotion, no matter how vile, can be informative. But if envy is systematically connected to our flourishing, because it concerns things we genuinely and centrally care about, perhaps its informational value is more precious than that of other fitting attitudes, such as humour. Furthermore, fitting aggressive envy has some prudential value, even if it’s immoral. Sometimes stealing the good is an effective way of increasing personal well-­being, and we should be aware of this, especially if we want to prevent aggressive envy’s harmful effects. Again, envy connects us to goodness in a peculiar way. Its affectively aversive nature makes it more motivating than pleasant emotions (Vaish et al. 2008), so we should acknowledge its power and value, even when we ultimately think that moral considerations defeat all the others. This is particularly important when thinking about emotions at the species level, as sometimes what is good for the individual conflicts with what is good for the species. In Idea for a Universal History from a Cosmopolitan Perspective (1784) Kant presents a thesis that starkly contrasts with his claims about envy in other works. There, he suggests that antisocial passions, including envious competitiveness, can help to develop humanity’s full potential. He compares human beings who accept the hardship of a civic union to trees, which ‘by seeking to take air and light from all the others around them, compel each other to look for air and light above themselves and thus grow up straight and beautiful, while those that live apart from others and sprout their branches freely grow stunted, crooked, and bent’ (8:22, Kant 2006: 8). In a civic union, individuals are bound by norms that limit their aggressive tendencies and that allow them to compete in a regulated manner, in a way that allows humankind to develop culture and art. Without the ‘unsociable’ tendencies that drive humans to compete and fight with each other, they would remain in an Arcadian state of contentment and mutual affection that would, however, leave them passive, slothful, and ‘as good-­natured as the sheep that they put out to pasture’ (8:21, Kant 2006: 7), their excellent natural capacities forever dormant. In this essay, Kant concedes that envy and competitive tendencies push us towards the good, towards developing our potential as humans. While it is pos­ sible that sometimes fitting envy should not be felt, expressed, or acted upon, because of countervailing reasons against it, and while some individuals might be

306 Fittingness better served by their envious tendencies than others, for all sorts of contextual and idiosyncratic reasons, envy offers to all humans the possibility of being attuned to distinctively human goodness, and the push to excel in domains they care about. This is such a fundamental feature of human reality, that even Kant cannot deny it.

References Aristotle. 2007. On Rhetoric: A Theory of Civic Discourse, trans. and ed. G. A. Kennedy. New York: Oxford University Press. Bers, Susan A., and Judith Rodin. 1984. ‘Social-comparison jealousy: A developmental and motivational study.’ Journal of Personality and Social Psychology 47 (4): 766–79. Bloomfield, Paul. 2001. Moral Reality. New York: Oxford University Press. Bradford, Gwen. 2015. Achievement. Oxford: Oxford University Press. Brighouse, Harry, and Adam Swift. 2006. ‘Equality, priority, and positional goods.’ Ethics 116 (3): 471–97. Corcoran, Katja, Jan Crusius, and Thomas Mussweiler. 2011. ‘Social Comparison: Motives, Standards, and Mechanisms.’ D’Arms, Justin. 2017. ‘Envy.’ In The Stanford Encyclopedia of Philosophy (spring 2017 Eedition). Edited by Edward N. Zalta. https://plato.stanford.edu/archives/spr2017/ entries/envy/s D’Arms, Justin, and Daniel Jacobson. 2000. ‘The moralistic fallacy: On the “appropriateness” of emotions.’ Philosophical and Phenomenological Research 61 (1): 65–90. D’Arms, Justin, and Daniel Jacobson. 2003. ‘The significance of recalcitrant emotion (or, anti-quasijudgmentalism).’ Philosophy and the Emotions 52: 127–45. D’Arms, Justin, and Daniel Jacobson. 2006. ‘Anthropocentric constraints on human value.’ Oxford Studies in Methaethics 1: 99–126. Darwall, Stephen. 2013. ‘Respect as Honor and as Accountability.’ In Honor, History, and Relationship: Essays in Second-Personal Ethics II, 11–29. Oxford: Oxford University Press. Fink, Bernhard, Nick Neave, John T. Manning, and Karl Grammer. 2006. ‘Facial symmetry and judgements of attractiveness, health and personality.’ Personality and Individual Differences 41(3): 491–9. Fiske, Susan T. 2011. Envy Up, Scorn Down: How Status Divides Us. New York: Russell Sage Foundation. Frank, Robert H. 1999. Luxury Fever: Why Money Fails to Satisfy in an Era of Excess. New York: Free Press. Hillesheim, Inga, and Mario Mechtel. 2013. ‘How much do others matter? Explaining positional concerns for different goods and personal characteristics.’ Journal of Economic Psychology 34: 61–77.

The Things We Envy: Fitting Envy and Human Goodness  307 Jones, Benedict  C., Anthony  C.  Little, Ian  S.  Penton-Voak, Bernard  P.  Tiddeman, D. Michael Burt, and David I. Perrett. 2001. ‘Facial symmetry and judgements of apparent health: Support for a “good genes” explanation of the attractiveness–­ symmetry relationship.’ Evolution and Human Behavior 22 (6): 417–29. Kant, Immanuel. (1797) 1996. The Metaphysics of Morals, trans. and ed. Mary J. Gregor. New York: Cambridge University Press. Kant, Immanuel. (1793) 1998. Religion within the Boundaries of Mere Reason, trans. George Di Giovanni, ed. Allen W. Wood. New York: Cambridge University Press. Kant, Immanuel. (1784) 2007a. Idea for a Universal History with a Cosmopolitan Aim, trans. Allen  W.  Wood, ed. G.  Zöller and R.  B.  Louden. Cambridge: Cambridge University Press. Kant, Immanuel. (1803) 2007b. Lectures on Pedagogy, trans. R.  B.  Louden, ed. G. Zöller and R. B. Louden. Cambridge: Cambridge University Press. Kant, Immanuel. 2006. Toward Perpetual Peace and Other Writings on Politics, Peace, and History, trans. David  L.  Colclasure, ed. Pauline Kleingeld, Contrib. David L. Colclasure, Jeremy Waldron, Michael W. Doyle, and Allen W. Wood. New Haven, CT: Yale University Press. La Caze, Marguerite. 2001. ‘Envy and resentment.’ Philosophical Explorations 4 (1): 31–45. Lange, Jens, and Jan Crusius. 2015. ‘Dispositional envy revisited: Unraveling the mo­ tiv­ ation­ al dynamics of benign and malicious envy.’ Personality and Social Psychology Bulletin 41 (2): 284–94. Miceli, Maria, and Cristiano Castelfranchi. 2007. ‘The envious mind.’ Cognition and Emotion 21 (3): 449–79. Murphy, Dominic. 2020. ‘Concepts of Disease and Health.’ In The Stanford Encyclopedia of Philosophy (Summer 2020 Edition), Edited by Edward  N.  Zalta. https://plato.stanford.edu/archives/sum2020/entries/health-disease/. Mussweiler, Thomas. 2003. ‘Comparison processes in social judgment: Mechanisms and consequences.’ Psychological Review, 110 (3): 472. Nussbaum, Martha. C. 1986/2001. The Fragility of Goodness: Luck and Ethics in Greek Tragedy and Philosophy. New York: Cambridge University Press. Parrott, W.  Gerald, and Richard  H.  Smith. 1993. ‘Distinguishing the experiences of envy and jealousy.’ Journal of Personality and Social Psychology 64: 906–20. Pegors, Teresa  K., Marcelo  G.  Mattar, Peter  B.  Bryan, and Russell  A.  Epstein. 2015. ‘Simultaneous perceptual and response biases on sequential face attractiveness judgments.’ Journal of Experimental Psychology 144 (3): 664–773. Protasi, Sara. 2016. ‘Varieties of envy.’ Philosophical Psychology 29 (4): 535–49. Protasi, Sara. 2021. The Philosophy of Envy. Cambridge: Cambridge University Press. Smith, Richard  H. and Sung Hee Kim. 2007. ‘Comprehending envy.’ Psychological Bulletin 133: 46–64. Solnick, Sara J. and David Hemenway. 1998. ‘Is more always better? A survey on pos­ ition­al concerns.’ Journal of Economic Behavior and Organization 37 (3): 373–83.

308 Fittingness Solnick, Sara J. and David Hemenway. 2005. ‘Are positional concerns stronger in some domains than in others?’ American Economic Review 95 (2): 147–51. Thomason, Krista  K. 2015. ‘The moral value of envy.’ The Southern Journal of Philosophy 53 (1): 36–53. Vaish, Amrisha, Tobias Grossmann, and Amanda Woodward. 2008. ‘Not all emotions are created equal: The negativity bias in social-emotional development.’ Psychological Bulletin 134 (3): 383–403. van de Ven, Niels. 2016. ‘Envy and its consequences: Why it’s useful to distinguish between benign and malicious envy.’ Social and Personality Psychology Compass 10 (6): 337–49. Young, Steven G., Donald F. Sacco, and Kurt Hugenberg. 2011. ‘Vulnerability to disease is associated with a domain-specific preference for symmetrical faces relative to symmetrical non-face stimuli.’ European Journal of Social Psychology 41 (5): 558–63.

14 Response-­Dependence and Aesthetic Theory Alex King

14.1  The Standard View What has the aesthetician historically wanted out of a meta-­aesthetic theory?1 Well, the aesthetician typically wants two things that are somewhat at odds. Hume (1777/1987) articulates them nicely. On the one hand, he says: Beauty is no quality in things themselves: It exists merely in the mind which contemplates them; and each mind perceives a different beauty.  (230)

And on the other: Whoever would assert an equality of genius and elegance between Ogilby and Milton, or Bunyan and Addison, would be thought to defend no less an extravagance, than if he had maintained a mole-­hill to be as high as Teneriffe, or a pond as extensive as the ocean. . . . [I]t appears an extravagant paradox, or rather a palp­able absurdity, where objects so disproportioned are compared together. (230–1)

At the crudest level, the aesthetician wants both subjectivity and objectivity. To be a bit more refined, they want, on the one hand, something like metaphysical un-­ spookiness (‘no quality in things themselves’, ‘exists merely in the mind’) and interpersonal variability in taste and preference (‘each mind perceives a different beauty’). But on the other, they also want a standard of taste that ranks some aesthetic objects as being genuinely above others, regardless of individual preference and inclination. To capture both sides, Hume defends an ideal observer theory: ‘If, in the sound state of the organ, there be an entire or a considerable uniformity of sentiment 1  This chapter has benefited enormously from earlier feedback. Thanks to audiences at Boston University, University of Leeds, the London Aesthetics Forum, and the UBC/SFU working group, the Fit Fest conference, and especially to Chris Howard and Rich Rowland. Alex King, Response-­Dependence and Aesthetic Theory In: Fittingness: Essays in the Philosophy of Normativity. Edited by: Christopher Howard and R. A. Rowland, Oxford University Press. © Alex King 2022. DOI: 10.1093/oso/9780192895882.003.0014

310 

Fittingness

among men, we may thence derive an idea of the perfect beauty’ (234). He goes on to articulate five criteria (delicacy of taste, practice, comparison, freedom from prejudice, and good sense) of an ideal observer and says that the ‘joint verdict of such . . . is the true standard of taste and beauty’ (241). This account falls under the umbrella of response-­dependence theories. This label has been used in different ways and to different ends, but here I’ll take it to label a class of theories according to which what makes it the case that an object possesses a certain property is that property’s relationship to human responses. For a canonical example, a response-­dependence theory of colour would attribute redness to an apple in virtue of the fact that humans experience the apple as red under normal conditions. Notice that the response can be idealized in some respects. Here, normal conditions involve certain lighting, a functional visual system, and so on. Thus Hume’s account, though it involves idealized observers and thus idealization conditions, falls under the response-­dependence umbrella. Such accounts offer a tantalizing resolution to the above paradox by allowing for interpersonal variability (some do not in fact experience the apple as red) while presenting a non-­spooky standard (they would if the conditions were right). In the colour case, colour qualia are clearly the relevant responses. But what are the relevant responses in the aesthetic case? We can group species of aesthetic response-­dependence accounts by their answer to this further question. The most natural thought, and the historically most popular one, takes pleasure of some kind to be the relevant response. Let’s call these hedonic response-­dependence accounts. Although Hume does not strictly commit himself to hedonism,2 he without a doubt thinks of pleasure as playing an important role in determining the ideal judges’ verdicts. Kant, similarly plagued by how to balance interpersonal variability with genuine standards of taste, argues that pleasure—­specifically, disinterested pleasure—­plays a central role in determining the beautiful. These two massive figures in the history of aesthetics have set the tone for much theorizing since then, and so it is unsurprising that hedonic response-­dependence in one version or another has become the default view in aesthetics. A final note about such accounts: they are virtually always accounts of the beautiful, rather than the delicious, funny, charming, elegant, sublime, or aesthetically good (which I take to be the broadest term of aesthetic praise). Aestheticians have thought that hedonic response-­dependence theories were the best or only way to resolve Hume’s tension. This chapter will examine the prospects for response-­dependence theories. I will argue that they need not be accompanied by hedonism and will point out a few commonly overlooked aesthetic properties that pose a problem for response-­dependence. This chapter is not meant as a definitive rejection of response-­dependence. It is rather meant to 2  The joint verdict could, for example, be based on preferences rather than the pleasure responses of ideal judges.

Response-Dependence and Aesthetic Theory  311 raise a few problems for response-­dependence, and thus raise the costs (and to an extent, question the benefits) of accepting such theories, even in aesthetics where such accounts have historically looked very appealing.

14.2  Two Versions of Response-­Dependence: Dispositions and Fit As I characterize the response-­dependence umbrella, it makes sense to distinguish within it two strands: the dispositional and the fittingness (or fitting attitude) theories. The most common version of hedonic response-­dependence defended in aesthetics is dispositional.3 That is, hedonic response-­dependence is defended as the view that what it is for something to be beautiful is for it to be disposed to produce pleasure under certain naturalistically specifiable conditions. Fittingness theories, by contrast, are those that deem as beautiful those objects in response to which it would be fitting (or appropriate or merited) to experience pleasure.4 Let’s first look at dispositional theories, which often rely on the colour analogy discussed above.5 This analogy is attractive because aesthetic properties, like ­colour properties, seem closely related to perception, sensation, and experience. Furthermore, not only is there a tempting broad analogy, but there seems to be a way to move from a dispositional theory of colour properties to a dispositional theory of full-­blown aesthetic properties that involves incremental steps, none of which introduces any significant differences. These steps take us from colour perception through gustation—­literal taste—­and on to canonical aesthetic properties. If we start with gustation, we can say that what makes something bitter is its disposition to taste bitter under normal conditions. Then, we could extend that slightly to say that what makes something delicious is its disposition to produce (gustatory) pleasure under normal conditions. Then, we could extend that slightly to say that what makes a painting beautiful is its disposition to produce (visual? disinterested? aesthetic?) pleasure under normal conditions. If successful, this argument, or sequence of arguments, will show that dispositional theories are just as plausible for colours as for full-­blown aesthetic properties like beauty. But how plausible is each step? Gustation is central to the first two steps because it seems to live at the intersection of the aesthetic and the purely sensory. The first step consists of purely descriptive gustatory properties. These range from the more literal, such as sweet 3 In addition to Hume (1777/1987) and Kant (1790/2000), see Sibley (1968), Wiggins (1988), Matherne (2020). Watkins and Shelley (2012) and Hanson (2018) both discuss this as the standard view, though both object to it. 4  Recently defended in D’Arms and Jacobson (2000b), Jacobson (2011), Gorodeisky (2021). See also Patridge and Jordan (2018), who defend a fittingness theory for the funniness of jokes. 5  For discussion, see also Schellekens (2006) and Simoniti (2017).

312 

Fittingness

or bitter, through the less literal, such as floral, metallic, or creamy, to the clearly metaphorical, such as sharp or light. Thinking about colours can make it seem as though purely descriptive sensory properties are also simple or obvious. But it might take a bit of training and cultivation of sensibility to determine whether a  certain wine is light or to detect the difference in metallic flavour drunk from bottled versus canned Coca-­Cola. But surely this makes no difference to the plausi­bil­ity of a dispositional theory. To the extent that such an account of colour properties is plausible, a corresponding account of descriptive gustatory properties is also extremely plausible. The next step takes us to evaluative gustatory properties. We can helpfully divide these into thicker properties like balanced, refreshing, bland, cloying, and thinner ones like delicious or tasty. Thick properties we could attempt to specify as those disposed to produce gustatory pleasure, but with respect to some particular aspect of their taste. For example, what makes a dish balanced is that its different flavours and their respective strengths combine in a way that is disposed to produce gustatory pleasure under normal conditions. With thin properties, we could say that what makes a dish delicious is that it is disposed to produce gustatory pleasure under normal conditions. We could be more specific here—­maybe it’s not just that it produces some gustatory pleasure but that it produces a certain minimum amount, or that it does not also produce gustatory displeasure. These modifications don’t pose any serious problem. What would the next step in our proposed sequence look like? It would take the dispositional analysis of evaluative gustatory properties and translate it into other, non-­gustatory aesthetic properties. What makes a joke funny is that it is disposed to produce amusement under normal conditions. What makes a scene charming is that it is disposed to charm an observer under normal conditions. What makes a landscape sublime is that it is disposed to produce awe under normal conditions. And what makes a painting beautiful is that it is disposed to prod­uce (disinterested) pleasure under normal conditions. These might not sound so bad. And they might not seem like any real leap from the evaluative gustatory cases we were considering above. However, something important changed when we moved from purely descriptive to evaluative gustatory properties. The dispositionalist has to say that all of the conditions in terms of which they seek to define aesthetic properties are naturalistically specifiable. How to do this is relatively (though perhaps not completely) clear in the colour and descriptive gustatory cases—­we must not be jaundiced, or have our tongues coated with the residue of toothpaste or grapefruit juice, and so on. But it is more difficult for evaluative properties. One might think it doesn’t simply require normal eating conditions and a well-­functioning tongue (and nose), but also discernment and sensitivity and good judgement, where these things cannot be given a purely naturalistic specification. We must, instead, add something normative into the conditions. This problem is even more apparent for non-­gustatory

Response-Dependence and Aesthetic Theory  313 aesthetic properties. Plausibly, a landscape is not sublime merely when it in fact produces awe under normal conditions. It seems to require someone to be perceptive, sensitive, and reflective in order to recognize the sublimity—­conditions which do not sound wholly descriptive. The view I’m suggesting sounds very much like an ideal observer theory. Indeed, it sounds very much like Hume’s ideal observer theory. His account suggests, at least on one reasonable interpretation, conditions that are not purely descriptive: delicacy of taste, freedom from prejudice, sound judgement and good sense. To the extent that we build anything normative into the conditions, we move away from dispositional theories as I have characterized them here. Let’s call ideal observer theories that build something normative into the idealizing conditions normative ideal observer theories and reserve the term descriptive ideal observer theories for those whose idealizing conditions are purely descriptive. One motivation for normative ideal observer theories is that it is plausible or at least possible that a great many perfectly normal people under perfectly normal conditions are mistaken about what in fact bears the evaluative properties in question. They can be mistaken about what is in fact balanced or what is in fact delicious. And a great many perfectly normal people under perfectly normal conditions can be mistaken about what is in fact beautiful, even if they experience disinterested pleasure in response to the object. In contrast, it would be surprising if a great many perfectly normal people under perfectly normal conditions were mistaken about what things are red or what things are bitter. In fact, we typically find it disturbing when there is widespread experiential divergence on these matters (think of the infamous black-­and-­blue/white-­and-­gold dress), unlike in the case of evaluative properties. Furthermore, in the case of evaluative properties, we can often offer justifying or normative reasons in support of our aesthetic verdicts, where in the typical dispositional cases, we can only point to an explanation. In contrast to dispositionalism, a fittingness account of ‘balanced’ will say that what makes a dish balanced is that it would be fitting to feel gustatory pleasure in response to its different flavours and the way their respective strengths combine. It will say that what makes a dish delicious is that it would be fitting to feel gustatory pleasure in response to eating it. Similarly, a joke’s funniness ultimately depends on the fittingness of amusement, charmingness on the fittingness of charm, sublimity on the fittingness of awe, and beauty on the fittingness of (disinterested) pleasure. Although they may look quite different, normative ideal observer theories are a species of fittingness theories. A normative ideal observer theory, in spelling out some or all of the conditions for when an observer counts as an ideal one, simply offers a bit more explanatory bulk than the very schematic fittingness theory. It tells us a little more about when responses are fitting. But it, too, ultimately says that the gustatory properties obtain in virtue of whether certain responses are fitting. And though different fittingness theories will cast, for example, the pleasure or

314 

Fittingness

the ‘fit’ relation differently, this is certainly true of variations in normative ideal observer theories, too. None of these variations are deep structural differences. Thus, as I present them, descriptive ideal observer theories are dispositionalist theories and normative ideal observer theories are fittingness theories. Both are response-­dependence theories.6 This is important because, though the distinction between dispositionalist and fittingness theories is readily apparent in their most popular formulations, the distinction between descriptive and normative ideal observer theories is not. So far, fittingness theories have an advantage over dispositional theories. Fittingness theories capture ways that people can be mistaken which dispositional theories cannot. But there is one important disadvantage that fittingness theories face. It is not an objection to the theories as such, but a way in which they are less useful than dispositional theories in solving the Humean tension. As I characterize them, fittingness theories explain aesthetic properties in terms of the fittingness of certain responses (and normative ideal observer theories are one way of fleshing this out). These theories do not attempt to give an analysis or discuss the metaphysical status of fittingness itself. They only push the normativity around. So whereas dispositionalism necessarily offers a descriptive analysis of normative properties like beauty, fittingness accounts do not do this because the very ­concept of fit is normative. Such a theory therefore leaves open an important metanormative question about the status of the (normative) fittingness facts. Is the fact that it would be fitting to feel pleasure itself non-­naturalistic? Or can we give it a naturalistic reduction? One might go either way here. One might be a (reductive) naturalist about fittingness facts—­but such a fittingness view will reduce to a descriptive dispositional theory. Alternatively, crucially, one might also be a non-­naturalist (or, perhaps, a non-­reductive naturalist) about fittingness facts. On this latter view, fittingness accounts are not purely descriptive. This means that fittingness accounts do not and cannot by themselves satisfy the Humean demand for un-­spookiness.7 For the aesthetician who likes response-­dependence and is motivated by the Humean tension above, the fittingness theory may not be wrong, but it can only be a partial solution. However, if I’m right that a key motivation for response-­ dependence was its promised resolution of this tension, then this conclusion about fittingness theories is no small matter. Though they fare better than dispositional theories in their analysis of evaluative aesthetic concepts, they also lose a primary raison d’être. 6  More precisely, the dispositional theories we look at here, which depend on responses, are versions of response-­dependence. Also, I take no stance here on whether it is possible to offer a descriptive or normative ideal observer theory that is equivalent to any dispositional or fittingness theory. 7  Compare Gorodeisky (2021), who combines her hedonic fittingness account of aesthetic value alongside a tentatively ‘primitivist’ account of that value: ‘a value that need not be analysed further’ (278).

Response-Dependence and Aesthetic Theory  315

14.3  Response-­Dependence and Hedonism There are additional worries for response-­dependence accounts. In this section, we will examine whether hedonism’s frequent pairing with such views poses any serious problems for them. It will not, though this common pairing also, maybe surprisingly, favours fittingness theories over dispositionalism. Then, in the next section, we will examine a very different class of aesthetic properties—­one that does pose a problem for response-­dependence. Both theories have been thought to work best when supplemented by hedonism. Of course, some of the above analyses don’t involve pleasure at all, but more robust emotional or emotion-­like responses such as amusement or awe. But relatively thinner aesthetic properties like deliciousness and beauty have often been analysed in terms of hedonic responses. The problem here is that aesthetic hedonism is not obviously correct. Aesthetic hedonism holds that the characteristic, constitutive, or appropriate response to beauty is some kind of pleasure.8 This might be disinterested pleasure (i.e. pleasure divorced from any relationship to one’s personal or pre-­existing ends), sensory pleasure (of the kind one might feel when drinking a cool glass of water or stepping into a warm bath), formal pleasure (i.e. pleasure occasioned by the formal properties of the object, such as shapes and colours or pitches and duration), or pleasure individuated in some other way. I take these to be at least intensionally different ways to pick out aesthetic pleasure, if not also extensionally divergent. But though hedonism is historically popular, recently a spate of objections has been raised against the view.9 To briefly note a few, there is the paradox of tragedy or the paradox of painful art: that some art is good or beautiful despite not producing pleasure in normal or idealized conditions (Smuts (2007)). There is the worry that it sounds completely sensible to say both that something is beautiful and that one doesn’t like it (Hanson (2018)). And there is the worry that what makes a beautiful thing good or reason-­giving cannot be fully explained by reference to the goodness or normativity of pleasure (Shelley (2010)). This is not the place to present a thorough refutation of hedonism, but I want to illustrate that hedonism is not as obvious as it may at first appear, and it would be good for response-­dependence theories not to be committed to it. Fortunately, they are not. The most promising way to retain response-­dependence while moving past aesthetic hedonism, I’ll argue, are with theories that invoke pro-­attitudes. But before I come to those, I want to briefly examine two other possibilities. First, we might say that what makes something beautiful is that it disposes one to or is fitting to prefer it. But because this version builds in preference, it only offers comparative judgements rather than absolute ones. That is, we cannot explain the 8  See fn. 2, as well as Beardsley (1982), Mothersill (1984), Matthen (2017), Gorodeisky (2021). 9  See Van der Berg (2019) for an overview of recent defences of and challenges to aesthetic hedonism.

316 

Fittingness

property beauty but only the relational property more beautiful than—­or if we can explain beauty, we can only do so ultimately by way of more beautiful than. Second, there is a common variety of fittingness theory that analyses value in terms of desires. According to these theories, what makes something good is that it is fitting to desire. These desire-­based accounts fare slightly worse than other theories, I think, when it comes to aesthetic value. Is it true that what makes an aesthetic object beautiful or good that it is fitting to have some corresponding desire? What would it be a desire for? Possession doesn’t seem right; we don’t want to possess all beautiful objects. Perhaps a desire for a state of affairs in which that object exists or continues to exist? But there are transient works whose transience makes them all the better, and even then it’s the state of affairs that comes out as good rather than the object itself. Perhaps a desire to engage with the object? But features other than beauty or aesthetic value can make desire for engagement fitting, it seems. More promising for aesthetics are response-­dependence theories that invoke pro-­attitudes. The characteristic or fitting response might be something like appreciation. Such an account would hold that what it is for something to be beautiful is for it to either dispose audiences to appreciate it (dispositional) or for it to be fitting to appreciate it (fittingness).10 We would of course have to spell out exactly what kind of appreciation is relevant, since if we want it to account for beauty or aesthetic goodness, we have to know how it is aesthetic as opposed to, say, moral.11 And we have to say exactly what we mean by appreciation, since on some accounts, appreciation is more action-­like and less response-­like.12 But that aside, this seems pretty good so far. If we plump for this response-­dependence theory, can we say anything about the choice between dispositional and fittingness strands, or between descriptive and normative ideal observer theories? I think so. The dispositional strand will say that what it is for something to be beautiful (say) is for it to dispose one to appreciate it under certain, descriptively specifiable circumstances. This looks even worse than the above dispositional theories. While it is somewhat plausible that an object’s disposing us to feel pleasure could explain what makes it good, it is much less plausible to say that an object’s disposing us to appreciate it could be what makes it good. Imagine an artwork that flatters its audience in some way, and so they are disposed to appreciate it. The dispositionalist now faces a dilemma. If  the artwork is not good, the analysis fails because the disposition is present without the corresponding aesthetic property. If the artwork is good, we have the 10 Here I’m assuming the standard use of ‘appreciation’ in aesthetics on which it is positively valenced, so that one does not ‘appreciate’ a disgusting meal in correctly recognizing it as disgusting. 11  See Harold (2008). 12  See, e.g. Cross (2017). But notice, too, that if appreciation is the core aesthetic response and it is fundamentally action-­like, then the desire theory might look comparatively more attractive. Then, one might analyse beauty in terms of what it would dispose one or be fit for one to want to do.

Response-Dependence and Aesthetic Theory  317 right answer but the wrong reason for it. Surely a more flattering artwork is not thereby a better one, but this view has a hard time explaining why. The general problem is that the dispositionalist relies on correlations between the object and one’s reactions, but many things dispose us to appreciation that are irrelevant to quality. Fittingness theories improve on this by requiring more than correlation: appreciation must fit the object. We’ve seen that fittingness theories face fewer problems than dispositionalism. Dispositionalism loses plausibility when unmoored from hedonism, but hedonism itself is questionable. But this does not mean that fittingness theories are without fault. They face a related worry that has been addressed in detail elsewhere, but which I want to flag. The biconditional—­that something is beautiful if and only if it is fitting to appreciate it—­is very hard to deny. (Again, this puts them a step ahead of dispositionalism, whose corresponding biconditional—­that something is beautiful if and only if it disposes us to appreciate it—­is easier to deny.) No-­priority views reject the priority of either side of this biconditional over the other. But fittingness views endorse the priority of the right-­hand side: whenever something is beautiful, that is because it is fitting to appreciate. And that is something that we may want to reject. A familiar refrain from existing critiques of fittingness theories is that the value is not to be explained in terms of the fittingness of the response, but that the fittingness of the response is to be explained in terms of the value (or the properties that give rise to the value). In other words, there’s a kind of Euthyphro problem here. Is appreciation fitting because the object is beautiful, or is the object beautiful because appreciation is fitting? The former seems the right answer, but fittingness theories appear to answer with the latter. One reply that its defenders sometimes offer is this: it is the lower-­level properties or the nature of the beautiful object that explains why it is beautiful, as well as why it is fitting to appreciate it (Scanlon (1998: 391), Rowland (2019: 65ff.)). Though I think this is a quite powerful worry, it has received attention already in metaethical debates, and I think that the aesthetic domain has neither more nor fewer worries than ethics on this score. I will thus set this concern aside, since what I am most interested in are advantages or disadvantages for response-­dependence theories that are distinctive to aesthetics.

14.4  A New Problem: Insight and Perspective In many respects, the advantages and disadvantages of response-­dependence theories are comparable in ethics and aesthetics. Many of the traditional worries for response-­dependence theories translate seamlessly to aesthetics: the problem dispositional theories have accounting for evaluative properties and the Euthyphro problem, for instance. But aesthetics strikes most people as more amen­able than ethics to response-­ dependence theories, to the point that some

318 

Fittingness

response-­dependence views (whether for ethical value or for all value) are ex­pli­cit­ly motivated by aesthetic examples (Wiggins (1998), D’Arms and Jacobson (2000b)). I suspect this is because it seems to many that aesthetics has to do with the senses, where ethics doesn’t; that aesthetics is focused on experiences, where morality is focused on action; that aesthetics leans more subjective than ethics does, and that this collection of features renders aesthetic response-­dependence more intuitively appealing. However, these assumptions aren’t totally uncontroversial, and I and others have argued against these ways of thinking. But even if these aren’t genuine disanalogies between ethics and aesthetics, that would only put aesthetics on a par with ethics as far as response-­dependence theories go. Here, I am interested in a different kind of problem for response-­dependence theories in aesthetics, in both its dispositional and fittingness forms. This problem does not, I think, have a ready equivalent in ethics. To see it, we need to look at two classes of more complex properties of artworks. Let’s call the first class of properties insight properties. In addition to being funny or beautiful, artworks can be insightful. They can help us understand ourselves and our relationships more profoundly, add dimension and texture to other­wise flat pictures of the world, reveal patterns and depths that we hadn’t previously imagined. They can also be provocative, challenging our views and preconceptions of the world, and they can be thought-­provoking, stimulating us to reflect on things previously unconsidered or overlooked. I group them together because they are all shades of the same phenomenon, where art offers us something cognitive that isn’t quite reducible to propositional knowledge. Each of these is also a way in which art can be valuable. Although ‘provocative’ and ‘thought-­provoking’ aren’t obviously evaluative terms, we do in many instances use them to bestow aesthetic praise (in a way that we wouldn’t simply, for ex­ample, point out how symmetrical or how blue a painting is). This isn’t to say that art that has these properties is always good, all things considered, and certainly not that such art is always beautiful. But when present, they can be good-­making features of artworks. The second class of properties are perspective properties. Artworks can give us new perspectives and shift our existing perspectives. I take the concept of perspective from Elisabeth Camp (2017a, 2017b). She develops the notion in the context of metaphor and language, and I here want to broaden it to an important feature of art more generally. In describing perspectives, she offers the slogan that they ‘are tools for thought, not thoughts in themselves’ (2017a: 79). We have, at any given time, a way of parsing and processing the information that surrounds us. We read people’s gestures and sift emails for tone and significance, noticing and emphasizing some features while ignoring or downplaying others. But art can frame experiences in such a way that we are given new ways to process information. One might leave a theatre after having watched some French new wave cinema and feel the smallest, most quotidian things to be suddenly infused with symbolic

Response-Dependence and Aesthetic Theory  319 import and existential pregnancy. Or one might watch a lot of romcoms and start to see the world in terms of meet-­cutes and dramatic declarations of love in the pouring rain. Less narratively rich artworks, too, may do this: Barbara Hepworth’s sculptures help us see the spaces between objects, just as John Cage helps us hear the spaces between sounds.13 A perspective, as Camp has it, is an interpretive toolkit, a hermeneutical lens. Art can also be valuable by possessing these perspective-­shifting and perspective-­enhancing properties. Indeed, like insightfulness, it is one of the things we take to be most centrally valuable about artworks. Art can express and communicate feeling; art can help us see through another’s eyes. Platitudes like this can be read as articulating the fundamental (and fundamentally artistic) value of these perspective properties. These two classes of properties pose a serious obstacle to response-­dependence. In short, it is hard to see them corresponding to certain responses when we understand responses in the relevant sense. The remainder of this section will flesh out this argument, and the final section will point out ways that a response-­ dependence theorist might be able to move past this. Before we look at the real problem, there is a prima facie worry: there is not always a single characteristic or most fitting insight or perspective to come away with. Compare awe and the sublime. Supposing that the characteristic response to the sublime is awe is perfectly compatible with the sublime also disposing us or making it fitting for us to have some other response. But those other responses cannot be part of what it is for something to be sublime, or else they too would appear in the analysis. Furthermore, assuming this analysis, the sublime cannot dispose or make it fitting for us to have a response that is contrary to or precludes awe. Notice how different this is for perspective properties. A French new wave film like Jean-­Luc Godard’s Breathless may turn us towards a world suffused with ex­ist­en­tial gravity, or it may, alternately, turn us towards a world that is suddenly lighter and sillier than before. A romcom like Nora Ephron’s You’ve Got Mail may make us see a world of drama and destiny against which everyday love pales by comparison, or it may make us see a world where even those simplest loves can be the greatest (after all it’s just two ordinary AOL users, not Romeo and Juliet or Paris and Helen of Troy). A novel like Toni Morrison’s Beloved may leave you seeing the world more spiritually and ephemerally, thinking of things past and lives beyond this one, or it may leave you seeing the world more materially and earthly than before, attending to real, final deaths in the real, final ground. Artworks contain multitudes, and this isn’t just a comment on the plurality of possible in­ter­pret­ations, but on the visions of the world they offer and teach us to see with. Similar remarks apply to insight as well: existential works can help us recognize

13  Thanks to Samantha Matherne for raising this issue.

320 

Fittingness

the true fallaciousness of human ambition, or instead recognize human resilience and ability to forge meaning. This might seem to be a problem because not only do these examples involve multiple equally dispositionally likely or equally fitting responses, but they are  opposed responses. Response-­dependence theories necessarily rely on the characteristic or fitting attitudes to help unify and isolate the target properties. Something seems to have gone wrong if there are two equally fitting but contrary responses. In reply, the response-­dependence theorist might analyse the property disjunctively, in terms of either eligible response. I think this reply is fine but has limited power to explain richer cases where there is no specifiable set of responses. A better reply holds that the property in question is not perspective-­p1-­producing, but perspective-­producing; not insight-­i1-­producing, but insightful. That is, the property in question is compatible with the artwork producing any change in perspective at all or any insight at all. It’s no problem that some licensed perspectives are incompatible, or that some licensed insights are incompatible. There is a more serious problem, however, and it has to do with what counts as a response. There is a use of ‘response’ where we say that we respond to a question by answering it or respond to injustice by protesting it. This is not a response in the relevant sense. Response-­dependence theories typically characterize responses as (a) attitudinal mental states that are (b) directed at the object in question. Paradigm response-­dependence involves fearing the fearful, being ashamed of the shameful. They don’t have to involve cognate terms, of course—­we can also be amused by the funny and admire the good—­but those are tidy examples. Notice that these are all psychological states (many are emotions) rather than actions, and that they are directed at the object in question. What does one fear? The object of fear, the fearful. What is one amused by? The funny joke. We can also extend this to mental states like desire, admiration, and appreciation. All of these attitudes are meant to be directed at the thing that is good or valuable. Insight and perspective don’t work quite this way. Insight is at least a mental state; our having an insight is obviously a thing that happens in the head, even if it is in some sense factive. We might also squint a bit and be able to see insight as a psychological attitude in the vein of the above examples. However, the other condition is not met. The insights delivered are not insights concerning the work itself. Though it’s true that the work is the cause of the insight, and we therefore call the work (rather than the world) insightful, the insights themselves are directed towards the world outside the work. So we have a violation of condition (b). Perspective, too, is plausibly in the head, even if it’s very complicated. That said, to adopt a particular perspective is not exactly to enter into a particular mental state. It is rather to enter into what we might think of as a genus of mental states. Metaphorically, it is a frame of mind rather than a particular image in the frame. This itself might not be such a serious problem, but what it means is that

Response-Dependence and Aesthetic Theory  321 these states, too, are not directed at the work itself but at the world outside it, and again a violation of (b). This is the first wave of the problem. In reply, the response-­dependence the­or­ ist may deny that they are committed to condition (b). Paradigm responses may happen to satisfy it, but that doesn’t mean that responses of the relevant kind must do so. How then does the response-­dependence theorist distinguish the preferred responses from everything else? This brings us to the second wave of the problem. Let’s focus first on fittingness theories.14 An important distinction in thinking about fittingness is between two kinds of reasons to have a response. Some are reasons of the right kind, and some are of the wrong kind. Admiring a tyrant because he will hurt you unless you admire him is a wrong kind of reason to admire the tyrant. Admiring a tyrant because, though tyrannical, he is very loyal to his partner is a right kind of reason to admire him. Intuitively, the latter contributes to his being worthy of admiration (i.e. to his possessing the value in question), while the former doesn’t—­it simply makes it to one’s advantage to admire him. Take also the case of morally dubious jokes. Justin D’Arms and Daniel Jacobson (2000a) argue that the moral reasons to not be amused by the joke do not affect how funny it is, since a joke can be hysterically funny but morally suspect. Their conclusion about jokes may not be right, but on this view a moral reason to be amused by a joke is the wrong kind of reason to be amused; it does not make amusement fitting. Again, this distinction is important for fittingness theories because it provides necessary clarification to when responses are fitting. And this matters because the fittingness of relevant responses determines when an object has the value property in question.15 Now, when a work is insightful (or offers us a new perspective), our reason to appreciate the artwork is because it will provide us with certain insights (or offer us a new perspective). In appreciating the work, I gain a new insight or perspective, and this is one thing that gives me reason to appreciate it. But this looks exactly parallel to paradigmatic examples of the wrong kind of reasons.16 As with 14  Thanks to Rich Rowland for helpful discussion of this point. 15  To clarify, the ‘reasons’ terminology is not essential to fittingness theories, and thus neither is the letter of this distinction. Still, the spirit of it is essential, since a fittingness theorist needs some way of distinguishing fitting from unfitting (but otherwise good or desirable) responses. This is a common way of doing so, and the idea that guides the distinction is apparent enough, even if purists might prefer it framed without reference to reasons. Cf. Chappell (2012), Howard (2019). 16  There are more technical, though controversial, ways of drawing the distinction that may be helpful for some readers. If one prefers to distinguish object- and state-­given reasons à la Schroeder (2012), the fact that an artwork can provide insight and perspective speaks to a beneficial state it can produce, not to the object itself. Thus, they look like state-­given reasons, which are the wrong kind of reasons. If we instead use the dual role conditions that Rabinowicz and Rønnow-­Rasmussen (2004) offer, then the insight or perspective we gain occurs in the intentional content of our attitude. Thus, it again looks like the wrong kind of reason.

322 

Fittingness

the tyrant, something about what the appreciation does for me is what drives my appreciation. If this is right, then the fact that an artwork is insightful or can prod­uce a change in perspective is the wrong kind of reason to appreciate it. Insight and perspective properties therefore, on this view, cannot be ways in which an artwork is valuable. But surely this cannot be correct. Still, you could admire the tyrant, not directly because he will harm you if you don’t, but because of his grit or his moxie, which is evidenced in his threats. Maybe insight and perspective properties are like that. I don’t think this is quite right. Part of why you (rightly) appreciate the artwork is that you gain insight or that you gain a shift in perspective. Of course you may also appreciate the fact that the artwork is such as to afford new insights or perspectives, but that is not all that’s going on. Imagine admiring a friend who will give you a nice gift if you recognize their generosity. A right kind of reason to admire the friend, it seems, is that they are generous (supposing that they indeed are). A wrong kind of reason to admire the friend is the very fact that they gave you a nice gift if you do so. But with insight and perspective properties, this is just the kind of analysis that I dispute. Here, the seeming wrong kind of reason really does look like a right kind of reason to appreciate the artwork. The problem, in sum, is that it is just quite hard for fittingness theories to capture the full range of aesthetic values. Are dispositional theories better off? They do not make, nor do they need, any such distinction between the right and wrong kinds of reasons. But the cor­res­ pond­ing problem for dispositional theories is that, in order to remain responsedependence theories, they need to explain when something counts as a response to the object as opposed to simply an effect the object produces. Think of a complex piece like Sleep No More, an interactive retelling of Macbeth in a New York warehouse-­turned-­theatre. In this piece, participants enter the building and move freely through different rooms to uncover and piece together parts of the story. Responses to this might be discomfort, excitement, or appreciation. This is unlike non-­response effects of this, which might include the feeling of cold, the experience of having one’s vision obscured, or the sensation of having one’s wrist grabbed by an actor. Without a condition that bears on the directedness of one’s response towards the work, it is hard to see how to distinguish these. Compare colour properties, the usual model of dispositional theories in aesthetics. What it is for something to be red is for it to produce the redness sensation in us, but in such a way that this sensation is directed at it. We wouldn’t call a pill red just because ingesting it caused a redness sensation in us. We only call the pill red if it produces the redness sensation in us in by virtue of its surface reflectance properties under normal lighting conditions, i.e. only if it looks red. And one natural way to capture that is by saying that our red perception is directed at the pill or takes the pill as its object. So the dispositionalist must also embrace directedness, but directedness is what generates the problem that insight and perspective properties cannot be valuable. Thus, dispositionalism too faces a form of the insight and perspective problem.

Response-Dependence and Aesthetic Theory  323 A final note is in order. I have characterized all fittingness theories and many dispositional theories as species of response-­dependence. But it may appear that, somewhere along the line, I lost the essence of what made response-­dependence attractive to the aesthetician in the first place: the plausibility that aesthetic properties and therefore aesthetic value correspond to emotion or emotion-­like, object-­directed attitudes. This simple (but perhaps not naïve) view is the his­tor­ic­ al­ly popular one, and it is for this reason that I have dismissed accounts that some metaethical response-­dependence theorists find appealing: views on which actions count as responses. However, in discussing preference, desires, and appreciation the aesthetician may still see us as having strayed from that essence of plausibility in response-­dependence. I do want to insist that the present theories are versions of response-­dependence and that they are ultimately more de­fens­ ible, albeit imperfect. But maybe what this reaction reveals is another point where these theories, once made more sound, end up losing something that motivated their existence in the first place.

14.5  Ways Forward for Response-­Dependence This chapter has examined the prospects for response-­dependence theories in aesthetics, suggesting that fittingness theories have an advantage over dispositional theories and pointing out which fittingness theories are best suited to explain aesthetic value. But the last section saw some aesthetic properties that make the case for aesthetic response-­dependence look worse than the case for response-­ dependence in ethics. This is surprising in itself, since it is a way that aesthetic response-­dependence, which seems intuitively even more attractive than response-­ dependence in ethics, is actually more difficult to accommodate. For those very captivated by the promise of response-­dependence, there are three broad strategies for getting past insight and perspective properties. First, the response-­dependence theorist might just deny that insight and perspective properties are ways in which an artwork can be valuable. To lessen the blow, they might say that they’re ways in which the artwork can be instrumentally valuable (akin to admiring the tyrant so he won’t harm you), but not ways in which the artwork is intrinsically or finally valuable. While technically sound, this solution gives up the very strong intuition that these really are things that make the artwork better qua artwork and seems therefore best to avoid if possible. One might, however, admit that these are ways that the artwork could be better qua artwork but hold that it is not thereby aesthetically better. They contribute to the work’s artistic value, but not its aesthetic value. Such views are promising, but they rely on some of the same controversial assumptions and commitments that the D’Arms and Jacobson (2000a) view of jokes relies on. In particular, D’Arms and Jacobson commit themselves to a version of aesthetic autonomism, the view that the moral properties of a work cannot affect its aesthetic value.

324 

Fittingness

Response-­dependence views that avail themselves of the current proposal have a cor­res­pond­ing but stronger commitment, one which says that the cognitive properties of a work cannot affect its aesthetic value. On this view, there would be a core of aesthetic value that is best accounted for by response-­dependence, and everything else is an effect of that core set of valuable features. But they are not ways in which the artwork is itself valuable. This is, it should be noted, an even more extreme version of autonomism. Most philosophers of art are likely to reject this out of hand because this trail of thinking leads naturally (though perhaps not in­ex­or­ably) to extreme formalism, the view that the only genuinely aesthetic properties are purely formal ones. This view faces problems so serious that virtually nobody holds it anymore.17 As such, this option is also best avoided if possible. Both of the above strategies work by arguing that insight and perspective properties do not contribute directly to aesthetic value. According to these options, response-­dependence does explain all aesthetic value(s), but they achieve that result by shrinking the domain of the aesthetic, so that insight and perspective properties are no longer actually aesthetic properties. An alternative strategy instead shrinks the domain of response-­dependence so that it remains the best theory for some—­but not all—­aesthetic value(s). This strategy is compatible with response-­dependence being the best account of some properties, like the sublime (it disposes us to feel awe or merits awe), but it is at heart pluralistic, holding that response-­dependence must work alongside other theories to explain the value of art and other aesthetic objects. Interestingly, it may be compatible with response-­ dependence being the best account of thin aesthetic properties, such as aesthetic goodness and perhaps beauty, but not all of the thicker evaluative aesthetic properties. Whether this approach works depends on how the thick aesthetic properties are thought to factor into the thin aesthetic properties, and whether we can specify the appropriate response to thin aesthetic properties (e.g. appreciation) without essentially appealing to the other range of aesthetic emotions and responses to thicker aesthetic properties. (For example, whether we can talk of appreciating the thunderstorm without feeling awe at its sublimity, or of appreciating the novel without having our perspective shifted by it.) Regardless of what we say about thin aesthetic properties, this strategy gives up on the promise that Hume-­inspired response-­dependence theories had of accounting for all aesthetic value and aesthetic normativity. Because of the serious problems facing the first two strategies—­which shrink the scope of the aesthetic in unappealing ways, it seems to me that the best way forward for the response-­dependence theorist leaves us embracing response-­ dependence in at most a limited way. This may be unsatisfying to the Humean 17  A quick tour of those problems includes, for example, that it’s very difficult to determine when a property counts as a formal one, that it’s unclear what the formal properties are for art forms like literature, and that such restrictions on non-­visual and non-­auditory art forms are generally implausible.

Response-Dependence and Aesthetic Theory  325 aesthetician who sought to resolve the tension between the spookiness and subjectivity through response-­dependence. This does not mean that any standard of taste must in fact be spooky, but it means that, if there’s a thoroughly non-­ spooky solution, it won’t be response-­dependence alone.

References Beardsley, Monroe (1982), The Aesthetic Point of View. Ithaca, NY: Cornell University Press. Camp, Elisabeth (2017a), ‘Perspectives in Imaginative Engagement with Fiction,’ Philosophical Perspectives 31: 73–102. Camp, Elisabeth (2017b), ‘Why Metaphors Make Good Insults: Perspectives, Presupposition, and Pragmatics,’ Philosophical Studies 174: 47–64. Chappell, Richard Yetter (2012), ‘Fittingness: The Sole Normative Primitive,’ The Philosophical Quarterly 62(249): 684–704. Cross, Anthony (2017), ‘Art Criticism as Practical Reasoning,’ British Journal of Aesthetics 57(3): 299–317. D’Arms, Justin and Daniel Jacobson (2000a), ‘The Moralistic Fallacy: On the “Appropriateness” of Emotions,’ Philosophy and Phenomenological Research 61(1): 65–90. D’Arms, Justin and Daniel Jacobson (2000b), ‘Sentiment and Value,’ Ethics 110: 722–48. Gorodeisky, Keren (2021), ‘On Liking Aesthetic Value,’ Philosophy and Phenomenological Research 102: 261–80. Hanson, Louise (2018), ‘Moral Realism, Aesthetic Realism, and the Asymmetry Claim,’ Ethics 129: 39–69. Harold, James (2008), ‘Can Expressivists Tell the Difference between Beauty and Moral Goodness?’ American Philosophical Quarterly 45(3): 287–98. Howard, Christopher (2019), ‘The Fundamentality of Fit,’ in R. Shafer-Landau (ed.), Oxford Studies in Metaethics 14. Oxford: Oxford University Press, 216–36. Hume, David (1777/1987), ‘Of the Standard of Taste,’ in E.  F.  Miller (ed.), Essays, Moral, Political, and Literary. Indianapolis: Liberty Fund, 226–49. Jacobson, Daniel (2011), ‘Fitting Attitude Theories of Value,’ in Edward N. Zalta (ed.), Stanford Encyclopedia of Philosophy. Available at: Kant, Immanuel (1790/2000). Critique of the Power of Judgment. Cambridge: Cambridge University Press. Matherne, Samantha (2020), ‘Edith Landmann-Kalischer on Aesthetic Demarcation and Normativity,’ British Journal of Aesthetics 60(3): 315–34. Matthen, Mohan (2017), ‘The Pleasure of Art,’ Australasian Philosophical Review 1: 6–28.

326 

Fittingness

Mothersill, Mary (1984), Beauty Restored. Oxford: Clarendon Press. Patridge, Stephanie and Andrew Jordan (2018), ‘Fitting Attitude Theory and the Normativity of Jokes,’ Erkenntnis 83: 1303–20. Rabinowicz, Wlodek and Toni Rønnow-Rasmussen (2004), ‘The Strike of the Demon: On Fitting Pro-attitudes and Value,’ Ethics 114: 391–423. Rowland, R.A. (2019), The Normative and the Evaluative. Oxford: Oxford University Press. Scanlon, T.  M. (1998), What We Owe to Each Other. Cambridge, Mass: Harvard University Press. Schellekens, Elisabeth (2006), ‘Towards a Reasonable Objectivism for Aesthetic Judgments,’ British Journal of Aesthetics 46(2): 163–77. Schroeder, Mark (2012), ‘The Ubiquity of State-Given Reasons,’ Ethics 122(3): 457–88. Shelley, James (2010), ‘Against Value Empiricism in Aesthetics,’ Australasian Journal of Philosophy 88(4): 707–20. Sibley (1968), ‘Objectivity and Aesthetics,’ Proceedings of the Aristotelian Society: Supplementary Volume 42(1): 31–54. Simoniti, Vid (2017), ‘Aesthetic Properties as Powers,’ European Journal of Philosophy 25(4): 1434–53. Smuts, Aaron (2007), ‘The Paradox of Painful Art,’ The Journal of Aesthetic Education 41(3): 59–76. Van der Berg, Servaas (2019), ‘Aesthetic Hedonism and its Critics,’ Philosophy Compass 15: 1–15. Watkins, Eric and James Shelley (2012), ‘Response-Dependence about Aesthetic Value,’ Pacific Philosophical Quarterly 93: 338–52. Wiggins, David (1998), ‘A Sensible Subjectivism?’ in his Needs, Values, Truth: Essays in the Philosophy of Value. Oxford: Blackwell, 185–214.

SECTION FOUR

F IT T INGN ES S A ND R E SP ON SIBI LI T Y

15 Fittingness as a Pitiful Intellectualist Trinket? Michael McKenna

15.1 Introduction In diagnosing the alleged failure of his libertarian adversaries, P.  F.  Strawson (1962) once remarked that they sought to fill a gap in their proposal with an in­tu­ition of fittingness, which he characterized as a pitiful intellectualist trinket. Really? Pitiful? Despite my great admiration for Strawson’s essay, I’ve always found that particular remark to be a cheap shot, an especially ungracious dismissal of what instead should have been charitably explored. Indeed, I regard it as the worst thing about that otherwise inspiring essay. In what follows, my primary goal is to resist Strawson’s caricature of (an intuition of) fittingness and show that in the domain with which he was most directly concerned—­moral responsibility—­ fittingness plays an important role. I will proceed in three steps. First, I will offer a general characterization of fittingness, one that applies to a range of domains, such as the credible, desirable, and humorous. Second, I will scrutinize Strawson’s disparaging assessment of appealing to an intuition of fittingness and consider what specifically about the debate among his contemporaries might have led him to be so harsh in this regard. It is illuminating that at roughly the same time Strawson published ‘Freedom and Resentment’ (1962), in ‘Justice and Personal Desert’ (1963) Joel Feinberg instead approvingly compared fittingness and desert, including the arena of desert that was of central concern to Strawson, rewards and punishments as well as praise and blame. Why such different responses? Third, building on what we can learn from Strawson and Feinberg, I will turn to contemporary debates about free will and moral responsibility, showing what role fittingness currently plays in this domain. In doing so, I will argue that desert as it bears on blame and punishment is best construed as a species of fittingness, but also, if no one deserves blame or punishment because no one has free will, certain responses to wrongdoers will remain fitting. In this way, fittingness is a more basic and encompassing relation.

Michael McKenna, Fittingness as a Pitiful Intellectualist Trinket? In: Fittingness: Essays in the Philosophy of Normativity. Edited by: Christopher Howard and R. A. Rowland, Oxford University Press. © Michael McKenna 2022. DOI: 10.1093/oso/9780192895882.003.0015

330 

Fittingness

15.2 Fittingness Consider first appropriateness as featured in the following principle proposed by R. Jay Wallace in Responsibility and the Moral Sentiments: (N)  S is morally responsible (for an action x) if and only if it would be ­appropriate to hold s morally responsible (for action x). (1994: 91) Wallace argues that the burden of a theory of responsibility is to ‘specify the norms by reference to which the appropriateness of that stance [of holding morally responsible] is to be gauged’ ([my brackets]: 92). Treating correctness in similar fashion, Wallace remarks: The terms ‘appropriate’ and ‘correct’ are bland and noncommittal terms of general appraisal. To render the normative interpretation [(N)] more determinate, it will be necessary to specify particular substantive norms by reference to which the question of the appropriateness of holding people responsible might be answered.  ([my brackets]: 92)

I will understand the terms apt, appropriate, and correct as Wallace does.1 Whereas I will treat fittingness and fit as terms that name a distinct, substantive normative specification of appropriateness. So, where Wallace himself unpacks appropriateness as it figured in (N) in terms of fairness (1994), we might consider fittingness as a competitor proposal in a theory of moral responsibility, as has been recently argued by David Shoemaker (2015 and 2017). For those who would balk at this proposal, note that some think of desert as a species of fittingness (Feinberg, 1963 and McKenna, 2019, forthcoming), and while it might on its face seem unsatisfying to account for the normative basis for holding morally responsible exclusively in terms of fittingness, it is regarded by many as canonical do so in terms of desert (e.g. Carlsson, 2017; Clarke, 2013, 2016; Feinberg, 1963; McKenna, 2012, 2019, 2020; Nelkin, 2013; Pereboom, 2001, 2014; and Scanlon, 2008, 2013). So, what is fittingness? Although it is a matter of considerable controversy just what fittingness comes to (e.g. see Howard, 2018), I propose that we begin here: fittingness identifies a sui generis normative relation whereby a response fits an object in such a way that the response accurately represents its object (e.g. D’Arms and Jacobson, 2000 and Rosen, 2015). This explains why admiration is a fitting response to its object just in case that object is admirable, and likewise desire is a 1 See Feinberg (1970: 56–7), who follows the same convention. Note that other authors (e.g. Howard, 2018) instead treat apt and appropriate as synonyms for fit and fittingness. So far as I can tell, there is no substantive dispute here but rather just a choice of nomenclature. However, because in philosophical discussions appropriateness is used so liberally, it seems wise to reserve fit and fittingness for a more precise normative relation.

Fittingness as a Pitiful Intellectualist Trinket?  331 fitting response to its object just in case that object is desirable. A similar gloss applies when we say that humour is a fitting response to the humorous, belief is a fitting response to the believable,2 fear a fitting response to the fearsome, and so on. So too on such a proposal, and more directly to the issue of moral responsibility, blame is a fitting response to the blameworthy, praise is a fitting response to the praiseworthy, and guilt is a fitting response to being guilty. Fittingness also supplies a distinctive sort of normative force, varying across different domains. It assesses its object by reference to standards of evaluation internal to its domain (cf. Feinberg, on desert 1970: 61). In doing so, it supplies reasons for distinctive responses expressive of the appraisals it features. Thus, a morally offensive joke can be positively evaluated for being humorous insofar as it is fitting to be amused by it—­because it really is funny—­even if a moral assessment is negative (e.g. see D’Arms and Jacobson, 2000). Likewise, my belief that my son thinks his father is a tool is fitting in light of my credible evidence. This is so despite my desiring that he does not think this. Any reasons supplied by my desire that he not think of me in this way have no bearing on it being fitting that I believe this, what with all his eye-­rolling and so on. Reasons supplied by considerations of fittingness are pro tanto and cannot assure all-­things-­considered reasons to respond as fittingness counsels (cf. Feinberg on desert, 1970: 60). So, for instance, the morally offensive features of a joke could outweigh the pro tanto reasons to express amusement in response to fittingly finding it funny. Of course, though few note this, that cuts both ways. Maybe sometimes the reasons of funniness for expressing amusement through laughter outweigh the moral reasons not to do so. After all, maybe the joke was very funny and the moral considerations were not that weighty. This accounts for the scope of what are often referred to as wrong kinds of ­reasons (cf. Feinberg on desert, 1970: 59). The morally offensive nature of a joke is a wrong kind of reason to conclude that the joke is not funny, and the undesirability of my belief that my son thinks I am a tool is the wrong kind of reason for me not to believe this. Wrong kinds of reasons, then, are irrelevant to fittingness, and yet such reasons can bear either positively or negatively on the all-­things-­considered reasons to respond or refrain from responding in accord with fitting reasons. That wrong kinds of reasons can play a supporting role in relation to fitting reasons often goes unnoticed. While there can be moral reasons not to respond to a

2  As Chris Howard and Richard Rowland correctly note, some will protest here. Suppose there are two norms bearing on belief, a truth norm and a warrant norm. If fittingness is a matter of accurate representation, shouldn’t belief be a fitting response to what is true, not what is believable? In short, shouldn’t fittingness track the truth norm, not the warrant norm? I don’t think so, but cannot explore the matter here. Note, however, that this turns on whether we can make sense of a believer accurately representing a belief as believable. I think we can. (In correspondence, Justin D’Arms and Dan Jacobson raised the same concern.)

332 

Fittingness

humorous joke with laughter, there can also be moral reasons to so respond. It might signal a much-­needed expression of solidarity, for instance. The preceding point has special significance for those working on moral responsibility who are impressed by Strawson’s (1962) own reference to wrong kinds of reasons. Strawson scolded his compatibilist contemporaries who favoured utilitarian theories more generally for justifying our costly practices of blaming and punishing exclusively in terms of social utility. These sorts of consequentialist considerations were, Strawson complained, the wrong kinds of r­ easons for resenting or being indignant with those whom we blame or punish. This was key to Strawson’s own efforts to do better. But of course, reasons of utility—­at least in the central cases—­often favour our persisting in blaming and punishing those whom it was, by the standards Strawson endorsed, appropriate to blame or punish. Indeed, Strawson even went out of his way to emphasize that there was no reason to dispute that utility has some important justificatory role: ‘It is far from wrong to emphasize the efficacy of all those practices which express or manifest our moral attitudes, in regulating behaviour in ways considered desirable’ (1962, as reprinted in Watson, 2003: 93). Here I propose one further feature of fittingness that perhaps is too idiosyncratic—­but I don’t think so. A particular fitting response fits its objects in a case-­specific way. How so? It cannot be settled exclusively by appeal to general principles or algorithms (cf. Feinberg on desert, 1963: 57–8, and 82). For instance, as I have noted elsewhere (2012, 2019), a conversational reply to an interlocutor can be fitting or unfitting. That evaluation can only be settled by the particular meaning of an interlocutor’s conversational contribution within the context of the assumptions shared in that conversation. Consider instead belief. If p is believable or credible, then believing that p is fitting. Note just how case-­specific that relation of fit is. A believer’s believing that p is manifested in the myriad particular sentences she would be willing to assent to in virtue of her belief that p, and in the inferences a rational believer would take to be licensed by the content of p rather than some other equally believable proposition q. Indeed, we can get even more case-­specific than that. Two believers, S and R with nonoverlapping sets of beliefs might both come to believe p, but given S’s other beliefs where they differ from R’s, S might be licensed in making inferences unavailable to R. So too for R; R might be licensed in making inferences unavailable to S. The particular way S’s or R’s belief that p is fitting will thus vary.3 To the best of my knowledge, this case-­specific feature highlighted here is not explicitly recognized among philosophers who make use of fittingness. Hence, I am prepared to treat it as no more than a stipulation of the term as I alone intend to use it. Nevertheless, my strong suspicion is that this further feature is often

3  I offer further illustration of this feature of fittingness in subsequent discussion.

Fittingness as a Pitiful Intellectualist Trinket?  333 implicit in the way fittingness is understood (e.g. see Feinberg, 1963).4 Regardless, I commend it to readers as a useful notion. If some think of it as too far a de­part­ ure from the industry standard, call it fittingness*. This completes my presentation of fittingness. I offer one preliminary remark regarding its application in the arena of moral responsibility. Many contend that the proper way to unpack Wallace’s principle (N) is in terms of desert, rather than as Wallace does in terms of fairness (e.g. Pereboom, 2001, 2014). Below we will explore the relationship between fittingness and desert. At this point, all I will note is that every point identified so far regarding a feature of fittingness also applies to desert, at least as it is used by philosophers like Joel Feinberg (1963). This is key and will figure crucially in subsequent discussion.

15.3  Feinberg and Strawson on Fittingness and Our Moral Responsibility Responses Now consider first Feinberg’s and then Strawson’s mention of fittingness in those two prominent papers from the early 1960s. Bear in mind that Feinberg’s focus was importantly different. He was interested in an analysis of desert itself and in its application to a range of domains, including awards and prizes, assignments of grades, reparation, as well as issues bearing directly on moral responsibility—­ praise and blame, and reward and punishment. He didn’t mess with the metaphysics of free will. Strawson made specific mention of fittingness only as it bore on a particular metaphysical solution to the free will problem, one supporting a libertarian diagnosis. Thus, any putative relation between fittingness and moral responsibility for Strawson was only by way of his philosophical adversaries’ linking moral responsibility to a particular solution to the free will problem. What I  now want to show is that, in contrast with Strawson’s disparaging reference, Feinberg’s edifying appeal to fittingness to help understand desert was available to Strawson. Indeed, appreciating this helps us with an interpretive problem in understanding Strawson. Charitably construed, Strawson was, in all but name, embracing a relation of fittingness—­even if suggesting the label might have driven him to a state of apoplexy.

15.3.1  Feinberg on Fittingness In arguing that utility is not a desert basis for deserved treatment,5 Feinberg noted that reasonable people ‘naturally entertain certain responsive attitudes toward 4  On this point, I am indebted to conversations with Mark Timmons. 5  For essentially the same reasons that Strawson argued that it was not a proper basis for our blame-­and-­praise-­constituting reactive attitudes.

334 

Fittingness

various actions, qualities, and achievements’ (1970: 81), and he treated these as the proper candidates for deserved responses. Included in his list of responsive attitudes were two that were central to Strawson’s focus, gratitude and resentment. Each kind of response, Feinberg contended, has its own appropriate target, and these can be assessed in terms of fittingness. Hence, the target of a response might be unfitting, and in this way lack a ‘special kind of propriety’—such as glee in response to another’s suffering (1970: 82). Comparing these judgements to aesthetic judgements, he wrote: If this is so, then the kind of propriety characteristic of personal desert is not only to be contrasted . . . with qualification under a rule or regulation; it is also to be likened to, or even identified with, a kind of “fittingness” between one person’s actions or qualities and another person’s responsive attitudes. This suggests in turn that responsive attitudes are the basic things persons deserve and the “modes of treatment” are deserved only in a derivative way, insofar perhaps as they are the natural or convenient means of expressing the morally fitting attitudes. That punishment, for example, might be deserved by the criminal only because it is the customary way of expressing the resentment or reprobation he “has coming.”  (Feinberg, 1963, as appearing in 1970: 82)

Note two things about this passage, each of which is crucial for what is to follow. First, Feinberg entertains but does not commit to the thesis that desert is a species of fittingness. Instead, maybe desert should only be compared with fittingness. He says nothing else to settle the matter. Moving forward, I will take the liberty of assuming that for Feinberg desert is a species of fittingness, and in section four I will shore up this position. Second, Feinberg endorses a thesis that I will eventually reject: what a deserving agent most basically deserves are responsive attitudes (what Strawson would call ‘reactive attitudes’) rather than various modes of treatment. Focus for now on Feinberg’s contention that desert is a species of fittingness. Recall that, as noted in section two, all of the properties identified for fittingness are featured in Feinberg’s formulation of desert. I’ll not rehearse each point of comparison here. One, however, bears highlighting before we press on, the case-­ specific feature of fittingness. Feinberg contends that when a response is deserved, it is so just in virtue of a relation between a response and its target object’s desert base (1970: 58–61), where the desert base features reasons bearing only on features internal to the domain of evaluation. He calls these basal reasons (59). A person deserves to win a prize, for instance, exclusively in virtue of basal reasons regarding a winner’s ‘preeminent possession of the skill singled out’ (64). As an upshot, any particular prize for any specific exercise of any particular skill will be deserved or not in virtue of the reasons supplied by the particular skilful per­form­ance. Hold this thought, and now let us turn to Strawson.

Fittingness as a Pitiful Intellectualist Trinket?  335

15.3.2  Strawson’s Disparaging Dismissal of Fittingness Strawson called the compatibilists he criticized optimists. What they were optimistic about, as he framed it, was the prospects for retaining the legitimacy of our moral responsibility practices were we to discover that determinism is true. Of particular concern was the justification for potentially harmful practices of blaming and punishing. Strawson’s compatibilist contemporaries argued that the social utility of blaming and punishing were justified exclusively in virtue of regulating behaviour for moral improvement, and this sort of justification was compatible with determinism (e.g. Schlick, 1939, and Smart, 1961). They then reverse-­engineered, so to speak, the freedom conditions for moral responsibility so that freedom simply amounted to whatever features of agency are such than an agent is susceptible to being influenced by blame and punishment, as well as praise and reward. Talk about buying compatibilism on the cheap! Strawson’s criticism of his compatibilist contemporaries just was a wrong-­ kind-­of-­reason objection: You turn towards me first the negative, and then the positive, faces of a freedom which nobody challenges. But the only reason you have given for the practices of moral condemnation and punishment in cases where this freedom is present is the efficacy of these practices in regulating behaviour in socially desirable ways. But this is not a sufficient basis, it is not even the right sort of basis, for these practices as we understand them.  (2003, my italics: 74)6

Well, then, what will supply us with the right kinds of reasons? Strawson contended that if we focus upon the reactive attitudes we will exhaust all that is needed for a justification when we blame or punish a person whom we take to be blameworthy. Attention to the attitudes themselves, he argued, fills in the explanatory and the justificatory gap left by these optimistic utilitarian compatibilists. Unfortunately, Strawson protested, his libertarian contemporaries were blind to this insight. Strawson called these libertarians pessimists. What they were pessimistic about were the prospects for retaining the legitimacy of our moral responsibility practices were we to discover that determinism is true. Failing to register the salience of the reactive attitudes themselves, Strawson explained, they argued that more was needed to supply the justification lacking in the optimists account, something that requires heavy-­duty metaphysics—­not cheapo reverse-­engineering. Here, in light of this dialectical context, is the crucial passage:

6  Strawson was giving voice to libertarians who objected to this compatibilist proposal. But he clearly agreed.

336 

Fittingness

The pessimist does not lose sight of these attitudes, but is unable to accept the fact that it is just these attitudes themselves which fill the gap in the optimist’s account. Because of this, he thinks the gap can be filled only if some general metaphysical proposition is repeatedly verified, verified in all instances where it is appropriate to attribute moral responsibility. This proposition he finds it as difficult to state coherently and with intelligible relevance as its deterministic contradictory. Even when a formula has been found (‘contra-­causal freedom’ or something of the kind) there still seems to be a gap between its applicability in particular cases and its supposed moral consequences. Sometimes he plugs this gap with an intuition of fittingness—­a pitiful intellectualist trinket for a philosopher to wear as a charm against his own humanity. (Strawson, 1962, as appearing in Watson, 2003: 92)

Now why exactly did Strawson describe any libertarian appeal to an intuition of fittingness as a pitiful intellectualist trinket? It is difficult to say. The passage is cryptic. Strawson specifies no concrete proposal by any of his contemporaries. Instead he describes their efforts whatever they might come to derisively, and parenthetically alludes to the expression ‘contra-­causal freedom’. Here is my assessment of what Strawson must have meant. Let’s begin with some historical context. In roughly the fifty years leading up to Strawson’s essay, the viable options for libertarian freedom were either by appeal to agent causation of the sort developed by C. A. Campbell (1951),7 Roderick Chisholm (1964), and Richard Taylor (1966), or instead a noncausal theory of freedom espoused most notably by Henri Bergson (1910) and Jean-­Paul Sartre (1948), but also found in Wittgensteinian approaches to action theory more generally, not just free will, such as A. I. Meldon (1961). One of these strategies featured a distinctive kind of causal relation between agent-­as-­irreducible-­substance and action. The other featured a noncausal relation between agent and action. Either way, what both libertarian proposals required as a condition for an exercise of free will is the falsity of the proposition that antecedent to any basically free act, conditions obtained that casually necessitated the act itself. Both strategies accepted that the mere absence of a deterministic cause is not enough for the act to be free. The negative condition alone affords the agent no control over the act whereby she initiates it rather than it merely being something that just happened. This, I assume, is the metaphysical proposition that, Strawson complained, the libertarians of his time had difficulty stating coherently. It involved some positive account—­with noncausal or agent-­causal resources—­of the agent’s bringing about a free act in the absence of garden-­variety event causal production by necessitating prior events.

7  Campbell makes liberal use of the expression ‘contra-­causal’ in his formulation—­an ugly and misleading term for a view that features causation by agents.

Fittingness as a Pitiful Intellectualist Trinket?  337 We are not yet at the target of the ‘pitiful trinket’ dig. Suppose, somehow, these libertarians were able to offer a coherent and intelligible story about the nondeterministic freedom-­conferring relation between every agent and every basically free action that amounts to an exercise of free will. Strawson grudgingly entertained the possibility that some libertarian thesis might succeed. This is the stage Strawson envisages when in the passage quoted above he writes, ‘Even when a formula has been found . . .’ The worry now is this: Who cares? Why does this matter? Recall, the requirement for a justification involves a normative con­sid­er­ ation, not just a metaphysical explanation. What do we need to be able to justify harmful blaming and punishing practices if as it turns out appeal to utilitarian considerations only generate the wrong kinds of reasons? This, I take it, is what Strawson meant by complaining that ‘there still seems to be a gap between its applicability in particular cases and its supposed moral consequences.’ Now we can begin to see what Strawson had in mind by objecting to the idea that right here the libertarian can invoke an intuition of fittingness. The libertarian contends that when we get the metaphysics coherently stated, we can have a direct intuition of fittingness supplying a normative justification for blame and punishment. Such a justification was pitiful, Strawson thought. But why precisely? We still don’t have a full accounting of Strawson’s remark. Strawson characterized this libertarian’s appeal to an intuition of fittingness as a pitiful intellectualist trinket to wear as a charm against his own humanity. Why intellectualist? And why against? Why not as an expression of his own humanity? This, I take it, was what Strawson found especially pitiful. As indicated in the passage quoted above, Strawson’s objection to these libertarians was that they failed to appreciate the crucial facts about the relevance of the reactive attitudes. They did not lose sight of them, he told us, but they did fail to appreciate their significance. Strawson argued that these attitudes could supply the resources to fill the gap left by the failure of the utilitarian-­compatibilist justification. This sort of libertarian, as Strawson put it, ‘rushes beyond these facts’ (89) to excessive metaphysics. Here, I propose, is the specific focus of Strawson’s searing indictment. Suppose, as Strawson did, that the reactive attitudes and our expression of them in our adult interpersonal lives are manifestations of our humanity. The fittingness called forth by these libertarian metaphysicians, Strawson suggested, cast the freedom they identified as a justification for blaming despite these natural human responses rather than because of them. Their appeal to an intuition of fittingness is a charm against our humanity; it is by no means an expression of it, much less a commit to it. It is in this sense that Strawson as a naturalist protests to an ‘intellectualist’ libertarian metaphysics that he saw as in some way nonnatural.8 His complaint, then, in this biting remark, was that the 8  Some might even say supernatural. Recall the libertarian Chisholm’s (1964) remark that if persons have free will they must have a power normally only attributed to God, to be a prime mover unmoved.

338 

Fittingness

libertarians’ appeal to an intuition of fittingness as they used it in this specific dialectical context was not only normatively unhelpful, but actually positively opposed to those features of our humanity—­our actual emotional lives—­that are so in­tim­ate­ly connected with how in fact we hold people to account. It therefore would not answer any more effectively the moral question that the utilitarian compatibilists of his day were unable to answer. If this is correct, we do not have much reason to think Strawson was opposed to fittingness per se, but rather to the way that the libertarians of his day made use of it.

15.3.3  Why Strawson Needs Fittingness So, how does Strawson fill the gap? How does his attention to the reactive attitudes alone do the work that both his compatibilist and libertarian peers failed to do by looking beyond them? His reasoning here is not at all transparent, and it has led philosophers to go in one of two different directions in interpreting him. Strawson proceeded by focusing on the scope of our excuses, justification, and exemptions, wherein reactive attitudes of moral anger (resentment and indignation) are not appropriate. If we extrapolate away from such cases, as Gary Watson (1987) has observed in characterizing Strawson’s view, we are then left with the cases where there is not a defeater to the proper application of our moral anger. This moral anger for Strawson is the medium by which we hold to account when we blame and punish. So, this means there is a not a defeater to blame or punishment. But we need more. Strawson’s contemporaries were trying to show what positive justification we might identify when we blame and punish. With the resources he provided, he showed us when, according to our practices, we should not blame or punish. But what is the positive normative glue that justifies our doing so when no excuse, justification, or exemption applies? What licenses us being angry with and harming others with these potentially costly practices? Here is one in­ter­pret­ ation of Strawson that, if correct, is rightly damning: When an agent does act with ill will or an insufficient degree of good will, and when a potential blamer perceives this, by a brute bit of nature, she is just disposed, causally, to have a reactive emotional response. There is no more normativity to be found. There are just psychological causes.

This is a dispositionalist interpretation of the proper extension of being blameworthy and the conditions for blaming and punishing. This is just how we are built. End of story. Various critics have entertained the prospect that this dispositionalist in­ter­ pret­ation was Strawson’s view, a crude form of naturalism, built with no more than primitive psychological facts (e.g. Fischer and Ravizza, 1993: 18–9 and

Fittingness as a Pitiful Intellectualist Trinket?  339 Wallace, 1994: 85). Interpreting him in this way, they have proposed that what is needed is an upgrade by instead advancing a normative interpretation, one that departed from what Strawson intended. Indeed, this is Wallace’s (N) quoted above. But the textual evidence to support the dispositional interpretation of Strawson over the normative one is thin. Strawson is to be faulted for remaining silent where he should have said more. He needed to offer some positive account of the normativity that relates an angry blaming response to perceived ill will in a wrongdoer. Nevertheless, Strawson’s commitment to some positive basis is implicit in his remark that his compatibilist contemporaries supplied the wrong kinds of ­reasons for blaming and punishing by attending to considerations of social utility. If there are wrong kinds of reasons, aren’t there right ones? These reasons would offer the requisite normative justification. Moreover, when criticizing libertarians, he claimed that they sought to fill a justificatory gap left by the utilitarian-­ compatibilists (optimists), and his claim was not that there was no gap, but that attention to the reactive attitudes was sufficient to fill it. So, as I read Strawson—­ and admittedly I am offering a highly charitable interpretation—­he was implicitly operating with a normative understanding and took calling attention to the reactive attitudes as responses to good and ill will as providing those normative resources; he was not saying we don’t need any justification. Many Strawsonians will strenuously protest at this point. Strawson was insistent that all his contemporaries who were participants to the debate overintellectualized the facts by seeking some external justification. Perhaps this is why Wallace (1994) should be read as departing radically from Strawson’s naturalism. Despite his contention to the contrary, it might be argued, Wallace did offer an external normative (rather than metaphysical) justification for our harmful blaming and punishing practices; they are justified by reference to principles of fairness. Maybe so. Maybe we should understand Wallace as standing opposed to Strawson in terms of providing an external justification for the appropriateness of our reactive attitudes and their attendant responsibility practices. Regardless, there is no reason to think that the normative interpretation of Strawson I am suggesting runs afoul of his prescription that we not seek an external justification. Denial of an external justification does not change the fact that when we do become morally angry with someone who treats us or others with ill will, in treating them poorly in response, we ought to be able to supply some justification of the propriety of our doing so. Strawson was cool with justifications internal to our practices. This is just what fittingness of the sort Feinberg identified would do for Strawson. It would supply a justification internal to our standards for moral anger as a response to ill will. And it would do so in just the case-­specific way I highlighted in my presentation of Feinberg’s proposal. Here is what Strawson should have said in order offer a complete answer to his contemporaries regarding the justificatory gap between blameworthy wrongdoers and blaming and punishing responses. When we react with anger to the ill will or

340 

Fittingness

lack of sufficient good will perceived in the wrongdoing of another, our anger and our outward expressions of it by blaming or punishing are not just brute automatic reactions. They are appraisals of the wrongdoer’s regard for us or others.9 When no justification, excuse, or exemption applies, they are accurate appraisals. By standards of appraisal internal to the domain of their kind—­the domain regarding expectations of good will—­they supply us with reasons to fit our angry response to the particular objectionable nature of the offender’s will. This just is fittingness as I have specified it above.10 What I would say here, boldly and with woefully inadequate textual evidence, is that this is what Strawson (probably?) had in mind. Insofar as he did, far from eschewing the role of fittingness in philosophical theorizing, in all but name he was relying upon it.

15.3.4  Comparing Strawson and Feinberg on Fittingness and Desert Of course, Feinberg did not specify fittingness as the normative relation bearing on blame and punishment. He relied upon desert. Shouldn’t we interpret Strawson likewise? Shouldn’t the preceding section be rewritten in terms of it being desert that Strawson needed? That depends on just what distinguishes fittingness from desert. As I have read him, Feinberg’s thesis is that desert is a species of fittingness. On his view, what more is there to desert than there is to mere fittingness? Feinberg does not say. All of the logical properties he assigned to desert in his essay apply to fittingness as I characterized it above (sect. 2). If there are no differences, this question amounts to no more than a choice of nomenclature. Desert is just another name for fittingness, and Feinberg was wrong to contend that one was a species of the other; one just is the other. But if there is more to the relation of desert than there is to fittingness, it’s an open question how best to understand both Feinberg and Strawson. For all Feinberg said, maybe he was right to think of desert as a species of fittingness, but wrong to identify the relation he described in his essay in terms of desert. Maybe he should have just relied on the more generic notion of fittingness. If so, Feinberg is to be faulted for failing to put his finger on what is especially unique about desert. As I will argue below, there is a substantive issue here, a crucial point at which desert departs from mere fittingness. Granting this, it is unsettled how to chart a path forward for Strawsonians in our own time. A more modest strategy would

9  Compare Feinberg when discussing gratitude and resentment: ‘These attitudes are not mere automatic responses to stimuli, but self-­conscious responses to desert bases, not mere “reactions to” but “requitals for”’ (1970: 70). 10  This chain of reasoning is so close to Feinberg’s that I encourage readers to see his 1970, pp. 80–2.

Fittingness as a Pitiful Intellectualist Trinket?  341 attempt to make do with just fittingness, whereas a more ambitious approach would take on desert as a more robust normative relation.

15.3.5  Diagnosing an Alleged Mistake and the Current State of the Debate Before wrapping up our assessment of both Feinberg and Strawson on fittingness, it is worth reflecting on why Strawson’s harsh criticism of his libertarian contemporaries seemed so compelling. If he was correct, they were making a bad mistake. Well, were they? Our question is not just academic. I intend to use it as a lesson for our time. I have argued that Strawson’s dismissal of fittingness was limited to a particular application of it to supply the normative basis for blame and punishment, linking exercises of free will directly to appropriate blame and punishment. Grant to Strawson that these libertarians made this appeal to an intuition of fittingness in a way that extracted it from the context of the relevant moral emotions. As he saw it, they ignored the salience of these attitudes for their justificatory enterprise. These metaphysicians thereby applied this normative relation of fittingness outside the context of any domain that could help guide an application of the relation. Recall, as set out above, fittingness identifies a relation within a domain that contains its own standards of evaluation. These moral emotions provide that domain. Without such a domain, fittingness has no anchor, no substance to inform applications of it. Now if Strawson’s description of his libertarian contemporaries were accurate— ­if indeed they had looked past the crucial relevance of these morally reactive emotions—­he would have been dead right to protest that their appeal to an in­tu­ ition of fittingness was a pitiful intellectualist trinket, divorced from their own humanity. But I am doubtful. Here I am not interested in an historical exploration of his contemporaries. Instead we can use this as an occasion to transition to the contemporary debate. Regardless of what the libertarians of his day thought, it was open to a thoughtful critic to resist Strawson then, as it clearly is in our time, by pointing out that on a normative interpretation of his proposal, it is unsettled what sort of basal reasons pertain to fitting expressions of moral anger. Mightn’t there after all be ingredients internal to standards of appraisal for our anger that raise questions about an agent’s freedom? If so, the libertarian and other incompatibilists might find a basis for their metaphysical conditions on freedom that is internal to our pertinent interpersonal norms regarding expression of these emotions. With this, we can leap forward a good half century. Derk Pereboom (2001, 2014: esp. 128–9) contends, and many others agree, that a presupposition of moral anger and angry blame, at least in many familiar cases, is that the target of one’s

342 

Fittingness

anger deserves be a recipient of unwelcome treatment, issuing in harm or pain. A further presupposition of deserving such treatment is that its target acted freely or of her own free will, where the freedom at issue is metaphysically demanding. In short, angry blame presupposes the freedom of those at whom it is directed. Others deny that the anger pertinent to theorizing about moral blame and blameworthiness has such a strong desert presupposition built into it. (e.g. Shoemaker, 2015 and Vargas, 2013). In light of our contemporary understanding of the debate, Strawson’s easy dismissal of his libertarian adversaries was far too quick. We can embrace along with him the centrality of the reactive attitudes in theorizing about moral responsibility while nevertheless seriously entertaining the possibility that the satisfaction conditions for fitting anger involve freedom of the sort specified by libertarians. It’s an open question.

15.4  Desert as a Species of Fittingness: Two Paths forward for Strawsonians I will now argue that desert is a species of fittingness. With this in place, I’ll turn to the options for contemporary Strawsonians regarding how best to theorize about moral responsibility.

15.4.1  Desert Is a Species of Fittingness When we assert that the guilty deserve to feel guilt, the culpable deserve punishment, the blameworthy deserve blame, and that the praiseworthy deserve praise, we can assert with confidence that it is fitting for the guilty to feel guilt, fitting to punish the culpable, fitting to blame the blameworthy, and fitting to praise the praiseworthy. This is not limited to the responsibility concepts that are our focus here. If Jane deserves to have won the race, then her winning it is fitting. Indeed, whenever there is a true sentence of the form x deserves y, there is a true sentence of the form it is fitting for x to y, or instead, y is a fitting response to x. But not so in reverse.11 Consider above the examples of belief and desire. It is fitting to believe the believable, and fitting to desire the desirable. Yet there are many instances of belief wherein it is fitting to believe a credible proposition, while no consideration of desert is salient. The same for the desirable. Peanut butter and jelly sandwiches are obviously desirable (duh!). Surely then it is fitting to desire them. But they do not deserve to be desired.

11  Howard renders this crystal clear (2018: 7). My argument here builds on his insights.

Fittingness as a Pitiful Intellectualist Trinket?  343 Moreover, if we think of desert as a form of appraisal falling within the domain of justice, as I am now proposing we should, following Feinberg (1970: 55), the preceding point is even clearer. We do no injustice to peanut butter and jelly sandwiches in failing to desire them. Nor do our undergraduates do an injustice to modus ponens when they fail to believe its validity. Not so for failing to praise the praiseworthy or punish the culpable. In such cases, questions of potential injustice are at least intelligible, even if sometimes there are reasons to deny an injustice has been done in not giving a person what she deserves. Why the hedge? An example from Feinberg illustrates the point (1970: 64). A runner who is clearly fleetest of foot might deserve to win the race and receive the medal, but because she pulls a muscle and turns up lame someone else crosses the finish line first. In that case, no injustice is done in giving the less-­deserving runner the first-­ place prize.12 The preceding point needs to be handled with some care. Some deserved response, like blame, are optional, and there is no question of injustice if one fails to blame a person who deserves it. Still, other questions of justice still bear on these sorts of responses that don’t for mere fitting responses like belief or desire. An injustice is done when one is blamed who deserves not to be blamed, whereas no injustice is done when one unfittingly believes that it is possible to square the circle or that no human has walked on the moon. Have we arrived at a substantive basis for distinguishing desert as a species of fittingness? No, and for two reasons. First, for all that has been said, maybe the distinction rests on no more than linguistic convention. Maybe we just call fittingness within the domain of justice ‘desert’, but there’s no more to it than that. Second, maybe there is partial overlap in the extension of two distinct relations, so that they do not stand to each other as species to genus. Desert is one thing; fittingness is another. Taking the latter challenge first, here is one reason to think desert is a species of fittingness. When one of the properties of fittingness identified above fails to apply to a case, rendering it unfitting to respond in a certain way, it provides grounds for the judgement that it is undeserved. It thereby helps to explain assessments of desert. If my blaming Mati for Njeri’s poor conduct is unfitting because my blame does not accurately represent the wrong as properly attributed to Njeri and not Mati, that is the same reason why blaming Mati is not deserved. Consider instead the case-­specific feature of fittingness I have highlighted, and reflect on the familiar claim that the punishment should fit the crime. Here are two cases. Geraldine embezzles a modest amount of money from a homeless

12  Feinberg (1970: 57) explains this in terms of qualifying conditions or rules used to help set the stage for assessing questions of desert. The rule says that the first to cross the finish line wins the medal. But the aim of the rule is to thereby award the medal to the most deserving. Sometimes pro­ ced­ures for achieving this aim fall short.

344 

Fittingness

shelter, and Leslie gets drunk and throws trash about the streets in his home town. Both being culpable, suppose both deserve a mild punishment that causes only two units of harm.13 A judge considers two punishments, each of which, as it happens, would be received by both Geraldine and Leslie as equally unwelcome to the tune of those two units. One punishment involves spending three Saturdays cleaning the local streets. The other involves three Saturdays preparing food in the kitchen at the local homeless shelter. Were the judge to dole out as punishment kitchen work at the homeless shelter to Leslie, who trashed the streets of his hometown, and the street cleaning to Geraldine who embezzled money from the homeless, her allocation would in some manner be unfitting. It would also not optimally supply either with what each deserved. Now consider the former concern, that there is no real substance to the distinction. Desert just is fittingness, and in the sphere of justice we use the more limited term. It’s just linguistic convention. How might we resist this charge? One possibility builds on the contention that evaluations of desert entail judgements of noninstrumental goodness, whereas evaluations of fittingness do not. It is noninstrumentally good that the praiseworthy are praised, and noninstrumentally good that the culpable receive punishment and the pain or harm accompanying it (a central assumption for some versions of retributivism). Suppose this controversial thesis is true; desert entails noninstrumental goodness; the world is rendered better when the deserving get what they deserve. Regrettably, this won’t help distinguish desert from mere fittingness. There are cases of mere fittingness that plausibly entail noninstrumental goodness. For instance, it is fitting to respond with grief to the loss of a loved one, and while this response and the pain accompanying it might be noninstrumentally good (e.g. McKenna, 2019), the pain of grief is not deserved (Carlsson, 2017; Nelkin, 2013, 2019). Perhaps there is a principled way to show that reasons of desert are more forceful than reasons of fit, or have more demanding logical properties in relation to other reasons, so that we get a natural parsing wherein one is a species of the other. Consider, for instance, the debate over the role of the reasons supplied by promising, what some have characterized as exclusionary reasons, as in contrast with other reasons (e.g. Wallace, 2019 and Raz, 1977). One who has promised now has reasons that do not just outweigh reasons for other courses of action, but rather discounts or excludes them. If I have promised you we’d share dinner Saturday night, then a later chance to see a good movie with another friend is not just outweighed by my reasons for having dinner with you. They don’t count at all; they’re to be excluded from my deliberations. I’m committed by virtue of my promise. Perhaps we can make use of a similar distinction.

13  I don’t endorse this sort of hedonistic calculus. I use it as a toy model just to make a simple point.

Fittingness as a Pitiful Intellectualist Trinket?  345 Here is a first pass at such a strategy. Focus first on negative evaluations of desert, when what is deserved is unwelcome or harmful, as in the cases of (outwardly directed) blame and punishment. Desert supplies reasons to discount or silence—­and not merely outweigh—­moral considerations that otherwise apply. Grant that morality demands of us that when interacting with the moral community we do not, as Strawson himself expressed it, ‘acquiesce in the infliction of suffering’ (1962, as appearing in Watson, 2003: 90), at least not unless it can be justified for consequentialist reasons. Reasons of desert supply reasons to discount or silence that sort of moral reason and thus support forms of treatment otherwise opposed to morality’s counsel. Mere reasons of fit do not work that way. When reasons of humour outweigh moral reasons not to express our amusement at the funny joke, they do not silence reasons of morality. They just outweigh them. Not so for desert—­or so I contend. That, on the proposal I am exploring here, is the distinguishing feature of desert that sets it apart from mere fittingness. Will this do? Perhaps, but there are details that need to be worked out. For instance, blame differs from punishment in that in the case of blame there is no pertinent strict requirement or duty not to cause the harm to others (often) characteristic of blame. Granted, overt blame directed at the blameworthy often involves some harm or the risk of it. But in a wide range of cases there is no moral requirement bearing on these sorts of harms that we can clearly identify as the reason silenced by considerations of deserved blame. Why? Suppose blame is, at least in some cases, no more than a partial withdrawal of good will towards those at whom we direct any variety of angry blame; it involves withdrawal of a degree of good will that is otherwise expected of those who are in good standing within the moral community (they are not blameworthy). Arguably, we do not owe our good will or, say, our friendly relations to others.14 There is no requirement here. So, what is the moral reason discounted or silenced in cases of deserved blame where there are not clear moral requirements upon us? In reply, distinguish between requiring reasons and favouring reasons. Morality broadly construed can supply both. There are favouring reasons for us to show others some reasonable degree of good will in our interactions with them, even if we are not morally bound to do so. Desert can supply reasons to discount or silence these as well. Yet another challenge concerns positive deserved responses, such as praise and reward. Reasons of desert in positive cases do not discount moral reasons for not treating people with more good will than might be expected, or heaping a prize or reward on them. In the absence of a consideration of desert, no harm or wrong would be done to one who is praised or rewarded even if it is not deserved. True. But there are other reasons in the offing. It is commonly thought that gratitude is 14  Scanlon, for instance, focuses on this point about our interpersonal relations (2008). When, motivated by blame, we withdraw our previously friendly relations with others, our doing so (at least often) violates nothing that we owe to one whom we might blame. Morality allows us leeway here.

346 

Fittingness

owed towards those who have acted in praiseworthy ways and so benefited us, and that rewards are owed to those meritorious persons who have acted in morally heroic ways. Were we not to praise or reward them, we would fail to do something morality requires or at least favours. In the case of positive desert, then, we do not owe it to others to treat them as we do those who deserve our gratitude or our reward. The reasons discounted, then, are reasons about the moral neutrality or optionality of the sort of treatment involved in showing gratitude and giving awards. There are now requiring or favouring reasons for treating the deserving in positive ways. The distinction proposed here between the normative force of desert and the normative force of mere fit helps correct a mistake in Feinberg’s proposal. Feinberg contends that it is attitudes such as resentment or gratitude that are ba­sic­ally deserved, and various modes of treatment, construed as conventional means of expressing the attitudes, are only derivatively deserved. That can’t be right. If Jones deserves punishment and gets only clues of our resentment or indignation—­genuine boiling moral anger—­but receives no punishment, there is no sense to thinking, ‘Boy oh boy did she get what she deserved!’ Contrary to what Feinberg contends, I propose that a further difference between mere fittingness and desert is that in cases of the mere fittingness of pertinent attitudes, what is  basically fitting is some attitude. The fittingness of modes of treatment as expressions of those attitudes are a derivative matter. In the case of desert, what is basically deserved is a mode of treatment that expresses the pertinent attitude. Of course, the attitude is still essential. If two people are locked in a room for three days, one accidentally because she shut the door and the handle broke so she could not leave, and the other is locked in a room as an expression of moral anger via retributive punishment, both might be guilty and deserve three days in the pokey, but only one gets what she deserved. The other is just unlucky; karma got the better of her. This difference regarding the basicness of attitudes versus treatment is supported by the difference identified in the normative force of reasons as between fittingness and desert. When fittingness gives us reasons to express our amusement, it does seem that what is most basically fitting is the attitude, the finding humorous. Outward expressions of humour are then only derivatively fitting. But note that the reasons supplied by the fittingness of humour do not discount moral reasons not to express amusement even if the reasons of humour outweigh the moral reasons. The moral reasons in these cases are hardly silenced. On the other hand, supposing reasons of desert do operate by discounting or silencing other sorts of moral reasons, these reasons of desert, so to speak, ‘reach outside’ their domain of applicability and can defeat other normative evaluations. In this way, they recommend not just attitudes we might have, but ways we might act as an expression of these attitudes despite what would otherwise be reasons against

Fittingness as a Pitiful Intellectualist Trinket?  347 doing so. Thus, reasons of desert supply stronger reasons for acting than the sort offered by mere fittingness, and this helps explain why for desert but not for mere fittingness, what is basic is a mode of expressing an attitude in action. To avoid any confusion, the point here is not that desert supplies all-­things-­ considered reasons. Granting all that has been said, a person might deserve punishment, and not merely the angry attitudes of others, and still the reasons to punish her are only pro tanto. If some great tragedy would unfold by punishing her, like the destruction of the world, then for consequentialist reasons, all-­ things-­considered she ought not be punished. The point is only that reasons of desert exclude or cancel some other range of reasons that otherwise would supply pro tanto reasons to act in opposition to pertinent reasons of desert. Admittedly, the preceding proposal is just a sketch. It is an effort to find a way to argue for the contention that reasons of desert are in some respect more forceful than reasons of fit. It is also an attempt to show how and why it is that in the case of desert, the most basic forms of response are forms of treatment and not mere attitudes, as in contrast with mere fittingness for a range of attitudes. Suppose careful inspection proves the preceding sketch inadequate. Another option is to treat the thesis I defend here as a brute fact. Reasons of desert just are more normatively forceful than reasons of fit, and one of their distinguishing features is just that the basic responses that are deserved are forms of treatment and not mere attitudes. I leave the matter unsettled and press on.

15.4.2  Desert in the Domain of Moral Responsibility My proposal can be distinguished from the view that desert is a species of fittingness pertaining only to matters of responsibility (e.g. Miller, 1999 and Pojman, 1997).15 The problem with this view is that it excludes too much (cf. Feinberg, 1970: 56). Suppose no one is morally responsible for anything because, for instance, no one has free will, so no one deserves praise or blame, reward or punishment (e.g. Pereboom, 2001, 2014). Still desert remains. The swiftest runner still deserves to win the race and the slowest lose it, the good and bad students each deserve their good and bad grades, and so forth. In short, it is not plausible that desert is limited just to the domain of moral responsibility. So how should we situate moral responsibility in relation to desert? Just as fittingness applies differentially to the differing domains of the desirable, the cred­ible, the humorous, or the disgusting, so too desert applies differentially to different domains. Moral responsibility is but one domain among others. As noted 15  I too have just recently proposed such a view (McKenna, forthcoming), but for the reasons I set out here, I now reject this view.

348 

Fittingness

above, fittingness evaluates different objects by standards internal to different domains. So does desert. As Feinberg noted, the desert base for these different domains varies, and the basal reasons applying to them vary. Hence, just as ­reasons of fit that bear on something’s being desirable differ from those bearing on something’s being believable, so too do reasons of desert bearing on con­sid­er­ations of responsibility differ from reasons of desert bearing on grading or awarding prizes in a competition. Feinberg mentions five different varieties of  desert: awards and prizes; assignment of grades; rewards and punishments; praise, blame, and other informal responses; reparation, liability, and other modes of compensation (1970: 62). In doing so he made ‘no claim to taxonomic precision or completeness’ (62). To dial in on the desert distinctly bearing on the domain of moral responsibility, and to return to our reflections of Strawson’s relevance to our own time, we need to focus upon the desert base for deserved blame and punishment.16 This in turn needs to be settled in terms of the basal reasons in virtue of which assessments of deserved blame or punishment are settled. We can begin where Strawson (1962) began, and others before him, for ex­ample, Austin (1956), in ‘A Plea for Excuses’, with pleas meant to defeat judgements of blameworthiness. Exemptions show that a person is not a competent candidate for being able to comply with moral demands at all. Justifications show no wrong has been done. Excuses naturally divide into those that show a person was non-­culpably ignorant of wrongdoing, and those that show that she did not act freely. From these clues, we can construct a credible desert base consisting of a collection of basal reasons. When no plea applies—­that is, when a person is morally responsible in the sense of being blameworthy for something or other—­all of the following considerations bear on her deserving some sort of blame or punishment. First, she is a competent person capable of guiding her intentional conduct by reasons, grasping moral demands, and suffering from no major impairments of agency of the sort involved in severe mental illness. Second, she actually performed an act (or omission) that was morally wrong or in some manner morally objectionable. Third, she did so either knowingly or from culpable ignorance. Fourth and finally, she did it freely. Something like this set of basal reasons figures in capturing the desert base in virtue of which an agent might deserve blame or punishment.17

16  Praise and reward too, but we’ll not attend to these here. A further issue as well is the desert bearing on a person’s status as a morally responsible agent. The young man or woman who is no longer a child might speak plain truth to his or her father or mother by demanding that he or she deserves to be treated like a responsible adult. 17  I will work with this, although in accord with arguments I have made elsewhere (2012), a further Strawson-­inspired condition would also seem to be required—­that an agent acted with a morally objectionable quality of will. My commitments here don’t bear on the arguments to follow, so I set them aside.

Fittingness as a Pitiful Intellectualist Trinket?  349

15.4.3  Robust Freedom in the Distinctive Desert Base for Moral Responsibility Now focus just on the freedom condition. It is but one condition figuring in the total set of basal reasons in virtue of which an agent might deserve harmful blame or punishment. On the current proposal, the desert base provided by the full set of basal reasons for deserved blame or punishment must provide reasons of a kind that can discount or silence, and not merely outweigh, some (but not necessarily all) reasons not to cause persons harm or suffering. Given this, it is plaus­ ible to suppose that the freedom condition offered as part of the desert base should involve a relatively industrial-­strength degree of agential control over one’s conduct. Freedom and control come in degrees, and lesser freedoms seem in­ad­ equate. Dogs and lesser evolved creatures have some control over and so some freedom with respect to their behaviour. Notably, they can be influenced with the aim of improvement by rewards and punishments in the manner highlighted by the utilitarian-­compatibilists of Strawson’s time—­cheapo freedom. Clearly, a more substantial sort of freedom is called for. What might it be? I propose the following as a robust freedom condition to help ground deserved blame or punishment when either involve expressions of our moral anger. Suppose all of the other essential basal reasons are in place for an agent’s deserving harmful blame or punishment but for the freedom condition. Grant that absent the agent’s freedom, she does not deserve harmful blame or punishment for any untoward conduct. Now add the following. When an agent exercises freedom of the sort at issue here—­call it freedom of the will—­she acts in such a way that she settles whether she deserves hard treatment. What such freedom supplies are the resources that make it true that when she acts, it is up to her whether she deserves this sort of treatment. When she does, she thereby brings on herself this unwelcome treatment. These three features of an industrial-­strength freedom—­up to one, settling, and bringing on oneself—­do the grounding work required for deserved harmful blame and punishment: By agential resources that are up to her, an agent who acts of her own free will settles whether she deserves harmful blame or punishment, and in doing so she thereby brings upon herself any forms of unwelcome treatment as expressions of the moral anger had by those who would hold her to account. Far from an intuition of fittingness offering no more than a pitiful trinket—­ something irrelevant to the justificatory role involved in harmful blame or punishment as an expression of our moral anger—­the preceding proposal offers a  robust freedom condition capturing something rooted deep in our common-­ sense understanding of moral responsibility. Indeed, when reflecting on free will so understood, I confess to having . . . wait for it . . . an intuition of fittingness.18 18  For what it’s worth, I might add that in response to Strawson’s derisive remark, in surveying my own psychic state I cannot find any evidence of feeling even the least bit panicky.

350 

Fittingness

15.4.4  Desert versus Mere Fittingness: Two Paths forward We now have before us, first, an argument for the thesis that desert is a species of fittingness (sect. 4.1), second, an account of the desert base unique to the domain of moral responsibility (sect. 4.2), and third the specification of a robust freedom condition figuring in that desert base (sect. 4.3). Given these resources, return to our assessment of Feinberg and Strawson. As explained above (sect.  3.4), it is unclear how we should develop their views. One path forward involves fittingness; another involves desert. Moreover, on a normative interpretation of Strawson’s view, I argued (sect.  3.5) that it is an open question whether there is a robust freedom condition internal to our normative standards for moral anger. If so, Strawson was just wrong to think he could handily dispense with libertarian metaphysical presuppositions. Bearing in mind these two issues, we can now size up two options for contemporary Strawsonians. Here is the more ambitious way forward. The normative resources deployed to justify the gap Strawson sought to fill between morally objectionable behaviour and our reactive anger is a matter of desert. When our moral anger is a deserved response to the moral wrongdoing of another, we are at least pro tanto justified in expressing our anger through modes of unwelcome treatment via blame or punishment. In these cases, what is most basically fitting—­because deserved—­is the treatment and not just the reactive emotion of anger which that treatment expresses. In being so justified, moral reasons against inflicting pertinent forms of harm or suffering on others are silenced or defeated.19 However, the desert base for such a normatively weighty relation involves basal reasons featuring a robust sort of freedom—­freedom of the will. Equipped with this freedom, a person who deserves blame or punishment settles whether such treatment of her is justified, because it is up to her whether she acts in morally objectionable ways, and when she does, she brings it on herself that she is the deserving recipient of such unwelcome treatment. Of course, this more ambitious path is philosophically riskier because it brings with it taxing metaphysical burdens. Doesn’t libertarian freedom just fall right out of the proposal? No. That is an open question too. The compatibilists’ burden on this approach is more demanding, but many contemporary compatibilists take on precisely the project of offering an account of free will that can accommodate the requirements of basic-­desert-­entailing moral responsibility.20 If some contemporary Strawsonians wishing to retain their commitment to compatibilism reject this more ambitious path only because they assume that in taking it on they

19  On some accounts of desert (e.g. Bennett, 2002 and McKenna, 2019), bringing upon the deserving the harm attendant with blame or punishment is noninstrumentally good, and this goodness provides reasons to favour even if not require causing such harms. 20  For example, Fischer and Ravizza (1998), McKenna (2013), Nelkin (2011), and Sartorio (2016).

Fittingness as a Pitiful Intellectualist Trinket?  351 simply must accept libertarianism, they are mistaken. The metaphysics can be industrial-­strength sans libertarianism—­or so sayz yours truly. The more cautious way forward for Strawsonians might after all be the wiser. Committing only to mere fittingness, a Strawsonian might argue that she can avoid the weighty burdens of any commitment to a robust freedom condition. When one’s anger is a justified response to the morally objectionable conduct of another, it is simply fitting. In being fitting, reactive anger still accurately represents conduct as, for example, issuing from ill will. In doing so, it appraises the conduct negatively by standards internal to its domain, and so on. But any r­ easons it supplies for expressing fitting moral anger in outward modes of treatment do not silence or defeat competing moral reasons against inflicting harm on others. Of course, it still might override those reasons, just like, as I noted earlier, fitting reasons for expressing amusement might outweigh moral reasons not to laugh at a mildly morally offensive joke. But it is consistent with the mere fittingness of reactive moral anger that the weight of any reasons of fit to treat harshly a blameworthy person are never strong enough on their own to outweigh the moral reasons not to inflict such harm. This is key. On this more cautious proposal, there will still remain basal reasons for fitting anger. These will figure in what might be called a fittingness base, rather than what Feinberg called a desert base. And these basal reasons will track roughly the kinds of reasons also bearing on a desert-­based approach, as I set it out above. It will remain unfitting to respond with an attitude of moral anger to one who is not a competent person, who did nothing morally wrong, who was non-­culpably ig­nor­ant of any wrongdoing, or who didn’t act freely. But the freedom condition here can be down-­graded considerably. It will be enough if the agent did what she wanted to do, was not coerced, was not acting from a powerful compulsion, and so on. Key to this less robust freedom condition is that the basal reasons for fittingness only directly provide reasons for a reactive attitude—­moral anger. On this view, that is all that will be basically fitting. Any modes of treatment as an expression of that anger will only be derivatively justified, if justified at all. Although several philosophers working in the Strawsonian tradition seem to opt for this more cautious strategy,21 with one exception, none have made explicit their commitment to it. David Shoemaker, however, clearly has, arguing that in a theory of moral responsibility, our aim should be to account for the fittingness and not the desert of our reactive moral emotions (Shoemaker, 2015: 221–3). He even points out that questions about harsh treatment ought to be treated as a moral concern extrinsic to both the fittingness of a sentimental response, and to theorizing about moral responsibility:

21  For example, see Graham (2014), Hieronymi (2004), and Smith (2007).

352 

Fittingness

I want to insist on the crucial distinction . . . between the fittingness of various sentimental responses and the appropriateness of harsh treatment of offenders. The latter is a distinctively moral response and it simply does not bear at all on whether we have reasons of fit for the former. The latter does of course bear on whether we have all-­things-­considered reasons to respond to certain people in certain ways, where these are a weighted function of reasons of fit and moral reasons . . . The real question, then, is whether the practices of harsh treatment to offenders are also necessarily and essentially responsibility responses, alongside the sentiments I have discussed. But I cannot see that they are, once this distinction has been highlighted.  (2015: 223)

Shoemaker’s admirably executed proposal is clearly more cautious than one that opts for an appeal to desert. It is easy to appreciate how one could handily skirt any taxing metaphysical freedom requirements. All that is at issue in advancing a theory of moral responsibility is the mere fittingness of an emotional response to a wrongdoer (where that response need not be expressed in any treatment of a culpable party). Nevertheless, the appeal to mere fittingness also has a con­sid­er­ able cost. It risks changing the subject. Strawson (1962) originally proposed that we focus upon the reactive attitudes as a way to fill the lacuna he claimed to find in the justification offered by his contemporaries for our moral responsibility practices, including our practices of blaming and punishing. My own view, which I cannot defend here, is that the proper way forward for Strawsonians is to take the more ambitious path: appeal to desert as the species of fittingness pertaining to the appropriateness of our morally reactive attitudes. Admittedly, we cannot achieve what Strawson thought we could by taking this path, which is an effortless end-­run around the thorny metaphysical problems that the libertarians of his day attempted to face head on. But in paying this price, we afford ourselves a theory of responsibility that treats as a central burden the justification of our interpersonal practices of blaming and punishing the blameworthy. It must, however, be granted that fittingness is a more basic and encompassing relation. In this respect it has a virtue that desert lacks in that it is more reliable or stable. How so? If, as it turns out, no one does deserve blame or punishment because no one has free will, as sceptics like Derk Pereboom (2001, 2014) and Galen Strawson (1986) argue, there appears to be no reason why the Strawsonian could not retreat to the more cautious position and so retain the highly plausible, Strawson-­inspired naturalistic contention that come what may there will remain some reasonable sense of the appropriateness of responding with at the very least an attitude of moral anger to the morally objectionable behaviour of fellow members of the moral community. In the spirit of Strawson’s ‘Freedom and Resentment’, this truly is an inescapable part of our nature, isn’t it?

Fittingness as a Pitiful Intellectualist Trinket?  353

15.5 Conclusion My main goal in this essay has been to explore with great care the relevance of  fittingness to theorizing about moral responsibility, especially as a tonic to Strawson’s derisive and ungenerous name-­calling. I have argued here that we can make good sense of fittingness as a distinctive and informative normative relation in developing a theory of moral responsibility. In the form of desert, it might after all give rise to metaphysical questions about free will Strawson would have rather avoided. But there is nothing opaque or incoherent or unintelligible here, nor anything panicky. There are just hard philosophical problems that merit serious answers. Were the ghost of Strawson to appear before me now and, with not so much as a single argument focused on any concrete proposal, accuse me of panicky metaphysics, incoherence, or whatnot, I have an intuition that the following would be a perfectly fitting reply to such name-­calling: I’m rubber; you’re glue; whatever you say bounces off me and sticks to you.22

References Austin, J. L. 1956–7. ‘A Plea for Excuses.’ Proceedings of the Aristotelian Society. 57: 1–30. Bennett, Christopher. 2002. ‘The Varieties of Retributive Experience.’ Philosophical Quarterly 52(207): 145–63. Bergson, Henri. 1889/1910. Essai sur les donnes immediates de la conscience, Paris: F. Alcan, 1889; translated as Time and Free Will, tr. F. L. Pogson. London: Allen and Unwin, 1910. Campbell, C. A. 1951. ‘Is Free Will a Pseudo Problem?’ Mind 60: 446–65. Carlsson, Andreas Brekke. 2017. ‘Blameworthiness as Deserved Guilt.’ Journal of Ethics 21: 89–115. Chisholm, Roderick. 1964. ‘Human Freedom and the Self.’ The Lindley Lectures. Copyright by the Department of Philosophy, University of Kansas. 22  I would like to thank Chris Howard along with his coeditor Richard Rowland for inviting me to contribute to this volume. I presented earlier drafts of this chapter to seminarians at a workshop sponsored by the Lund-­Gothenburg Responsibility Project at Lund University, in Lund, Sweden, on 27 January 2021, and also to seminarians at the Murphy Institute at Tulane University on 23 April 2021. I profited enormously from both sessions. For helpful comments, I am indebted to Nathan Biebel, Gunnar Bjornsson, Olle Blomberg, Eric Brown, Randy Clarke, Stew Cohen, Juan Comesana, Justin D’Arms, Alison Denham, Anton Emilsson, Bob Hartman, Dan Jacobson, Marta Johansson, Tim Kearl, Matt King, Andrew Lichter, Per Milan, Dana Nelkin, Derk Pereboom, Abelard Podgorski, Travis Quigley, Wlodek Rabinowicz, Paul Russell, Nicholas Sars, Carolina Sartorio, David Shoemaker, Matt Talbert, Mark Timmons, Alexander Velichkov, and Robert Wallace. I would also like to note that I have profited considerably from Howard’s article ‘Fittingness’ (2018), which is a model of clarity and insight on this topic. Finally, and most importantly, I want to thank both Howard and Rowland for an outstanding set of thorough and engaging comments on an earlier draft.

354 

Fittingness

Clarke, Randolph. 2013. ‘Some Theses on Desert.’ Philosophical Explorations 16: 153–64. Clarke, Randolph. 2016. ‘Moral Responsibility, Guilt, and Retribution.’ Journal of Ethics 20: 121–37. D’Arms, Justin and Jacobson, Daniel. 2000. ‘The Moralistic Fallacy: On the “Appropriateness” of the Emotions.’ Philosophy and Phenomenological Research 61: 65–90. Feinberg, Joel. 1963. ‘Justice and Personal Desert.’ In Carl  J.  Friedrich and John W. Chapman, eds, Nomos VI: Justice. New York: Atherton Press: 69–97. Feinberg, Joel. 1970. Doing and Deserving. Princeton: Princeton University Press. Fischer, John Martin, and Mark Ravizza, eds. 1993. Perspectives on Moral Responsibility. Ithaca: Cornell University Press. Fischer, John Martin, and Mark Ravizza. 1998. Responsibility and Control: An Essay on Moral Responsibility. Cambridge: Cambridge University Press. Graham, Peter. 2014. ‘A Sketch of a Theory of Blameworthiness.’ Philosophy and Phenomenological Research 88(2): 388–409. Hieronymi, Pamela. 2004. ‘The Force and Fairness of Blame.’ Philosophical Perspectives 18(1): 115–48. Howard, Christopher. 2018. ‘Fittingness.’ Philosophy Compass. Vol. 13 (11): e12542 McKenna, Michael. 2012. Conversation and Responsibility. New York: Oxford University Press. McKenna, Michael. 2013. ‘Reasons-Responsiveness, Agents, and Mechanisms’ In David Shoemaker, ed., Oxford Studies in Agency and Responsibility, Vol. 1. New York: Oxford University Press: 151–84. McKenna, Michael. 2019. ‘Basically Deserved Blame and its Value.’ Journal of Ethics and Social Philosophy 15(3): 255–82. McKenna, Michael. 2020. ‘Punishment and the Value of Deserved Suffering.’ Public Affairs Quarterly 34(2): 97–123. McKenna, Michael. Forthcoming. ‘Guilt and Self-Blame within a Conversational Theory of Moral Responsibility.’ In A.  Carlsson, ed., Self-Blame and Moral Responsibility. Cambridge: Cambridge University Press. Meldon, A. I. 1961. Free Action. New York: Routledge and Kegan Paul. Miller, D. 1999. Principles of Social Justice. Cambridge, MA: Harvard University Press. Nelkin, Dana Kay. 2011. Making Sense of Freedom and Responsibility. Oxford: Oxford University Press. Nelkin, Dana Kay. 2013. ‘Desert, Fairness, and Resentment.’ Philosophical Explorations 16: 117–32. Nelkin, Dana Kay. 2016. ‘Accountability and Desert.’ The Journal of Ethics 20(1): 173–89. Nelkin, Dana Kay. 2019. ‘Guilt, Grief, and the Good.’ Social Philosophy and Policy. 36(1): 173–91. Pereboom, Derk. 2001. Living without Free Will. Cambridge: Cambridge University Press.

Fittingness as a Pitiful Intellectualist Trinket?  355 Pereboom, Derk. 2014. Free Will, Agency, and Meaning in Life. New York: Oxford University Press. Pojman, L. 1997. ‘Equality and Desert.’ Philosophy 72: 549–70. Raz, Joseph. 1977. ‘Promises and Obligations.’ In P. M. S. Hacker and J. Raz, eds, Law, Morality, and Society: Essays in Honor of H. L. A. Hart. Oxford: Clarendon Press. Rosen, Gideon. 2015. ‘The Alethic Conception of Moral Responsibility.’ In R. Clarke, M. McKenna, and A. M. Smith, eds, The Nature of Moral Responsibility: New Essays. Oxford: Oxford University Press: 65–88. Sartorio, Carolina. 2016. Causation and Free Will. Oxford: Oxford University Press. Sartre, Jean-Paul. 1948. Being and Nothingness, tr. Hazel  E.  Barnes. New York: Philosophical Library. Scanlon, T. M. 2008. Moral Dimensions: Permissibility, Meaning, Blame. Cambridge, MA: Belknap Harvard Press. Scanlon, T. M. 2013. ‘Giving Desert its Due.’ Philosophical Explorations 16: 101–16. Schlick, Moritz. 1939. ‘When Is a Man Responsible?’ In M. Schlick, Problems of Ethics. Engelwood Cliffs, NJ: Prentice-Hall: 143–56. Shoemaker, David. 2015. Responsibility from the Margins. Oxford: Oxford University Press. Shoemaker, David. 2017. ‘Response-Dependent Responsibility.’ Philosophical Review 126: 481–527. Smart, J.J.C. 1961. ‘Free Will, Praise, and Blame.’ Mind 70: 291–306. Smith, Angela. 2007. ‘On Being and Holding Responsible.’ Journal of Ethics 11: 465–84. Strawson, Galen. 1986. Freedom and Belief. Oxford: Clarendon Press. Strawson, P.  F. 1962. ‘Freedom and Resentment.’ Proceedings of the British Academy 48: 187–211. Taylor, Richard. 1966. Action and Purpose. Engelwood Cliffs, NJ: Prentice-Hall. Vargas, Manuel. 2013. Building Better Beings. Oxford: Oxford University Press. Wallace, R.  Jay. 1994. Responsibility and the Moral Sentiments. Cambridge, MA: Harvard University Press. Wallace, R. Jay. 2019. The Moral Nexus. Princeton, NJ: Princeton University Press. Watson, Gary. 1987. ‘Responsibility and the Limits of Evil: Variations on a Strawsonian Theme.’ In Ferdinand Schoeman, ed. Responsibility, Character, and the Emotions: New Essays in Moral Psychology. Cambridge: Cambridge University Press: 256–86. Watson, Gary, ed. 2003. Free Will, 2nd ed. New York: Oxford University Press.

16 Blame’s Commitment to Its Own Fittingness Rachel Achs

16.1  Introduction Blame is ubiquitous in our day-­to-­day moral lives. Yet even if we explicitly limit our focus to occurrent interpersonal cases,1 it can be difficult to make progress when theorizing about blame’s nature because the types of blame are multifarious. At a very general level, blame is a response to someone in light of some (apparent) norm violation on that person’s part—­it’s a reaction people have to one another’s wrongdoings.2 But the ways of responding to wrongdoing that serve as examples of blame are very diverse. Public calling out, stony silence, private seething, sadly deciding to stop relying on a friend, dispassionately unfriending someone on Facebook—­these all arguably count as instances of blame.3 Correspondingly, philosophers who offer accounts of blame are in dispute about more than just minor details. Rather, they have been unable to agree even on what kind of response blame is—­a judgement? a way of behaving? an emotion? a set of dispositions? some combination of these things?—and have held that blame may be comprised either entirely or partially by components that together make up an ontologically multitudinous list.4 1  My topic here is interpersonal blame—­the type of blaming that counts as a way of holding a person responsible. I thus set aside the ‘causal’ way in which one may ‘blame’ the rainfall for the flood. 2  A wrongdoing, as I think of it, is the violation of a practical requirement: an impermissible action or omission over which a person had voluntary control. I will assume here that what people ultimately blame one another for are (putative) wrongdoings because I’ve defended this view elsewhere (Achs 2020). However, anyone who disagrees should, when reading future mentions of ‘wrongdoing’, substitute in his or her own preferred characterization of the type of (putative) violation blame is ultimately for (such as, for example, exhibiting an ill quality of will). 3  The last two examples come from Scanlon 2008: 136 and Smith 2013: 32, respectively. 4  Among other things, philosophers have proposed that blame consists in (1) some disapproving or sanctioning behaviour (Schlick 1962); (2) a judgement that the blameworthy party has diminished her moral credit or moral worth by her actions (Glover 1970; Zimmerman 1988); (3) the experience of reactive emotions such as resentment, indignation, or guilt—­in turn comprised of sensory, behavioural, or physiological dispositions; physiological changes and bodily feelings; and cognitive elements (Wallace 1994; Pickard 2013; Menges 2017); (4) a judgement that some person has acted badly or is a bad person, conjoined with a desire that she not have done so and the affective and behavioural dis­ posi­tions to which the frustration of that desire gives rise (Sher 2006); and (5) a revision of one’s Rachel Achs, Blame’s Commitment to Its Own Fittingness In: Fittingness: Essays in the Philosophy of Normativity. Edited by: Christopher Howard and R. A. Rowland, Oxford University Press. © Rachel Achs 2022. DOI: 10.1093/oso/9780192895882.003.0016

Blame ’ s Commitment to Its Own Fittingness  357 In the face of such a variety of opinions and purported examples, it is probably wise to be ecumenical and allow that blame really can come in a wide range of response-­types. But if so, then before delving deeply into the study of blame, it is worth reflecting on whether and why we should ever want to theorize about blame considered as a single kind of thing. A common core merely of being responses to wrongdoing is not sufficient to render the many modes of blame meaningfully unified. After all, there are responses to wrongdoing that seem the polar opposite of one another: one might react to a beloved son’s commission of murder by becoming enraged, but also by showing him extra love and affection in anticipation of the hateful treatment he’ll receive from others.5 Yet, aside from all sharing the property of being reactions to wrongdoing, the types of response that can apparently be involved in blame are a motley collection of feelings (attitudes? actions? changes?), indeed. So do the diverse modes of blame actually have anything significant enough in common to give us reason to ask questions about blame? Or are they rather, as Martha Nussbaum (2016: 259) suspects, just ‘descriptions of different phenomena somewhat misleadingly grouped under a single rubric’? If the latter, then perhaps blame is only ever properly subject to more piecemeal styles of investigation. The thesis I want to defend in this chapter is that blame involves a particular, formal element of commitment to its own fittingness. I will argue that attributing this element to blame can illuminate several of blame’s more elusive features. First, it can shed light on blame’s directedness: on the way in which blame is targeted towards someone for some (putative) violation on her part. Indeed, I will argue that properly characterizing blame’s directedness requires attributing this element to blame. Second, this element of blame can help to account for a felt quality that blame often has and that spans through many of its guises. I’ll describe this quality in more depth later, but for now I’ll just gloss it as a feeling of assuredness in one’s blame. Finally, this element of blame can answer the question just posed about whether anything interesting ties the various types of blame together. It is a property common to blame’s diverse modes which should make us think of blame as a phenomenon with unity, and one worth investigating as such. In a sentence which will require explication, the formal element I claim blame possesses is this: when S blames T, S takes her present way of reacting to be fitting directly on the basis of T’s having done something wrong.6 I will call this blame’s element of reflexive endorsement.

attitudes, intentions, and expectations towards the blameworthy party in a way that reflects that ­person’s possession of relationship-­impairing attitudes (Scanlon 2008; Smith 2013). 5  Compare Smith 2013: 38. 6  A clarificatory note: ‘T’s having done something wrong’ should be read as referring to the (putative) fact that some particular φ-ing of T’s is a wrongdoing—­as opposed to the existence, in general, of something which is a wrongdoing of T’s. How the blamer’s commitment refers to the φ-ing of which it predicates wrongdoing (e.g. whether under a particular description, or by mental ostension) may vary

358 Fittingness Such reflexive endorsement is ‘formal’ in the sense that a great many types of reaction can conform to its structure; many types of reaction can be ones that subjects take to be fitting directly on the basis of someone’s wrongdoing. Indeed, I will proceed under the assumption that any of the ways of reacting that philo­ sophers have cited as components of blame—­behaviours, judgements, dis­pos­ itional changes, bodily feelings, modifications of intentions and expectations, etc.—can figure as the material content of a blaming reaction. Reflexive endorsement is ‘reflexive’ in the sense that it is self-­referential, although I also hold that it can be ‘reflexive’ in the sense of being produced by automatic or inattentive mental processes. (It needn’t be ‘reflective’, in the sense of representing a subject’s considered judgement.) It is nevertheless an ‘endorsement’ because it involves a subject taking her reaction to be supported by a particular kind of justificatory basis. Here is how my argument will be structured: in Section 16.2, I’ll make an initial case for the claim that blaming involves an endorsement of the way that one is presently reacting by considering blame’s directedness more closely. In Section 16.3, I’ll issue several clarifications about the justificatory relation at issue in this element of blame, with the goal of explaining why I think that, while there are many ways of responding to wrongdoing, only blaming responses involve taking a wrongdoing to, in and of itself (i.e. directly), make one’s present response fitting. In Section 16.4, I’ll raise a problem for the picture of blame that I’ve drawn thus far, which I’ll then solve by elaborating on the sense in which reflexive endorsement is self-­referential. This will allow me both to finish defending my view that attributing reflexive endorsement to blame is necessary for accounting for blame’s directedness and to say more about the assured feeling that I claim is typical to blame, which reflexive endorsement helps account for too. In Section 16.5, I’ll argue that blame’s element of reflexive endorsement gives us reason to at least sometimes inquire about blame as a unified entity. I’ll then close by considering whether we should think that human responses beyond blame exhibit analogous commitments to their own fittingness.

from case to case. How the blamer’s commitment represents the property of being a wrongdoing may also vary, so long as it predicates that property of T’s φ-ing. But also remember, if you think that wrongdoings are not really what people blame one another for, that it is fine to substitute in a different sort of norm violation here. What’s important is just that whatever sort of (apparent) norm violation blame responds to go in the place of ‘T’s having done something wrong’. Indeed, I think the arguments that follow would go through even if one just subbed in ‘T’s having violated a norm’. Theorists of blame generally agree that blame responds to some sort of (apparent) norm violation (Shoemaker and Vargas 2021: 582). Further clarifications, including on what I mean by ‘directly’, will follow in due course.

Blame ’ s Commitment to Its Own Fittingness  359

16.2 Blaming Someone for Although it seems that a host of different elements may be involved in blaming, there is at least one respect in which blame—­no matter what its elements—­bears a resemblance to an emotion. Blame is a directed way of responding to one’s en­vir­on­ment, just as many emotions seem to be. Just as one might be afraid of commitment, or grateful for yoga, so too does blame consist in a reaction that one experiences as concerned with some particular ‘object’. More specifically, like the object of some, but not all, emotions, blame’s object is comprised of two parts: someone it is targeted towards and something it is for. One always blames someone for something that person has (apparently) done wrong. An initial reason to believe that blame always involves an element of endorsement is that doing so seems necessary for capturing the phenomenal character of blame’s relation to its object. What it is like to blame someone involves seeing that person’s (apparent) violation as justifying one’s present way of reacting to him.7 Thinking through the various ways in which we might try to describe the relation between blame and its object can help one to recognize this aspect of what blaming is like. Remember, we’re supposing that blame may involve a variety of possible parts: behavioural or physiological dispositions, bodily feelings, modifications of intention, behaviours, judgements, etc. Let us call an episode of blame—­whichever one, or many, of these things it might involve—­a ‘blaming reaction’. Let us refer to the person blame is directed towards and the particular wrongdoing it is for as its ‘target’ and ‘focus’, respectively. Our question is: what is the relation between a blaming reaction, the target it is directed at, and the violation that reaction is for? Pre-­theoretically one might think that a blaming reaction is always caused by the target’s wrongdoing. But that can’t be right. For one thing, subjects sometimes blame targets who haven’t actually done anything wrong. And for another, supposing merely that a target’s wrongdoing causes a blaming reaction won’t capture the relation between a blaming reaction and its object, since one person’s wrong­ doing can cause another to experience a reaction directed at or about any number of other things. For example, John’s betraying my secret might cause me to be angry about something else entirely, such as John’s father’s having neglected to be a better role model for him.8 Thus, many in the blame literature suggest that we can capture blame’s directedness by holding that blame partially consists in a subject’s judging that the target has done something wrong (or in some other way instantiates blameworthy-­making

7  To avoid clunkiness I’ll sometimes drop the ‘(apparent)’ in what follows. 8  See Solomon 1973: 21. I assume the anger in this example is an instance of blame. In general, I think ‘blame’ and ‘anger’ have overlapping extensions: some, but not all, blame is anger, and some, but not all, anger is blame.

360 Fittingness properties).9 The idea is that blaming consists of such a judgement plus some further way of reacting (be it a physiological response, or a behavioural one, or what have you). However, saying merely that blaming reactions have this judgemental component won’t suffice to capture blame’s directedness either. For, whatever they consist in precisely, it seems that the other, non-­judgemental components of a blaming reaction can also be directed at a target’s violation. If one person blames another by yelling, for instance, then we say that the blamer is yelling at the target for that person’s wrongdoing. In blame, a decision to no longer trust someone seems to be a response to her for what she’s done. (‘Well, I guess I won’t confide in her again.’) If a blamer feels miffed, then she is miffed at the target for the target’s behaviour. Indeed, if a subject’s blame involves awareness of changes in her body, then those changes too can be experienced as concerned with the target’s wrong. (‘My blood was just boiling at what he said!’)10 Thus, if a blaming reaction does involve a judgement that some person has done wrong, it will be crucial to specify the right relation between this judgemental component and whatever else that reaction may consist in in order to fully describe the relation between a blaming reaction and its object. In short: we must capture the way in which non-judgemental components of blame are directed at its object, too. So, assuming for the moment that blaming does involve the relevant judgemental component, what could the relation between that judgement and the rest of a blaming reaction be? Again, we can’t just say that a judgement of wrongdoing causes whatever else is involved in a blaming reaction. My, say, urge to slam my fist down may be just as much about John’s father’s neglect when it is caused by the judgement that John has betrayed my secret as when it is caused by his wrongful betrayal itself. We could try to build up the way in which the rest of the blaming reaction is about its object out of a reliable causal connection between it and a judgement of wrongdoing. But I think one should worry that, if we try to do this, we’ll face a problem of deviant causal chains. After all, insofar as a subject does judge that some person has acted wrongly, and then (say) lashes out at him for it, surely that person’s wrongdoing is not just the cause of her lashing out, but rather also seems to the subject to be the reason that her lashing out is merited. (And something more sophisticated than a mere reliable causal connection seems required to capture what it is for the target’s wrongdoing to appear as the subject’s reason.)11 9  Wallace (1994), Sher (2006), Scanlon (2008), and Smith (2013) all suppose that blaming involves a judgemental component. 10  See Goldie 2000: 55, 2002: 248 for a similar point that all components of an emotion have directedness, and Na’aman (this volume) for defence of the thought that the physiological components of an emotion do. 11  Say that you reliably feel angry when insulted. Say, further, that a mad scientist rigs some wires to your brain so that pressing a certain button has the following effect: it induces you to forget everything that has occurred in the last five minutes, and then to exhibit exactly the syndrome of physiological,

Blame ’ s Commitment to Its Own Fittingness  361 But that the target’s wrongdoing has this appearance of a reason is precisely the point I wanted to draw out. If we suppose that the way in which blaming reactions are directed at individuals for their wrongdoing is explained by those reactions consisting, in part, in a judgement that some person has done wrong, then it is only natural to understand the subject as viewing the behaviour which that judgement concerns as providing justification for the rest of her blaming reaction. I submit that it is only natural to form this picture because what it is like to blame someone, and to blame him for some wrongdoing, involves regarding that person’s wrongdoing as justifying the way one is presently reacting to him. Perhaps, however, an episode of blame needn’t involve a representation that some person has done wrong as a component part at all. An alternative way of approaching the problem of the relation between a blaming reaction and its object is to propose not that it involves a judgement, but rather that a blaming reaction fully consists in a representation of a person and his wrongdoing. This is how many psychologists and philosophers of emotion seem to conceive of the way in which emotions are directed. Emotional reactions, they say, are themselves a mode of representation: perhaps judgements about their objects, or perhaps perceptual or perception-­like experiences of their objects as bearing some evalu­ative significance (Solomon 1973; Döring 2009; Tappolet 2012, 2016). Or, say the more empirically minded, they are ‘appraisals’ of their objects, where this typ­ic­al­ly means that emotional reactions function to represent an organism’s relationship to some object(s) in its environment as conferring harms or benefits (Lazarus 1991: 820).12 For example, D’Arms and Jacobson (2003: 139) describe fear as a syndrome of directed attention, physiological changes, affect, and motivation that can be functionally understood as constituting a kind of appraisal of the circumstances.

It is usually assumed, in adopting this approach, that the various types of emotion are differentiated from one another by the object of an emotional reaction appearing to the subject to possess evaluative properties that emotions of that type represent. (Some call such properties the emotion’s ‘formal object’.) For example, it is assumed that a reactive syndrome that is an episode of fear will be one in which some object appears to the subject as fearsome; that an episode of contempt will motivational, and attention-­directing effects that is typically involved in feeling angry. The scientist then proceeds to amuse himself thus: he invites a series of your most tactless family members into the room, and every time one of them says something insulting to you, he waits to make sure you have registered the wrongfulness of this insult and then presses the button. Thus, it is your belief that you have been wrongfully insulted which causes him to press the button, which, in turn, causes you to forget the insult and then to undergo the syndrome of physiological, motivational, and attention-­ directing effects that are regularly caused by violations against you. Intuitively, this ‘anger-­syndrome’ is not one that is about being insulted—­at least not in the way that anger is normally about an insult. 12  D’Arms and Jacobson (2003), Prinz (2004), Moors (2013), Pickard (2013), and Menges (2017) also speak of emotional (or blaming) reactions as ‘appraisals’.

362 Fittingness be one in which some object appears to the subject as contemptible; and so forth. Adopting this approach for the case of blame, we would suppose that a blaming reaction—­whatever it consists in—­itself represents some object as blameworthy. Or rather, to more accurately accommodate blame’s dyadic object, we might suppose that blaming reactions represent some target as blameworthy and represent some focus (that target’s wrongdoing) as making him so.13 To say this is not yet to suppose that blaming must involve a subject’s taking the target’s wrongdoing to justify her present way of reacting. It is to suppose that blaming involves representing some target’s wrongdoing as making that person blameworthy, but one might imagine that this occurs without the relation between the target and the subject’s present way of reacting to him actually figuring in the content of the subject’s representation. To be sure, the subject represents some target person as bearing some evaluative badness, and the name that we have for this sort of badness is ‘blameworthy’—but this doesn’t necessarily mean that the subject must represent the target as meriting the subject’s emotional reaction itself. And yet, it seems to me that saying anything less would leave out something about the sort of appraisal that is involved in both emotions and blame. For emotional experiences—­and also blame experiences—­do seem to involve at least some registration of the relation between an emotion’s object and the way one is reacting to that object. Consider: although emotions are about particular objects, they are not ‘transparent’ vehicles of representation in the way that judgements and perceptions are. In introspecting on emotional experience, one doesn’t simply ‘see through’ that experience to the object one’s emotion is directed at (Deonna and Teroni 2012: 69). Rather, one also becomes aware of how one feels—­because emotional episodes aren’t just experiences of objects, but rather experiences of responding to objects. But once we say this, I think we are pressed to acknowledge that the sense in which emotions represent their objects does involve representing those objects as meriting our feelings. For the connection between our emotional feelings and their objects certainly doesn’t seem arbitrary from the inside. Nor does the connection here seem to be merely causal.14 Moreover, the preposition we use to describe the relation of a blaming reaction to its focus further supports the view that the target’s wrongdoing presents to the subject as providing normative support for her blame. We say that the blame is ‘for’ the wrongdoing, as when we give a gift to a person ‘for’ something they’ve 13  Behaviours can remain a component part of blame on this model too, provided we allow behavioural expressions to be part of what is doing the representational work of blame. The idea that behaviours may represent doesn’t seem odd if we recall the many behaviours that do. Ballet, for instance, represents. 14  Cowan (2016: 74) proposes that emotional opacity supports the thought that emotions themselves may figure in the representational content of emotions. Relatedly, Mitchell (2019: 373) argues that we are blocked from fully distinguishing attitude and content in the case of emotions, because emotions represent the evaluative properties of their objects as ‘having the power to intelligibly motivate’ a component of emotional experience itself.

Blame ’ s Commitment to Its Own Fittingness  363 done to merit it. Assuming now that a blaming reaction is a representation of its object, it is hard to imagine why we would speak this way if representing some target as blameworthy didn’t also involve representing our blaming reaction itself as made appropriate by the target’s wrongdoing. After all, it is not normally the case that when I represent an object’s possession of some property as obtaining in virtue of some other fact, I speak of myself as having a reaction that is for that latter fact. I submit that, in the case of blame, I do regard my reaction as for its target’s wrongdoing precisely because the way I represent a person when blaming him does involve representing his wrongdoing as making him worthy of my way of reacting. Thus it seems to me that, no matter what its other elements are precisely, blaming always involves taking some person’s wrongdoing to justify one’s present way of reacting, because this is what must be said if one wants to accurately capture what it’s like to blame someone, and to blame him for some wrongdoing. Let me also add two clarificatory notes about this point: first, once we recognize that a blaming subject views the target’s behaviour as providing reason for her present way of reacting, I think it becomes apparent that the subject doesn’t take the target’s behaviour to provide justification merely by virtue of just anything about it. A blamer doesn’t take the target’s wrongdoing to justify her blaming reaction merely by virtue of that behaviour’s being, say, an utterance. Rather she takes that behaviour to provide justification by virtue of its making the target worthy of blame. And whatever precisely need be the case in order for the target’s behaviour to do this, it includes, at the very least, that behaviour’s being the type of thing that makes people worthy of blame—­which is wrongdoing. So what I actually think we’re in a position to conclude at this moment, put more carefully, is that the blaming subject sees the target’s wrongdoing qua wrongdoing as justifying her present way of reacting to him.15 From here on, this qualifier should be understood as implicit whenever I say that the subject takes the target’s wrong­ doing to justify her present way of reacting. Second, I must make a clarification that I already alluded to briefly in the introduction: although the commitment we’ve located in blame is an endorsement in that it involves the subject taking her present reaction to be justified, the endorsement at issue needn’t be a reflective one, in the sense of representing the subject’s considered judgement. In holding that blaming subjects take a target’s wrong­doing to make their blame justified, in other words, I by no means deny that a person might blame while simultaneously judging that her blame isn’t justified, perhaps because she thinks the target hasn’t actually done anything wrong. The idea is rather that, because one always blames a target for that person’s 15  Or, in other words, that she takes the (putative) fact that the target’s φ-ing was a wrongdoing to justify her present way of reacting. Although, again, anyone who disagrees with me about what type of thing people blame one another for may make appropriate substitutions.

364 Fittingness wrong­doing, blame internally involves a commitment to that wrongdoing’s justifying one’s present reaction, and thus internally involves an endorsement of one’s blaming response. I take it that episodes of blame, which include commitments to the subject’s present way of reacting being justified, may sometimes be triggered automatically by stimuli other than a subject’s reflectively judging that the target has behaved wrongfully. Thus, blame may sometimes conflict with and be recalcitrant to a subject’s considered beliefs—­and thus be regarded as irrational by her own lights.16

16.3  Justificatory Relation I’ll now continue to sharpen the picture of blame’s formal element that I’m advancing, focusing in this section on the justificatory relation involved in reflexive endorsement. Specifically, I want to speak carefully about the dimension along which the subject takes her reaction to be justified; the type of justificatory support at issue in the subject’s commitment; and the distinctive justificatory role played by the target’s wrongdoing in the subject’s thought. My aim in this section is twofold: in addition to staving off any potential confusion about the content of the commitment that blame incorporates, getting more precise about the justificatory relation invoked in blame’s element of reflexive endorsement will also help me to point out the features that make me think this commitment is unique to blame. As an initial clarification, we’ll distinguish between two dimensions along which a response might have the status of being ‘justified’: the dimension of fittingness and the dimension of that attitude’s being good to have (D’Arms and Jacobson 2000a, 2000b). Pascal’s wager purportedly demonstrates that belief in God can be good to have even if it isn’t fitting. An offer of a million dollars to 16  One may, if one wishes, hold that the cognitive commitment internal to blame is not a judgement, but rather what Rosen (2015: 71) calls a ‘seeming’, or what Roberts (1988: 191) calls a ‘construal’ which possess a certain ‘verisimilitude’, and has ‘for the construer, the appearance of truth’. For my own part, what I think is important is just to hold that blame, including the endorsement internal to it, can be the upshot of the sort of mental processing psychologists call ‘autonomous’. An autonomous process ‘initiates and completes outside of deliberate control’ (Pennycook 2018: 8); such processes are ‘mandatory when their triggering stimuli are encountered’ (Evans and Stanovich 2013: 236). It seems to me that, so long as such a process can produce blame, and can be triggered by something other than a subject’s considered judgement that a target is blameworthy, we can account for the existence of recalcitrant blame. There is an interesting question which may be raised here about how best to account for the level of felt irrationality involved in recalcitrant blame and in recalcitrant emotions. While experiences of recalcitrant emotion and blame feel more irrational than illusory perceptual experiences, they also feel less incoherent than experiences of making two directly contradictory judgements. (For discussion, see Döring 2009 and Tappolet 2012.) Some philosophers attempt to account for the level of felt irrationality present in recalcitrance by assimilating the type of appraisal involved in emotions and blame to a type of perception that we strive to manage over time (D’Arms and Jacobson 2003; Tappolet 2012). I am personally unconvinced that this is the best strategy—­but I won’t pursue that issue here.

Blame ’ s Commitment to Its Own Fittingness  365 desire a saucer of mud may make that desire very good to have, even if saucers of mud are never fitting objects of desire. In general, while whether a reaction is good to have can speak for or against bringing it about that one has it, ‘fit’ is the normative relation that obtains between responses and things that merit, call for, or are worthy of those responses: between desire and the desirable, amusement and the amusing, contempt and the contemptible, etc.17 As with all responses that we deem to be fitting, there is always something a blaming subject takes her present reaction to be appropriate to, namely its target. Moreover, that a target has violated some requirement bears most directly on whether that person is worthy of a subject’s blame—­in and of itself such a violation does not tell us whether it would be a good thing for a subject to blame that person. So, when I say that blame involves a subject taking some target’s violation to ‘justify’ the way she is reacting, I mean ‘justified’ with respect to whether it fits its target, not with respect to whether it is good to have. I must also clarify the sense in which a reflexively endorsing subject takes the target’s wrongdoing to provide support for the fittingness of her present way of reacting. It is not simply that she regards the target’s wrongdoing as a normative reason for that reaction. Rather, she regards the target’s wrongdoing as providing support for how she reacts from a first-­personal perspective which implicates both justification and explanation at once. By ‘normative reason,’ I refer to a consideration that can provide support for the fittingness of a subject’s reacting in a certain way regardless of how she came to be reacting in that way, and, indeed, regardless of whether she even is reacting in that way.18 The relation of justificatory support invoked by the reflexively endorsing subject isn’t merely that she takes the target’s wrongdoing to be a normative reason to react in the way that she is. After all, a subject may take the target’s wrongdoing to be a normative reason (or even to provide her with sufficient normative reason) to react in the way that she presently is without that wrong­ doing being the focus of her blame. To illustrate, imagine you encounter me slamming my fist down in anger (about John’s father’s neglect), and you tell me that John’s betrayal of my secret makes this behaviour fitting. I may concur wholeheartedly, but my anger still needn’t be about John’s betrayal. What would be missing if we held reflexive endorsement merely to invoke the normative reason relation is that the blaming subject also takes the target’s 17  Reasons that count towards a reaction’s fittingness are sometimes called ‘right-­kind’, while con­ sid­er­ations that favour having a reaction merely by counting towards bringing it about that one reacts in that way are called ‘wrong-­kind’ reasons. Some philosophers think that wrong-­kind considerations in favour of attitudes aren’t really reasons for those attitudes, because those attitudes can’t be directly based on such considerations (Kelly 2002; Shah 2006). Maguire (2018) argues that right-­kind reasons for affective attitudes aren’t really reasons because they are neither gradable nor contributory in the way that reasons for action are. I refer to both types of consideration as reasons, although from here on out I’ll really only be talking about right-­kind (i.e. fit-­related) reasons anyway. 18  This is the analogue of some consideration’s supplying ‘propositional justification’ in epistemology.

366 Fittingness wrong­doing to play a role in the genesis of her blaming response—­to explain why she is reacting in that way. But of course the sense of ‘explanation’ here is not merely causal. It is instead that the blaming subject regards the target’s wrong­ doing as we regard the rational bases of our attitudes from the inside. She takes that wrong­doing to explain her present way of reacting and to justify it, and, indeed, to explain it because of the justification it provides.19 She takes it to render her reaction justificatorily well supported. As I initially characterized her, the reflexively endorsing subject ‘takes her present way of reacting to be fitting on the basis of the target’s wrongdoing’. But what I mean by this phrase is open to misunderstandings that need warding off. One may hear me as asserting simply that the subject takes the target’s wrongdoing to provide normative support for her present reaction. But I don’t mean this; the subject also takes the target’s wrongdoing to be a reason for which she reacts. Alternatively, one may hear me as asserting that the target’s wrongdoing is the rational basis of the subject’s taking, as opposed to what she takes to be the rational bases of her reaction. But I don’t mean this either.20 Rather, I mean that the subject takes the target’s wrongdoing to render her reaction justificatorily well supported, with respect to fit. The complication in clarifying the sort of justificatory support at issue in the content of reflexive endorsement has at least has one simplifying effect. We need no longer ask whether the blaming subject merely takes herself to have some reason to think that her attitude is fitting or whether she takes herself to have sufficient reason for its fittingness. Since the idea is not just that the subject takes herself to have justificatory support for her reaction, which is something that a person may have in part but not in full, but rather that she takes her reaction to be justificatorily well supported, which a reaction simply either is or isn’t, it follows that the subject takes herself to have sufficient reason for her reaction’s fittingness. This doesn’t mean, however, that she takes the target’s wrongdoing to provide sufficient reason for her reaction’s fittingness no matter what. Indeed, I think we can infer, simply from thinking about what it takes to make a blaming reaction fitting, that, if blame involves reflexive endorsement, then blaming subjects are also at least implicitly committed to certain enabling conditions on the fittingness of their present reactions being met. For, although the fact that some person has behaved impermissibly can make that person worthy of blame by some possible person, it is not sufficient to ensure that that person warrants that particular blamer’s blame. Even when a target is blameworthy in the sense of satisfying the criteria that a person must to be the fitting target of some

19  This is the analogue of some consideration’s supplying ‘doxastic justification’. 20  Although a blaming subject must take the target’s having done something wrong to be a reason for which she blames, I make no claim that the target’s wrongdoing must be a rational basis either of her blame or of her reflexive endorsement.

Blame ’ s Commitment to Its Own Fittingness  367 possible person’s blame, intuitively a subject must also have requisite standing if the target is really to merit blame from her. She must, for instance, not have committed the same wrongdoing without having reformed, or else her blame would be hypocritical. Thus, since reflexive endorsement does involve a subject’s taking her present way of reacting to be justificatorily well supported, we can infer that the blaming subject must also be (perhaps more implicitly) taking criteria beyond the target’s having wronged to be satisfied, such as the criterion that she possesses standing to blame.21 Let’s turn now to the distinctive justificatory role played by the target’s wrong­ doing in a blaming subject’s thought. To see the special role of wrongdoing that I  want to highlight, consider first that it is possible to take one’s non-blaming reaction to be made fitting by a target’s wrongdoing. For example, if I’ve been coaching you not to be such a goody two-­shoes, then I may be proud when you finally do something wrong and take your wrongdoing to justify my pride. But consider also a further feature of this sort of case: if I take your wrongdoing to provide justificatory support for my pride, I must also take additional con­sid­er­ ations to link your wrongdoing to my reaction’s fittingness. Indeed, if we roughly gloss pride’s fittingness conditions following Philippa Foot’s (2002: 76) suggestion that pride responds to achievements that are ‘in some way splendid and in some way one’s own’, then we will straightaway recognize that the satisfaction of those conditions play a mediating role in our coaching case. Because acting wrongly is precisely what I’ve been trying to coach you to do, I regard your wrongdoing as my own splendid achievement, and my prideful reaction as fitting in virtue of my own success. I thus take your wrongdoing to make my prideful reaction fitting precisely because your wrongdoing demonstrates that I’ve achieved something splendid—­my expert coaching has finally borne fruit!—which, in turn, renders my prideful reaction fitting. Moreover, I can’t take your wrongdoing to provide justificatory support for my pride without also taking that wrongdoing to demonstrate that I’ve achieved something splendid, because wrongdoings are not, in and of themselves, prideworthy. In contrast, a blaming subject needn’t see the target’s wrongdoing as calling for her blaming response via demonstrating that anything else is the case. It is wrong­ doing, after all, which makes a person blameworthy.22 Thus, a blaming subject takes the target’s wrongdoing to directly justify her blaming response. In other words, regardless of whether her reaction involves getting in the target’s face or merely distancing herself from that person, a blaming subject takes the target’s 21  For discussion of what contributes to standing, see Cohen 2006 and Todd 2019. My suggestion here is that lacking standing can render a person’s blame unfitting. Thus, philosophers who argue that there is nothing morally wrong with hypocritical blame do not directly address what I take to be at least one problem with blaming hypocritically. 22  Although background conditions may need to be satisfied for a target’s wrongdoing to do so. And, again, one may make appropriate substitutions for ‘wrongdoing’ if one wishes.

368 Fittingness wrongdoing to justify her present response without having to also demonstrate that some further set of conditions has been satisfied. Indeed, this is why a blaming subject often does take her response to a target to be justified without also taking that person’s wrongdoing to demonstrate anything further. For example, I  might take your wrongdoing to justify my decision to stop speaking to you, without actually thinking that your wrongdoing makes giving you the silent treatment worth doing. Rather, I might take your wrongdoing simply to make you worthy of the silent treatment. Having noticed the directness of the justificatory role a blaming subject at­tri­ butes to the target’s wrongdoing, it is thus worth building this directness into our understanding of the commitment involved in blame. It is worth saying (as, indeed, I did initially), that a reflexively endorsing subject takes her present way of reacting to be fitting directly on the basis of the target’s wrongdoing—­where what I mean is that the subject takes the target’s wrongdoing to (i) justify her blaming response and (ii) to justify it without having to demonstrate that something else is the case which, in turn, justifies her response. It is worth building this directness condition into blame’s element of reflexive endorsement because this condition ensures that the reflexively endorsing subject (perhaps more implicitly) takes her present way of reacting to be justifiable directly by wrongdoing. It thus ensures that she (at least implicitly) takes her present way of reacting to have justification conditions that, in human life anyway, appear to be unique to blame. Indeed, it bears emphasizing here that the specific justificatory relation invoked in blame’s element of reflexive endorsement is what makes it plausible that this commitment is the formal element of blame, in particular. In the introduction, I explained that merely having in common that they are all responses to a person’s wrongdoing does not suffice to render the many types of blame a unified phenomenon. But, although there are many ways of responding to wrongdoing, and, although several of those ways may perhaps involve a subject taking someone’s wrongdoing to provide justificatory support for her response, they don’t all involve a subject taking her response to be made fitting to the target, and made so by the target’s wrongdoing, in and of itself. Some responses to wrongdoing, while they may involve commitments to their own justification, don’t involve a subject taking a target’s wrongdoing to provide justificatory support for that response’s fittingness at all: a mother who decides to show her murderous son extra affection in order to compensate for the treatment he’ll engender from others is not taking her son’s wrongdoing to make him merit her extra affection. Rather, insofar as her son’s wrongdoing is a reason for her response, it is a consideration that makes her affectionate response good to have: it indicates that her son is likely in for harsh treatment, and that compensatory affection may thus do him some good. Other responses to wrongdoing may be ones that a subject regards as made fitting by the target’s wrongdoing, but made fitting by his wrongdoing only indirectly. One could take a person’s wrongdoing to render him worthy of affection, for instance, if one

Blame ’ s Commitment to Its Own Fittingness  369 has particular reason to view his wrongdoing as demonstrating that he satisfies the fittingness conditions of affection. Perhaps, having coached you to behave badly, I now regard the fact that you’ve done something wrong as indirectly warranting my affectionate response by demonstrating that you have affection-­ worthy qualities: you’re a special, fun person, say, with a dare-­devil streak—­not one of those lame goody two-­shoes after all. But in blame, a subject takes the target’s wrongdoing to, in and of itself, make her present response fitting—­that is, to make it fitting directly.23 Non-­blaming responses aren’t made fitting by wrongdoing in and of itself, but rather, insofar as they can be fitting at all, made fitting directly by objects possessing other sorts of properties. This is why I think that, although there are many ways in which humans respond to wrongdoing, we find the commitment I’ve described just in those among them that are blame.

16.4  Self-­Reference I have supposed that a blaming reaction may contain any of the diverse ways of reacting that philosophers have thought involved in interpersonal blame, although I’ve also argued that each episode of blame must involve a subject’s taking her present way of reacting to be fitting directly on the basis of the target’s wrongdoing. So, as things stand, one should understand blame as involving a subject’s reflexive endorsement plus a material reaction, where the latter might consist in a way of behaving, a physiological response, or any number of other things. But there is a final twist that must be added to this picture, which will help elucidate how reflexive endorsement makes blame what it is. To see why a final twist is needed, consider how the reference in reflexive endorsement to the blaming subject’s ‘present way of reacting’ is to be understood. Which aspects of what the subject is presently doing count for her as the way of reacting she is endorsing? A natural thought is that what the subject takes to be justified is just whatever the material reaction component of her blaming consists in. For instance, if the material element of her blame consists in not speaking to the person it’s directed at, then what she takes to be justified is her refusal to speak to that person. But this thought may seem strange given that I’ve allowed that the material reaction in an episode of blame can consist in any of the diverse reactions that philo­ sophers have held blame to involve. In particular, it may seem strange given that I’ve allowed that the material element of blame might consist in things like physio­logic­al changes and dispositions, such as, say, one’s blood pressure rising or 23  Or, again, if one wishes to make substitutions: something unique that is very close to wrong­ doing—­a norm violation of some sort.

370 Fittingness a disposition to scream and make certain facial expressions. The idea that these sorts of reactions could be what the blaming subject takes to be justified may seem dubious, since physiological changes and dispositions aren’t really the sort of thing that can be characterized as justified or unjustified. Taking these ways of reacting to be justified, one may worry, would be like taking having a stomach-­ ache to be justified. And yet, the phenomenological considerations raised in Section  16.2 would seem to suggest that even material elements of blame that we would normally consider arational are what the subject takes to be justified. The feeling of ‘boiling blood’, I suggested, can seem to be concerned with a target’s wrongdoing, and postulating a merely causal connection—­even a reliable one—­between a judgement that some target has wronged and a subject’s body temperature rising seems insufficient for capturing what that’s like. Moreover, insofar as emotionally blaming someone seems not merely to involve representing him and his wrongdoing, but also that wrongdoing as justifying how one feels, it’s hard to imagine this ‘how one feels’ bit as constituted by any type of thing that we normally take to be rationally justifiable. It seems as if anything left over in emotional experience once we’ve subtracted out representation of an object is just awareness of the physiological and dispositional changes involved in one’s body readying itself to act in certain ways. Indeed, although the trouble seems starkest for arational material elements of blame, there is actually an issue for the rationally justifiable material elements of blame too. I have claimed that, in reflexive endorsement, a subject takes her present way of reacting to be made fitting directly by the target’s wrongdoing. Yet the mental states that philosophers generally hold to be the rationally justifiable components of blame—­states like beliefs, desires, intentions, and actions—­do not seem to be the sort of thing that normally can be justified directly by wrongdoing. After all, these states all have their own justification conditions. Thus, one would think that the only way for a target’s wrongdoing to justify these sorts of states would be via demonstrating that their justification conditions had been satisfied—­in which case that wrongdoing wouldn’t justify them directly.24 But then we face a problem: it seems as if accounting for what it’s like for blame to be directed requires holding that blamers take the material elements of their blame to be ­justified directly by the target’s wrongdoing. Yet, if neither the arational nor 24  Consider intention. Normally considerations that can serve as reasons for intending to φ are considerations that demonstrate φ-ing worth doing (either as means to our ends or as ends in themselves). So how can a blaming subject take deciding not to speak to someone—­forming an intention—­to be justified directly by a target’s wrongdoing, rather than justified by that wrongdoing merely because it demonstrates giving the target the silent treatment to be worth doing? And yet I have claimed that a blaming subject sometimes does take responses like deciding not to speak to the target to be justified without also taking giving the silent treatment to be worth doing. So she must take the target’s wrongdoing to directly justify this sort of response. How is it possible for her to do this?

Blame ’ s Commitment to Its Own Fittingness  371 rationally justifiable material components of blame are generally justifiable directly by a target’s wrongdoing, then how can blamers take the material components of blame to be justified in this way? Here’s one thought. If it is right to consider blame a way of representing some target’s wrongdoing as making her blameworthy, then it seems reasonable to hold that when a blaming subject takes her present way of reacting to be fitting on the basis of the target’s wrongdoing, what she regards as her ‘present way of reacting’ is blame. Moreover, what I’ve been arguing all along is that reflexive endorsement is an element of blame. Thus, I suggest that, whenever a blaming subject takes her present way of reacting to be fitting, she never takes merely the material components of blame to be fitting. Rather, she also regards her reflexive endorsement itself as part of her present way of reacting that she takes to be justified. The proposal, then, is that blaming involves taking one’s present way of reacting to be justified, where one’s present way of reacting is constituted both by particular physiological, behavioural, or cognitive responses one is presently manifesting and by one’s representation of what one is presently doing as justified. On this proposal, blaming’s element of reflexive endorsement becomes truly reflexive, in the sense of being self-­referential. It is as if the blamer is pointing at her present reaction and saying ‘this way of reacting is fitting’—while understanding that ‘this way of reacting’ includes that very pointing at her present reaction and asserting that it is fitting. This is the final twist to our picture of blame. Does this picture solve the problem of how blamers can take the material elem­ ents of blame to be justified directly by the target’s wrongdoing? Do we not now just have a picture on which our blaming subjects take the target’s wrongdoing to directly justify several things, many—­if not all—­of which can’t be justified in this way? I think we have solved our problem, so long as we don’t view blame’s ma­ter­ ial elements as fully separable from her endorsement itself in the subject’s mind. In endorsing the material elements of her reaction together with her endorsement itself, as I want us to see it, it is not that the subject endorses separate things (despite my having, for ease of exposition, spoken of blame’s elements as if they were separable). Rather, we must hold that the material elements of her blame, by virtue of being represented by the subject’s endorsement as one with that endorsement itself, are transformed into something that they wouldn’t otherwise be. They are transformed, via the union within endorsement with endorsement, into parts of blame, and it is only as parts of a whole unity—­her blame—­that the subject takes blame’s material elements to be justified. In what sense does the subject’s reflexive endorsement transform blame’s ma­ter­ial elements? In the sense that it renders blame’s material elements directed. My idea, in other words, is that material elements of blame which otherwise wouldn’t necessarily be about a person’s wrongdoing—­ranging, potentially, from physio­logic­al changes to overt behaviours—­are made to be about a person’s

372 Fittingness wrongdoing by being united with the subject’s endorsement within that endorsement. Thus, in endorsing the unity born of that very endorsement, a blamer never takes merely any material elements of blame to be justified. Here is my best attempt to flesh out why reflexive endorsement can make blame’s material elements directed. We give shape to our experiences, at least in part, by how we interpret them. And so there seems something plausible in the idea that I could render a material response one that is concerned with a particular person’s particular behaviour by taking that response to be called for by that person’s behaviour: that what could make my reaction directed towards you for what you’ve done is my taking it to be justified—­in the sense of fitting to you—­by what you’ve done. The same idea about giving shape to our own experiences also makes it seem plausible that part of what makes a response have particular justification conditions is viewing it as capable of being justified by particular reasons—­ such that taking a response to be justified directly by someone’s wrongdoing could be part of what makes it the sort of response that can be justified directly by someone’s wrongdoing. A response’s being justifiable directly by wrongdoing, moreover, would also seem to make it a type of response that, whenever it’s ex­peri­enced as for someone’s behaviour, would cast that behaviour as a wrong­ doing. So our ability to give shape to our own experiences through how we interpret them makes it plausible that taking some material response to be made fitting directly by a target’s wrongdoing could make that response have the directedness characteristic of blame. Yet there seems a problem with taking any response that isn’t already relevantly directed to be justified directly by a target’s wrongdoing. How can I take your behaviour to be rendering my response justificatorily well supported unless it’s already a response to you for your behaviour? And how can I take that response to be made fitting directly by your behaviour’s being wrong unless it’s already the type of response that can be made fitting directly by someone’s wrongdoing? My thought is that this is where the reflexivity helps. By endorsing my material responses as parts of a united whole which includes this very endorsement, I can simultaneously take the material parts of blame to be merited by you directly for your wrongdoing and conceive of those parts as already ones that I take to be justified in this way—­and thus as already relevantly directed. So I can make all the material parts of blame directed, and conceive of those parts as I must to bestow that directedness, all in one go. I argued in Section  16.2 that capturing the phenomenal character of blame’s directedness requires holding that blame involves an element of endorsement. In fact, it requires holding that blame involves an element of reflexive endorsement. Holding that blame involves an endorsement is necessary for capturing what it’s like for all the material parts of blame that seem directed to seem this way. Holding that blame involves a reflexive endorsement is necessary for explaining how a subject’s endorsement of blame’s material components is possible. Thus,

Blame ’ s Commitment to Its Own Fittingness  373 attributing reflexive endorsement to blame is necessary for properly characterizing blame’s directedness—­since we can’t properly characterize blame’s directedness without accommodating its phenomenal character. And, indeed, attributing reflexive endorsement to blame isn’t merely necessary for capturing what it’s like for blame to be directed. Rather, blame’s element of reflexive endorsement is what gives blame its characteristic directedness. Blame’s formal element really does form. .  .  . Thinking of blame as involving this sort of transformative self-­reference solves the problem of how a blaming subject can take the material parts of her blame to be justified directly by the target’s wrongdoing. But, additionally, thinking of the element of reflexive endorsement as self-­referential also provides us with further reason to favour the view that blame involves this element. This is because reflexive endorsement, so understood, can help to explain something else about blame. It can help to explain why blaming often involves a feeling that transcends the phenomenal character of its directedness—­a way of feeling that I roughly glossed earlier as one of assuredness in one’s blame. The idea that an attitude’s involving a self-­referential endorsement may, in some cases, imbue that attitude with a more holistic feeling is one I take from a place that may seem far afield of our present discussion: from Kant’s account of aesthetic appreciation of the beautiful. Nevertheless, I think there is something to be learned from briefly reflecting on Kant’s view. As Kant understood it, appreciating a beautiful object involves regarding it disinterestedly—­without any ulterior motive or plan to use the object for some personal purpose—­and then taking one’s current way of regarding that object to be appropriate or legitimate to that object (‘universally valid’) (1790/1987: 58). Kant took this element of endorsement involved in aesthetic appreciation to be self-­referential in the same way I take reflexive endorsement in blame to be. As Ginsborg (1991: 299–300) describes Kant’s view, aesthetic appreciation of the beautiful involves ‘self-­referentially judging that one’s mental state in that very act of judging is universally communicable or universally valid with respect to an object.’ Moreover, Kant thought the self-­referential character of this type of judging gave it a sort of self-­perpetuating, or self-­maintaining, quality and held that our consciousness of this quality is the distinctive feeling of aesthetic pleasure. As Ginsborg explains it, the consciousness that I ought to be in the very same mental state as that in which I presently find myself . . . qualifies as a feeling of disinterested pleasure. For although it involves no interest or desire, the consciousness that I ought to be in my present state of mind supports or maintains itself by serving as a ground or justification for being in that very state of mind.  (302)

374 Fittingness We might think of it thus: to approve of one’s present state of mind, as one does on Kant’s picture of aesthetic appreciation, involves approving of that very state of approval. But then this, in turn, amounts to approving of that very state of approving of one’s present state of approval. Which, in turn, amounts to approving of that very state of approving of approving of one’s present state of approval . . . and so on. This view may seem implausible if we picture the aesthetic appreciator as occurrently thinking that she approves of her approving, etc., but we needn’t picture her in this way. It is rather that the content of her judgement implicitly contains an infinite number of iterations, such that, were she to unpack it, she would find herself committed to approving yet again. We might say that she is in a state of mind that seems to invite its own renewal; each time she approves of what she is doing, the content of that very approval invites her to approve of what she is doing once more. And for Kant, that aesthetic appreciation is a state that continuously invites the subject into another iteration of that very same state explains why aesthetic appreciation feels pleasurable in a distinctive way. The feeling of such a state of engagement with an object is the feeling of being captivated by it. Of course, blaming does not feel like being captivated by a beautiful object. But then there are already dissimilarities we can point to between aesthetic ap­pre­ci­ ation and blame. Although blame’s material components are diverse, I think it fair to say that blame is not ever constituted by strictly disinterested contemplation. Additionally, blame is a way of reacting that the subject takes to be made fitting by another’s wrongdoing, while of course aesthetic appreciation is not. Moreover, I’ve said, being in such a state involves being committed to having a certain standing. But blaming does, on my view, like aesthetic appreciation as Kant conceived of it, involve a self-­referential type of approval. And I think we can recognize a phenomenological characteristic typical to blaming that, together with these other important factors, this self-­referential approval can help to account for. Blaming, I claim, involves taking the way I am reacting right now to be fitting, where the way I am reacting right now includes taking the way I am reacting right now to be fitting. Hence, blaming also amounts to taking taking the way I am reacting right now to be fitting to itself be fitting. But then my blaming must also amount to taking taking taking the way I am reacting right now to be fitting to itself be fitting to itself be fitting . . . and so on. Again, my thought is not that blaming must involve thinking about how the commitment it consists in iterates. Rather, it is that, like aesthetic engagement with beauty on Kant’s view, to blame is to be in a state that, in virtue of its content, invites itself to self-­perpetuate. It is to be in a state that involves determining what’s right, determining oneself to have a certain standing with respect to another, and also approving of one’s making those very determinations—­such that one feels right to be determining what’s right, and perhaps even right to feel right to be determining what’s right . . . and so on. Thus, I think we do well to propose that blame involves a self-­referential commitment to its own fittingness. This commitment helps explain why blamers can feel so

Blame ’ s Commitment to Its Own Fittingness  375 justified in pronouncing judgement, and even so justified in feeling justified, so (goddamned, one wants to say) self-­righteous.25

16.5  Unity in Diversity Postulating reflexive endorsement as blame’s formal element can help to account for the self-­righteous phenomenology that often accompanies blame and explains what makes blame directed. Moreover, if I’m right that reflexive endorsement makes blame directed, then reflexive endorsement is common to all forms of blame—­because one always blames someone for something. But if reflexive endorsement is a necessary and, indeed, essential feature of blame, then there is something significant enough to make blame itself worth inquiring after that ties all the types of blame together. What cooler, more judgemental, private blame has in common with a furious and public dressing down is that they are both instances of the type of response that involves a commitment to its own fittingness directly on the basis of a target’s wrongdoing. Such a commitment is capable of taking diverse forms of reaction and giving them all a, well, blamey shape. There are several questions we might want to ask about blame, so conceived. For instance, while we do reflexively endorse a wide variety of responses, it doesn’t seem as if just anything goes; being in the grips of blame doesn’t seem, as a matter of fact, to involve reflexively endorsing just any old material component. So one thing we may wish to know is why, as a matter of fact, blaming involves reflexively endorsing certain material components but not others. (Why not, for example, just jumping twice?) Is it something in the nature of wrongdoing? Something in our own nature? A matter of convention? Recall also that I have proposed that, because not everyone’s blame is fitting in response to a particular target’s wrong, since blame involves reflexive endorsement, it must also involve various background commitments—­such as commitments to the subject’s standing to blame. So we may also wish to inquire further about which background commitments blame involves, and about the contributions they make to its meaning. Note that the absence of standing to blame a target for a particular wrong doesn’t seem to undermine the fittingness of only certain types of blame. A person who has violated a norm that you, yourself, violate happily all the time merits neither your high-­handed speech nor your silent disapproval.

25  As far as I know, the only other philosopher who has claimed to explain why blaming feels self-­ righteous is Pickard (2013). On her view, blame consists in a first-­order emotional appraisal of some target (e.g. anger towards it), plus a second-­order feeling of entitlement to that first-­order appraisal. But my point has been that taking oneself to be entitled to undergo the material responses involved in blame, in the way characteristic of blame—­that is, taking one’s present response to be fitting, directly on the basis of the target’s wrongdoing—­requires taking that very entitlement to be part of what one is entitled to.

376 Fittingness So when we ask what background commitments concerning standing are involved in blame, we are asking about blame considered as a unity. And such questions certainly seem worth asking. None of this is to say that specific types of blame don’t also sometimes deserve their own attention apart from other types. Blame’s diversity surely demands that. Indeed, part of what I like about the picture that I’ve drawn here is that it allows for blame to be diverse. Something that I, in any case, find interesting about blame is that blame is like an emotion insofar as it’s directed, and yet is also an especially mutable form of response, capable of incorporating a range of other mental and physical states and events—­almost as if blame is an expanded emotion. The thought that blame is like an emotion brings me to the question I want to consider briefly in conclusion. Much of my argument has appealed to blame’s directedness—­a feature it bears in common with emotions. Thus, the question naturally arises: should we say that emotions all involve an element analogous to reflexive endorsement? Should we say, for example, that fear involves taking one’s present way of reacting to be fitting directly on the basis of something dangerous, or that grief involves taking one’s present way of reacting to be fitting directly on the basis of a significant loss? I am content to remain officially neutral on this issue. I hope to have convinced the reader that, in blame, the target’s wrongdoing has the appearance of a reason, and thus that blame involves an endorsement of one’s present way of reacting. So long as these claims are accepted, I won’t be too bothered by anyone who wants to deny that the objects of all emotions have a similar appearance from the ‘inside’, and thus wants to deny the stronger claim that all emotions involve an element analogous to reflexive endorsement. After all, discovering and carefully describing this sort of element is especially important in the case of blame. For while it is true in general that the tokens of a particular emotion-­type vary in their feel and intensity, it isn’t true that emotions in general are as diverse as blame. Blame doesn’t just vary in feel and intensity, but rather also in the very category of response it seems to involve—­so much so that there is reason to be sceptical about whether blame itself, considered as a unified phenomenon, is worth investigating. Holding that blame involves a commitment to its own fittingness directly on the basis of a target’s wrongdoing helps us to see that cases of blame which don’t involve many of the paradigmatic features of emotion—­and that we might therefore hesitate to even call ‘emotion’—still have something significant in common with cases that do. And this makes it particularly illuminating to attribute a distinctive commitment to its own fittingness to blame. Yet, although I would be satisfied to have made my case just for blame, I do believe, for my own part, that all emotions—­or, at least, all my (adult, human) emotions—­involve something analogous to reflexive endorsement. I believe this precisely because I can identify a similarity in what it’s like for blame to be for wrongdoing and what it’s like for emotions to have directedness generally. The

Blame ’ s Commitment to Its Own Fittingness  377 objects of my fear do seem to both explain and justify my fear because of their dangerousness; the objects of my grief do seem to both explain and justify my grief by virtue of being significant losses; etc. But reflexive endorsement, even if it can be unreflective, seems cognitively sophisticated. So those concerned to at­tri­ bute emotions to less cognitively sophisticated animals will likely balk at the suggestion that analogous formal elements are required for capturing the way in which our emotions in general are about objects. Such a view might seem to (unacceptably) imply that many animals don’t have emotions that are about things. What I believe such a view would imply, however, is not that less cognitively sophisticated animals don’t have emotions that represent objects in some sense. (After all, one can always hold that any sort of emotional response performs a representational function.) Rather, what it would imply is that some such animals don’t have emotions that are directed at objects in the same way that ours are, and thus that what it is like for such animals for their fear to be about things is different from what it is like for ours to be. Some philosophers will dislike even this concession, but personally I don’t mind making it. Indeed, I’d be unsurprised to learn that the rational capacities of we creatures who possess them infuse the way we think so thoroughly that they transform even our most primitive emotional experiences.26 Say, though, that I am right to think that all adult human emotions have a component analogous to reflexive endorsement. Would such an element, then, in every case have an effect—­as Kant thought it did for aesthetic appreciation—­on our emotional feeling itself? I admit to being somewhat unsure what to say, in part because I do think that how each emotion feels must be mediated largely by both its material components and by what particular considerations make that emotion fitting. But I do not think it obviously wrong to say that the way in which emotional experiences seem, from the inside, to be called for is part of what explains the feeling of being taken with the objects of our emotions—­part of what makes emotional experiences feel enthralling. On such a picture of emotions, reflexive endorsement might be thought to explain why valuing or disvaluing something feels the way it does.27

26  For this sort of picture of rational minds, see Boyle 2012 and 2016. 27  Research for this work was supported by an ERC advanced grant project, Roots of Responsibility: Metaphysics, Humanity, and Society. This project has received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (grant agreement no. 789270). This chapter has also benefited enormously from feedback on earlier drafts. For written comments, discussion, or both, I thank Diana Acosta Navas, Selim Berker, Sandy Diehl, Lidal Dror, Aaron Elliot, David Enoch, John Hyman, Christine Korsgaard, Jed Lewinsohn, Max Lewis, Jeff McDonough, Oded Na’aman, Yuuki Ohta, Tim Scanlon, Heather Spradley, Christine Tappolet, Daniel Telech, Mark Thomson, and Preston Werner. Thanks also to the editors of this volume, Chris Howard and Rich Rowland, for extremely helpful notes and suggestions.

378 Fittingness

References Achs, Rachel. 2020. ‘Blameworthiness and Constitutive Control.’ Philosophical Studies 177: 3695–715. Boyle, Matthew. 2012. ‘Essentially Rational Animals.’ In Rethinking Epistemology, Vol. 2, edited by G. Abel and J. Conant, 395–428. Boston, MA: De Gruyter. Boyle, Matthew. 2016. ‘Additive Theories of Rationality: A Critique.’ European Journal of Philosophy 24: 527–55. Cohen, G. A. 2006. ‘Casting the First Stone: Who Can, and Who Can’t, Condemn the Terrorists?’ Royal Institute of Philosophy Supplements 58: 113–36. Cowan, Robert. 2016. ‘Epistemic Perceptualism and Neo-sentimentalist Objections.’ Canadian Journal of Philosophy 46: 59–81. D’Arms, Justin, and Daniel Jacobson. 2000a. ‘The Moralistic Fallacy: On the “Appropriateness” of the Emotions.’ Philosophy and Phenomenological Research 61: 65–90. D’Arms, Justin, and Daniel Jacobson. 2000b. ‘Sentiment and Value.’ Ethics 110: 722–48. D’Arms, Justin, and Daniel Jacobson. 2003. ‘The Significance of Recalcitrant Emotions (or, Anti-quasijudgmentalism).’ Royal Institute of Philosophy Supplements 52: 127–45. Deonna, Julien and Fabrice Teroni. 2012. The Emotions: A Philosophical Introduction. New York: Routledge. Döring, Sabine. 2009. ‘The Logic of Emotional Experience: Noninferentiality and the Problem of Conflict without Contradiction.’ Emotion Review 1: 240–7. Evans, Jonathan, and Keith Stanovich. 2013. ‘Dual-Process Theories of Higher Cognition: Advancing the Debate.’ Perspectives on Psychological Science 8: 223–41. Foot, Philippa. 2002. ‘Hume on Moral Judgement.’ In Virtues and Vices, 74–80. Oxford: Oxford University Press. Ginsborg, Hannah. 1991. ‘On the Key to Kant’s Critique of Taste.’ Pacific Philosophical Quarterly 72: 290–313. Glover, Jonathan. 1970. Responsibility. London: Routledge & Kegan Paul. Goldie, Peter. 2000. The Emotions: A Philosophical Exploration. Oxford: Oxford University Press. Goldie, Peter. 2002. ‘Emotions, Feelings, and Intentionality.’ Phenomenology and the Cognitive Sciences 1: 235–54. Kant, Immanuel. 1790/1987. Critique of Judgment. Translated by W.  Pluhar. Indianapolis/Cambridge: Hackett Publishing Company. Kelly, Thomas. 2002. ‘The Rationality of Belief and Some Other Propositional Attitudes.’ Philosophical Studies 110: 163–96. Lazarus, Richard. 1991. ‘Progress on a Cognitive-Motivational-Relational Theory of Emotion.’ American Psychologist 46: 819–34. Maguire, Barry. 2018. ‘There Are No Reasons for Affective Attitudes.’ Mind 127: 779–805.

Blame ’ s Commitment to Its Own Fittingness  379 Menges, Leonhard. 2017. ‘The Emotion Account of Blame.’ Philosophical Studies 174: 257–73. Mitchell, Jonathan. 2019. ‘Emotional Intentionality and the Attitude-Content Distinction.’ Pacific Philosophical Quarterly 100: 359–86. Moors, Agnes. 2013. ‘On the Causal Role of Appraisal in Emotion.’ Emotion Review 5: 132–40. Nussbaum, Martha. 2016. Anger and Forgiveness: Resentment, Generosity, Justice. New York: Oxford University Press. Pennycook, Gordon. 2018. ‘A Perspective on the Theoretical Foundation of Dual Process Models.’ In Dual Process Theory 2.0, edited by W. De Neys, 5–27. New York: Routledge. Pickard, Hanna. 2013. ‘Irrational Blame.’ Analysis 73: 613–26. Prinz, Jesse. 2004. Gut Reactions. Oxford: Oxford University Press. Roberts, Robert. 1988. ‘What an Emotion Is: A Sketch.’ Philosophical Review 97: 183–209. Scanlon, T. M. 2008. Moral Dimensions: Permissibility, Meaning, Blame. Cambridge, MA: Harvard University Press. Schlick, Moritz. 1962. ‘When Is a Man Responsible?’ In Problems of Ethics, translated by David Rynin, 143–58. New York: Dover Publications. Shah, Nishi. 2006. ‘A New Argument for Evidentialism.’ Philosophical Quarterly 56: 481–98. Sher, George. 2006. In Praise of Blame. Oxford: Oxford University Press. Shoemaker, David and Manuel Vargas. 2021. ‘Moral Torch-Fishing: A Signaling Theory of Blame.’ Noûs 55: 581–602. Smith, Angela. 2013. ‘Moral Blame and Moral Protest.’ In Blame: Its Nature and Norms, edited by D.  J.  Coates and N.  Tognazzini, 27–48. Oxford: Oxford University Press. Solomon, Robert. 1973. ‘Emotions and Choice.’ Review of Metaphysics 27: 20–41. Tappolet, Christine. 2012. ‘Emotions, Perceptions, and Emotional Illusions.’ In Perceptual Illusions: Philosophical and Psychological Essays, edited by C.  Calabi, 207–24. New York: Palgrave-Macmillan. Tappolet, Christine. 2016. Emotions, Values, and Agency. New York: Oxford University Press. Todd, Patrick. 2019. ‘A Unified Account of the Moral Standing to Blame.’ Noûs 53: 347–74. Wallace, R.  Jay. 1994. Responsibility and the Moral Sentiments. Cambridge, MA: Harvard University Press. Wolf, Susan. 1990. Freedom within Reason. Oxford: Oxford University Press. Zimmerman, Michael. 1988. An Essay on Moral Responsibility. Totowa, NJ: Rowman & Littlefield.

17 Making Amends How to Alter the Fittingness of Blame Hannah Tierney

17.1 Introduction It is typically thought that all and only blameworthy agents are the fitting targets of blame.1 And on one popular approach to blameworthiness, agents are blame­ worthy in virtue of culpably performing some wrong action. But this view faces a puzzle. If an agent culpably performs some wrong action, then this fact will always be true of them. So, an agent, once blameworthy, will be the fitting target of blame, forever (or, to be more precise, until they cease to exist). Many find this to be counter-­intuitive—­we typically judge that agents who culpably perform wrong actions are not the fitting targets of blame in perpetuity. Most would see no reason to continue to blame themselves or others for decades-­old wrongs, espe­ cially once amends are made and apologies accepted.2 To accommodate the in­tu­ ition that blame is not fitting forever, philosophers have made a distinction between accounts of synchronic blameworthiness—­the set of conditions under which an agent is blameworthy for an action at the time of the action—­and accounts of diachronic blameworthiness—­the set of conditions under which an  agent is blameworthy for an action at some later time (Khoury 2013, 2022; 1  In this chapter, I assume that blameworthiness is necessary and sufficient for blame to be fitting, though I’ll remain as neutral as possible about the nature of the fittingness relation and focus instead on the conditions under which blame is fitting. In assuming that blameworthiness is both necessary and sufficient for the fittingness of blame, I part ways with others who defend views of blameworthi­ ness over time, particularly Khoury, who, in recent work, commits only to the necessity claim (forth­ coming: fn. 1). Though this is a significant difference between our views, it will not affect my criticism of Khoury and Matheson’s (2018) account of diachronic blameworthiness, which only relies on blameworthiness being necessary for blame to be fitting. 2  Discussions of the fittingness of blame over time are closely related to recent debates regarding the fittingness of emotions over time (Callard 2017; Howard 2022; Marušić 2018; Moller 2017; Na’aman 2020). But how closely these debates are related will depend on one’s view of blame. They will overlap to a much greater extent if one defends an emotion-­based view of blame as opposed to a cog­ nitive account, for example. In this chapter, I try to remain neutral on the nature of blame (though I  briefly make reference to an account of blame’s evaluative and communicative content in sec­ tion 17.4.3.) such that it will be possible to answer the question of whether blame can become less fit­ ting over time without also supplying an answer to the question of whether emotions can become less fitting over time (and vice versa). Hannah Tierney, Making Amends: How to Alter the Fittingness of Blame In: Fittingness: Essays in the Philosophy of Normativity. Edited by: Christopher Howard and R. A. Rowland, Oxford University Press. © Hannah Tierney 2022. DOI: 10.1093/oso/9780192895882.003.0017

Making Amends: How to Alter the Fittingness of Blame  381 Matheson 2014; Khoury and Matheson 2018). While accounts of synchronic blameworthiness typically feature conditions for culpable wrongdoing, such as control and epistemic conditions,3 those who wish to argue that blame can become less fitting over time offer accounts of diachronic blameworthiness that go beyond culpability.4 For example, Andrew Khoury and Benjamin Matheson (2018) have developed a view of diachronic blameworthiness in terms of psychological connectedness. On their view, if an individual bears no, or a diminished number of, relevant psy­ cho­logic­al connections to a past individual who culpably committed a wrong action, then the current individual is not blameworthy, or less blameworthy, for that action. This is so even if the current individual is personally identical to the past individual. While Khoury and Matheson allow for a whole host of psy­cho­ logic­al changes to impact diachronic blameworthiness, others focus on more specific psychological states.5 For example, Andreas Carlsson (2017; 2022) and Douglas Portmore (2019; 2022) argue that blameworthy agents deserve to suffer the pains of guilt to a particular degree, and are thus no longer blameworthy, or the fitting targets of blame, once they’ve experienced the requisite amount of guilt. While these accounts of diachronic blameworthiness have many virtues, including their ability to capture the intuition that blame can become less fitting over time, I’ll argue that both views are incomplete. Though changes in psy­cho­ logic­al connectedness may be necessary to diminish diachronic blameworthiness, these changes are not sufficient to render agents entirely unblameworthy for past wrongs. If they were, then privately reformed wrongdoers—­agents who undergo significant psychological changes without making any form of amends to their victims—­wouldn’t be blameworthy for former wrongs. And though experiencing guilt may be necessary for agents to stop being blameworthy for past wrongs, it is not sufficient. If it were, then privately remorseful wrongdoers—­agents who feel the requisite amount of guilt without offering amends—­would not be the fitting targets of blame. I contend that these counter-­intuitive implications indicate that there is more to diachronic blameworthiness and fitting blame than psychological connectedness and deserved guilt. In particular, I’ll defend a reparative account of diachronic blameworthiness, according to which blameworthy agents have 3  Other examples of conditions for culpable wrongdoing include quality of will conditions (Arpaly 2003, Arpaly and Schroeder 2013; McKenna 2012), ownership conditions (Fischer 2012), and tracing conditions (Fischer and Tognazzini 2009). 4  Another way to accommodate this intuition is to make a distinction between being blameworthy and being the fitting target of blame. On such a view, an agent, once blameworthy, will be blame­ worthy forever but it can become less fitting to blame them over time. I don’t take this route, since I take being blameworthy to be necessary and sufficient for blame to be fitting, but if one prefers this way of carving up the conceptual space, then one can simply take the accounts of diachronic blame­ worthiness discussed in this chapter to be an account of the fittingness of blame over time. Thanks to Oded Na’aman for raising this concern. 5  Per-­Erik Milam (forthcoming) also defends a view according to which an agent’s change of heart is sufficient for a reduction in their diachronic blameworthiness.

382 Fittingness rep­ara­tive obligations to their victims and remain blameworthy, and the fitting target of blame, until these obligations are fulfilled. I’ll conclude by defending the rep­ara­tive account from the objection that reparations are irrelevant to diachronic blameworthiness.

17.2  Psychological Connectedness and Guilt-­Based Accounts of Diachronic Blameworthiness The view that blame, if fitting, is fitting forever is an accepted implication of many accounts of blameworthiness. Ledger theorists, for example, like Joel Feinberg, are happy to embrace this consequence.6 He argues: Moral responsibility, so conceived, is a liability to charges and credits on some ideal record, liability to credit or blame (in the sense of “blame” that implies no action). Just as it is, as we say, “forever to the credit” of a hero or saint that he performed some noble act, so a man can forever be “to blame” for his faults. (Feinberg 1970: 30)

On this view, while it may not be appropriate to punish, sanction, or even openly criticize an individual for past wrongs, these wrongs will always reflect poorly on the agents who performed them, and in this sense they will be blameworthy, and the fitting targets of blame, forever. There is very little difference between syn­ chronic and diachronic blameworthiness on such an approach—­ all that is required to be diachronically blameworthy for an action is to be personally iden­ tical to an agent who was synchronically blameworthy for that action. Khoury and Matheson present a helpful formulation of this view of diachronic blameworthiness: Blameworthiness is Forever (BIF): If a person A at t1 is blameworthy to degree d for committing act X at t1, then an individual B at t2 is blameworthy to degree d for act X if A is personally identical with B.  (2018: 207)

Recently, BIF has come under attack and philosophers have begun to develop accounts of diachronic blameworthiness that can compete with it. Khoury and Matheson (2018) argue that synchronic blameworthiness coupled with personal identity is not sufficient for diachronic blameworthiness, for the diminishment of psychological connectedness alone can impact diachronic blameworthiness. And Carlsson (2022) and Portmore (2019; 2022) argue that an agent can cease to be 6 Other theorists who defend the view that blameworthiness is forever include: Haji 1998; Tognazzini 2010; Zimmerman 1988.

Making Amends: How to Alter the Fittingness of Blame  383 blameworthy even if they are identical to an agent who performed a past wrong, so long as the agent experiences the deserved amount of guilt in virtue of their culpability. In this section, I present the psychological connectedness and guilt-­ based approaches to diachronic blameworthiness.

17.2.1  Psychological Connectedness and Diachronic Blameworthiness In arguing that personal identity is not sufficient for diachronic blameworthiness, Khoury and Matheson highlight an important feature of this relation: it is pos­ sible for an individual to be personally identical to a past individual despite shar­ ing no distinctive psychological features with that individual. On Khoury and Matheson’s view, distinctive psychological features are those that tend to vary across persons—­memories, desires, character traits, beliefs, etc. (2018: 211). And even on views of personal identity that take these states to play an important role in persistence, like the psychological approach to personal identity, it will be pos­ sible for a present individual to be identical to a past individual despite having none of the same memories, desires, character traits, beliefs, etc.7 This is because the psychological approach analyses personal identity in terms of psychological continuity, which consists in overlapping chains of strong psychological connect­ edness (Parfit 1984). And while individuals who are strongly psychologically con­ nected will share a substantial number of distinctive psychological features, overlapping chains of strong psychological connectedness can form between individuals who don’t possess any of the same psychological features. For ex­ample, I now am strongly psychologically connected to one-­year-­ago me—­both individuals possess many of the same memories, desires, character traits, and beliefs. And one-­year-­ago me is strongly psychologically connected to two-­years-­ ago me, who is strongly psychologically connected to three-­years-­ago me, and so on and so forth all the way back to thirty-­years-­ago me. So, I now am psy­cho­ logic­ally continuous with thirty-­years-­ago me, since there is an overlapping chain of strong psychological connectedness between us, and this is true even though I  now have none of the same memories, desires, character traits or beliefs as thirty-­years-­ago me. The fact that agents can be personally identical to individuals they are not psy­ cho­logic­ally connected to is not objectionable—­we want our theories of personal identity to explain how an individual can persist while also changing over time. But what works for a theory of persistence over time does not necessarily work 7  According to biological accounts of personal identity (e.g. DeGrazia 2005; Olson 1997), personal identity involves the persistence of biological animals, and this does not require the persistence of any psychological features at all.

384 Fittingness for a theory of blameworthiness over time. And according to BIF, agents will be blameworthy for the wrongdoings of individuals they are personally identical to, despite sharing no psychological features with these individuals. This, according to Khoury and Matheson, is counter-­intuitive. Consider an agent who performs a wrong action early in their life, and is synchronically blameworthy for this wrong, but then goes on to develop into a moral saint who has no memory of this earlier wrongdoing, shares no objectionable character traits with their youthful counter­ part, and fails to possess any moral failings of any kind. Surely the moral saint is not blameworthy for the youthful indiscretion. Khoury and Matheson argue: It seems that whether blame is appropriately directed at a person at a time t must depend in some way on what that person is like at t . . . blameworthiness seems to involve a particular kind of criticizability—­that is, a blameworthy individual is criticizable in light of a particular kind of flaw. A moral saint, however, is not so criticizable because she, by definition, lacks any such flaws.  (2018: 214)

On Khoury and Matheson’s view, being blameworthy involves being flawed in a particular way. But the moral saint doesn’t possess the flawed psychological fea­ tures that gave rise to the earlier performance of a wrong action. Indeed, they don’t possess any moral flaws at all. Thus, the moral saint is not blameworthy for the wrongdoing they committed in their youth because they share no relevant psychological features with their younger self. This indicates that psychological connectedness, and not personal identity, is the relation relevant for diachronic blameworthiness, which leads Khoury and Matheson to develop the psy­cho­ logic­al connectedness account of diachronic blameworthiness (PCA). First, they argue that maximal psychological connectedness is sufficient for diachronic blameworthiness: PCA’s sufficiency claim: If a person A at t1 is blameworthy to degree d for act X which occurs at t1, then an individual B at t2 is blameworthy for X to degree d if B is maximally psychologically connected to A. (2018: 216)8

They go on to argue that some degree of psychological connectedness is necessary for diachronic blameworthiness and mitigated psychological connectedness also mitigates diachronic blameworthiness: ‘When there is no distinctive psy­cho­ logic­al connectedness, there is no diachronic blameworthiness . . . Furthermore, if 8  Unlike personal identity, psychological connectedness is not a one-­to-­one relation, and it’s pos­ sible for an individual at t1 to be maximally psychologically connected to two or more individuals at tn (as in cases of fission, for example). It’s not clear whether Khoury and Matheson would be happy to accept the implication of their view that the individuals at tn would both be diachronically blame­ worthy, and the fitting targets of blame, for wrongs that the individual at t1 committed or if they’d prefer to amend their view to claim that maximal and unique psychological connectedness is sufficient for diachronic blameworthiness.

Making Amends: How to Alter the Fittingness of Blame  385 there is diminished relevant psychological connectedness, then we think there is diminished blameworthiness’ (2018: 217). We can represent these claims in the following way: PCA’s necessity claim:  If a person A at t1 is blameworthy to degree d for act X which occurs at t1, then an individual B at t2 will not be blameworthy to any degree for X if B shares no distinctive psychological connections to A. PCA’s scalar claim:  If a person A at t1 is blameworthy to degree d for act X which occurs at t1, then an individual B at t2 will be blameworthy to a degree less than d for X if B is not maximally relevantly psychologically connected to A. These claims are squarely at odds with BIF, for they indicate that an agent can fail to be diachronically blameworthy, or will be less blameworthy, for past wrongs even if the agent is identical to the individual who committed these wrongs. And because most agents’ distinctive psychological features change over time, it will be rare (though not impossible) for an agent to be maximally diachronically blame­ worthy for past wrongs for their whole lives. Thus, on PCA, blame will rarely be fitting forever.

17.2.2  Guilt and Diachronic Blameworthiness Not everyone who rejects BIF accepts PCA. Carlsson (2022), for example, argues that psychological connectedness can be diminished in ways that do not diminish the degree to which agents are diachronically blameworthy. If an agent acts wrongly, and then their distinctive psychological features are altered by accident, or because they took a pill to purposefully diminish the degree to which they are psychologically connected to the agent that committed the wrong act, then the agent would not intuitively be less blameworthy for their wrongdoing (Carlsson 2022: 182). For Carlsson, diachronic blameworthiness does not depend on the persistence of an agent’s distinctive psychological features. Rather, it depends on the experience of a particular psychological state: guilt. In a series of recent papers, both Carlsson (2017; 2022) and Portmore (2019; 2022) present guilt-­based accounts of blameworthiness (GBA) that give rise to an interesting solution to the puzzle of blameworthiness over time. Though there are important differences between their accounts of blameworthiness, both argue that an agent is blameworthy for performing some action if and only if they deserve to feel guilty (or regretful or remorseful) for performing the action (Carlsson 2022: 190; Portmore 2022: 63). So, if a blameworthy agent experiences none of the guilt they deserve to experience in virtue of being blameworthy for a past action, then it will continue to be true that they deserve to feel guilty for this action and will continue to be blameworthy for it. And, once a blameworthy agent

386 Fittingness experiences the deserved amount of guilt for acting wrongly, then it will be false that they continue to deserve to feel guilty for their wrongdoing and they will cease to be blameworthy. Both theorists also argue that the degree to which an agent is blameworthy over time depends on the degree of deserved guilt they experience—­the more deservedly guilty an agent feels, the less blameworthy they will be (Carlsson 2022: 193; Portmore 2022: 70). We can represent these claims in the following way: GBA’s sufficiency claim:  If a person A at t1 is blameworthy to degree d for act X which occurs at t1, then A at t2 will be blameworthy for X to degree d if A experi­ ences none of the guilt they deserve to experience in virtue of being blameworthy to degree d for X at t1. GBA’s necessity claim:  If a person A at t1 is blameworthy to degree d for act X which occurs at t1, then A at t2 will not be blameworthy for X if A experiences all of the guilt they deserve to experience in virtue of being blameworthy to degree d for X at t1. GBA’s scalar claim:  If a person A at t1 is blameworthy to degree d for X which occurs at t1, then A at t2 will be blameworthy to a degree less than d for X if A experiences some of the guilt they deserve to experience in virtue of being blame­ worthy to degree d for X at t1. GBA can easily make sense of the intuition that blame becomes less fitting over time, since blameworthy agents often feel at least somewhat guilty about their past wrongdoings, particularly once they are blamed by those they have wronged. This places GBA at odds with BIF. According to GBA, an agent can cease being blameworthy for past wrongs by feeling deservedly guilty for these wrongs even if they remain personally identical to the individual who committed these wrongful actions. And because the experience of guilt comes apart from psychological con­ nectedness, GBA and PCA will issue different verdicts about diachronic blame­ worthiness in certain cases. For example, an agent who becomes less psychologically connected over time without ever experiencing any degree of guilt will be less diachronically blameworthy for past wrongs on PCA but remain equally blameworthy across time on GBA.

17.2.3  Reformed and Remorseful Wrongdoers Though these accounts differ in important ways, PCA and GBA will issue the same judgements of diachronic blameworthiness in a wide range of situations. Take the following pair of cases:

Making Amends: How to Alter the Fittingness of Blame  387 Reformed and Remorseful Gardener:  Edie asks her friend Jackie to care for her houseplants while she’s travelling for work. Edie loves her houseplants very much, and writes up detailed instructions for Jackie on how to care for them. But Jackie never reads the instructions, exhibiting a culpable degree of carelessness. Jackie waters the plants twice a day, which she would have known is far too often if she had read Edie’s instructions. By the time Edie returns home, most of the plants have died due to overwatering. When it comes to light that Jackie over­watered the plants because she didn’t read Edie’s instructions, Edie blames Jackie for being so careless. Jackie realizes that she was wrong to be careless, feels guilty about her behaviour, and apologizes profusely to Edie. She offers to buy her new house­ plants and promises to be more attentive in the future. Over time, Edie succeeds in becoming a more attentive person and never treats her friends’ possessions carelessly again. Unchanging and Unrepentant Gardener:  A year later, Edie asks another friend, Lee, to look after her plants while she travels for work. However, Lee, like Jackie, also fails to read Edie’s detailed instructions due to culpable carelessness. And, like Jackie, Lee overwaters the plants, killing most of them by the time Edie comes home. When it comes to light that Lee overwatered the plants because she didn’t read Edie’s instructions, Edie blames Lee for being careless, just as she had blamed Jackie. However, unlike Jackie, Lee does not realize that she was wrong to be careless or feel guilty about her behaviour, even after being blamed by Edie. She does not apologize, offer to buy Edie new houseplants, or promise to be a more attentive plant sitter in the future. Over time, Lee remains exactly as careless as she was with Edie’s houseplants. Though both Jackie and Lee performed the same wrong action, and are equally synchronically blameworthy when they perform it, Jackie is intuitively less blameworthy than Lee as a result of the way she responds to Edie’s blame. Many would find it unfitting for Edie to continue to blame Jackie to the same degree after Jackie expressed remorse for her action, apologized, and bought Edie new houseplants. It certainly seems permissible for Jackie to object to being the target of Edie’s continued blame once she recognized the error in her ways, made amends, and did all she could to become a more attentive friend. In contrast, there would be nothing objectionable about Edie continuing to blame Lee long after the incident occurred. As long as Lee remains an unrepentant and careless person, it will intuitively be fitting for Edie to blame her. Both PCA and GBA can easily make sense of the judgement that Jackie becomes less blameworthy than Lee over time, though these views provide dis­ tinct explanations for why this is so. According to PCA, agents become less blameworthy for past wrongs as they become less distinctively psychologically

388 Fittingness connected to the agents that committed these wrongs, and once they share no distinctive psychological features with their former selves, they are no longer blameworthy at all. So, as Jackie works to reform herself into a less careless per­ son, she becomes less diachronically blameworthy, and when she eventually gets rid of the flawed psychological features that gave rise to her careless action, she will cease to be blameworthy entirely.9 In contrast, the flawed psychological fea­ tures that gave rise to Lee performing a wrong action persist over time—­Lee remains exactly as careless as she was when she overwatered Edie’s plants. Thus, on PCA, Lee will remain exactly as blameworthy as she was when she committed the careless action. According to GBA, agents become less blameworthy for past wrongs as they experience the guilt they deserve to feel and once they feel precisely the degree of guilt they deserve to feel, they are no longer blameworthy. On this view, Jackie becomes less blameworthy in the wake of being blamed by Edie not because she becomes a more attentive person but because she begins to feel guilty for her culp­ably careless behaviour. And Lee remains equally diachronically blame­ worthy not because she fails to become a better, less careless person, but because she never feels guilty for her carelessness. In this way, PCA and GBA both provide promising accounts of diachronic blameworthiness, each offering an intuitively plausible explanation of how blame can become less fitting over time. However, in the next section, I’ll argue that PCA and GBA fail to offer a complete account of how and why agents can become less blameworthy over time. While psychological change and the experience of guilt may contribute to, and even be necessary for, the diminishment of ­diachronic blameworthiness, these factors are neither independently nor jointly sufficient.

17.3  Privately Reformed and Remorseful Wrongdoers Though PCA and GBA can each capture the intuitive differences in blameworthi­ ness between the reformed and remorseful and unchanging and unrepentant gardener cases, both views also imply that there are differences between cases where there is intuitively none to be found. Consider a modified version of the reformed and remorseful gardener case:10 9  In this chapter, I’m assuming that it is the flawed psychological features that gave rise to a syn­ chronically blameworthy action that must change in order for agents to become less blameworthy, which coheres well with Khoury and Matheson’s discussion of why the moral saint is not blameworthy for their youthful indiscretions. However, Khoury and Matheson intend to only give a model of dia­ chronic blameworthiness in this paper, and leave open the possibility that a change in any psy­cho­ logic­al connection could impact diachronic blameworthiness (2018: 219), though they take stronger stances on the nature of the relevant psychological connections in other work (Khoury 2013, 2022; Matheson 2014). 10  I consider similar cases in ‘Expanding Moral Understanding’ (2021a) and ‘Guilty Confessions’ (2021b).

Making Amends: How to Alter the Fittingness of Blame  389 Privately Reformed and Remorseful Gardener:  Edie asks her friend Jackie* to care for her houseplants while she’s travelling for work. Edie loves her house­ plants very much, and writes up detailed instructions for Jackie* on how to care for them. But Jackie* never reads the instructions, exhibiting a culpable degree of carelessness. Jackie* waters the plants twice a day, which she would have known is far too often if she had read Edie’s instructions. As the plants start to die from overwatering, Jackie* realizes the error in her ways and feels incredibly guilty for being so careless with her friend’s plants. She disavows her careless ways and takes immediate action towards becoming a more attentive friend. By the time Edie returns home, Jackie* has experienced precisely the amount of guilt she deserved to feel in virtue of her culpable carelessness and has evolved into an attentive and caring person. When it comes to light that Jackie* overwatered the plants because she didn’t read Edie’s instructions, Edie blames Jackie* for being so careless. In response, Jackie* apologizes profusely, offers to buy Edie new houseplants, and promises to be a more attentive plant sitter in the future. According to both PCA and BGA, there is a significant difference between Jackie and Jackie* at the time Edie blames them. Because Jackie does not undergo psy­ cho­logic­al change or feel guilty prior to Edie’s arrival, she is blameworthy, and the fitting target of blame, when Edie blames her. But Jackie* has already eliminated the relevantly flawed features of her psychology and experienced the exact amount of guilt she deserves to feel before being blamed by Edie. So, according to both PCA and GBA, Jackie* is not blameworthy at the time Edie blames her—­the blame is unfitting. But this is counter-­intuitive. Edie’s blame is intuitively fitting in the privately reformed and remorseful gardener case, just as it is in the reformed and remorseful gardener case. Jackie* certainly takes Edie’s blame to be fitting—­ she accepts and responds to Edie’s blame just as Jackie does: by apologizing pro­ fusely, offering to buy her new houseplants, and promising to do better in the future. And this seems like the morally appropriate response in both cases. It would be bizarre if Jackie*, upon being confronted for the first time by Edie, objected to her blame, arguing that she had already changed and/or felt guilty enough such that Edie’s blame is unfitting. Notice that this is so even though Jackie* is intimately familiar with the fact that she has changed and suffered suf­ ficiently for her wrongdoing. Of course, just because blame is unfitting doesn’t entail that it is all-­things-­ considered inappropriate. And PCA and GBA theorists could attempt to accom­ modate the intuition that Edie’s blame is appropriate in the immediately reformed and remorseful gardener case without granting that it is fitting by appealing to epistemic considerations. After all, Edie doesn’t know that Jackie* has become a sufficiently attentive person or suffered to the deserved degree prior to being blamed. And Jackie* knows that Edie doesn’t know the degree to which she’s changed and suffered. This could explain both why we take Edie to be justified in

390 Fittingness blaming Jackie* and Jackie’s* acceptance of the blame as appropriate. Just as a belief can be justified without being true, an expression of blame can be justified without being fitting. Edie has good reasons to think that Jackie* is blameworthy: she performed a wrong action, possessed the agential capacities to be able to avoid performing the action, and had the epistemic capacities to be able to know its moral status. This information could justify Edie’s blaming response without rendering it fitting. And, because Jackie* knows that Edie’s blame is justified, this could explain why it would be inappropriate for Jackie* to object to it. I do not deny that epistemic considerations can render unfitting blame all-­ things-­considered appropriate. It is certainly true that we are sometimes unaware, or only dimly cognizant, of others’ psychological changes and emotional states, and this can lead to a whole host of justifiable but unfitting reactions, including blame. But this isn’t always the case. People will often let us know how they’ve changed and what emotions they are experiencing, and reform and remorse are no exception. In fact, we can imagine a version of the privately reformed and remorseful gardener case where this is true. Expressively Reformed and Remorseful Gardner:  Edie asks her friend Jackie** to care for her houseplants while she’s travelling for work. Edie loves her house­ plants very much, and writes up detailed instructions for Jackie** on how to care for them. But Jackie** never reads the instructions, exhibiting a culpable degree of carelessness. Jackie** waters the plants twice a day, which she would have known is far too often if she had read Edie’s instructions. As the plants start to die from overwatering, Jackie** realizes the error in her ways and feels incredibly guilty for being so careless with her friend’s plants. She disavows her careless ways and takes immediate action towards becoming a more attentive friend. By the time Edie returns home, Jackie** has experienced precisely the right amount of guilt she deserved to feel in virtue of her carelessness and has evolved into an attentive and caring person. As soon as Edie walks in the door, Jackie** tells Edie that she’s overwatered the plants because she carelessly didn’t read Edie’s instruc­ tions and that she feels incredibly guilty for doing so. She also tells her how much more attentive and caring she’s become since Edie left town. Upon hearing Jackie**’s confession, Edie blames Jackie** for being so careless. In response, Jackie** apologizes profusely, offers to buy her new houseplants, and promises to be a more attentive plant sitter in the future. In this case, Edie is quite aware that Jackie** has significantly reformed and ex­peri­enced the amount of guilt she deserves to feel for what she’s done. So, Edie’s blame cannot be justified by appeal to epistemic considerations. Still, it strikes me that Edie’s blame of Jackie** is fitting in this case. We often continue to blame those who wrong us, even when we know they have changed significantly and feel guilty for what they’ve done; this is particularly true when we first learn that

Making Amends: How to Alter the Fittingness of Blame  391 someone has wronged us through their expression of reform and remorse. And just as it would be bizarre for Jackie* to object to Edie’s blame in the privately reformed and remorseful gardener case, it would also be bizarre if Jackie** objected to being blamed in the expressively reformed and remorseful gardener case. Often, moral reform prompts us to confess to those we’ve wronged, and express guilt for what we’ve done, in order to give them the opportunity to blame us, not to shield ourselves from their blame. Presumably, this is because we think it is fitting for our victims to blame us. In this way, reform and remorse often lead us to see others’ blame as fitting, contrary to what PCA and GBA would predict. And when agents reference their own reform or remorse in an attempt to deflect others’ blame, this is rarely taken as a reason to think that the blame is unfitting. These observations indicate that Edie’s blame is appropriate not because of epi­ stem­ic considerations, but because it is fitting. And this is so despite the fact that Jackie** is no longer a careless person and has suffered the deserved degree of guilt prior to being blamed by Edie. If this is right, then PCA and GBA are incom­ plete and there is likely more to the reduction of blameworthiness over time than reform and remorse. In the next section, I propose a competing account of dia­ chronic blameworthiness.

17.4  Repair and Diachronic Blameworthiness Why do we think that Jackie* and Jackie** are blameworthy and Edie’s blame is fitting in the privately and expressively reformed and remorseful gardener cases? I suspect that it is because Jackie* and Jackie** have yet to do all that is required of them in virtue of their culpable wrongdoing and Edie’s blame both reflects this fact and calls upon them to fulfil their obligations. Although Jackie* and Jackie** are sufficiently reformed and experience the deserved degree of guilt, it isn’t until they are blamed that they apologize, offer to buy Edie new houseplants, and promise to be more attentive. Indeed, in the privately reformed and remorseful gardener case, Jackie* doesn’t even admit to wrongdoing prior to being blamed by Edie. But acknowledgement, apology, and restitution are all things we typically think blameworthy agents owe to those they have wronged. Though PCA and GBA theorists focus only on what occurs inside blameworthy agents’ heads, the idea that blameworthy agents also have duties to their victims and must take action to fulfil them is familiar in the literature (Nelkin 2013; Radzik 2009; Walker 2006; Wallace 2019). And just as many theorists take blameworthy agents to owe something to their victims, they also take blame to express a demand or call for these obligations to be fulfilled (Macnamara 2013; McGeer 2013; Smith 2013). But these familiar features of both blameworthiness and blame are missing from PCA and GBA. It’s possible for an agent to undergo significant psychological change and experience a deserved amount of guilt without attempting to fulfil

392 Fittingness any of their reparative duties to their victims. And on these views, once an indi­ vidual is sufficiently psychologically disconnected to the agent who performed a wrong action, or feels as guilty as they deserve to feel, they are no longer blame­ worthy and are thus neither subject to the reparative obligations that come along with blameworthiness nor the fitting targets of blaming reactions that call for the agent to fulfil these obligations.11 But this is deeply counter-­intuitive. Jackie* and Jackie** owe Edie an apology, and this remains true even after they become atten­ tive people and experience a deserved amount of guilt. Thus, in order to make sense of the full range of gardener cases, a theory of diachronic blameworthiness must focus not only on what happens in blameworthy agents’ heads but also the kinds of obligations blameworthy agents must fulfil. Such a view will also provide a satisfying account of how blame can become less fitting over time. I will present a sketch of just such a view in this section.

17.4.1  Obligations to Repair Why think that blameworthy agents have reparative duties to those they wrong? An answer to this question can be found in Linda Radzik’s (2009) work on atone­ ment. Radzik builds on the view that blameworthy actions express a kind of dis­ respect and/or threat to their victims and negatively impact the wrongdoers and victims’ moral relationships. These views of blameworthy action are quite com­ mon in the literature. For example, Jeffrie Murphy argues ‘[moral] injuries are also messages—­symbolic communications. They are ways a wrongdoer has of say­ ing to us, “I count but you do not,” “I can use you for my purposes,” or “I am up here on high and you are down there below”’ (Murphy and Hampton 1988: 25). And quality of will theorists like Nomy Arpaly (2003) and Michael McKenna (2012) take blameworthy actions to express ill will (or insufficient good will) towards their victims, while Pamela Hieronymi argues that past wrongdoings persist as threats if certain conditions are not met: I suggest that a past wrong against you, standing in your history without apology, atonement, retribution, punishment, restitution, condemnation, or anything else that might recognize it as a wrong, makes a claim. It says, in effect, that you can be treated this way, and that such treatment is acceptable. That—­that claim—­is what you resent. It poses a threat.  (2001: 546)

Finally, T. M. Scanlon defends a relational view of blameworthy actions, accord­ ing to which blameworthy actions are those that impair an agent’s relationship 11  Defenders of PCA and GBA can attempt to accommodate to existence of reparative obligations and the role they play in our blaming practices by arguing that something other than blameworthiness confers these obligations. I consider this possible line of defence in section 17.5.

Making Amends: How to Alter the Fittingness of Blame  393 with others (2008; 2013). Building upon the idea that blameworthy actions express disrespect, pose a threat, and damage moral relationships, Radzik argues that blameworthy agents have an obligation to eliminate the disrespect, defuse the threat, and repair the damage to their victims and relationships that their actions have caused.12 But what, precisely, do these reparative obligations amount to? According to Radzik, blameworthy agents are required to achieve three goals: First, the wrongdoer must morally improve herself. It is not enough to convince herself or others that she is trustworthy. She must actually become trustworthy. Second, the wrongdoer must communicate with the victim and in some cases the community in a way that withdraws the insult and the threat that the wrong­ ful act expressed. Third, the wrongdoer must make reparation for the various sorts of harms she created.  (2009: 85)

Interestingly, reform and remorse both have roles to play in Radzik’s reparative account. The first goal of moral improvement can only be achieved if the agent undergoes psychological change and eliminates the flawed psychological features that gave rise to their blameworthy action. And, according to Radzik, ex­peri­en­ cing guilt and remorse (as well as shame and humility) are necessary to achieve the first goal of moral improvement as well (2009: 87–90).13 Reform and remorse also play a role in achieving the second goal, for communicating that these changes have occurred and expressing one’s guilt and remorse are important aspects of eliminating the disrespect and withdrawing the threat that a previous wrong action generated (2009: 88). But, for Radzik, there is much more to the reparative process than experiencing and expressing psychological change and self-­critical emotions. Blameworthy agents must also make reparations for the harms they’ve caused, be they material, physical, psychological, or relational. Importantly, moral improvement and the experience of guilt alone cannot ful­ fil the third goal of making reparations. Morally improving and feeling guilty for past wrongdoing will likely motivate agents to repair the harms they’ve caused, but reform and remorse do not in and of themselves constitute a form of rep­ar­ ation. And informing others that one has morally improved and feels guilty for past wrongdoing will likely communicate that the agent has disavowed the action

12  Radzik also discusses the ways in which blameworthy actions damage wrongdoers’ relationships with the moral community and themselves (2009: 80–3). This will no doubt give rise to a distinct set of reparative obligations, but in order to be concise I focus only on the obligations blameworthy agents have to their victims in this chapter. 13  However, Radzik argues that these emotions are required not because they generate deserved suffering (though she does not deny that they in fact are painful), but rather because they fulfil an important reparative function. In this chapter, I’ll remain neutral on whether blameworthy agents deserve to feel guilty or are simply required to do so in order to complete the reparative process.

394 Fittingness and defused it as a threat, but again, this alone cannot serve as an act of rep­ar­ ation. As I’ve argued in earlier work: Being told by the person who wronged you that they feel very bad for what they have done is a lot like a third-­party report on the wrongdoer’s guilt. While it might give a victim some comfort to know that the person who wronged them blames themselves for their behaviour, this alone cannot repair the relationship. Victims should be given moral attention and care in light of being wronged, but expressions of self-­blame that focus only on the negative affect of feeling guilt cannot do this.  (2022: 129).

The same is true for moral reform. Being told by a person that they no longer possess the same character traits, values, or desires that caused them to wrong you in the past is valuable in that it may give you good reason to think that they won’t wrong you in a relevantly similar way in the future. But this doesn’t address the harm that was generated by the past wrong. Knowing that a friend will not steal from you in the future does not replace what they stole from you in the past, just as knowing that a friend is no longer a liar will not repair the psychological and relational harms that their past lies caused. In order for blameworthy agents to make reparations, they must do more than report to their victims the extent to which they have changed or suffered in light of their wrongdoing. They must also provide their victims with special forms of moral attention and concern. While the precise form of attention that is owed will vary with context, typical forms of re-­address include apology, material restitu­ tion, acceptance of blame, and perhaps even punishment.14 And this duty cannot be fulfilled by a blameworthy agent’s becoming psychologically disconnected to the person they were when they did wrong or experiencing any amount of guilt. This is so even if the blameworthy agent informs their victim of the degree to which they have reformed and felt remorse.

17.4.2  The Reparative Account of Diachronic Blameworthiness I am now in a position to offer a reparative account (RPA) of diachronic blame­ worthiness that can compete with PCA and GBA. On this view, agents who are blameworthy for performing wrong actions must fulfil a set of reparative obliga­ tions in virtue of their culpable wrongdoing and agents who must fulfil reparative obligations in virtue of culpably doing wrong are blameworthy for this wrong. So, 14  In future work, I plan to go into more detail on how blameworthy agents can fulfil their rep­ara­ tive obligations to their victims. In this chapter, it will be sufficient to argue that blameworthy agents in fact have reparative obligations to their victims and the mere experience, and even expression, of psychological change and guilt cannot discharge these obligations.

Making Amends: How to Alter the Fittingness of Blame  395 if an agent is blameworthy for a past wrong, then they will continue to be blame­ worthy at future times if it continues to be true that they must fulfil their repara­ tive obligations. And if an agent who is blameworthy for a past wrong continues to be blameworthy for this wrong into the future, then this agent will continue to be required to fulfil their reparative obligations. However, once an agent fulfils these reparative obligations, it will be false that they must continue to fulfil them, and they will cease to be blameworthy and the fitting target of blame. So, the degree to which blameworthy agents fulfil their reparative obligations will impact the degree to which they are blameworthy, such that the more one does to fulfil these obligations, the less blameworthy they become. We can represent these views of diachronic blameworthiness in the following way: RPA’s sufficiency claim:  If a person A at t1 is blameworthy to degree d for act X at t1, then A at t2 will be blameworthy for X to degree d if A must fulfil all of the reparative obligations they possess in virtue of being blameworthy to degree d for X at t1. RPA’s necessity claim:  If a person A at t1 is blameworthy to degree d for act X at t1, then A at t2 will not be blameworthy for X if A completely fulfils the reparative obligations they possess in virtue of being blameworthy to degree d for X at t1. RPA’s scalar claim:  If a person A at t1 is blameworthy to degree d for act X at t1, then A at t2 will be blameworthy for X to a degree less than d if A partially fulfils the reparative obligations they possess in virtue of being blameworthy to degree d for X at t1. With RPA in hand, let’s return to the series of gardener cases. Like PCA and GBA, RPA can explain why Jackie in the reformed and remorseful gardener case is less blameworthy than Lee in the unchanging and unrepentant gardener case once they have each responded to Edie’s blame. According to RPA, Jackie is less blame­ worthy than Lee after responding to Edie’s blame because she begins to fulfil her reparative obligations to Edie while Lee does not. Jackie makes progress towards all three reparative goals: the remorse she feels contributes to her moral improve­ ment, her apology withdraws the disrespect her past wrong communicates and the threat it poses, and her offer to buy new houseplants and her promise to become more attentive begins to repair the harm done to Edie and their relation­ ship. In contrast, Lee doesn’t take any steps to fulfil her reparative obligations—­ she doesn’t morally improve, or withdraw the disrespect and threat her past action communicates, or offer any reparations to Edie. And this is why Lee remains blameworthy for her culpably careless behaviour even after Edie blames her. RPA can also accommodate the intuition that Jackie* and Jackie** are blame­ worthy in the privately and expressively reformed and remorseful gardener cases, unlike its competitors. Though Jackie* and Jackie** undergo significant psy­cho­ logic­al change and feel exactly as guilty as they deserve to feel for being culpably

396 Fittingness careless with Edie’s plants, neither offers reparations prior to being blamed by Edie. While Jackie** acknowledges her wrongdoing, expresses how much she’s changed and how guilty she feels, these cannot function as a form of reparation and she fails to apologize, offer to buy Edie more houseplants, or promise to do better in the future until Edie blames her. And according to RPA, blameworthy agents remain at least somewhat blameworthy until all of their reparative obliga­ tions are fulfilled. So, on this view, Jackie* is blameworthy at the time that Edie blames her. In contrast with PCA and CBA, then, RPA indicates that there is nothing inappropriate about Edie’s response to Jackie* and Jackie**—in fact, it is perfectly fitting. Before exploring the implications of RPA, it will be useful to clarify a few fea­ tures of the view. First, RPA is largely compatible with PCA’s and GBA’s scalar claims. Because one of the reparative goals is moral improvement, and moral improvement requires psychological change and the experience of guilt, these things will decrease the degree to which agents are diachronically blameworthy on RPA, just as they do on PCA and GBA.15 Likewise, if a blameworthy agent experiences no guilt or psychological change, then this agent will be at least some­ what diachronically blameworthy on RPA. However, such agents will not neces­ sarily be maximally diachronically blameworthy on this view, since it’s possible (though unlikely) that such an agent could fulfil some of their reparative obliga­ tions, like providing material restitution or accepting punishment, without undergoing a change in their relevant psychological features or experiencing any degree of guilt. In this way, RPA rejects both PCA’s and GBA’s sufficiency claims. RPA is also incompatible with these views’ necessity claims. An agent could be diachronically blameworthy for a past wrong on RPA even if they are not psy­cho­ logic­ally connected to the agent who performed the wrong action and even if they experienced the deserved amount of guilt for performing this action. Of course, such agents will be less blameworthy than they would otherwise be if they didn’t morally reform or feel remorse, but they will still be somewhat blame­ worthy if they fail to make reparations for the harm their culpably wrong actions caused.

17.4.3  Implications and Advantages RPA has several implications and advantages that are worth highlighting. First, it will be possible for two agents to be equally synchronically blameworthy for

15  Not all psychological changes will impact diachronic blameworthiness according to RPA, only those that contribute to their moral improvement, though this is compatible with the interpretation of PCA I’m operating with in this chapter. And, again, I’m remaining neutral on whether blameworthy agents deserve to feel guilty or are simply required to experience guilt as part of the reparative process.

Making Amends: How to Alter the Fittingness of Blame  397 performing relevantly similar wrongs, but for one agent to become less dia­chron­ ic­al­ly blameworthy than the other over time. This will occur when one agent does more to fulfil the reparative obligations they possess in virtue of culpably per­ forming the past wrong than the other agent, as in the cases of Jackie and Lee. I take this to be an intuitive implication of the view—­we typically think it is fitting to blame those who fail to reform, feel remorse, or make reparations for much longer than agents who address their wrongs. And while it will likely be rare, it is possible for agents to remain diachronically blameworthy, and the fitting target of blame, for as long as they exist on RPA. This will occur when a blameworthy agent never takes steps to fulfil their reparative obligations.16 Notice that this can occur for even minor wrongs. Again, I take this to be an intuitive implication of the view. Consider Lee, the plant sitter who never acknowledges that she was wrong to treat Edie’s plants carelessly, never feels guilty or remorseful, and never apologizes or attempts to make amends. It strikes me as entirely fitting for Edie to blame Lee in perpetuity. After all, Lee in fact has rep­ ara­tive obligations to Edie, so it is entirely fitting to demand, or call for, Lee to fulfil these obligations. Of course, this isn’t to say that it will always be all-­things-­ considered appropriate for Edie to openly blame Lee—­there may be pragmatic reasons for Edie to stop expressing blame after a certain point. In addition to capturing a wide range of intuitions about the fittingness of blame over time, RPA can also explain why our blaming practices are valuable and important. Recall that one way of taking steps to fulfil the reparative obliga­ tions that come with being blameworthy is to open yourself up to, and accept, blame from others, particularly those you have wronged. There are several ­reasons for this. First, when a victim blames those who wrong them, they are able to express respect for themselves. In previous work, I’ve argued that expressions of blame can communicate self-­respect because they convey that an agent takes themselves to: (1) be the kind of agent who ought not be shown ill will, or who should not be treated in a way that is blameworthy, or whose legitimate demands should be respected by others. And,

16  One might also wonder about cases in which a wrongdoer is unable to fulfil their reparative obligations because their victim is dead or otherwise unreachable. There is much to say about these kinds of cases, and much will depend on the details of the case. But briefly, many reparative obliga­ tions don’t require one to engage in any way with one’s victims (i.e. expressing remorse to the moral community, morally improving, etc.), so even wrongdoers who are unable to interact with their vic­ tims can become less blameworthy over time by fulfilling at least some of their reparative duties. But even in these cases, such agents will remain at least somewhat blameworthy, since it will always be true of them that they were unable to make reparations to their victim. I take this to be an unfortu­ nate, but not counter-­intuitive, implication of RPA. I plan to explore these cases in more depth in future work.

398 Fittingness (2) deserve moral attention, care, or concern in light of being shown ill will, or being treated in a way that is blameworthy, or having her legitimate demands being disrespected.  (Tierney 2021b: 193)17 These messages of self-­respect are important for blameworthy agents to receive, particularly when they are attempting to counteract the disrespect expressed by their previous wrongs and repair the psychic and relational damage that these wrongs imposed. Expressions of blame are also important to the reparative pro­ cess because they allow the victim to make clear what forms of reparation they take to be appropriate. While all blameworthy agents have reparative duties to their victims, there are a variety of ways an agent can fulfil these duties, and what kinds of action are appropriate will depend, at least partially, on what the victim takes to be appropriate. Material restitution will do little to repair a blameworthy agent’s relationship with their victim if the victim doesn’t care about material res­ titution and instead wants an apology. And a public apology will likely do more harm than good if the victim would have preferred a private apology.18 Given the important role our practices of blame play in the reparative process, it will be unlikely that an agent will be able to fulfil all their reparative duties without ever being blamed by others. Thus, an agent will likely remain at least somewhat blameworthy until they receive and respond to others’ blame. Of course, there are bound to be exceptions. Perhaps partners or friends who have come to an agreement on how culpable wrongs be addressed in the context of their relationship could successfully discharge their reparative obligations without being the target of the other’s blame. There may even be good reasons to form such an arrangement, since experiencing and expressing blame are not without costs. These agreements will be rare (and even more rarely successful), since they require agents to have a long history of trust and open communica­ tion.19 These situations are also precisely the kinds of contexts in which blame is not valuable, since both parties respect one another, know that the other respects them as well as themselves, and are confident in how they can repair the damage their culpable wrongs cause. In this way, RPA can account for when our blaming practices are valuable and when they are not. This is not so for its competitors. On 17  Notice that this view of the communicative and evaluative content of blame provides independ­ ent support for RPA—­if blame’s evaluative content represents its targets as owing reparations to their victims in virtue of culpably wronging them, then blame will only be fitting if the target in fact owes reparations in virtue of their culpability. But it would take me too far afield to argue for this view of blame in this chapter, so I’ll limit my defence of RPA to the considerations above. 18  I also discuss this important feature of blame in Tierney 2021a. 19  There will also be a limited range of wrongs for which such arrangements can be made. While we can plausibly form arrangements about how to handle ‘everyday’ expressions of ill will and disregard that most of us, unfortunately, engage in from time to time, there are many wrongs that are so out of the ordinary and so serious that we cannot reasonably or advisably expect them to be resolved without expressions of blame. In these situations, receiving and responding to others’ blame will be necessary to diminish one’s blameworthiness. Thanks to Oded Na’aman for discussion on this point.

Making Amends: How to Alter the Fittingness of Blame  399 PCA and GBA, there will be far more situations in which an agent will cease being blameworthy before ever being blamed by another agent, even their victims. In these scenarios, blame will be unfitting, even though it could contribute to the reparative process. Not only is this counter-­intuitive, it also fails to capture one of the key values of our blaming practices, which many take to be an important desideratum of theorizing about blameworthiness (e.g. Carlsson 2017).

17.5  Objection: Reparations without Blameworthiness According to RPA, there is a tight connection between blameworthiness and reparative obligations: agents who are blameworthy for performing wrong actions must fulfil a set of reparative obligations in virtue of their culpable wrongdoing and such agents will remain blameworthy until these obligations are fulfilled. So, one could object to the view by challenging this connection. One way to do this would be to argue that reparative obligations are not unique to blameworthy agents.20 Indeed, defenders of both PCA and GBA explicitly argue that agents can possess reparative duties even if they are not blameworthy. Khoury and Matheson contend that agents can be required to fulfil compensatory duties in virtue of benefiting from a past wrong action (2018: 215–16). On this view, an agent can fail to be blameworthy for a past action because they are not psychologically con­ nected to the agent who performed it, but nevertheless continue to benefit from this action and thus owe some form of reparations to its victims. And Carlsson argues that we can treat others poorly, and owe them special moral attention, care, or concern in virtue of doing so, without being blameworthy for treating them this way. He writes: ‘Suppose I forgot my brother’s birthday. It may be the case that I don’t deserve to feel guilty for this. Nevertheless, it seems quite clear to me that my brother ought not to be treated this way, and that he deserves moral attention, care and so on in light of my omission’ (Carlsson 2022: 195).21 But if agents can be required to fulfil reparative duties for actions they are no longer, or were never, blameworthy for, then it would be a mistake to argue that the posses­ sion of reparative obligations is sufficient for diachronic blameworthiness. First, it’s important to note that RPA is not committed to the view that the pos­ session of any kind of reparative duty entails that an agent is blameworthy. Rather, 20  Another way to do this would be to argue that an agent can discharge their reparative obligations yet continue to be blameworthy. But because discharging one’s reparative obligations includes psy­cho­ logic­al change and the experience of guilt, both of which are sufficient to eliminate diachronic blame­ worthiness on PCA and GBA, it is unlikely that defenders of these views will take this route. 21  Though I won’t press the point, I do not share the intuition that these kinds of omissions are non-­culpable. Forgetting a sibling’s birthday strikes me as precisely the kind of expression of insuffi­ cient good will, and failure to meet a legitimate demand, that is blameworthy and that we ought to hold others accountable for.

400 Fittingness RPA’s sufficiency claim indicates that agents who continue to possess reparative obligations in virtue of culpably performing some past action will continue to be blameworthy for this action. In this way, RPA does not deny that agents can pos­ sess reparative obligations in virtue of performing actions that they are not blameworthy for. The view is entirely compatible with agents owing reparations to others due to strict liability, non-­culpable negligence, etc. The view also doesn’t deny that agents who are no longer blameworthy for past actions could still possess reparative obligations related to their wrongs for ­reasons not having to do with their culpability. Consider Khoury and Matheson’s observa­ tion that agents can be required to make reparations for past wrongs for which they are no longer blameworthy if they continue to benefit from these wrongs. This is entirely compatible with RPA, which claims only that an agent will cease being blameworthy for a past wrong when they fulfil the reparative obligations that they possess in virtue of culpably performing this action. It’s possible for an agent to engage in culpable wrongdoing, then go on to fully discharge their rep­ ara­tive obligations—­i.e. by morally reforming, withdrawing the disrespect/threat the wrong posed, and making amends—­yet continue to benefit from this wrong action. On RPA, the agent in question would no longer be blameworthy, but they could be required to continue to fulfil compensatory obligations because of the benefits they continue to accrue. In this way, pointing out that agents can be required to fulfil reparative obligations for actions that they are not, or are no longer, blameworthy for does not undermine the reparative view. Thus, in order to sever the connection between reparations and blameworthi­ ness that RPA is committed to, one must argue that the reparative obligations we take blameworthy agents to possess are not in fact possessed in virtue of these agents being blameworthy. If defenders of PCA and GBA could defend this claim, not only would this successfully challenge RPA, it could also allow these views to address many of the challenges they face. Take the privately and expressively reformed and remorseful gardener cases. I objected to PCA and GBA on the grounds that they render Edie’s blame of Jackie* and Jackie** unfitting in these cases, despite the fact that Jackie* and Jackie** fail to make reparations to Edie. But if Jackie* and Jackie** could owe reparations to Edie without being blame­ worthy, then this could justify Edie’s blame on the grounds that it could motivate Jackie* and Jackie** to fulfil her reparative obligations. If this is right, then PCA and GBA could make sense of the intuition that it is appropriate for Edie to blame Jackie**, even if unfitting, which could render these views less objectionable than previously thought. But if Jackie* and Jackie** don’t possess their reparative obligations in virtue of being blameworthy, why do they owe these reparations? Neither Jackie* nor Jackie** continue to benefit from their past wrongs, so they can’t owe Edie re­par­ ations for this reason. Perhaps defenders of PCA and GBA could argue that

Making Amends: How to Alter the Fittingness of Blame  401 Jackie* and Jackie** owe reparations to Edie because they harm her when they carelessly overwater her plants. After all, agents can harm others without being blameworthy and harm is also something that calls out for repair. This coheres nicely with Carlsson’s claim that we can owe agents special moral attention when we treat them in ways that they ought not be treated, without being blameworthy for such treatment. While I don’t deny that non-­blameworthy harm could ground a host of rep­ara­ tive obligations, it is unlikely that generating harm can give rise to the kinds of reparative obligations we take to be associated with blameworthy agents. This is because the kinds of reparative obligations that arise from acting in ways that harm others will depend on whether the agent is blameworthy for their action. Non-­culpably harming someone does not communicate disrespect or stand as a threat in the same way that culpable wrongdoing does, nor does it inflict the same kind of psychic and relational damage. Thus, non-­culpable harms will generate different kinds of reparative obligations than culpable wrongs. Imagine a version of the reformed and remorseful gardener case in which Jackie*** was careless with Edie’s plants due to an unknown side effect of a lifesaving medication she recently began taking. In this case, Jackie*** is not blameworthy for carelessly overwatering Edie’s plants, though her action still harms Edie, and may well give rise to some reparative obligations. But Jackie*** wouldn’t be required to engage in moral reform in order to repair her relationship with Edie, since her careless­ ness wasn’t the result of a moral flaw. And withdrawing the disrespect or threat her action communicated would only require an explanation of what occurred, not an expression of guilt or remorse. And while offering to buy Edie new plants may be required to repair the harm the action caused, apologies and promises to be morally better in the future are not necessary. In contrast, Jackie* and Jackie** are required to morally improve in order to repair their relationships with Edie, just as they are required to feel and express guilt, and attempt to make amends and apologize. But the only difference between Jackie*, Jackie**, and Jackie*** is that Jackie* and Jackie** are blameworthy for carelessly overwatering Edie’s plants, while Jackie*** is not blameworthy for this action. So, the best explanation for why Jackie* and Jackie** possess the reparative obligations they do is that they are blameworthy for their careless actions. Thus, it is unlikely that PCA and GBA can make sense of the kinds of reparations Jackie* and Jackie** owe Edie, or why it is fitting for Edie to blame them, without granting that they possess these obligations in virtue of being blameworthy. If the above reasoning is right, then the connection between reparations and blameworthiness at the heart of RPA remains unscathed—­blameworthy agents possess a distinctive set of reparative obligations in virtue of culpably performing wrong actions and such agents will remain blameworthy until these obligations are fulfilled.

402 Fittingness

17.6 Conclusion In this chapter, I’ve defended a reparative account of diachronic blameworthiness, RPA, according to which blameworthy agents have reparative obligations to their victims and remain blameworthy, and the fitting targets of blame, until these obli­ gations are fulfilled. I’ve argued that RPA outperforms its competitors because it captures a wider range of intuitions about blame’s fittingness over time and better accounts for the value and importance of our blaming practices. While there is much more to say about the particular kinds of reparative obligations blame­ worthy agents possess, and how they can successfully fulfil them, I hope that the reflections here will serve as a useful foundation for these future discussions.22

References Arpaly, N. (2003) Unprincipled Virtue: An Inquiry into Moral Agency. Oxford University Press. Arpaly, N. & T. Schroeder (2013) In Praise of Desire. Oxford University Press. Callard, A. (2017) ‘The Reason to Be Angry Forever,’ in M. Cherry and O. Flanagan (eds) The Moral Psychology of Anger. Rowman & Littlefield: 123–37. Carlsson, A. (2017) ‘Blameworthiness as Deserved Guilt,’ Journal of Ethics 21: 89–115. Carlsson, A. (2022) ‘Guilt and Blameworthiness over Time,’ in A. Carlsson (ed.) SelfBlame and Moral Responsibility. Cambridge University Press: 175–197. DeGrazia, D. (2005) Human Identity and Bioethics. Cambridge University Press. Feinberg, J. (1970) Doing and Deserving: Essays in the Theory of Responsibility. Princeton University Press. Fischer, M. (2012) Deep Control: Essays on Free Will and Value. Oxford University Press. Fischer, M. and N. Tognazzini (2009) ‘The Truth about Tracing,’ Noûs 43: 531–56. Haji, I. (1998) Moral Appraisability. Oxford University Press. Hieronymi, P. (2001) ‘Articulating an Uncompromising Forgiveness,’ Philosophy and Phenomenological Research 62: 529–55. Howard, C. (2022) ‘Forever Fitting Feelings.’ Philosophy and Phenomenological Research: 1–19. Khoury, A. C. (2013) ‘Synchronic and Diachronic Responsibility,’ Philosophical Studies 165: 735–52. Khoury, A.  C. (2022) ‘Forgiveness, Repentance, and Diachronic Blameworthiness,’ Journal of the American Philosophical Association: 1–21. 22  I’m very grateful to David Glick, Christopher Howard, Oded Na’aman, and the DaGERS reading group for providing helpful feedback on earlier drafts of this chapter.

Making Amends: How to Alter the Fittingness of Blame  403 Khoury, A. C. and B. Matheson (2018) ‘Is Blameworthiness Forever?’ Journal of the American Philosophical Association 4(2): 204–24. Macnamara, C. (2013) ‘Taking Demands out of Blame,’ in D.  J.  Coates and N. Tognazzini (eds) Blame: Its Nature and Norms. Oxford University Press. Marušić, M. (2018) ‘Do Reasons Expire? An Essay on Grief,’ Philosophers’ Imprint 18: 1–21. Matheson, B. (2014) ‘Compatibilism and Personal Identity,’ Philosophical Studies 170: 317–34. McGeer, V. (2013) ‘Civilizing Blame,’ in D. J. Coates and N. Tognazzini (eds) Blame: Its Nature and Norms. Oxford University Press. McKenna, M. (2012) Conversation and Responsibility. Oxford University Press. Milam, P-E. (forthcoming) ‘Forgiveness and Ceasing to Blame,’ in P.  Satne and K. M. Scheiter (eds) Conflict and Resolution: The Ethics of Forgiveness, Revenge, and Punishment. Springer. Moller, D. (2017) ‘Love and the Rationality of Grief,’ in C. Grau and A. Smuts (eds) The Oxford Handbook of Love. Oxford University Press. Murphy, J. and J. Hampton (1988) Forgiveness and Mercy. Cambridge University Press. Na’aman, O. (2020) ‘The Fitting Resolution of Anger,’ Philosophical Studies 177: 2417–30. Nelkin, D. (2013) ‘Freedom and Forgiveness,’ in I. Haji and J. Cauette (eds) Free Will and Moral Responsibility. Cambridge Scholars Publishing: 165–88. Olson, E. (1997) The Human Animal: Personal Identity without Psychology. Oxford University Press. Parfit, D. (1984) Reasons and Persons. Oxford University Press. Portmore, D. (2019) ‘Control, Attitudes, and Accountability,’ in D.  Shoemaker (ed.) Oxford Studies in Agency and Responsibility, Volume 6. Oxford University Press: 7–32. Portmore, D. (2022) ‘A Comprehensive Account of Blame,’ in A. Carlsson (ed.) SelfBlame and Moral Responsibility. Cambridge University Press: 48–76. Radzik, L. (2009) Making Amends: Atonement in Morality, Law, and Politics. Oxford University Press. Rosen, G. (2015) ‘The Alethic Conception of Moral Responsibility,’ in R.  Clarke, M. McKenna, and A. M. Smith (eds) The Nature of Moral Responsibility: New Essays. Oxford University Press: 65–87. Scanlon, T. M. (2008) Moral Dimensions: Permissibility, Meaning, and Blame. Belknap Press of Harvard University Press. Scanlon, T.  M. (2013) ‘Interpreting Blame,’ in D.  Coates and N.  Tognazzini (eds) Blame: Its Nature and Norms. Oxford University Press: 84–99. Smith, A. (2013) ‘Moral Blame and Moral Protest’, in D.  Coates and N.  Tognazzini (eds) Blame: Its Nature and Norms. Oxford University Press: 27–48.

404 Fittingness Tierney, H. (2021a) ‘Expanding Moral Understanding,’ Australasian Philosophical Review 3: 318–23. Tierney, H. (2021b) ‘Guilty Confessions,’ in D.  Shoemaker (ed.) Oxford Studies in Agency and Responsibility, Volume 7. Oxford University Press: 182–204. Tierney, H. (2022) ‘Don’t Suffer in Silence: A Self-Help Guide to Self-Blame,’ in A. Carlsson (ed.) Self-Blame and Moral Responsibility. Cambridge University Press: 117–133. Tognazzini, N. (2010) ‘Persistence and Responsibility,’ in J. Campbell, M. O’Rourke, and H. Silverstein (eds) Time and Identity. MIT Press: 149–63. Wallace, R. J. (2019) The Moral Nexus. Princeton University Press. Walker, M. U. (2006) Moral Repair: Reconstructing Moral Relations after Wrongdoing. Cambridge University Press. Zimmerman, M. (1988) An Essay on Moral Responsibility. Rowman & Littlefield.

Index Note: Tables are indicated by an italic “t” and notes are indicated by “n” following the page numbers. For the benefit of digital users, indexed terms that span two pages (e.g., 52–53) may, on occasion, appear on only one of those pages. Achs, Rachel  11, 356–77 adequate understanding, its justificatory role 132–3 aesthetics 309–11 and fittingness  314 see also fittingness terms in general and Hume  309–11 and Kant  373–4 see also Kant, Immanuel appreciation 373 hedonism 315–17 see also well-being insight properties  318 normativity of  373–4 perspective properties  318–19 relation to ethics  317–18 response-dependence  310–11, 315–17, 320–2 theory of  309 aesthetic value  315–16, 322–5 alternatives dependence  29, 29t, 44–5, 48t, 51–2 aretaic categories  50 Aristotle  207, 291–3 atonement 392 attitudes intentionality of  93–8, 225, 274 judgement-sensitive 88–9 authenticity  268, 281–2 badness  23–31, 40, 45, 48t beauty  309–17, 324 Berker, Selim  23, 63, 332–3 blame  10–12, 356–77, 380–402 alethic view of  112 ascriptions of  359–60 blameworthiness  31, 50n.39, 53–4 cases discussing  386–99 diachronic  380–1, 394–6 directedness  359–64, 376 fitting duration of  381–2 forever fitting  382 nature of  356–9, 374–5 phenomenology of  375–8 reflexive endorsement  369–75 relation to guilt  385–6 relation to justification  364–9 relation to wrongdoing  363, 370–1 relationship to guilt  380–2 Strawson’s view of  338–40 see also Strawson, P. F.

synchronic 380–1 Brandt, Richard  32–4, 37–9, 42, 44n.32, 49–50, see also fittingness, history of; fittingness, nature of Brentano, Franz  2–3, 109–10, 109n.9, 179, see also fittingness, history of; fittingness, nature of Broad, C. D.  13–14, 58–9, 66–7, 74–5, 136, 200, 228 Broome, John  76–8, 206 buck-passing account  176–8, 227, 245, 252–3, see also fitting attitude accounts of value Camp, Elizabeth  318–19 Chudnoff, E.  137–8 choiceworthiness  31, 33, 35, 42n.29, 43n.31, 53–4 cognitivism  111–13, 117–21 and interpretivism  118–19 colour  310–12, 322 consequentialism  29–31, 45, 47, 54, 75, 182, 194–5 rule-consequentialism  182–3, 194–5 constitutive correctness  12, 32, 124n.29, 203–4, 226–32 correctness  31, 40–2, 44n.32, 47–8 Crisp, Roger  6–7, 187n.21 Cullity, Garrett  63, 151, 200 Dancy, Jonathan  133, 135–9 D’Arms and Jacobson  200, 291, 298–9, 304, 321, 364–5 D’Arms Justin  33–4, 45n.34, 105, 200, 298–9 Darwin, Charles  80 deliberation 205–13 and advice  214–15 constitutive aim of  203–5, 208, 212 reasons to deliberate  211 deontic categories  23–31, 35–45, 48–53 and definitional duals  26 and extensional duals  25 and duality  25–7, 29t, 37–8, 40–1, 48t, 51–2 deontic concepts  24–31, 35–45, 73–8 deontic facts  64–8, 190–1 deontology  54, 77–8, 181–2 directed duties  24 undirected duties  24 dispositions 311–13 doctrine of double effect  131–2

406 Index duties, see also deontic categories prima facie  51, 73–5, 77 pro tanto  51 reparative  398, 401 emotion and narrative  98–100 and reflexive endorsement  357–8, 363–9, 375–7 individuation of  361–2 judgmentalism 97–8 perceptualism  97–8, 133, 273–5 envy  192–4, 290 aggressive 294–5 and children  299–301 and fittingness  291–6, 298–9, 304–6 and goodness  225, 291–2, 299–301, 303–4 benign 293 definition of  292, 294 inert 294–5 Kant’s critique of  296–8, 305–6 malicious  224, 293 spiteful 294–5 epistemology and intuitions  133, 143–4 and justification  133–4 rationalist view of  131–4 experience machine  280, 287n.18 euthyphro problem  169–70, 317 evaluative categories  23–31, 45–7, 52 evaluative facts  2, 24–31, 71 relation to fittingness  68–73 Ewing, A. C.  2–3, 58–9, 134, 136, 200 see also fittingness, history of; fittingness, nature of family of categories  23 feasibility and advice  214–15 and deliberation  202–5 and dispositions  202 and fact-dependence  205–9 and fittingness  200–2 and gradability  201n.2 and objectivity  205–9 and possibility  202, 204–5 and probability  202, 204–5 functional status of  202 modal status of  202 non-normativity of  209–13 Feinberg, Joel  10, 329, 333–5, 340–1, 347–9, 382 fittingness 31–50 correct representation (or alethic) view of  12–13, 72, 112, 117–27, 158, 252, 276–7 and action  136, 158–9 and advice  214–15 see also feasibility and advice and belief  180 and constitutive correctness  222, 226–32 and correctness  12–13, 276–7 and deontic facts  23–4, 35–45, 64–8, 190–1 and desires  178, 292, 342 and emotion  3–4, 71, 97–101, 105–8, 111–13, 138–40, 192–4 and evaluative facts  2, 24–31, 45–7, 71

and fit-making facts  187, 237–8 and gender  16 and gradability  41–4, 48t, 49–50, 52–3, 73–8, 232 and intentionality  94, 97–8, 105–6, 225 and intentions  80, 105–6 and permittedness  27–8, 35–45, 47–8 and political philosophy  15, 200 and proportionality  107 and reasons  151–2, 156–7 and requiredness  29, 35–45, 47–8, 157 and response-dependency  108, 310–11, 315–17, 320–2 and response-independence  111–13 and responsibility  10–12, 329, 333–42, 347–9 and the partiality problem  8, 245, 250–2, 254–6, 263–4 and the “wrong kind of reason problem”  95 see also wrong kind of reason and value theory  187 and weight  13, 232, 235–6 criticism of  10, 64–8, 134 definition of  1, 31, 106, 157, 227, 298, 330–1 deontic view of  151n.2 duration of  11–12, 380–1 epistemology of  130–4 evaluative view of  121–7 fittingness-first  65, 152–6, 176, 190, 196–7 see also value-first; reasons-first fundamentality of  59, 65, 151–2 see also value-first; reasons-first history of  24–31, 157, 200, 228, see also Brandt, Richard; Brentano, Franz; Ewing, A. C., Moore, G. E., Ross, W. D. nature of  23–4, 71, 107, 121–7, 134–5, 157 normativity of  71, 122–3 rationalist view of  158 relation to belief  122–3 skepticism of  134 thin vs. thick fittingness properties  31–2 fittingness terms in general  31–5 acceptable  31, 35, 392 admirable  1, 12, 31, 160, 163, 235–6 adorable 3–4 agreeable 31 ambiguous 134 amusing  31, 94, 321 appropriate  32, 70, 105, 112, 330–3 apt  31, 130 beautiful  9–10, 311, 315 boredom  223, 225–6 called for  31, 228 charming  9–10, 310, 312 contemptable  31, 361–2 credible  31, 135, 229–30, 332 delicious  9–10, 311–13, 315 delightful 247 depression 225–6 desert  10–11, 333–5, 340–7 desirable  12, 35, 42, 46n.36, 60–1, 68–9, 157–8, 179, 254, 315–16, 324 despicable 236

Index  407 different types of  135–6 disgraceful  31, 135 dreadful  31, 180 enviable  71–2, 180, 224–5, 290 excusable  35, 338, 348 fearsome  31, 192, 273–4 funny  31, 312, 321 grief  11–12, 86–91, 262 happiness  9, 247, 267–70, 277–8 inappropriate  31, 290, 389–90 incorrect  12, 286–7 justified  31, 53–4, 87 loveable  12, 37, 221, 251, 276 merited  45n.33, 70, 251 persuasive 31 praiseworthiness  31, 157–8 regretful  11–12, 80, 87 schadenfreude 223 shameful  3–4, 31, 115–23 suitable  12, 31, 61, 75 tasty 31 trustworthy  31, 36, 393 vague 134 Foot, Philippa  111, 367 free will  329, 349, 352 libertarian theories of  336–8 skepticism about  352 goodness  23–31, 35–6, 40–1, 45–7, 48t, 51–4 attributive  195–6, 246 goodness simpliciter  8, 178, 178n.6, 181, 189, 195–6, 246, 250 see also Rowland, R. A. goodness-for  9, 246 relation to fittingness  2, 24–31, 45–7, 71 gradability  28, 29t; see also fittingness, gradability of; reasons, gradability of happiness  9, 247, 267–70, 277–8 and the ill-being objection  284–7 and the inauthenticity objection  281–2 and the lack of value objection  280–1 and the passivity objection  282–3 and the shallowness objection  279–80 Haybron, Daniel  270–3 hedonism  267, 310–11, 315–17 Hieronymi, Pamela  313, 392 Howard, Chris  71–2, 221, 251, 330–1 Hume, David  291–3, 309–11, 315–17 humour/funny  312–13, 323–4 Hurka, Thomas  58–9, 134, 179 hypological categories  50n.39 ideal observer theory  309–10, 313–14 insight  318, 320–2, 324 instrumental value  323 intentionality  93–8, 105–6, 225 internalism moral judgement internalism  136 intuition internalism  137 intrinsic value  323 intuitions practical intuitions  137

emotional intuitions  136 intellectual intuitions  133, 136 Jacobson, Daniel  33–4, 45n.34 Kant, Immanuel  296–8, 310, 373–4 Kaupinnen, A.  136–7 Kiesewetter, Benjamin  5–6, 228 King, Alex  9–10, 15, 309 Korsgaard, Christine  209–10, 248n.9 Leary, Stephanie  7, 221 Maguire, Barry  3–4, 42–4, 232–8 McDowell, John  68, 228 McHugh, Connor  2–3, 63, 165–6 McKenna, Michael  329 mere permissibility, see optionality moods  9, 85, 225, 269, 274 Moore, G. E.  49n.38, 58–9, 70, 179, see also fittingness, history of; fittingness, nature of Na’aman, Oded  3, 80, 130 naturalism and non-naturalism  314 neutrality  24, 26, 28–9, 46, 48t, 52–3 Nietzsche, Friedrich  182–3 normativity 23n.1 formal  188, 191 robust 188–93 normative categories in general  23–4, 23n.1, 29, 31, 47–50 and polar opposition  27–8, 29t, 48t, 52–3 and privative opposition  27–8, 46 Nozick, Robert  179, 287n.18 Nussbaum, Martha  111–12, 303, 357 optionality  24, 28 ought all things considered  73–8, 190–1, 331–2, 389–90 implies can  202 epistemic 63 partiality  3, 250–6 Parfit, Derek  7, 95n.13, 141n.29, 152n.4 Pereboom, Derk  341–2, 352 permissibility  23–31, 35–45, 48t, 48–51 perspective  318–22, 324 personal identity  383–4 phenomenology  115, 143–6, 284 positional goods  297, 299, 301–3 Prichard, H. A.  58–9, 76–7, 142 Protasi, Sara  290 Radzick, Linda  392–4 rational agent  140–1 rational evaluability  82–6 conditions of  82–5, 95 Rawls, John  54, 280 Raz, Joseph  158n.35, 344 reasons  39n.25, 50–4 agent-neutral  8, 252–6 agent-relative 252–6

408 Index reasons (cont.) authoritative  188–9, 222, 226 basal  334, 348, 351 decisive 51 enticing 135 exclusionary 344 explanatory  62–3, 365–6 externalism 135 gradability of  51–3 see also gradability internalism 135 motivating 62–3 see also motivating reasons normative  62–3, 73, 151–6, 161, 165–6, 183, 187, 365 see also normative reasons overall  232, 247 pragmatic  7, 62–3 prima facie  66–7, 73–8, 130 see also reasons, prima facie; duties, prima facie primitivism 155–6 pro tanto  51–2, 73–8, 135, 247 see also reasons, pro tanto; duties, tanto reasons fundamentalism  176–8 relationship to fittingness  152–7 relationship to value  176–8, 187 right-kind  7, 51n.40, 53n.44, 158, 170–2, 221, 224, 252, 321 see also right-kind reasons skepticism about right-kind reasons  7, 221–2 see also right-kind reasons skepticism about wrong-kind reasons  7, 95, 221 see also wrong-kind reasons weighing of  232–8 wrong-kind  7, 158, 191, 221, 249, 321 see also wrong-kind reasons reasons fundamentalism  2–3, 63, 76, 85n.5, 152–6, 176–8 see also Broome, John; value-first; fittingness-first reasons-first  2–3, 76, 76n.5, 152–6, 176–8 see also Broome, John; value-first; fittingness-first response, what is a  320, 323 response dependence  34n.20 requiredeness  23–32, 35–45, 45n.34, 47–54, 48t rightness  23, 48–9, 54 Rinard, Susana  226 Rosen, Gideon  12–13, 112, 230–1 Ross, W. D.  49n.38, 51, 58–9, 66–7, 74–7, 130, 134, 136, 139, 182–3, see also fittingness, history of; fittingness, nature of Rossi Mauro  9, 126, 267 Rowland, R. A.  4–5, 8–9, 16, 176, 178n.6, 181n.14, 187n.21, 225–6 Scanlon, T. M.  39n.25, 140–1, 200, 392–3 Schroeder, Mark  13–14, 228–9 seemings 133 practical seemings  136–7 self-evidence 131 sentimentalism  33–4, 105–17 Shoemaker, David  10, 330, 351–2 Sidgwick, Henry  58–63 skepticism

about free will  352 skepticism about right-kind reasons  7, 221, 226 see also right-kind reasons skepticism about wrong-kind reasons  7, 221 see also wrong-kind reasons and fittingness  134 about fitting envy  304 see also envy Sosa, Ernest  143–4 Southwood, Nicholas  6, 200–19 Stratton-Lake, Phillip  130–46 Strawson, P. F.  10, 329, 332–3, 335–40 see also blame historical context of his views  336 sufficient reason  51, 53n.44 see also reasons Tappolet, Christine  12–13, 267, 361 taste (gustation)  311–13 thick ethical concepts  32, 65–8 thin ethical concepts  65 Tierney, Hannah  11–12, 380–402 trolley problem  131–2 value aesthetic  315–16, 322–4 buck-passing account of  176–8, 227, 245, 252–3 derivative 247 final  178, 186, 247 fitting-attitude theories of  35–6, 45–7 instrumental  179, 194–5 intrinsic  65, 69, 179 norm attitude account of  245–50 primitivism 247–8 relation to fittingness  108–11, 176–8 see also fittingness relation to reasons  176–8, 187 see also reasons subjectivism 247–8 theory of  7–10 vice 50 virtue  23n.1, 50 Value-first view  176–8, 192–4, see also fittingness-first; reasons-first direct 176–8 indirect  176–8, 182–5, 194, 196–7 volitional capacity  204–5 Wallace, R. Jay  290, 338–9 see also Strawson, P. F. Way, Jonathan  165–6 well-being 267–70 fitting happiness theory of  269–78 hybrid happiness-based theory of  268 objective list theory of  268 perfectionist theory of  268 satisfactionist theory of  268 standard happiness-based theory of  268 Williams, Bernard  15, 68 Wiggins, David  228, 317–18 wrongness 48–9 wrong kind of reason  7, 158, 191, 221, 249, 321–2, see also reasons, wrong-kind Zimmerman, Michael J.  50n.39, 256n.24