Attachment and Character: Attachment Theory, Ethics, and the Developmental Psychology of Vice and Virtue 0192898124, 9780192898128

Citation preview

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 26/07/21, SPi

Attachment and Character

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 26/07/21, SPi

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 26/07/21, SPi

Attachment and Character Attachment Theory, Ethics, and the Developmental Psychology of Vice and Virtue Edited by

E DWA R D HA R C O U RT

1

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 26/07/21, SPi

1 Great Clarendon Street, Oxford, OX2 6DP, United Kingdom Oxford University Press is a department of the University of Oxford. It furthers the University’s objective of excellence in research, scholarship, and education by publishing worldwide. Oxford is a registered trade mark of Oxford University Press in the UK and in certain other countries © the several contributors 2021 The moral rights of the authors have been asserted First Edition published in 2021 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: 2021936011 ISBN 978–0–19–289812–8 DOI: 10.1093/oso/9780192898128.001.0001 Printed and bound by CPI Group (UK) Ltd, Croydon, CR0 4YY Links to third party websites are provided by Oxford in good faith and for information only. Oxford disclaims any responsibility for the materials contained in any third party website referenced in this work.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 26/07/21, SPi

Table of Contents Acknowledgements List of Contributors

vii ix

Introduction1 Edward Harcourt 1. Attachment Theory and Moral Development Ross A. Thompson 2. Dimensions, Determinants, and Development of Prosocial Behaviour: A Differential Susceptibility Hypothesis on Attachment and Moral Character Marian J. Bakermans-­Kranenburg and Marinus H. van IJzendoorn 3. Attachment, the Virtues, and the Vices: Some Developmental and Conceptual Considerations Jonathan Hill 4. The Evolved Nest, Virtue, and Vice Darcia Narvaez

18

44

71 87

5. The Virtues of Attachment Keith S. Cox and Micah Lott

105

6. Beyond Virtue: The Development of Reproductive Strategies Jay Belsky

123

7. Two Routes from Secure Attachment to Virtue Edward Harcourt

137

8. Education for Virtue: Aquinas on Passions and Attention Terence Irwin

154

9. Attachment, Virtue, and the Second Person Katrien Schaubroeck

174

10. An Aetiology of Recognition: Empathy, Attachment, and Moral Competence A. E. Denham

195

11. Attachment, Addiction, and Vices of Valuing Monique Wonderly

224

Index

247

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 26/07/21, SPi

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 26/07/21, SPi

Acknowledgements The origins of this collection of papers go back some years. As long ago as 2011, I  hosted a preliminary workshop in Oxford with my colleague Terry Irwin—a rewarding meeting intellectually, and I hope participants will remember the fact that one afternoon it got very cold and dark as part of the fun. I’m grateful to the Oxford University John Fell Fund and to Keble College, Oxford, for financial support. Christof Rapp, who took part in that original workshop, then invited me to co-­organize with him a conference at the Centre for Advanced Studies at the Ludwig-­Maximilians-­Universität Munich, Attachment, Virtue and the Regulation of Emotion, which took place in 2016. Warm thanks are due to the Centre for Advanced Study for financial and administrative support, and to Christof for contributing his considerable enthusiasm and intellectual energy towards making that meeting and what subsequently came of it a success. Even earlier than the Keble workshop, Jen Lexmond invited me to Demos to give a philosopher’s take on the then politically topical discussions about character and the early years; subsequent conversations with her helped to form some of the ideas behind this collection. She also pointed me towards Jay Belsky, then still at Birkbeck in London. Our initial conversation confirmed my sense that philo­ sophers and attachment theorists could have a productive time talking to each other. I also found my way to Marinus van IJzendoorn and Marian Bakermans-­ Kranenburg, with whom several conversations subsequently have helped my own ideas on attachment and the virtues to take shape. The LMU Munich conference became part of a three-­conference series with funding for the other two—one back at Keble, the other at the University of California, Davis—made available through a research network award to me from the Arts and Humanities Research Council (grant reference AH/N004043/1). Working across research communities which belong to different disciplines and lack any prehistory of collaboration requires patience and effort, and the series was critical to clarifying ideas and creating intellectual partnerships which lie at the foundation of the papers collected here. I’m very grateful to AHRC for the award, and to all of those who took part in those three meetings. I’d also like to thank administrative colleagues at both Oxford and UC Davis for shouldering many of the burdens inevitably generated by events of this kind; Rhian Harris, then Director of the V&A Museum of Childhood, for hosting an event Set up for Life? in 2017, which brought some of the ideas generated by the conference series to a wider public; and Peter Momtchiloff at OUP for his confidence in the project and his wise advice en route to publication.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 26/07/21, SPi

viii Acknowledgements A great many more people have given generously of their time and expertise to help this collection on its way. Unfortunately, the longer such a thing is in the making, the more likely it is that memory will fail. So with apologies for the probable omissions, in addition to those already mentioned thanks are due to at least the following: Jan Bransen, Gustavo Carlo, Jude Cassidy, Daniel Ewon Choe, Keith Cox, Roger Crisp, Alison Denham, Julien Deonna, Pasco Fearon, Liza Fior, Matt Grist, Jonathan Hill, Jeremy Holmes, Jim Hopkins, Claire Hughes, Kate Humphreys, Fiona Jelley, Rachana Kamtekar, Richard Kraut, Michael Little, Micah Lott, Mario Mikulincer, Darcia Narvaez, Julia Peters, Nancy Potter, Tamar Schapiro, Katrien Schaubroeck, Ilina Singh, John Skorupski, Gopal Sreenivasan, Kathy Sylva, Mary Target, Ross Thompson, and Monique Wonderly.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 26/07/21, SPi

List of Contributors Marian J. Bakermans-­Kranenburg Professor of Neurobiology of Parenting and Child Development, Clinical Child and Family Studies, Vrije Universiteit Amsterdam Jay Belsky Robert M. and Natalie Reid Dorn Professor of Human Development, Human Development Graduate Group, University of California, Davis Keith S. Cox Department of Psychology, University of North Carolina, Asheville A. E. Denham Departments of Philosophy and Political Economy/The Murphy Institute, Tulane University; Associate and Emeritus, Faculty of Philosophy, University of Oxford and Senior Research Fellow, St Anne’s College, Oxford Edward Harcourt Professor of Philosophy, University of Oxford and Fellow, Keble College, Oxford Jonathan Hill Professor of Child and Adolescent Psychiatry, School of Psychology & Clinical Language Sciences, University of Reading; Honorary Consultant Psychiatrist, Oxford Health NHS Foundation Trust Terence Irwin Professor of Ancient Philosophy Emeritus, Faculty of Philosophy, University of Oxford and Fellow, Keble College, Oxford Marinus H. van IJzendoorn Professor of Human Development, Erasmus University Rotterdam Micah Lott Department of Philosophy, Boston College Darcia Narvaez Professor of Psychology Emerita, University of Notre Dame; Founder and Host, EvolvedNest.org Katrien Schaubroeck Department of Philosophy, University of Antwerp Ross A. Thompson Distinguished Professor of Psychology, University of California, Davis Monique Wonderly Department of Philosophy, University of California, San Diego

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 26/07/21, SPi

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

Introduction Edward Harcourt

The chapters in this volume have mostly been selected from papers given at a workshop and three conferences which brought together on the one hand a group of philosophers, most of whom were interested in one way or another in what has come to be known as ‘virtue ethics’—moral psychology in the wake of Aristotle— and, on the other, some developmental psychologists working, albeit in different ways, in an attachment paradigm. I organized these meetings partly because my own reading of attachment theory persuaded me there were a number of exciting points of contact between developmental psychology done this way and the kinds of questions Aristotle’s ethics raises, and which interest me; partly because almost no philosophers back then seemed even to have heard of attachment theory. This Introduction presents, inevitably through the eyes of a philosopher, what I take to be attachment theory’s main claims, and then tries to identify why philosophical moral psychologists should take it much more seriously than they have done to date—as I hope this volume itself will help them to do. Attachment theory was first formulated by John Bowlby, a dissenting psycho­ analyst (Bowlby 1976–­7/2005).1 Partly inspired by the effects of maternal de­priv­ ation on children brought about by wartime evacuation, he criticized Freudian ideas about human nature, rejecting for example the ‘secondary-­drive theory’, the claim that infants form attachments to particular others simply as a means to get­ ting food or warmth. The theory was developed by two North American women, Mary Ainsworth and Mary Main (Ainsworth et al. 1978), and has developed fur­ ther since. It now has its own journals and is part of the academic mainstream. As to its main claims, attachment theorists distinguish between the generic and the ‘individual-­difference’ components of the theory, that is, between the part of the theory which deals with what is common to all (or almost all) human beings and the part which aims to explain certain differences between us (Simpson and Belsky  2016). The generic component notes that newborns of various species, including our own, cannot survive unaided. Attachment theory’s hypothesis is that the ‘attachment system’ in infants serves, when activated, to maintain prox­ imity to an attachment figure who is able to protect the infant from threats to its 1 For a very brief history, see Holmes 2011: chs 3–5. For Bowlby and psychoanalysis, see Holmes 1996. Edward Harcourt, Introduction In: Attachment and Character: Attachment Theory, Ethics, and the Developmental Psychology of Vice and Virtue. Edited by: Edward Harcourt, Oxford University Press. © Edward Harcourt 2021. DOI: 10.1093/oso/9780192898128.003.0001

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

2  Edward Harcourt survival. Everyone has heard about the unignorable pitch of a baby’s screams, but don’t forget about the unignorable charm of a baby’s smile: the attachment sys­ tem’s behavioural expressions are various and maintain proximity in cor­res­pond­ ing­ ly various ways, including smiling and vocalizing, clinging, crying, and approaching and following. Thus we attach ourselves in infancy to a small set of special attachment figures and by doing so tie them to us, enlisting the attach­ ment figures’ responses; if the infant responds in turn to those responses, a virtu­ ous cycle is set in train that will help see to it that the infant survives. This is not to say that the attachment system is the only trait that serves to maintain proximity: a specific ‘caregiving system’ among parents, and non-­behavioural characteristics of infants such as endearing ‘babyish’ features (Lorenz 1971; Maestripieri 2004), may also play this role. Nonetheless maintaining proximity with the attachment figure enhances the offspring’s chances of making it through to reproductive age, and thus their genes’ chances of being replicated. So, plausibly, natural selection explains the emergence of a cluster of traits that, working together, maintain infants’ and indeed children’s proximity to attachment figures. But though almost all human infants form attachments, not all attachments are similar in quality, and this brings us to attachment theory’s ‘individual-­difference component’. The first measure of attachment quality was the ‘Strange Situation’, developed by Mary Ainsworth (Ainsworth and Wittig 1969). This is administered at either 12 or 18 months, to one parent–infant pair at a time—and note that at this age, at least, infants can fall into different attachment types with respect to different parents. To simplify, the Strange Situation proceeds as follows: the par­ ent enters an unfamiliar room with the infant and settles it down to play with some toys. A ‘stranger’ (an unfamiliar research assistant) then enters who starts to play with the infant, and after a short time the parent tells the infant he or she is leaving, leaves for three minutes, and then comes back. The separation and re­union is repeated, with the stranger absent (Solomon and George 2008: 386). The infant’s behaviour is recorded throughout. Observed infant behaviour falls into at least three recognizable patterns. In pattern B, the infant is overtly dis­ tressed when the parent leaves, then seeks proximity with him or her on return and is comforted by that, and then resumes playing. In pattern A, the infant does not express distress when the parent leaves, though it displays other signs of dis­ tress such as more rapid breathing and heart rate, suggesting the suppression of overt distress rather than indifference. The infant doesn’t show a preference in play as between the parent and the stranger, and is then also relatively indifferent when the parent returns (e.g. looks or turns away from them). In pattern C, the infant may be ‘clingy’ towards the parent and uninterested in the toys even before they leave, is immoderately distressed when they do leave, but when they come back doesn’t calm down and exhibits ‘furious clinging’—‘seeking contact, then resisting contact angrily once it is achieved’. These three attachment types are labelled ‘insecure-­avoidant’ (type A; also ‘deactivating’), ‘secure’ (type B), and

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

Introduction  3 ‘insecure-­ambivalent’ or ‘insecure-­resistant’ (type C; also ‘hyperactivating’).2 But though infants were the first to be systematically classified into attachment types, attachment theory does not apply only to infants: on the contrary, it is supposed to apply across the lifespan, and is tested using various observation and interview measures. A second aspect of attachment theory’s individual-­difference component con­ cerns the further characteristics with which secure and insecure attachment are associated. According to some recent attachment literature (Thompson  2016), infant attachment security predicts a good relationship at least a year later between the child and the attachment figure, considered in terms of ‘enthusiasm, compliance’, and ‘less frustration and aggression’ during shared tasks; secure attachment also predicts harmonious caregiver–child relations over longer ­periods in the presence of continued sensitive caregiver behaviour. Secure attach­ ment in adulthood, meanwhile (whether or not itself predicted by secure attach­ ment in infancy), is correlated with greater sensitivity to one’s own children’s needs, and ‘more warmth and appropriate structuring of learning tasks’ (Simpson and Belsky 2016: 145); and, in attachments to peers, a capacity inter alia to admit vulnerability and need for the other without ‘continually worrying about the attachment figure’s availability’ (Mikulincer and Shaver 2008: 507). Moving to the next widest outcome domain, that of other relationships, children with secure attachment histories have less conflictual relationships with peers from preschool to 7 years (Thompson  2016), are less dependent on teachers in preschool (Sroufe 1983), less dependent on counsellors at summer camp aged 10, and more sociable with unfamiliar adults (Thompson 2016). By contrast insecurely attached 4-­year-­old boys exhibit more ‘aggressive, assertive, controlling, and attention-­ seeking behaviour than their securely attached counterparts’. Finally, attachment theory argues for a connection between attachment security and broader person­ ality traits. The Minnesota Study of Risk and Adaptation from Birth to Adulthood (Sroufe 2005) argues for ‘significant associations between early attachment se­cur­ ity and personality characteristics throughout childhood and adoles­ cence . . . [including] self-­esteem, agency and self-­confidence, [and] positive affect’. Securely attached children aged 6 describe themselves generally in positive terms but are better at admitting flaws; the insecurely attached either are more negative about themselves or do not admit flaws. There is also an important con­tem­por­an­ eous association between secure attachment and the capacity for emotion-­ regulation, including in adulthood (Mikulincer and Shaver 2008); and between secure attachment and psychological understanding (they are more ‘proficient at  identifying emotions in others, . . . especially . . . negative emotions and mixed feelings’). 2  Further variant terminology is also used (Holmes 2011: 3). Quotations in this paragraph are from Weinfeld et al. 2008. See also Solomon and George 2008.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

4  Edward Harcourt As this—necessarily brief—summary shows, attachment theory seems to take us quite a long way into the explanation of evaluatively significant differences between people. It may seem strange, therefore, that philosophers working in moral psychology, or broadly in ‘virtue ethics’—better, the triangle of relation­ ships between the concepts of human nature, human excellence, and the best life (or lives) for human beings—have not paid more attention to attachment theory, especially given the widespread interest among such philosophers, stretching back at least a decade, in various other theoretical outlooks within empirical psych­ology (Lapsley and Power 2005; Russell 2009; Snow 2010; Miller 2013; Tiberius 2015; Masala and Webber 2016). Obviously for any empirical theory to make a serious claim on philosophers’ (or anyone else’s) attention, it needs to stand a good chance of being true. But the philosophers who have picked up, for example, positive psychology, personality psychology, or affective neuroscience and gone somewhere interesting with them certainly did not begin by in­de­pend­ ent­ly establishing the truth of these theories, nor would they have been compe­ tent to do so had they tried. When philosophers draw on empirical psychology—as Aristotle himself surely did—they make themselves vulnerable if the psy­cho­ logic­al theories in question are discredited; but before that happens, they have simply to take a great deal on trust. As one of today’s leading theoretical orienta­ tions in developmental psychology (Schaffer 2006: 160), attachment theory at least has a claim to be taken seriously by philosophers. Given that fact, this intro­ duction will now argue that philosophers working in moral psychology or in ‘vir­ tue ethics’—whether or not they are followers of Aristotle’s own version of that outlook—ought to take more of an interest in attachment theory than they have done to date. First of all, attachment theory is a theory of psychological development which, though it applies throughout life, starts in the early years, and Aristotle told us— surely correctly—that it makes no small difference . . . whether we form habits of one kind or of another from our very youth; it makes a very great difference, or rather all the difference. (Aristotle 1992: 1743)

This is not of course to say that attachment theory takes Aristotle’s view that good character is produced simply by habituation. But simply insofar as it has much to say about ‘our very youth’, attachment theory stands a chance of telling us what Aristotle tells us very little about, namely how we get from there to here, as well as merely that we sometimes do. And if its way of doing so draws on a richer con­ ceptual repertoire than habituation alone, that is surely all to the good. Secondly, the characteristics attachment theory is basically a developmental theory of—the various subvarieties of attachment—are themselves evaluatively

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

Introduction  5 inflected: to be securely attached to a parent, or to a special other in adulthood, isn’t just to have an attachment of type B, but to have what almost any observer would agree is a kind of attachment that makes for a good intimate relationship, one characterized by physical closeness, emotional sharing, an ability to tolerate the other’s independence (because you trust they’ll come back again if they go away), an ability to admit weakness and ask for help or comfort, and so on. But obviously the classification of human character in terms of the virtues—which­ ever characteristics exactly they are, but presumably including courage, generos­ ity, fairness, and so on—is evaluatively inflected too. So it would be strange if there were no story to be told about how these two sets of evaluatively inflected descriptions relate to one another—strange, that is, if they were not in some way descriptions of different parts of the same elephant. Anyone interested in the vir­ tues, therefore, needs to think about what that story might be. A very simple story says secure attachment predicts prosocial behaviour both contemporaneously and over time (an empirical claim), and prosocial behaviour just is the expression of the various virtues (a conceptual one). Moreover one can bolster the simple story by appealing to the role of secure attachment, to which I’ve already referred, in successful emotion regulation. For if we follow Aristotle at least, emotion regulation is central to virtue: [F]ear and confidence and appetite and anger and pity and in general pleasure and pain may be felt both too much and too little, and in both cases not well; but to feel them at the right times, with reference to the right objects, towards the right people, with the right aim, and in the right way, is what is both intermedi­ ate and best, and this is characteristic of excellence.  (Aristotle 1992: 1748)

This volume of essays certainly does not advocate the simple story, but the simple story serves as a helpful benchmark against which a number of questions can be raised. One such question concerns the relationship between prosocial behaviour and the virtues. Some attachment literature proceeds as if the notion of prosocial behaviour wears its sense on its face and, relatedly, tends to focus on a relatively narrow band of socially valuable behaviour such as cooperation in tasks, ‘helping’, and the expression of sympathy (Schroeder and Graziano 2015; cf. Bakermans-­ Kranenburg and van IJzendoorn, this volume, for discussion). But is ‘prosocial behaviour’ just one thing? If it is, that would seem to stand in the way of any sim­ ple mapping from secure attachment via prosocial behaviour to virtue, because virtue seemingly isn’t one thing, insofar as it is possible to be honest without being generous, to be brave without being just, and so on. But might there not be similar variety within the prosocial? It has often been argued that the virtue of justice is necessary for humans to be able to live together

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

6  Edward Harcourt (Foot  2001) so, taking the expression at face value, that would make justice ‘prosocial’. But a solitary rescuer protecting a would-­be victim against a group of bullying peers, though it may be required by justice, may look behaviourally not at all like cooperation in a task, and only ambiguously like ‘helping’. I suspect therefore that pursuing empirical evidence for the view that secure attachment favours virtue—a fascinating project—will need to be much more sensitive than it has been so far to differences between the virtues. Instead of investigating the links between secure attachment and prosocial behaviour and taking the pro­ social as an uninterrogated proxy for virtue, more philosophically informed future research will need to test directly for links between secure attachment and different virtues—kindness and compassion, but also fairness, truth-­telling, fidel­ ity to promises, and so on. Ideally one would also correlate this kind of data with naturalistic observational data of the same subjects; with parental and/or (say) schoolteacher reports; and—especially in older subjects—using the latter kinds of data sources to gather evidence as to why this or that kind of behaviour is dis­ played when it is (such as: merely to impress, fit in, or obey, or for the multifari­ ous reasons characteristic of the various virtues). Only then might we be in a position to say whether secure attachment predicts the virtues, or rather whether some of that variety within the virtues is more closely related to secure attach­ ment than the rest. Another—and now more conceptual—question which arises for the simple story is whether the concept of the prosocial, as it features within attachment theory, is even supposed to be a virtue-­like concept. There are at least two ­reasons for thinking it might not be—though there is a great deal to say about the question which I cannot say here. The first is that the concept is often applied to the very young. But can immature human beings be said to have virtues? Aristotle would say no: virtues, he says, are stable dispositions (1992: 1005b5–1105b12, 1746), but the dispositions of the young may be unstable—a child may switch from secure attachment to a variety of insecure attachment to the same parent if, for example, the parent develops post-­natal depression after the birth of a subsequent child, presumably with consequences for the child’s prosociality. Secondly—and this consideration may resonate with the modern ear more than the first, since we may be more comfortable than Aristotle was with the idea that even adults can gain and lose virtues—whether prosocial behaviour is virtu­ ous at all will arguably depend in part on the nature of the social environment. Cooperation may fall far short of virtue if one is cooperating with the secret police by collecting lists of dissidents; disruptive behaviour may (non-­virtuously) prevent other students from getting on with their work, but it may also virtuously disrupt the routine workings of an unjust or irresponsible bureaucracy. Seen in this light, prosocial behaviour (or its absence) might well be explained by attach­ ment type, but both concepts would belong at a theoretical level below that at

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

Introduction  7 which distinctions between virtue and vice are drawn—that is, they would be human psychological dispositions which only get hammered into virtues (or not) depending on further factors, including the nature of the social environment, which it may lie beyond the business of attachment theory itself to explain. This potentially interesting explanatory role for the social environment brings me to situationism, a third area—and an area of overlap between philosophy and psychology which has excited considerable recent interest—in which attachment theory ought to recommend itself to contemporary moral psychology, and to ‘vir­ tue ethics’. On the whole, situationism (Merritt, Doris, and Harman 2010) has been seen by philosophers as constituting a challenge to the very idea of a charac­ ter trait, and thence to that of a virtue or vice.3 Notwithstanding some popular caricatures of attachment theory which see it as sponsoring a version of ‘early years determinism’—the view that early nurture lays down a character for life—at least some attachment theorists are comfortable with the idea that attachment status can change thanks to environmental features (as in the example above about post-­natal depression and the subsequent child) and, conversely, that to be sustained over time, attachment status requires the cooperation of the environ­ ment (as in the claim above that secure attachment predicts harmonious care­ giver–child relations over longer periods in the presence of continued sensitive caregiver behaviour; cf. also Bakermans-­Kranenburg and van IJzendoorn’s discus­ sion of inherited ‘responsivity to concurrent environmental factors’, in their con­ tribution to this volume).4 So far, this should be welcome to the philosophical situationist. But, if the environment does cooperate, what appears to result is the sustaining over time of something which, even if not a virtue (or a vice) is a stable and potentially explanatory complex disposition—that is, a character trait. So attachment theory seems to offer the virtue-­minded philosopher the resources to give the situationist something of what she’s after, but without conceding the cen­ tral point. A fourth area—and the final one I shall mention—in which attachment theory ought to recommend itself to contemporary moral psychology, to ‘virtue ethics’ and perhaps beyond is thanks to its relevance to the complex of philosophical issues gathered up by the term ‘naturalism’. Attachment theory has found a persuasive way of linking up its own evalua­ tively inflected classifications with characteristics which, according to the theory, 3  The same is not true of situationists in psychology: the later work of Walter Mischel and col­ leagues can be seen as an extended defence of the idea of a character trait, using situationist data to mount it. See e.g. Shoda, Mischel, and Wright 1994, and Miller 2013. 4  This is not to say that attachment theory downplays the significance of the early years: there is evidence that a mother’s attachment status predicts her own infant’s attachment status to her (Fonagy and Target 1997), and that her infant’s attachment status to her predicts her attachment status to that child if and when it becomes and adult on whom she is dependent in old age (Magai et al. 2016). This calls attention to what should be obvious anyway, that attachment theorists do not all agree with one another on everything.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

8  Edward Harcourt we have thanks to natural selection: the individual-­difference component of the theory describes how the evaluatively inflected attachment classifications come about thanks to variations which environmental and other factors play on our common evolved endowment as members of our species, as described by the theory’s ‘generic component’ discussed above. In a vocabulary that is congenial to many philosophers in the ‘virtue ethics’ mould, that is, it connects up our ‘second nature’ with our ‘first nature’ (McDowell 1994). Moreover—at least if attachment classifications explain the presence or absence of virtue but are not themselves virtue (or its absence) under another name—attachment theory might thus con­ stitute a bridging level of theory between the unarguably natural and the ethical, thus supplying ‘an account of human beings which is to the greatest extent pos­ sible prior to ideas of the ethical’ (Williams 2000: 154). Thus attachment theory would help to articulate a version of naturalism which claims not, indeed, an identity between second and first nature—that is, a reductive naturalism—but rather seeks to locate the ethical unmysteriously in the world as capable of being made intelligible to us by natural science, by displaying the continuity between our second and our first, or between our ethical and our psychobiological natures.5 However, this modest version of ethical naturalism is by no means the only version of ethical naturalism which philosophers and philosophically minded psychologists have sought to support by appeal to attachment theory. Some attachment theorists have argued that secure attachment is not only better than the alternatives because, for example, it enables good close relationships, but because it is adaptive (Humber and Moss 2005), by which is meant a characteris­ tic we need in order to play a full part in human social life. Add in the ‘simple story’ sketched earlier and attachment theory can be made to join hands with one of the most exigent sub-­varieties of neo-­Aristotelian thinking currently around in ethics, the naturalism associated with Philippa Foot and her followers (Foot 2001). Once again, the proposed volume does not endorse that view—indeed both the view and some of its rivals are represented in it. But the mere fact that this view can so easily be articulated both within attachment theory and within ‘virtue eth­ ics’ shows how much the two fields have to gain from increased exposure to one another. I now turn to summarizing the chapters collected in the volume, though the summary will continually refer back to the points of contact between attachment

5  For senses of ‘naturalism’, see McDowell 1998 and Williams 2000: esp. 148–50. For the idea of second nature, which but for the term itself McDowell ascribes to Aristotle, see McDowell 1998: 184 and 1994: 84; I take it that to display the continuity of our second with our first natures would be to relate ‘human ethical life . . . to the rest of human nature’, i.e. to answer the ‘recurrent naturalist ques­ tion’ (‘can we explain . . . the phenomenon in question in terms of the rest of nature?’) in the ‘special form’ it takes in the case of our ethical lives (Williams 2000: 154, 150). See also Williams 1995: 203 (‘a conception of ethics . . . continuous with our understanding of human beings in other respects’).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

Introduction  9 theory and contemporary philosophy that I have already mentioned. The ­chapters can be divided into three groups. The first group of chapters in the volume sets the scene for the collection as a whole, both making the case for attachment theory as a credible developmental moral psychology for vice and virtue, and drawing attention to the complexity of the explanatory connections, and of how to weight (for example) early nurture as against other potential explanatory factors. As Ross Thompson argues, theories of moral development of a broadly Aristotelian kind see virtuous behaviour as stem­ ming from virtuous character, so the developmental prehistory of virtuous behav­ iour will be a developmental account of character. Attachment research has recently made significant contributions in this area, showing (according to Thompson) that the processes that contribute to the development of secure attachment appear to contribute to moral growth in the young child, and also that the achievement of a secure attachment is associated with morally relevant behav­ iours and dispositions. In a somewhat more sceptical vein, Bakermans-­ Kranenburg and van IJzendoorn provide a comprehensive survey of the evidence as to why there are individual differences in prosocial behaviour, weighing genetic and neuroscien­ tific explanations against parenting/attachment and concurrent environmental influences, including ‘nudges’. Still, this should not be taken simply as an expres­ sion of agnosticism: their conclusion is that our natural endowment may be a responsivity to concurrent environmental factors, especially to other human beings—a conclusion which might be congenial to the philosopher who regards ‘virtue ethics’ as on the whole too individualistic, and insufficiently concerned with the role in virtue of a ‘good enough world’ (Lear 1990). Hill’s chapter mean­ while uses empirical research in attachment theory and in child development more broadly to test the strength of the connection between attachment security and virtue, and attachment insecurity and vice. Two main questions emerge: how to make a convincing explanatory connection, given that attachment security is a relatively narrow-­band disposition, while virtue is (seemingly) a very broad-­band one? And, supposing that problem can be overcome, why should attachment in particular be the dominant explanatory factor, as opposed to (say) early traits (e.g. empathy) that are arguably closer to virtue (or some of the virtues) them­ selves? Hill’s questions echo discussions which continue within attachment theory as to whether the nature of attachment is an independent explanatory factor, or a proxy for something else (e.g. parental sensitivity). All three chapters call attention to the fact that future research on the role of attachment in the developmental history of vice and virtue also requires a more nuanced—that is to say, philosophically nuanced—notion of the prosocial, or a philosophically informed expansion of attachment theory’s focus beyond its usual territory of cooperation, ‘helping’, and so on to a focus on the variety of the

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

10  Edward Harcourt virtues as a whole. Honesty, the capacity for teamwork, compassion, fairness are all different traits. Though they all fall under the umbrella of the ‘prosocial’—and rightly so, as they are all traits which it’s good to have if we are to live peacefully together—they may have different psychologies and thus relate more or less closely to the traits that typically come under the attachment researcher’s micro­ scope. An attachment-­based developmental history of the virtues (and perhaps vices) would, I suspect, look different if some of the experimental and observa­ tional work we currently rely on were to be rerun in the light of a more philo­ sophically nuanced taxonomy of the prosocial. Moving beyond attachment theory itself, a second group of chapters engage with the central theme of ethical naturalism (in one of the several senses it has carried in recent philosophical discussion). According to Philippa Foot and her followers, the familiar virtues and vices are natural excellences and defects in a sense given by the idea of our common species life: certain traits (the virtues) are necessary to enable us to engage in this species life, and certain others (the vices) debar us from engaging in it. As the first three chapters make clear, attachment theory has often argued for a developmental connection between secure attach­ ment and virtue. Some attachment theorists, moreover, have argued that secure attachment is optimal not only in the sense that it arguably favours something good—virtue—but in the sense that it represents, independently, the optimal developmental pathway for human beings. So does attachment theory lend plausi­bil­ity to Foot-­style ethical naturalism? Darcia Narvaez argues emphatically for an affirmative answer. Human life, she claims, is getting worse, at least in advanced industrial nations—not just worse for the other species and for the planet around us, but worse for humans them­ selves. Moreover, we are getting worse morally speaking—more selfish, less cap­ able of cooperation, more destructive, and so on. Why? Many studies both within and beyond the attachment paradigm have, she argues, shown links between cer­ tain types of parenting and early years education—those which emphasize phys­ ic­al warmth, trust, and play—and prosocial attitudes, virtues indeed, in later life. So a reasonable hypothesis as to why human life is getting worse in the ways described is that—as Narvaez argues—these styles of parenting and early years education are hard to achieve, perhaps even frowned upon, in advanced societies, especially perhaps the USA. However, Narvaez wants to go further than simply pointing to a decline in these respects, and proposing an explanation of it. She also wants to suggest that the ‘human baseline’ for optimal development is set by our ‘evolved developmental niche’: roughly speaking the type of family and broader social setting which evolved (as humans did) over a very long period of time and which fits our human developmental needs. In other words, the de­vi­ ations from that evolved niche which have taken place since humanity’s hunter-­ gatherer days not only explain why we and life for us are (in advanced Western

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

Introduction  11 societies) getting worse. That would be simply to make a psychological-­cum-­ sociological claim. But while Narvaez does not herself refer to moral naturalism, there is, if we follow her, ammunition for the moral naturalist here too. For these deviations from our evolved niche display the goods which (if Narvaez is right) are going missing—both moral goods and goods for us—as distinctively natural goods: the way it’s best to develop (both morally and for us) is also our ‘species-­ typical human development’, the way we are supposed by nature to be, and in ways which interdisciplinary studies of human nature—ethology, developmental psychology, and so on—are well placed to show. Keith Cox and Micah Lott also see the relationship between attachment and the development of the virtues as a testing ground for Footian naturalism. They consider a disagreement over how best to parent—in a warm and attentive style, or in a more authoritarian style aimed at producing a high achiever. Could an appeal to attachment theory settle the disagreement? In particular, could it settle the disagreement by showing that being a high achiever—or perhaps, being a high achiever at the expense of generosity and compassion—isn’t really a virtue? The argument would be that because it’s not a trait brought about by warm and attentive nurture, it doesn’t feature in (as Narvaez would have it) the optimal developmental trajectory for human beings, but since what it is to be a virtue is to be a characteristic which human beings possess who have developed naturally as they should, being a high achiever can’t be one. While Cox and Lott see it as very much an open question how empirical science—including attachment theory— might inform our understanding of what it is to live well, they reject any very direct appeal to it to settle what does and does not count as a virtue, on the grounds that vindicating a trait as a virtue is in part a matter of displaying it as worthwhile, and this in turn is a matter not of science but of self-­interpretation. Naturalism—in the sense of the identification of the virtues by way of a devel­ opmental trajectory identified as optimal prior to bringing virtue or vice into the picture—is also the main preoccupation, though not in those very words, of Jay Belsky’s chapter. Belsky, however, is strongly opposed to naturalism in this spirit. It may well be that there is a close causal and predictive association between secure attachment and characteristics generally regarded as virtues (and between insecure attachment and characteristics generally regarded as vices). It may also be that what we (at least, the likely readers of this volume) all want for ourselves and our children is that we possess the former characteristics. However, Belsky argues, these considerations together are not enough to show that the develop­ mental pathway from secure attachment to such admired characteristics is in any further sense ‘optimal’, in particular not in any sense which shows that that devel­ opmental pathway is naturally favoured, or can be shown to be favoured by appeal to considerations to do with the proper development of our natures. Belsky’s spe­ cific focus is the theory of reproductive strategies. Early adversity, Belsky argues,

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

12  Edward Harcourt favours early sexual maturation and therefore early reproduction, a strategy that in that context is evolutionarily optimal (pass on your genes fast, since you might not survive for long) even if it’s also a strategy which much effort is spent getting young people not to adopt. That’s not to say that more conventionally admired reproductive strategies may not also coincide with the development of re­pro­ duct­ive fitness, but it takes a ‘safe and secure world’ for them to do so. What is nat­ur­al­ly—that is, in Belsky’s view, evolutionarily—optimal depends crucially on the environment, thus challenging the idea that reflection on the naturally opti­ mal will home in on just one privileged set of behavioural or character traits. Though Belsky himself does not press this point, his argument points once again to the thought that ‘virtue ethics’—whether informed by developmental psychology or not—has often been insufficiently interested in the political—or a ‘good enough world’, in the phrase I used earlier, and has thus arguably neglected the role of benign and stable political environments in making the seeming con­ nection between the good (virtue) and the good for us. To see this, consider the common claim that secure attachment is socially adaptive—though this should not be taken to imply that it is also evolutionarily adaptive, a property attachment theory arguably reserves to the disposition to form attachments simpliciter, not to form attachments of any particular type. For example it is said that it’s (socially) adaptive to form a secure attachment in year one of life, because it increases the likelihood that a second ‘positive adaptation’, such as ‘the capacity for cooperation with the mother to gain needed help in problem solving’, will develop in the second year of life (Londerville and Main 1981: 290). Or again, some attachment theorists say that the ‘adaptive goals’ are ‘the facilitation of social integration . . . problem solving ability, flexibility, . . . and the ability to use adult assistance’, or that ‘toddlers of 12 to 18 months of age who experience an attachment relationship that supports mastery competence are more adaptive than children who experience an attachment relationship in which exploration apart from the parent is difficult to achieve’ (Humber and Moss 2005). The adaptations just described are positive adaptations only relative to a familiar and benign social environment (that is, a preschool class where the teachers are aiming to open the children’s minds to learning, good citizenship, and so on). But switch the context to, say, ancient Sparta—of legend if not of fact—and it might very probably be more adaptive to be insecurely attached: think of the oft-­cited insecure-­avoidant trait of precocious self-­reliance, useful if one has to spend days on end on solitary sentry duty. One can make the same point for insecure-­ ambivalent attachment: in a war zone, insecure-­ambivalent unwillingness to allow distance from the attachment figure is the more adaptive characteristic because the threats to the attachment figure’s life, and thus to the child’s survival, if (say) they leave home in search of water, is all too real (Belsky 2008: 265). This is not to argue for moral relativism: one of the many things that make it dreadful

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

Introduction  13 to have to live in Sparta or in the war zone might be precisely the costs—here, in terms of the real goods of a good childhood, or of good intimate relations, though there are surely costs of many other kinds—which must be forgone in order to adapt. A trait that helps us survive in a hostile environment is in many ways not good for us. Nonetheless, if secure attachment is privileged in respect of suiting us better to social life, the privilege is surely relative to the more or less benign and stable circumstances in which we live. But now consider in this light the claim of the Foot-­inspired naturalist, who argues that the virtues are the traits necessary for us to live a characteristically human social life. Let’s also suppose—as the ‘simple story’ above would have it, but also as some chapters in this volume aspire, in more complex ways, to argue— that secure attachment is more closely related to the virtues than are the other attachment types. It looks as if it follows that the virtues are not necessary for us to live a characteristically human social life, since—in Sparta, in the war zone, and in an affluent university town in the UK or the United States—a char­ac­ter­is­ tic­al­ly human social life is being led one way or another, enabled by very different attachment characteristics. Indeed this point can be made perhaps more force­ fully if we set aside ancient Sparta and reflect on the complexity of our own soci­ ety, in which very different social micro-­ environments coexist in the same countries and cities, so the modus operandi of the social whole makes very differ­ ent adaptive demands on different members of a single population. Secure attach­ ment, and so—presumptively—the virtues are only needed to lead a good characteristically human social life: not a bad conclusion to reach, but a disap­ pointing one from the point of view of someone trying to show that the absence of virtue is a defect of our first natures. Alternatively, one might take seriously Foot’s idea that the virtues are the traits necessary for us to lead a char­ac­ter­is­tic­ al­ly human social life, but rather than assuming that ‘the virtues’ or the human excellences coincide with the familiar catalogue of honesty, justice, generosity, and so on, remain open-­minded as to which characteristics are netted by that phrase till the investigation is over (Harcourt 2016b). Following through Belsky’s line of thought, it might be argued that the disposition to form attachments simpliciter is the only natural human excellence in the neighbourhood, a different route to the conclusion that vindicating the familiar catalogue by appeal to con­ siderations of the proper development of our natures still has some work left to do. The third and final group of chapters deal with further issues in philosophical moral psychology, either thrown up by attachment theory or which attachment theory allegedly helps to address. Harcourt, Denham, Irwin, and Schaubroeck all take seriously the evidence that certain types or styles of parenting are associated with secure attachment, and these in turn with (at least some of) the virtues. However, supposing that this connection is empirically well attested, it is a further philosophical task to make

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

14  Edward Harcourt the connection intelligible: echoing a question raised by Hill, why—as Schaubroeck asks—would a certain kind of relationship to a small set of specific other persons (secure attachment) be the developmental forerunner of a far more generalized disposition? Or as Harcourt puts it, the usual explanation of the con­ nection between secure attachment and virtue assumes we can go from disposi­ tions to favour special others to dispositions to favour any others, but this type of explanation gives out at just the point at which it is most needed. Schaubroeck reviews various answers in the literature, and rejects them in favour of an ex­plan­ ation that exploits Darwall’s account of the essentially second-­personal nature of moral demands. Harcourt, on the other hand, makes the connection indirect, via the idea of virtue as openness to reasons, reasons as features of the environment, and the fact that the securely attached—because they can rely on their special others and so afford (other things being equal) to direct their attention elsewhere—are char­ac­ ter­is­tic­al­ly more open to the environment and to the reasons it has to offer than those belonging to other attachment types. Direction of attention is also central to Irwin’s chapter, who argues that Aquinas’ account of the relation between rea­ son and non-­rational states of mind—passions—helps to explain the putative connection between secure attachment and the virtues. On Aquinas’ account one way reason controls passions is indirectly, by shifting our attention away from one object onto other more suitable ones. But, if our attention is to be captured by other possible objects, we have to be curious enough about the world to survey it. Secure attachment favours curiosity and exploration, and so is an important pre­ condition of the possibility of the redirection of attention. Aquinas, on Irwin’s account, thus helps us towards a developmental story both in which secure attachment belongs in the preconditions of virtue, and which acknowledges the continuing role of the non-­rational part of us in a mature psychology of virtue. Denham meanwhile explores the connection between secure attachment and virtue via the notion of empathy. Intuitively, empathy and virtue are related, but a common answer to the question—that empathy plays a critical simultaneous role in moral judgement—is rejected, on the grounds that sound moral judgement doesn’t need, and is sometimes undermined by, empathy. Attachment theory sug­ gests instead that the right way to locate empathy in relation to at least part of virtue is diachronically, specifically in the relationship attachment theory insists on between empathically attuned child/caregiver relations and secure attachment: we have evolved so as to ‘arrive prepared for empathically attuned intimacy’, so recognition and regard for others in maturity is our ‘default position’. Finally, Wonderly appeals to attachment theory to illuminate the moral psych­ ology of addiction, seeing addiction as a kind of pathology of valuation. On the basis of behavioural and other similarities between addiction and certain forms of attachment to persons, it has been argued that these forms of attachment should be seen as addictions to their objects. Wonderly’s chapter explores the

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

Introduction  15 converse thesis, that addictions conventionally so called—to alcohol, drugs, etc.—are a certain type of attachment to their objects. This opens the way to an exploration of ‘vices of valuing’—problematic orientations in respect of valuing— which may characterize certain attachments, whether to non-­human objects or to persons. It also shows how the philosophical possibilities opened up by intro­du­ cing the concepts of attachment theory into the bloodstream of philosophical moral psychology are only just beginning to be explored.

References Ainsworth, M., and Wittig, B. 1969. ‘Attachment and exploratory behaviour in oneyear-olds in a strange situation’. In B.  M.  Foss (ed.), Determinants of Infant Behaviour, vol. 4, 111–36. London: Methuen. Aristotle. 1992. Nicomachean Ethics, trans. W.  D.  Ross, revised J.  O.  Urmson. In Aristotle: The Complete Works. Electronic edition. BOLLINGEN SERIES LXXI, vol. 2. Charlottesville, VA: InteLex Corporation. Belsky, J. 2008. ‘War, trauma and children's development: Observations from a mod­ ern evolutionary perspective’. International Journal of Behavioral Development 32. 260–71. Bowlby, J. 1976–7/2005. ‘The making and breaking of affectional bonds’. In J. Bowlby, The Making and Breaking of Affectional Bonds, 150–88. Abingdon: Routledge. Fonagy, P., and Target, M. 1997. ‘Attachment and reflective function: Their role in selforganization.’ Development and psychopathology 9:04, 679–700. Foot, P. 2001. Natural Goodness. Oxford: Oxford University Press. Harcourt, E. 2016b. ‘“Mental Health” and Human Excellence.’ Aristotelian Society Supplementary Volume XC. 217–235. Holmes, J. 2006. Attachment, Intimacy and Autonomy. New York: Jason Aronson. Holmes, J. 2011. Exploring in Security. London: Routledge. Humber, N., and Moss, E. 2005. ‘The relationship of preschool and early school age attachment to mother–child interaction’. American Journal of Orthopsychiatry 75, 128–41. Lapsley,  D., and F.  Clark Power (eds.). 2005. Character Psychology and Character Education. Notre Dame, IN: Notre Dame Press. Lear, J. 1990. Love and Its Place in Nature. London: Faber. Londerville, S., and M. Main. 1981. ‘Security of attachment, compliance, and maternal training methods in the second year of life’. Developmental Psychology 17, 289–99. Lorenz, K. 1971. ‘Part and parcel in animal and human societies’. In K. Lorenz, Studies in animal and human behaviour, Vol. II, 115–195. Cambridge, MA: Harvard University Press. McDowell, J. 1994. Mind and World. Cambridge, MA: Harvard University Press.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

16  Edward Harcourt McDowell, J. 1998. ‘Two Sorts of Naturalism’. In J. McDowell, Mind, Value and Reality, 167–197. Cambridge MA & London: Harvard University Press. Maestripieri, D. 2004. ‘Developmental and Evolutionary Aspects of Female Attraction to Babies.’ APA Psychological Science Agenda. https://www.apa.org/science/about/ psa/2004/01/maestripieri Magai C., M. T. Frías, and P. R. Shaver. 2016. ‘Attachment in Middle and Later Age.’ In In Cassidy, J., Shaver, P.  R. (Eds.), Handbook of attachment: Theory, research, and clinical applications (3rd ed.), 534–552. New York: Guilford Press. Masala, A., and J.  Webber (eds.). 2016. From Personality to Virtue: Essays on the Philosophy of Character. Oxford: Oxford University Press. Merritt, M.W., Doris, J.M., and Harman, G. 2010. ‘Character’. In J.M Doris & the Moral Psychology Research Group, Eds., The moral psychology handbook, 355–401. Oxford: Oxford University Press. Mikulincer, M., and Shaver, P.  R. 2008. ‘Adult attachment and affect regulation’. In J.  Cassidy and P.  Shaver (eds), Handbook of Attachment: Theory, Research and Clinical Applications, 503–31. New York/London: The Guilford Press. Miller, C.  B. 2013. Moral Character: An Empirical Theory. Oxford: Oxford University Press. Russell, D. 2009. Practical Intelligence and the Virtues. Oxford: Oxford University Press. Schaffer, H. R. 2006. Key Concepts in Developmental Psychology. London: Sage. Schroeder,  D., and W.  Graziano (eds.). 2015. The Oxford Handbook of Prosocial Behaviour. New York: Oxford University Press. Shoda Y., W. Mischel, and J. Wright. 1994. ‘Intraindividual Stability in the Organization and Patterning of Behaviour.’ J. Personality and Social Psychology 67, 674–87. Simpson, J. A., and Belsky, J. 2016. ‘Attachment theory within a modern evolutionary framework’. In J. Cassidy and P. R. Shaver (eds), Handbook of Attachment: Theory, Research and Clinical Applications (3rd edn), 91–116. New York/London: The Guilford Press. Snow, N. 2010. Virtue as Social Intelligence. New York: Routledge. Solomon, J., and George, C. 2008. ‘The measurement of attachment security’. In J.  Cassidy and P.  R.  Shaver (eds), Handbook of Attachment, 383–418. New York/ London: The Guilford Press. Sroufe, L.  A. 1983. ‘Infant–caregiver attachment and patterns of adaptation in pre­ school’. In M.  Perlmutter (ed.), Minnesota Symposia on Child Psychology, vol. 16, 41–83. Hillsdale, NJ: Erlbaum. Sroufe, L. A. 2005. ‘Attachment and Development: A prospective, longitudinal study from birth to adulthood’. In Attachment and Human Development 7 (4), 349–367. Thompson, R. A. 2016. ‘Early attachment and later development: reframing the ques­ tions’. In J. Cassidy and P. R. Shaver (eds), Handbook of Attachment: Theory, Research and Clinical Applications (3rd edn), 330–48. New York/London: The Guilford Press.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

Introduction  17 Tiberius, V. 2015. Moral Psychology: A Contemporary Introduction. New York: Routledge. Weinfeld, N. S., et al. 2008. ‘Individual differences in infant–caregiver attachment’. In J. Cassidy and P. Shaver (eds), Handbook of Attachment, 78–101. New York/London: The Guilford Press. Williams, B. 1995. ‘Replies’. In World, Mind and Ethics. Ed. J. E. J.  Altham and R. Harrison, 185–224. Cambridge: Cambridge University Press. Williams, B. 2000. ‘Naturalism and Genealogy’. In Morality, Reflection and Ideology, ed. E. Harcourt, 148–161. Oxford: Oxford University Press.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

1

Attachment Theory and Moral Development Ross A. Thompson

1. Introduction Moral development has been an enduring interest to philosophers and psychologists because morality is central to how people live cooperatively with each other. But moral development is complex, especially to those who study its development. Consider the following question: What develops in moral development? What, in  other words, are the fundamental causes of moral growth? Developmental ­psychology has offered several answers to this question over the years, and each builds on long-­standing perspectives from Western philosophy. One answer to this question is that moral development consists of children’s gradual internalization of rules and values from parents and other social au­thor­ ities. In this view, young children do not begin life as moral actors but, quite the opposite, as self-­interested and egocentric. Parental socialization is essential to redirecting children’s natural tendencies through rewards and praise for desirable conduct, sanctions for disapproved conduct, guided imitation of models of good behaviour, and other practices. Especially when these socialization efforts are accompanied by rational justifications for morally desirable behaviour, they contribute to the child’s progressive internalization of moral values so that the child eventually feels a personal commitment to upholding them. Simply stated, the egocentric young child develops into a responsible moral actor through the incentives afforded by parental socialization. Such a view would be familiar to Western philosophers like Hobbes, who argued that humans are self-­interested but will act consistently with the collective good when it is perceived as consistent with self-­interest. This approach to moral development is also consistent with (although does not necessarily imply) utilitarian or consequentialist approaches to morality. Another answer to this question is that moral development is cognitively constructed by the child through a combination of intellectual growth and social experience, and this results in progressively more adequate, and principled, forms of moral judgement. Moral development derives from more sophisticated moral Ross A. Thompson, Attachment Theory and Moral Development In: Attachment and Character: Attachment Theory, Ethics, and the Developmental Psychology of Vice and Virtue. Edited by: Edward Harcourt, Oxford University Press. © Ross A. Thompson 2021. DOI: 10.1093/oso/9780192898128.003.0002

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

Attachment Theory and Moral Development  19 reasoning. Both Jean Piaget (1935) and Lawrence Kohlberg (1969) argued that young children are initially motivated to comply because of external incentives, but with cognitive growth come diminishing egocentrism and greater capacities to understand the perspectives of others and to view oneself in the context of the rights and responsibilities of a broader social order. Further cognitive growth inspires reliance on moral principles to guide conduct, such as those concerning justice, respect for persons, and the value of human life. This developmental story would be congenial to Kant and his followers because of their concern with universalizable laws that guide moral conduct. Kant was a philosophical constructivist in much the same way that Piaget and Kohlberg were psychological or developmental constructivists. A third answer to the question, what develops in moral development?, is that it is the unfolding of a network of biological predispositions that have contributed to the survival of the human species. Drawing on a range of research literatures in cultural anthropology, primatology, neuroscience, and developmental psych­ ology, psychologists like Paul Bloom (2013) and Jonathan Haidt (2012) argue that moral development builds on moral intuitions that are biologically deeply rooted and derive from the intraspecies cooperation necessary for human survival. Viewed in this light, even infants and young children begin life with a range of basic predispositions toward fairness and cooperation that form the core of moral development and that eventually become incorporated into broader moral values and rules. Rousseau would find this a familiar view and might urge, consistently with some contemporary developmental psychologists, that children’s natural dispositions are the foundation on which constructive, prosocial, and compassionate responses to others are founded. Followers of the natural law tradition would similarly be drawn to this approach. Each of these psychological perspectives to the problem of moral development is valuable, but for three reasons, each is limited in light of contemporary thinking in developmental science (Thompson 2012). One reason is growing recognition of the influence of early experiences on developing personality and character, especially viewed in the context of the powerful reasoning capacities of young children. Developmental scientists have realized that young children are rapidly constructing a remarkably perceptive and nonegocentric understanding of ­people and the world based on everyday social experiences. A theory of moral development must take these early experiences into account if they help explain some of the foundations of moral growth. A second reason is that current approaches have little to say about the self as an important element of moral development, despite attention to moral character and moral identity in cognate fields such as education. Whether children perceive themselves as moral actors, and how they do so with increasing age, is an important element of their moral growth. A third reason is that current approaches provide limited understanding of the complex

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

20  Ross A. Thompson influence of close relationships on moral development and understanding, focusing on rewards and punishments, moral tutelage, and examples of moral responsibility in the family. Missing, however, is a view of relationships as forums for cultivating capacities for cooperation and compassion, developing trust and becoming a trustworthy partner, and developing traits and dispositions that undergird moral conduct. There is emerging, however, a fourth way of answering the question, what develops in moral development?, that begins to address these limitations. In this view, the essential quality of moral development is not the internalization of values, development of moral judgement, or elaboration of moral intuitions, but rather the development of personality qualities (or character) that yield cooperative, compassionate, and constructive moral conduct. These elements of personality begin to develop very early, are shaped by the quality of close relationships, and are influenced by young children’s rapidly growing understanding of other people, self, and how to relate to others. Essentially, moral conduct arises from developing character. Such a view is consistent, in some ways, with Aristotelian virtue ethics and the understanding that virtuous conduct derives from the development of character that arises from habitual dispositions and rational reflection. And consistent with the Aristotelian view, contemporary developmental scientists regard the quality of parent–child relationships as central forums for moral development and the growth of character and personality. Contemporary attachment theory is a leading instantiation of this fourth approach to moral development. Although attachment theory is not a moral development theory, its claims and the range of its applications extend to moral development and, equally importantly, it provides a portrayal of what a relationally oriented theory of the development of moral character might look like in light of contemporary developmental science. Considering moral development in the context of attachment theory also suggests some of the nuances that are ne­ces­sary in thinking about moral growth in a relational context, while also contributing new perspectives on the habituation of virtuous dispositions and the growth of rational reflection that might inform the field of virtue ethics. The purpose of this chapter is to appraise how far the ideas of attachment theory can take us toward a new and contemporary portrayal of moral ­development. In section 2, the basic ideas of attachment theory are outlined, especially those relevant to the development of secure and insecure parent– child relationships. Following this, research on the security of attachment is ­summarized to address the extent to which a secure attachment contributes to  the characteristics supporting moral growth and virtuous character. The research review then turns to examining the empirical evidence concerning the attributes of parenting leading to a secure attachment to determine whether these attributes also promote moral development. The chapter then offers some

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

Attachment Theory and Moral Development  21 reflections on what an attachment theory of moral development might look like, and finally some closing reflections.

2.  A Précis of Attachment Theory An attachment can be described as an enduring emotional tie between two ­people that connects them across distance and time (Bowlby  1969; Ainsworth  1973). Attachments thus can be viewed as characteristic of close relationships throughout life. To attachment theorists, however, the attachments that develop shortly after birth between young infants and their caregivers are distinct both bio­logic­ al­ly and psychologically. Biologically, attachment theorists believe, attachments evolved for this early developmental period to ensure the survival of dependent infants to reproductive maturity. In the environmental contexts of human evolution, risks from predation, abandonment, and other dangers posed formidable challenges to the survival of the young, which made the evolution of behaviours to promote protection and nurturance essential to the reproductive success of both parents and children. These evolved behaviours in children include seeking proximity to caregivers, especially when the child is distressed, alarmed, or in danger, and other behaviours like clinging, smiling, and crying to maintain proximity once it is achieved. These are complemented by behaviours in caregivers that entail sensitive awareness and quick responsiveness to these infant emotions and behaviours. Together, these complementary responses in both infant and parent help to promote the child’s survival, and establish a motivational basis for the strong emotional bonds between them. It is thus typical for infants to develop emotional attachments to those who care for them in normal circumstances (i.e. conditions that are not grossly neglectful or abusive). But because the capability of the infant to effectively solicit an adult’s assistance depends on the caregiver’s sensitive responsiveness, the attachments that develop may be secure or insecure. A secure attachment is more likely to develop when caregivers are sensitive and responsive to the child’s signals, and such an attachment enables the child to play and explore because of confidence that the adult will be helpful when needed. In contrast, an insecure attachment is more likely to develop when caregivers are insensitive, unresponsive, or inconsistently responsive, and these patterns of adult behaviour undermine the child’s confidence in assistance and result in various patterns of insecure behaviour, such as becoming preoccupied with the adult to the exclusion of exploration. It is important to note, however, that an insecure attachment is not the same thing as having no attachment at all. Even when a young child is insecure, there is still important emotional support derived from his or her attachment to an adult who provides care, especially compared to a child with no attachments to any adult.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

22  Ross A. Thompson The biological uniqueness of infant–parent attachment is complemented by its psychological uniqueness, according to attachment theorists. Simply stated, many central features of psychological growth are developing in concert with a secure or insecure attachment, and consequently the security of attachment can have a formative effect on early psychological development. The security of attachment is likely to influence, for example, how young children begin to construct and understand their relationships with other people (generalizing from experiences with the attachment figure), along with children’s expectations for the support they may or may not receive, and beliefs about what people are like and how to engage with them. Attachment might also, for this reason, influence emerging social dispositions, emotion and emotion regulation, and personality development. The security of attachment would also be expected to influence how young children construct an awareness of themselves as deserving care or expecting success in relating to other people. There are, in fact, a wide range of developing social, emotional, and personality characteristics that may be affected by a secure or insecure attachment, and this has been a topic of wide-­ranging research by developmental psychologists for many years. The breadth of the developmental influences of early attachment is underscored by the possibility that these influences exist unconsciously as well as consciously, and might have implications for developing mental health. Attachment theorists describe the psychological cor­ rel­ates of secure or insecure attachment in terms of ‘internal working models’ of relationships and the self. This term implies that these implicit concepts of self and relationships constitute models that guide the child’s interactions with others and that they are continuously in process. There are more elements of attachment theory and the large research literature it has inspired (see Thompson 2013, for a brief overview). The theoretical richness of attachment theory would unlikely have been so influential, however, but for the development of methods for measuring the security of attachment in infancy and at later ages, and with the creation and validation of these meth­od­ olo­gies the field of attachment has been a dominant influence on developmental psychology for more than fifty years. In general, research has confirmed many of the central claims of attachment theory, such as those concerning the association of attachment security with the sensitive responsiveness of adult caregivers, and the range of social, emotional, and personality characteristics that are associated with a secure or insecure attachment (Thompson  2017). Unsurprisingly, there have also emerged provocative debates concerning, for example, the influence of culture on attachment, whether and how the security of attachment changes over time, and the public policy implications of attachment for issues such as child custody following parental divorce and childcare (see contributors to Thompson, Simpson, and Berlin, in press). All of these are signs of a generative theory that has every potential of continuing to be a strong influence on future thinking about early development.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

Attachment Theory and Moral Development  23

3.  Security of Attachment and Moral Development This précis of attachment theory helps explain why there might be associations between the security of attachment and moral development. If the internal working models inspired by a secure or insecure attachment provide guidance for how to get along with people, for example, this would have implications for how children develop a sense of responsibility to others, caring and compassion, and a sense of self as a morally responsible being. Furthermore, if the parenting quality that leads to a secure attachment involves responsiveness to the child’s needs and signals, this might provide a model for the child’s responsiveness to others’ needs that is also morally relevant. At the outset, it is important to comment on the meaning of the term ‘moral development’ in psychology. In traditional theories of moral development, the concept has somewhat narrowly meant the development of internalized compliance with societal values and rules. While this is important to moral conduct, researchers have increasingly broadened their understanding of moral development to include emerging conceptions of distributive justice (such as how to share resources among several people varying in need, deservingness, or other characteristics), prosocial behaviour (behaviours such as helping or sharing which, according to most psychologists, are non-­obligatory on any given occasion), empathy and compassion for the distress of another (which may or may not be accompanied by assistance), and a developing sense of oneself as a moral actor (including moral identity in adolescence). In broadening the traditional conceptualization of moral development, researchers also expand the potential range of influences on these morally relevant behaviours, including those potentially from attachment relationships. Attachment researchers have not systematically examined the association of the security of attachment with the full breadth of influences on the development of moral conduct. Nor has there been adequate examination of the direct and indirect ways that attachment might be relevant to the development of morality, broadly defined. But research reviewed below provides some indicators of how attachment and moral development might be associated.

3.1  Self-­regulation and emotion regulation An important element of moral behaviour is that it is self-­regulated, which en­ables conduct to be guided by values and standards rather than impulse. There is evidence that young children in secure attachments to their mothers are more advanced in self-­regulation compared to insecurely attached children (Bernier et al.2015; von der Lippe et al. 2010). This conclusion comes from studies in which the security of attachment was assessed in infancy or early childhood, and

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

24  Ross A. Thompson behavioural measures of children’s self-­regulation, including skills in planning, impulse control, working memory, and shifting mental set, were subsequently obtained at age 5–6. These skills, commonly described as ‘executive functions’, are a critical mental foundation of self-­regulatory capability. Executive functions also contribute to developing emotion regulation, which is important to moral development because emotional outbursts can easily lead to spontaneous aggression and other antisocial responses. A number of studies have found that securely attached children are also stronger in emotion regulation, including developing capacities for anger management (see review by Zimmer-­Gembeck et al. 2017). Greater capacities for emotion self-­regulation also contribute to the coping skills of securely attached children facing difficulty or adversity. The growth of self-­regulation has a long developmental course lasting until early adulthood, which means that young children are still neophytes in self-­ control (Thompson, 2015). This means that they must rely on the co-­regulatory support provided by a parent, and attachment theorists believe that one reason securely attached children are stronger in self-­regulation is that they have parents who respond with sensitivity and support for their emerging self-­regulatory skills. There is research evidence that when parents are sensitive and responsive to their children’s needs, offspring develop stronger self-­regulatory capabilities (Calkins and Leerkes 2011).

3.2  Social problem-­solving and conflict avoidance with peers Moral development theorists like Piaget and Kohlberg argued that peer relationships are an important arena for the development of moral conduct and dispositions. They explained that the more egalitarian quality of peer relationships enables children to participate in creating and enforcing normative expectations, compared to the more authoritarian quality of parent–child relationships. In this light, it is noteworthy that securely attached children of various ages show greater social competence with peers (see review by Groh et al. 2014). In these studies, social competence was measured as constructive social skills, positive social interaction with peers, popularity, interpersonal awareness and other characteristics based on adult or peer ratings or observations of children. Although researchers have not elucidated the connections between the peer competence of securely attached children and their moral development, the ­reasons that securely attached children have this advantage suggest that they are developing a more constructive orientation toward other children. In their ana­ lysis of a large national sample of more than 1,000 children, Raikes and Thompson (2008a) found that securely attached children devised more competent solutions to hypothetical stories involving social problems with peers at age 4½ (such as making a request rather than grabbing another child’s toy when wanting to play

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

Attachment Theory and Moral Development  25 with it). At age 4½ and also at first grade, securely attached children also showed less bias toward attributing hostility to a story protagonist who acted with am­bigu­ous intent (such as when the protagonist ran past another child’s block tower which subsequently collapsed). Interestingly, these associations between the security of attachment and peer sociability were apparent over and above the influence of maternal sensitivity, suggesting that these dispositions have arisen in young children’s give-­and-­take experience of peer sociability. In another study of this sample, securely attached children were rated at age 4½, kindergarten, and first grade as showing lower levels of peer conflict compared to children with insecure attachments (Raikes et al.  2013). This developmental period typically witnesses a decline in peer conflict and this was true of the children in this sample as well, but securely attached children showed a steeper decline in peer conflict from age 4½ to first grade than did other children, and their greater social problem-­solving skills were one reason why. Taken together, these findings suggest that in early childhood a secure attachment is associated with greater skill in solving social problems, perhaps associated with their attributing the more benign motives to peers that would be consistent with the more positive internal working models of relationships that are believed to derive from a secure attachment.

3.3  Emotion understanding and empathy Emotion understanding is an important means by which young children establish a psychological connection to the experience of another, and thus it has the potential of becoming an early resource for moral emotion and motivation. Several studies have shown that securely attached children show greater emotion understanding compared with insecurely attached children, especially when nega­tive emotions are concerned (see review by Cooke et al. 2016). The assessments included measures of children’s understanding of emotions in other p ­ eople, hypothetical story characters, and themselves. Developmental researchers have found that developing emotion understanding is associated with how mothers talk about emotions with their young children. Mothers in secure relationships discuss emotion-­ related events in a descriptively richer, more elaborative manner than do mothers of insecurely attached young children (Laible  2004; Ontai and Thompson  2002). In other words, they provide explanations, context, and other information to help children comprehend the causes and outcomes of the emotions they feel or observe. Mothers in secure relationships also engage in greater coaching of emotion regulation, and provide greater validation and support to their child’s viewpoint even as they are expanding it (Raikes and Thompson  2008b; Thompson, Virmani, et  al.  2013). When children are upset, mothers in secure relationships are also

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

26  Ross A. Thompson more likely to accurately perceive and interpret their children’s emotions, and this may also contribute to the child’s emotion understanding (Waters et al.  2010). Taken together, it appears that there are many elements of the mother–child ­relationship that enable securely attached children to converse more comfortably about their feelings and the feelings of others, especially negative emotions, and to learn about them from the parent. This may contribute, in turn, to enhanced empathy for the emotions of others. There is evidence that securely attached young children respond with greater empathy to another’s distress (Kim and Kochanska  2017; Murphy and Laible  2013). In these studies, empathy was measured by the young child’s response to a parent’s instructed feigning of a minor injury (such as pretending that the child hit her finger while hammering pegs on a pounding block toy) or response to the recorded sound of a baby’s cry in an adjacent room. Empathy was typically indexed by the child’s expression of concern, whether or not it was accompanied by helping or comforting the victim. Greater empathic responding would be consistent with the more positive, constructive orientation to others of securely attached children that was previously discussed concerning peer relationships.

3.4  Conscience and the moral self As developmental psychologists use the term, conscience refers to the developing capacity to behave consistently with internal, generalizable standards of conduct. It is typically assessed in young children’s cooperative conduct and compliance with parental requests when the parent is absent. Several studies with preschool-­age and older children have found that securely attached children exhibit more advanced conscience development than children with insecure attachments (Kochanska 1995; Kochanska, Aksan, and Carlson 2005; Laible and Thompson 2000). Attachment theorists believe that secure parent–child relationships contribute to developing conscience in several ways. A secure attachment enhances the young child’s responsiveness to the parent’s guidance and instructions, it increases the child’s motivation to cooperate and comply, and for these reasons it reduces the parent’s reliance on assertive discipline practices in favour of gentler methods. Although these different avenues from attachment to conscience have not been systematically tested, there is some supportive evidence. When talking about past misbehaviour with their 2½-­year-­old children, for example, mothers in secure attachment relationships were more likely to make reference to people’s feelings, make morally evaluative statements (e.g., ‘that was a nice thing to do’), and were less likely to refer to rules and the consequences of breaking them than were mothers in insecure relationships with their children (Laible and Thompson 2000). These conversational elements concerning misbehaviour, along with their shared

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

Attachment Theory and Moral Development  27 positive emotion when interacting together, were related to children’s more advanced conscience development at age 3. Another study showed that although mothers in secure relationships with their 3-­year-­olds did not necessarily engage in less conflict with their children during home observations, their conflict episodes entailed greater maternal use of justifications and compromise, there were fewer maternal responses that aggravated conflict, and conflict episodes were more likely to be resolved compared to mothers in insecure relationships (Laible, Panfile, and Makariev 2008). Over time, these conversational elements may begin to influence how children perceive themselves. Although it was long believed that young children regard themselves exclusively in concrete terms (e.g., what they look like, what they can do), it is now apparent that young children have a sense of themselves as emotional and psychological beings as well: when interviewed appropriately, they describe themselves as sociable or shy, expressive or reserved, and in other psy­ cho­logic­al ways using the terms available to them. Research of this kind indicates that a secure attachment with the mother is associated with a more positive self-­ concept in young children (Goodvin et al. 2008). However, the evidence is less clear concerning the association of attachment with children’s perception of themselves as moral actors—that is, who perceive themselves as bothered by wrongdoing, apologizing or making amends, feeling sympathy for another, and trying to do the right thing. This collection of self-­reported characteristics has been called the ‘moral self ’ in young children (Kochanska 2002a). The only study to test this association reported that although there was no direct association between the security of attachment (measured at age 3½) and the ‘moral self ’ at age 7, there was an indirect association. Security of attachment predicted later measures of the child’s positive, responsive cooperation with the adult at age 7 which was, in turn, associated with measures of the ‘moral self ’ at the same age (Goffin, Boldt, and Kochanska  2018). More research on this question is certainly needed.

3.5  Prosocial behaviour The term prosocial to psychologists generally means behaviours that are intended to benefit another. In developmental study, prosocial behaviour is typically indexed by actions such as helping, sharing, and comforting another in distress. An important characteristic these behaviours share in common is that they are— with the qualification I’ve made already1—non-­obligatory: it is desirable to help and share with others, for example, but the failure to do so on a given occasion

1  See above, section 3.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

28  Ross A. Thompson does not constitute a moral violation. Another characteristic they share in ­common is that the child does not benefit, at least directly, by acting in these ways (instead, the child may benefit most by not sharing). Together, this makes the motivation for prosocial behaviour different from the motivational bases for other moral actions, and raises interesting questions concerning its development. Helping, sharing, and even simple comforting can be observed in toddlers as young as 18 months, although they are inconsistent and depend on the context and the recipient of assistance, and prosocial behaviour increases in frequency as children mature (Thompson and Newton 2013). Contrary to expectations, however, researchers have found only mixed and inconsistent evidence of an as­so­ci­ ation between the security of attachment and prosocial behaviour in children (see review by Gross et al.  2017). By contrast, studies of attachment in adolescence and adulthood yield much stronger evidence of the greater prosocial motivation of securely attached individuals (Shaver et al. 2016). One possibility, therefore, is that because it is nonobligatory and entails few direct benefits, prosocial behaviour becomes associated with the security of attachment only at later ages, when attachment may begin to motivate a broader prosocial disposition toward others. The association of prosocial behaviour with attachment may also be developmentally extended because these actions require greater psychological sophistication than simply cooperating with parental expectations. Helping requires awareness of how to assist someone in achieving their goals, sharing requires that the child give up resources for another’s benefit, and comforting requires understanding of the causes of another’s distress and how to remedy it. These elements of prosocial assistance can be challenging for young children, and it may require further growth in psychological understanding before these behaviours become influenced by attachment security. It is also possible that the association of prosocial behaviour with the security of attachment depends for young children on the recipient of assistance. Although researchers have long been aware that young children help and share with family members, most recent research has examined prosocial behaviour toward un­famil­iar adults who cannot reciprocate the assistance they receive. Young children are not, however, typically in situations affording assistance to unknown adults. This leads to the question of whether it is reasonable to expect a secure attachment to lead to a generalized prosocial disposition toward other people in childhood, or whether it should instead motivate helpful assistance toward those whom the child knows. Indeed, a generalized prosocial disposition could potentially put young children at risk if strangers are needy but malevolent. With increasing age, however, there may be stronger and more reliable associations between the security of attachment and prosocial behaviour as internal working models developed from family relationships become generalized to interactions with others, and as children become more discriminating social perceivers.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

Attachment Theory and Moral Development  29 These considerations underscore the need for new research in this field to explore these hypotheses. For present purposes, however, they also indicate (a) the need for careful thinking concerning the association of the security of attachment with moral development (some associations may be developmentally emergent), (b) the importance of resisting simple expectations like ‘all good things go together’ in the association of attachment with different aspects of moral development, and (c) the recognition that the potential outcomes of a secure or insecure attachment are likely to be influenced by many developmental processes besides attachment.

3.6  Interim conclusion Taken together, the research evidence on attachment and moral development offers thought-­provoking conclusions that are generally consistent with the formulations of attachment theory. Securely attached children have many of the characteristics that are associated with admirable moral conduct: greater self-­ regulation, emotion regulation and coping skills, greater emotion understanding and empathy, stronger conflict resolution and social problem-­solving skills with peers, and greater cooperation and compliance with the parent (i.e. conscience development). For more complex forms of morally relevant behaviour, most notably the development of a ‘moral self ’ and prosocial responding, the evidence is more ambiguous and supports calls for further research. A general expectation that secure attachment is associated with more advanced moral conduct is thus only partially supported, especially for young children. Equally important, this review underscores some of the nuances in conceptualizing the association between attachment and moral development, especially that some associations may emerge with increasing age and psychological sophistication, and that attachment intersects with other developmental influences on moral growth. This review also highlights the elements of moral development that are not well addressed by attachment research, such as moral emotion, the development of morally relevant thinking, and the growth of moral identity. There is the potential for the theory to be expanded in ways that encompass these developmental processes, but this work has not yet been done. Attachment theory provides more promising guidance concerning the development of personality or character, although the development of character is more than the development of habitual or automatic responses to people and situations, such as the securely attached child’s greater empathic responding or conflict-­ resolution skills. These are im­port­ant, and the idea that these routine ways of thinking and responding are both conscious and unconscious and become incorporated into the development of children’s internal working models of relationships and the self provides

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

30  Ross A. Thompson attachment theory with a powerful explanatory framework for thinking about these aspects of character formation (Thompson and Lavine 2016). But character development also involves processes of rational reflection that enable the individual to make thoughtful moral choices and decisions, and attachment theory has relatively little to contribute concerning this aspect of character that is also important to Aristotelian virtue ethics. Finally, the research discussed above denotes associations between one developmental outcome—security of attachment—and another—the development of moral character. But this does not well inform inquiry into how a secure attachment contributes to moral character. Concerning the latter, there is research evidence that many elements of moral development discussed above are influenced by parental practices that are also associated with a secure attachment. Young children whose parents are sensitive and responsive are more advanced in conscience development (Kochanska, Forman, Aksan, and Dunbar 2005; Kochanska and Murray 2000), are observed acting more prosocially toward peers or strangers (Lindsey, Creemens, and Caldera 2010; Newton, Thompson, and Goodman 2016), show greater empathy toward mother or a stranger (Spinrad and Stifter  2006; van der Mark, van IJzendoorn, and Bakermans-­ Kranenburg  2002), and distress at wrongdoing (Kochanska, Forman and Coy  1999), and are more emotionally self-­regulated (see review by Calkins and Leerkes 2011). In addition, research findings discussed earlier document how mothers in secure relationships with their children use responsive and low-­confrontational discipline practices (contributing to conscience development), discuss emotions in a descriptively richer, more elaborative fashion (contributing to emotion understanding), and coach emotion regulation skills (likely contributing to children’s emotion self-­regulation) compared to mothers in insecure relationships (Laible  2004; Laible and Thompson  2000; Thompson et al. 2013). These relational elements are likely to foster and support secure mother–child relationships as well as aspects of early moral growth. These findings suggest that the apparent effects of a secure attachment may derive instead from the influence of parental practices that contribute both to attachment security and to moral development as children mature. This would be reasonable in light of how a parent’s thoughtful and supportive responsiveness to offspring has many consequences as it guides social-­emotional learning, provides a model of helpful, constructive conduct that young children may generalize to their interactions with others, and inspires greater receptivity to the parent’s values and expectations. Viewed in this light, the central contribution to early moral growth is parental responsiveness, not a secure attachment. But it is also im­port­ ant to note that other studies have shown that the security of attachment is consequential over and above the influence of parental sensitivity. As earlier noted, for example, Raikes and Thompson (2008a) found that securely attached children exhibited greater social problem-­solving skills and were less likely to attribute

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

Attachment Theory and Moral Development  31 hostile intentions to someone acting with ambiguous intent compared to insecurely attached children, and that these influences were apparent even when the effects of prior maternal sensitivity were taken into account. Contributing interesting complexity to this issue are indications that the se­cur­ ity of attachment may also affect the impact of other parental practices on developing moral conduct. Attachment is one of multiple ways that parents affect moral development, of course, and the interaction of parental influences is sometimes greater than each considered alone. In one study, for example, a composite measure of the mother’s responsive, gentle parenting over 14–45 months was associated with the child’s conscience development at 56 months—but only for  children who were securely attached (Kochanska, Aksan, Knaack, and Rhines 2004; Kochanska, Woodard, et al. 2010). There was no association between responsive parenting and conscience for insecurely attached children. The op­pos­ ite is also true. In another sample, Kochanska and her group found that measures of parental power-­ assertive practices in early childhood predicted children’s resentful opposition to parents at age 52 months which, in turn, predicted ratings of children’s antisocial conduct at 67 months, but only for children who were insecurely attached. There was no association between these measures for securely attached children (Kochanska, Barry, Stellern, and O’Bleness 2009). These findings have led Kochanska and her colleagues to propose that early attachment inaugurates a cascading series of changes in parent–child interaction that move children toward positive or negative developmental trajectories in their socioemotional, moral, and behavioural development (Kochanska, Boldt, and Goffin  2019). In this respect, early security or insecurity has a ‘snowball effect’ because of its effects on the impact of other parental practices on children’s development. These ideas merit further exploration, especially for the perspective they offer on the habituation of character through early childhood family experience. They underscore the interaction and convergence of diverse relational experiences as a context for the development of virtuous character and moral conduct. In comparison to predominant theories of moral development, these studies also broaden understanding of the relational influences that contribute centrally to moral growth (beyond reward, punishment, instruction and imitation) and underscore how early in development these experiences are influential. In a manner that is important both to virtue ethics and attachment theory, moreover, they suggest that it is necessary not just to denote associations between secure attachment and moral development, but also to identify the underlying elements of young children’s relational experiences in the family that contribute both to attachment security and moral conduct. Although attachment theory is not a moral development theory, therefore, it has many unique contributions to offer in understanding the growth of character

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

32  Ross A. Thompson and morality. As an avenue to envisioning what a more fully developed attachment view might look like, an outline of an attachment perspective to moral development is proposed in section  4 (see Thompson  2012,  2015b for further detail). Importantly, it is not only concerned with the association of a secure attachment with moral conduct, but with the developmental and relational processes contributing to the growth of security and moral sensibility.

4.  Toward an Attachment Perspective to Moral Development Such a perspective builds on several theoretical foundations of attachment theory discussed earlier. One is the view that species evolution has conferred special significance to the communication of human emotions and needs because of their relevance to the survival of dependent infants. This helps to account for the unique attention adults devote to infant cries and smiling, but it also helps to explain young children’s special attention to the emotions and needs of other ­people, especially if they might be relevant to the child’s well-­being. Early emotional sensitivity may constitute an important resource, in turn, for emergent moral awareness. It is important to note that this view is different from those who propose that basic elements of moral understanding, such as valuing helping and defending victims, are innate to our species (e.g., Bloom  2013). The alternative view proposed here, which is consistent with attachment theory, is that emotional sensitivity may be the more important early constituent of developing moral awareness, and that moral intuitions are not necessary to this account. Second, just as attachment theory enlisted contemporary understanding of young children’s cognitive development to explain the growth of attachment, an attachment perspective to moral development must also enlist an updated understanding of early cognitive growth. This is important because there have been significant advances in knowledge of how young children learn about people and the mental causes of people’s actions—conventionally called ‘theory of mind’— that have important implications for their developing moral awareness. These early advances in psychological understanding provide young children with basic understandings about why people act as they do and the consequences of a person’s actions for another’s feelings, goals, and needs. Finally, an attachment perspective to moral development regards moral growth within a constructivist framework rather than an internalization model. In other words, moral development is not primarily a matter of young children internalizing standards and expectations from others but of constructing normative ex­pect­ ations for conduct as children apply their developing understanding of people, self, and relationships to everyday social experiences. The constructivist orientation of attachment theory is most evident in how internal working models are believed to be constructed from experiences of sensitive (or insensitive) parental

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

Attachment Theory and Moral Development  33 care as they interact with the infant’s emergent understanding of parents as ­people. Likewise, an attachment perspective to moral development proposes that early developing moral sensitivity derives initially from young children’s social experiences as they interact with emerging understanding of people, not primarily as the result of instruction and reinforcement from parents.

4.1  Developing a premoral sensibility Emotions are an important means of establishing a psychological connection to another person’s experience. Consequently it is not surprising that adults attend to an infant’s emotions, but it may be more surprising how much infants attend to an adult’s emotional expressions. One-­year-­olds perceive a range of emotions in others and try to understand the causes of those emotions (see Thompson and Lagattuta 2006, for a review). They are adept, for example, at using a person’s visual gaze to infer the reason for that person’s emotional expression, such as looking at the individual or object that the person feeling the emotion is looking at. In doing so, they derive information about whether the individual or object eliciting the person’s attention and emotion may be harmful or pleasant. In the second year, emotion understanding expands as toddlers grasp that another’s emotional expressions can be understood as deriving from other mental states, such as the person’s desires and preferences. Toddlers increasingly recognize, in other words, that people feel happy when they get what they desire, sad when they do not, and disappointed when they get what they do not want. In the next two years, young children expand their repertoire of emotion understanding to connect different emotions with the prototypical situations in which they might be elicited, broadening their capacity to link emotions not only to the fulfilment or frustration of desires and preferences but also to goal achievement or failure, loss or deprivation, threat or danger, physical pain, and other conditions directly relevant to well-­ being (Thompson and Lagattuta  2006). Somewhat later, young children also understand how emotions are associated with what people are thinking, such as their recall or anticipation of emotionally evocative events. Stated simply, early emotion understanding expands significantly in the early years based on developing acuity in discerning the psy­cho­ logic­al causes of the emotions the child observes. Developing emotion understanding contributes to empathy, which can be observed in the ‘concerned attention’ of 1-­year-­olds to another’s distress, and which increases significantly in the years that follow (see Davidov et al. 2013 for a review). Early empathy is rarely accompanied by verbal or behavioural comforting at this age but it is accompanied by efforts to understand the causes of another’s distress, such as by inspecting the source of harm or asking simple q ­ uestions. When young children later talk about matters of right and wrong, their evalu­ations

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

34  Ross A. Thompson are often based on their observation of people’s feelings and comments about their causes, along with concerns over another’s welfare. These readily lead to attributions of responsibility to the person or event that caused harm to another (Wright and Bartsch 2008). Emotion understanding is accompanied by other achievements in understanding other people and their mental states (or ‘theory of mind’). By the end of the first year and increasingly in the second, for example, infants are becoming aware that people act according to their intentions and goals (Woodward 2009). When they observe someone opening the door of a cabinet, for example, they perceive that act not as an end in itself but as a means of reaching and using another object. By the middle of the second year, toddlers imitate another’s actions based on inferred intentions: they act to accomplish a goal (such as dropping objects into a cup) that they had observed an adult trying, but failing, to accomplish (Carpenter, Call, and Tomasello  2005). Like emotion understanding, recognizing another’s goals establishes a psychological connection to another person’s experience. As illustrated by the preceding example, young children begin to share and participate in accomplishing another’s goals, a developing capacity called shared intentionality (Tomasello et al.  2005). Shared intentionality can be observed in cooperative game-­playing or problem-­solving, mutual conversation, as well as in simple forms of helping and sharing (Newton et al. 2016). Shared intentionality contributes, in short, to social connection and understanding. Young children’s sensitivity to another’s goals and intentions may also influence how they respond to third-­party observations, such as when they witness one actor help or hinder a second actor from accomplishing her goals. Observed in such staged scenarios, children in their second year have been found to give resources to a victim of harm, deny resources from one who had impeded another’s goals, and provide rewards to one who had acted helpfully to another (Hamlin et al. 2011; Vaish, Carpenter, and Tomasello 2009). Although these responses cannot ameliorate the effects of the prior helping or hindering that the child observed, they often have the effect of adjusting outcomes based on whether an actor’s goals were facilitated or impeded. In one study, for example, 3-­year-­olds protested the behaviour of a puppet who destroyed the artwork of another puppet, sometimes making reference to the feelings of the victim (e.g., ‘She will be sad then’), and then they assisted the victim and tattled on the perpetrator (Vaish, Missana, and Tomasello 2011). Three-­year-­olds even deny resources to an adult with harmful intentions toward another person, even when the adult’s intention to harm fails to be achieved (Vaish, Carpenter, and Tomasello 2010). Intentions matter, even to young children. And in a study of distributional justice, children age 3½ allocated the rewards of two story characters’ shared activity to the one who had contributed the most (Baumard, Mascaro, and Chevallier 2012). Taken together, young children are developing an early understanding of people in which intentions and goals, emotions, desires, and needs occupy ­

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

Attachment Theory and Moral Development  35 prominent place, consistent with their early developing theory of mind. Further, they respond to the actions of others in terms of their effects on these mental states, and these appraisals value actions that help others accomplish their goals, reach what they desire, experience positive emotions, and not experience negative ones. These intuitive appraisals seem to cause children, in turn, to act in ways to benefit victims, reward helpers, punish harmdoers, and call out harmful behaviour in another. Other studies show that 5- and 6-­year-­old children spon­tan­eous­ly protest violations of distributional justice, such as when greater resources are provided to a puppet who is rich rather than one who is poor (Wörle and Paulus 2018), and that preschoolers appreciate that sharing has emotionally positive benefits for both benefactors and recipients (Paulus and Moore 2015, 2017). Beyond these, young children also sometimes exhibit empathy, especially to another in distress, and show helping and sharing as expressions of shared intentionality. These behaviours are constituents of a premoral sensibility (Thompson  2012, 2015b) that enlists the rapidly developing psychological understanding of early childhood into simple, nonegocentric appraisals of right and wrong based on the consequences of behaviour for another’s goals, needs, and emotions. It leads to intuitive appraisals of desirable and disapproved conduct, actions that are punitive, indignant, rehabilitative, or rewarding, and sometimes attributions of character (e.g., niceness or meanness) to actors (Wright and Bartsch 2008). It is also associated with simple expressions of guilt and shame over misconduct during this period, especially when misconduct has consequences for people (rather than property) (see review by Thompson, in press). Within a broad definition of moral conduct, it causes young children to act with moral agency before a true moral sense has developed. That is because the reason that these judgements are characterized as ‘premoral’ is because they are intuitive rather than reasoned, and are not yet connected to a broader system of values. That requires the further influence of their relationships with parents.

4.2  From premoral to moral The development of a premoral sensibility in early childhood prepares young children for moral socialization, because parental messages concerning right and wrong conduct often build on that sensibility. Experiences that a young child may not represent in morally relevant ways become re-­represented in the context of interaction with a parent by which generalizable values concerning good and bad conduct are connected to another’s feelings or welfare. Mothers justify the enforcement of moral rules with their 2- and 3-­year-­old children, for example, on the basis of people’s needs and welfare, putting a human face on matters of right and wrong. By contrast, they justify social conventional rules in terms of social

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

36  Ross A. Thompson order and regulation (Smetana 1989). Consequently, by age 3 or 4, young ­children regard moral violations as more serious, and moral rules as irrevocable (i.e. valid regardless of social guidelines), justifying their judgements in terms of unfairness and the harm to others entailed in moral violations (Nucci and Weber  1995; Smetana  1989). The interaction of young children’s intuitive sensibility with parental explanations like these helps to connect moral standards to human needs and emotions, and contributes to a more humanistic (rather than self-­focused) orientation to moral conduct than has traditionally been credited young children. By their more frequent reference to people’s feelings in conversations about approved or disapproved conduct—and less frequent references simply to rules and the consequences of breaking them—mothers in secure attachment relationships with their children more often connect moral conduct to people’s feelings and needs compared to mothers in insecure relationships (Laible and Thompson  2000). In the context of their gentler discipline practices, moreover, these mothers also help children conceptualize their conduct in terms of its consequences for others rather than just for oneself (Kochanska et al.  2004,  2010). Their more frequent use of morally evaluative language (e.g., ‘good boy’ and ‘that was not nice’) also embeds the appraisal of conduct in attributional terms that make personal responsibility explicit (Wright and Bartsch  2008). Stated differently, the connection between people’s feelings and needs, the child’s actions, and values concerning right and wrong, building on the child’s premoral sensibility, become connected and incorporated into internal working models of people, the self, and relationships through parent–child conversation in secure relationships. Parent–child conversations like these often occur with reference to the child’s past conduct, when issues of right and wrong can be discussed outside of the heat of the discipline encounter. But they also occur with reference to the behaviour of other people (such as siblings, peers, or other adults), and in a variety of forums that are morally relevant, such as when parents explain the motivations behind their own choices to act, or when enlisting the child into certain shared activity (such as volunteer efforts or community activities). As children mature, more­ over, these conversations incorporate more complex elements of moral socialization, such as beliefs about personal responsibility, self-­reliance and dependency, deservingness, and other kinds of judgements that can support or undermine responsible moral conduct. Lapsley and Narvaez (2004) have proposed that these conversations are thus a critical forum for the development of moral character because they contribute to children developing chronically accessible, easily primed and readily activated moral schemas that cause them to appraise everyday situations in morally relevant ways and which guide their conduct accordingly. It is both the readiness to perceive common situations as morally relevant and the values that are mentally evoked in these situations that are likely to be shaped by diverse forms of parent–child conversation throughout childhood and contribute to the rational reflection that is essential to moral choice and decision-­making in a manner consistent with the Aristotelian model.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

Attachment Theory and Moral Development  37 Parent–child conversation is influential in a relational context, however. There is also evidence that the conceptual catalysts of parent–child conversation are supported by cooperative reciprocity in secure attachment relationships. Kochanska (2002b) calls this reciprocity a ‘mutually responsive orientation’ that also fosters moral growth. Young children are accustomed to the one-­way as­sist­ ance of others who care for them, of course. But as they become increasingly expected to contribute to others’ needs (such as in household tasks or co­oper­ ation with family members), their motivation for cooperating derives, at least in part, from the responsiveness of parents to their own needs. In support of this view, several studies have confirmed the association of the positive mutual responsiveness of parents and toddlers in home observations with laboratory measures of children’s moral conduct several years later (Kochanska, Aksan, and Koenig 1995; Kochanska and Murray 2000). The parent’s responsiveness has these effects because, according to Kochanska, the development of a mutually responsive orientation sensitizes children to the reciprocal responsibilities of relationships. Over time, this disposition is likely to become integrated into the child’s conceptions (or working models) of how to behave in relationships, and it may later develop into a more generalized disposition toward other people. An attachment perspective to moral development proposes, therefore, that the rapid advances of early psychological development sensitize young children to the feelings, goals, and needs of other people and the effects of another’s action on those mental states, and that moral development proceeds in earnest as parents enlist this premoral sensibility into their moral socialization efforts. These efforts include conversation that conveys values concerning approved and disapproved conduct, and relational interactions that tutor young children in the mutual responsibilities of close relationships. When parent–child relationships are secure, the constructive orientation toward others that has developed from the internal working models of the early years becomes generalized to other people, and elem­ents of secure relationships—from emotion-­focused discourse when discussing misbehaviour to a mutually responsive relational orientation—help to support this process. Over time, as secure relationships provide important catalysts to personality development, these elements of secure relationships also contribute to the growth of moral character, and to the habitual dispositions and rational choices entailed in virtuous moral conduct.

5.  Concluding Comments There is much more to an attachment perspective to moral development than this short outline can provide, of course. But this outline may be sufficient to illustrate three things. First, there is considerable need for a new theory of moral development that incorporates current ideas concerning the importance of early experiences, the

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

38  Ross A. Thompson astonishing advances in psychological understanding of the early years, a broad appreciation for the relational influences that shape the early development of moral conduct and character, and the child’s construction of values. No current psychological theory of moral development accomplishes these aims, and major philosophical perspectives on morality and character also fall short. Second, elements of attachment theory offer promising avenues for addressing some of these challenges. In particular, the view that early experience with parents contributes to the development of mental representations of self and relationships that become generalized to the child’s interactions with others and to self-­understanding offers potentially powerful tools for reconceptualizing fundamental aspects of moral development. It is not just the association of secure attachment with developing moral awareness that is important, but also the processes highlighted by attachment theory: the significance of sensitive, responsive care for developing orientations toward others and the self, the child’s construction of moral awareness from the interaction of social experience and developing understanding of people, and the significance of emotion as a gateway to connecting with another’s experience. Finally, the renewal of interest in Aristotelian virtue ethics within philosophy offers a remarkable opportunity to establish cross-­ cutting connections with developmental science to advance each field. Attachment theory provides an empirical forum for considering the growth of personality and character, while virtue ethicists have much to say about what virtuous character is and does. Attachment theory focuses primarily (although not exclusively) on childhood, whereas virtue ethicists rarely devote significant attention to habituation of virtue during the childhood years (but see Swanton 2016, for an exception). Both adopt an orientation of naturalism to the development of admirable character, but of very different forms (Harcourt 2013). There is, in short, much to be gained from their collaboration.

References Ainsworth, M.  D.  S. 1973. ‘The development of infant–mother attachment’. In B.  Caldwell and H.  Ricciuti (eds), Review of Child Development Research, vol. 3, 1–94. Chicago: University of Chicago Press. Baumard, N., Mascaro, O., and Chevallier, C. 2012. ‘Preschoolers are able to take merit into account when distributing goods’. Developmental Psychology 48, 492–8. Bernier, A., Beauchamp, M. H., Carlson, S. M., and Lalonde, G. 2015. ‘A secure base from which to regulate: attachment security in toddlerhood as a predictor of executive functioning at school entry’. Developmental Psychology 51, 1177–89. Bloom, P. 2013. Just Babies: The Origins of Good and Evil. New York: Crown. Bowlby, J. 1969. Attachment and Loss, vol. 1: Attachment. New York: Basic Books.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

Attachment Theory and Moral Development  39 Calkins, S. E., and Leerkes, E. M. 2011. ‘Early attachment processes and the development of emotional self-regulation’. In K.  D.  Vohs and R.  F.  Baumeister (eds), Handbook of Self-Regulation (2nd edn), 355–73. New York: Guilford. Carpenter, M., Call, J., and Tomasello, M. 2005. ‘Twelve- and 18-month-olds copy actions in terms of goals’. Developmental Science 8(1), F13–20. Cooke, J.  E., Stuart-Parrigon, K.  L., Movahed-Abtahi, M., Koehn, A.  J., and Kerns, K.  A. 2016. ‘Children’s emotion understanding and mother–child attachment: a meta-analysis’. Emotion 16, 1102–6. Davidov, M., Zahn-Waxler, C., Roth-Hanania, R., and Knafo, A. 2013. ‘Concern for others in the first year of life: theory, evidence, and avenues for research’. Child Development Perspectives 7, 126–31. Goffin, K.  C., Boldt, L.  J., and Kochanska, G. 2018. ‘A secure base from which to ­co­oper­ate: security, child and parent willing stance, and adaptive and maladaptive outcomes in two longitudinal studies’. Journal of Abnormal Child Psychology 46, 1061–75. Goodvin, R., Meyer, S., Thompson, R. A., and Hayes, R. 2008. ‘Self-understanding in early childhood: associations with attachment security and maternal emotional risk’. Attachment & Human Development 16, 103–36. Groh, A.  M., Fearon, R.  P., Bakermans-Kranenburg, M.  J., van IJzendoorn, M.  H., Steele, R. D., and Roisman, G. I. 2014. ‘The significance of attachment security for children’s social competence with peers: a meta-analytic study’. Attachment & Human Development 16, 103–36. Gross, J. T., Stern, J. A., Brett, B. E., and Cassidy, J. 2017. ‘The multifaceted nature of prosocial behavior in children: links with attachment theory and research’. Social Development 26(4), 661–78. Haidt, J. 2012. The Righteous Mind. New York: Vintage. Hamlin, J. K., Wynn, K., Bloom, P., and Mahajan, N. 2011. ‘How infants and toddlers react to antisocial others’. Proceedings of the National Academy of Sciences 108, 11931–6. Harcourt, E. 2013. ‘Attachment theory, character and naturalism’. In J.  Peters (ed.), Aristotelian Ethics in Contemporary Perspective, 114–29. London: Routledge. Kim, S., and Kochanska, G. 2017. ‘Relational antecedents and social implications of the emotion of empathy: evidence from three studies’. Emotion 17, 981–92. Kochanska, G. 1995. ‘Children’s temperament, mothers’ discipline, and security of attachment: multiple pathways to emerging internalization’. Child Development 66, 597–615. Kochanska, G. 2002a. ‘Committed compliance, moral self, and internalization: a mediational model’. Developmental Psychology 38, 339–51. Kochanska, G. 2002b. ‘Mutually responsive orientation between mothers and their young children: a context for the early development of conscience’. Current Directions in Psychological Science 11, 191–5. Kochanska, G., Aksan, N., and Carlson, J. J. 2005. ‘Temperament, relationships, and young children’s receptive cooperation with their parents’. Developmental Psychology 41, 648–60.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

40  Ross A. Thompson Kochanska, G., Aksan, N., Knaack, A., and Rhines, H. M. 2004. ‘Maternal parenting and children’s conscience: early security as moderator’. Child Development 75, 1229–42. Kochanska, G., Aksan, N., and Koenig, A. L. 1995. ‘A longitudinal study of the roots of preschoolers’ conscience: committed compliance and emerging internalization’. Child Development 66, 1752–69. Kochanska, G., Barry, R. A., Stellern, S. A., and O’Bleness, J. J. 2009. ‘Early attachment organization moderates the parent–child mutually coercive pathway to children’s antisocial conduct’. Child Development 80, 1288–300. Kochanska, G., Boldt, L. J., and Goffin, K. 2019. ‘Early relational experience: a foundation for the unfolding dynamics of parent–child socialization’. Child Development Perspectives 13, 41–7. Kochanska, G., Forman, D. R., Aksan, N., and Dunbar, S. B. 2005. ‘Pathways to conscience: early mother–child mutually responsive orientation and children’s moral emotion, conduct, and cognition’. Journal of Child Psychology and Psychiatry 46, 19–34. Kochanska, G., Forman, D. R., and Coy, K. C. 1999. ‘Implications of the mother–child relationships in infancy for socialization in the second year of life’. Infant Behavior & Development 22, 249–65. Kochanska, G., and Murray, K. T. 2000. ‘Mother–child mutually responsive orientation and conscience development: from toddler to early school age’. Child Development 71, 417–31. Kochanska, G., Woodard, J., Kim, S., Koenig, J. L., Yoon, J. E., and Barry, R. A. 2010. ‘Positive socialization mechanisms in secure and insecure parent–child dyads: two longitudinal studies’. Journal of Child Psychology and Psychiatry 51, 998–1009. Kohlberg, L. 1969. ‘Stage and sequence: the cognitive developmental approach to socialization’. In D.  Goslin (ed.), Handbook of Socialization Theory and Research, 347–480. New York: Rand McNally. Laible, D. 2004. ‘Mother–child discourse in two contexts: links with child temperament, attachment security, and socioemotional competence’. Developmental Psychology 40, 979–92. Laible, D., Panfile, T., and Makariev, D. 2008. ‘The quality and frequency of mother– toddler conflict: links with attachment and temperament’. Child Development 79, 426–43. Laible, D. L., and Thompson, R. A. 2000. ‘Mother–child discourse, attachment se­cur­ ity, shared positive affect, and early conscience development’. Child Development 71, 1424–40. Lapsley, D.  K., & Narvaez, D. 2004. ‘A Social-Cognitive Approach to the Moral Personality’. In D. K. Lapsley & D. Narvaez (eds.), Moral development, self, and identity, 189–212. New Jersey: Erlbaum. Lindsey, E. W., Cremeens, P. R., and Caldera, Y. M. 2010. ‘Mother–child and father–child mutuality in two contexts: consequences for young children’s peer relationships’. Infant and Child Development 19, 142–60.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

Attachment Theory and Moral Development  41 Murphy, T. P., and Laible, D. J. 2013. ‘The influence of attachment security on preschool children’s empathic concern’. International Journal of Behavioral Development 37, 436–40. Newton, E. K., Thompson, R. A., and Goodman, M. 2016. ‘Individual differences in toddlers’ prosociality: experiences in early relationships explain variability in prosocial behavior’. Child Development 87, 1715–26. Nucci, L., and Weber, E. K. 1995. ‘Social interactions in the home and the development of young children’s conceptions of the personal’. Child Development 66, 1438–52. Ontai, L.  L., and Thompson, R.  A. 2002. ‘Patterns of attachment and maternal discourse effects on children’s emotion understanding from 3- to 5-years of age’. Social Development 11, 433–50. Paulus, M., and Moore, C. 2015. ‘Preschool children’s anticipation of recipients’ emotions affects their resource allocation’. Social Development 24, 852–67. Paulus, M., and Moore, C. 2017. ‘Preschoolers’ generosity increases with understanding of the affective benefits of sharing’. Developmental Science 20, el2417. Piaget, J. 1932/1965. Moral Judgment of the Child. New York: Free Press. Raikes, H. A., and Thompson, R. A. 2008b. ‘Conversations about emotion in high-risk dyads’. Attachment & Human Development 10, 359–77. Raikes, H. A., and Thompson, R. A. 2008a. ‘Attachment security and parenting quality predict children’s problem-solving, attributions, and loneliness with peers’. Attachment & Human Development 10, 1–26. Raikes, H.  A., Virmani, E.  A., Thompson, R.  A., and Hatton, H. 2013. ‘Declines in peer conflict from preschool through first grade: influences from early attachment and social information processing’. Attachment & Human Development 15, 65–82. Shaver, P. R., Mikulincer, M., Gross, J. T., Stern, J. A., and Cassidy, J. 2016. ‘A lifespan perspective on attachment and care for others: empathy, altruism, and prosocial behavior’. In J. Cassidy and P. R. Shaver (eds), Handbook of Attachment (3rd edn), 878–916. New York: Guilford. Smetana, J. G. 1989. ‘Toddlers’ social interactions in the context of moral and conventional transgressions in the home’. Developmental Psychology 25, 499–508. Spinrad, T. L., and Stifter, C. A. 2006. ‘Toddlers’ empathy-related responding to distress: predictions from negative emotionality and maternal behavior in infancy’. Infancy 10, 97–121. Swanton, C. 2016. ‘Developmental virtue ethics’. In N.  Snow (ed.), Developing the Virtues: Integrating Perspectives, 116–34. New York: Oxford University Press. Thompson, R. A. 2012. ‘Wither the preconventional child? Toward a life-span moral development theory’. Child Development Perspectives 6, 423–9. Thompson, R.  A. 2013. ‘Attachment theory and research: precis and prospect’. In P.  Zelazo (ed.), Oxford Handbook of Developmental Psychology, vol. 2: Self and Others, 191–216. New York: Oxford University Press.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

42  Ross A. Thompson Thompson, R.  A. 2015a. ‘Relationships, regulation, and early development’. In R.  M.  Lerner (ed.), Handbook of Child Psychology and Developmental Science (7th edn), vol. 3: Social and Emotional Development. Ed. M. E. Lamb and C. Garcia Coll, 201–46. New York: Wiley. Thompson, R. A. 2015b. ‘The development of virtue: a perspective from developmental psychology’. In N.  E.  Snow (ed.), Cultivating Virtue: Perspectives from Philosophy, Theology, and Psychology, 279–306. New York: Oxford University Press. Thompson, R.  A. 2017. ‘Twenty-first century attachment theory’. In H.  Keller and K. A. Bard (eds), The Cultural Nature of Attachment: Contextualizing Relationships and Development. Strüngmann Forum Reports vol. 22, 301–19. Cambridge, MA: MIT Press. Thompson, R. A. 2020. ‘The development of moral self-awareness’. In L. Jensen (ed.), The Oxford Handbook of Moral Development: An Interdisciplinary Perspective. New York: Oxford University Press. Thompson, R. A., and Lagattuta, K. 2006. ‘Feeling and understanding: early emotional development’. In K.  McCartney and D.  Phillips (eds), The Blackwell Handbook of Early Childhood Development, 317–37. Oxford: Blackwell. Thompson, R.  A., and Lavine, A.  S. 2016. ‘The development of virtuous character: automatic and reflective dispositions’. In J.  Annas, D.  Narvaez, and N.  E.  Snow (eds), Developing Virtue: Integrating Perspectives, 95–115. New York: Oxford University Press. Thompson, R. A., and Newton, E. K. 2013. ‘Baby altruists? Examining the complexity of prosocial motivation in young children’. Infancy 18, 120–33. Thompson, R. A., Simpson, J. A., and Berlin, L. J. 2021. Attachment: The Fundamental Questions. New York: Guilford. Thompson, R.  A., Virmani, E., Waters, S.  F., Meyer, S., and Raikes, A. 2013. ‘The development of emotion self-regulation: the whole and the sum of the parts’. In K.  Barrett, N.  A.  Fox, G.  A.  Morgan, D.  J.  Fidler, and L.  A.  Daunhauer (eds), Handbook of Self-Regulatory Processes in Development, 5–26. New York: Taylor and Francis. Tomasello, M., Carpenter, M., Call, J., Behne, T., and Moll, H. 2005. Understanding and sharing intentions: the origins of cultural cognition. Behavioral and Brain Sciences 28, 675–735. Vaish, A., Carpenter, M., and Tomasello, M. 2009. ‘Sympathy through affective perspective taking and its relation to prosocial behavior in toddlers’. Developmental Psychology 45, 534–43. Vaish, A., Carpenter, M., and Tomasello, M. 2010. ‘Young children selectively avoid helping people with harmful intentions’. Child Development 81, 1661–9. Vaish, A., Missana, M., and Tomasello, M. 2011. ‘Three-year-old children intervene in third-party moral transgressions’. British Journal of Developmental Psychology 29, 124–30.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

Attachment Theory and Moral Development  43 Van der Mark, I. L., Van IJzendoorn, M. H., and Bakermans-Kranenburg, M. J. 2002. ‘Development of empathy in girls during the second year of life: associations with parenting, attachment and temperament’. Soc. Dev. 11(4), 451–68. von der Lippe, A., Eilertsen, D.  E., Hartmann, E., and Killen, K. 2010. ‘The role of maternal attachment in children’s attachment and cognitive executive functioning: a preliminary study’. Attachment & Human Development 12, 429–44. Waters, S., Virmani, E., Thompson, R. A., Meyer, S., Raikes, A., and Jochem, R. 2010. ‘Emotion regulation and attachment: unpacking two constructs and their as­so­ci­ ation’. Journal of Psychopathology and Behavioral Assessment 32, 37–47. Woodward, A. 2009. ‘Infants’ grasp of others’ intentions’. Current Directions in Psychological Science 18, 53–7. Wörle, M., and Paulus, M. 2018. ‘Normative expectations about fairness: the development of a charity norm in preschoolers’. Journal of Experimental Child Psychology 165, 66–84. Wright, J.  C., and Bartsch, K. 2008. ‘Portraits of early moral sensibility in two children’s everyday conversations’. Merrill-Palmer Quarterly 54, 56–85. Zimmer-Gembeck. M. J., Webb, H. J., Pepping, C. A., Swan, K., Merlo, O., Skinner, E. A., Avdagic, E., and Dunbar, M. 2017. ‘Review: is parent–child attachment a correlate of children’s emotion regulation and coping?’ International Journal of Behavioral Development 41, 75–93.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

2

Dimensions, Determinants, and Development of Prosocial Behaviour A Differential Susceptibility Hypothesis on Attachment and Moral Character Marian J. Bakermans-­Kranenburg and Marinus H. van IJzendoorn

1.  Introduction The idea that individuals with secure attachment relationships would be more inclined to behave prosocially than insecure individuals is attractive and makes intuitive sense: attachment security is related to better emotion regulation and would thus enable individuals to deal more adequately with another person’s distress and needs.1 Moreover, since a secure attachment relationship is the predicted outcome of a childhood history of sensitive caregiving experiences, securely attached individuals are likely to have experienced models who were responsive to others’ needs and distress. This may give secure children and adults a head-­start in responding to others in a sensitive way. In the current chapter we will review the evidence on this assumed link between attachment and prosocial behaviour. We first define prosocial behaviour. We describe various ways to assess prosocial behaviours and review how prosociality develops over time. Because empirical research has failed to demonstrate even medium levels of stability of prosocial behaviour across time and situations, we tentatively conclude that the construct of prosociality as an enduring character or personality trait is problematic. It does not follow, however, that there is no enduring trait that has been handed down over generations. Our conclusion is that this trait may be the susceptibility to environmental factors, particularly influences from conspecifics, that makes us act more or less prosocially depending on the nature of the context. Identifying determinants or predictors of prosocial behaviour remains essential, even when these determinants are environmental 1  We thank Edward Harcourt for the thoughtful and enjoyable discussions that inspired the conclusion of this chapter. Marian J. Bakermans-­Kranenburg and Marinus H. van Ijzendoorn, Dimensions, Determinants, and Development of Prosocial Behaviour: A Differential Susceptibility Hypothesis on Attachment and Moral Character In: Attachment and Character: Attachment Theory, Ethics, and the Developmental Psychology of Vice and Virtue. Edited by: Edward Harcourt, Oxford University Press. © Marian J. Bakermans-­Kranenburg and Marinus H. van Ijzendoorn 2021. DOI: 10.1093/oso/9780192898128.003.0003

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

Dimensions, Determinants, and Development of Prosocial  45 rather than intra-­individual. Once they have been revealed, programmes to stimulate prosociality can be developed.

2.  What is Prosocial Behaviour? Prosocial behaviour is defined as ‘any action that serves to benefit another person’ (The Oxford Handbook of Prosocial Behaviour). Human beings have the tendency to help others in need and to take care of unfortunate conspecifics. Situations that come to mind in daily life include helping elderly people cross the street, donating blood or money, voluntarily taking over tasks of a colleague who is late or ill, and taking care of the sick. Apparently, these prosocial behaviours belong to our innate repertoire. That is not to say that we understand the conditions in which such behaviours are present or absent, nor that we can predict who will act when to whom in a prosocial way, and what makes some individuals behave more prosocially than others. In fact, it is an unanswered question whether interindividual differences in prosociality, which would imply stability of prosocial behaviour across time and situations, do exist. Morally good character traits, or even distinct moral personalities, may populate theories more than reality, or so it can seem in empirical research (Merritt, Doris, and Harman 2010). In research contexts four broad dimensions of prosocial behaviour have been studied: helping, volunteerism, cooperation, and caregiving. Helping behaviour can be observed in children from a young age, although it requires cognitive ­competence and emotional empathy to help someone. The helper must understand the other’s problem and also possess the altruistic motivation to act on behalf of the other (Warneken and Tomasello 2007). Children as young as 12 months do show concern for others who are in distress and can show comforting behaviour (Eisenberg and Fabes 1998), though they may offer the comfort they would prefer for themselves (e.g. bringing their cuddly toy). Volunteerism includes self-­reported participation in non-­paid activities that benefit the society, but also includes donating time or money. Rushton and Wheelwright (1980) val­id­ated donating behaviour of 6- to 10-­year-­olds who were asked to donate tokens to a charity. Donations were related to teacher-­reported differences in altruism, and with ­children’s willingness to share scarce resources with their friends. Others showed that donating in the lab was related to spontaneous sharing of scarce resources or materials with peers in the preschool classroom (Eisenberg et al. 1999). It should be noted that in studies on prosocial behaviour donating money is measurable and thus preferred over self-­report of donating that carries the risk of biases ­presenting the reporter in a more favourable light. Cooperation constitutes a somewhat less evident class of prosocial behaviour, and we argue that in fact it does not represent prosocial behaviour. Cooperation, also referred to as reciprocal altruism, involves a win-­win situation; a wished-­for

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

46  M. J. Bakermans-Kranenburg and M. H. van IJzendoorn goal can be reached by joining forces. Cooperation can be found in all biological systems: natural selection has been argued to favour cooperation if the benefit of the altruistic act divided by the cost exceeds the average number of neighbours (Ohtsuki et al. 2006). Cooperation may benefit another person, but this is not the direct aim, and therefore fails to comply with the definition given above if we assume that the term ‘serve’ implies that the aim of the prosocial action is to bene­fit another person. Lastly, it is questionable whether caregiving shown to kin should be considered as prosocial behaviour or is primarily an investment in the survival of one’s own genes. Responding to one’s own crying infant, feeding one’s own but also a sister’s offspring contributes to one’s inclusive fitness (Trivers 1974), the chance that part of one’s gene pool will survive and make it to reproduction. Altruistic, ‘prosocial’ acts are much more common among parents vis-­à-­vis their children (and kin in general) than in non-­kin relationships. For this reason we consider parental sensitivity as paving the way for children’s prosocial development, and the evolutionary cradle of prosociality (see below), but not as prosocial behaviour in and of itself. At the same time, and following from this line of reason­ing, caregiving responses to unknown children do qualify as prosocial behaviour. Thus, rather than considering actions as prosocial when they serve to benefit another person, we advance that the litmus test of prosocial behaviour is its serving to benefit an unfamiliar person. This is a critical difference that we deem essential for thinking about the dimensions, determinants, and development of prosocial behaviour. As evident from the examples above, prosocial behaviour can be assessed with a variety of observations and measures, and results are usually not highly convergent. This already throws some doubt on the idea of a unitary prosocial personality. A first important distinction concerns the difference between prosocial behaviour and prosocial intentions or reasoning. In our own research, we have mainly focused on prosocial behaviour, arguing that helpful intentions and benevolent reasoning do not benefit another person as long as they are not accompanied by prosocial behaviour. Moreover, measures of intentions and reason­ ing may be more susceptible to social desirability than behavioural assessments. It can be argued—and it has been argued indeed—that prosocial behaviour may result from non-­social motives, varying from feeling good to the idea that one is positively evaluated by others or that one may receive a favour in return in the future (tit for tat). Rather ingenious studies have been conducted to distinguish between the wish to be recognized and the wish to care for others’ well-­ being. In one study, 2-­year-­old children’s internal arousal was measured when they were exposed to an adult who accidentally dropped an object. Higher arousal was associated with shorter delays in helping the adult. Importantly, their arousal decreased both when they themselves returned the object and when someone else did so, while it remained high when the adult was not helped at all (Hepach,

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

Dimensions, Determinants, and Development of Prosocial  47 Vaish, and Tomasello 2012). This suggests that the children’s concern was the adult in need, and not their own performance. Helping may also be driven by an underlying non-­social motive to see things restored, independent of the need of another person. Addressing this issue, 2-­year-­olds have been presented with an adult who was either making a drawing or hanging up pieces of cloth and dropped both cloth pins and crayons (Hepach et al. 2016). Children were sensitive to the adult’s needs: when the adult was drawing a picture, the first object they picked up was more often a crayon. In other words, they preferentially picked up objects that the adult needed, and they were not merely interested in restoring a situation without dropped objects. Similarly, in the Prosocial Cyberball game, a virtual ball tossing game in which participants can act prosocially by tossing more balls to an excluded co-­player (see below), participants’ responses appeared not dependent of the position of the excluded player (Wesselmann, Williams, and Hales 2013) countering preference for symmetry as an alternative explanation. In sum, although motivations for helping, donating, and other prosocial behaviours may be diverse and may not always be easily discernible, neither for the observer, nor for the actor, we argue that the motives are less important than the acts: the world would be a better place if its population always acted in a prosocial way—even if it were for selfish motives. Having the right feelings, the right character traits, or correct moral reasons disconnected from actual behaviour is irrelevant for situations in which moral choices matter, such as helping victims of genocidal regimes to flee or to hide, or discontinuing torturous treatment of prisoners when told to continue (Zimbardo 2007). The study of moral and prosocial development should focus on behaviour, not merely on ethically or morally sound but abstract reasoning. This is why our review will discuss pro­ social development from a behavioural perspective, and predominantly focus on studies with behavioural measures.

2.1  Evolutionary explanations of prosociality Prosocial and altruistic behaviours have been speculated to be evolutionary-­ based universal competences of human beings, and of many avian and mammalian species as well. They may have evolved from reciprocal altruism, as in-­group solidarity created elevated chances of survival in deadly struggles with out-­groups and predators (Darwin  1871/1981), and in tandem with capacities for social evaluation: humans are selective cooperators who readily assess others’ co­opera­ tive potential and choose social partners accordingly, avoiding or punishing non-­ cooperators (Hamlin  2014). From these competences the actual altruistic performance might follow, as well as individual and situational differences in altruistic behaviour.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

48  M. J. Bakermans-Kranenburg and M. H. van IJzendoorn De Waal (2008) argued that the basis for empathic concern and prosocial may have developed in the context of parental care long before the human species evolved. He refers to the evolutionary-­based attachment behaviour of offspring, and the better chances of survival and reproduction of those offspring who had parents who were sensitive to their needs. The neural circuits related to affiliative and caring behaviours are highly conserved in mammals: ancient subcortical circuits (brainstem, amygdala, hypothalamus, and basal ganglia) are involved in the perception and response to the distress of others (Tucker, Luu, and Derryberry 2005; Rilling and Mascaro  2017). The prolonged dependence of offspring in humans and primates made it essential for their mothers to detect signs of suffering and distress, and mothers who were good at detecting and responding to such signals ended up with more surviving offspring. Thus, over time a communication system developed in which children’s distress signals triggered maternal care (Decety et al. 2016; Witteman et al. 2019). Hrdy (2009), extending this model, noted that it takes 13 million calories to rear a human baby to nutritional independence. This is far more than a foraging mother could provide on her own, and help from her group members, so-­called alloparents, would be necessary. Furthermore, mortality of prospective mothers during labour and delivery was high as well, and exclusive dependence of the newborn on the biological mother would imply elevated risk of waste of inclusive fitness for kin and in-­group members. Cooperative breeding was thus essential for survival in an environment with scarce resources and high parent and child mortality, and selection would favour those individuals who contributed to co­opera­tive breeding. Prosocial behaviour, empathic concern, and sharing of resources would be rooted in the basic need for cooperative child rearing (see also De Waal 2008). Recent studies on humans and other primates show converging evidence that humans and chimpanzees share the motivation and skills necessary to help others (Melis 2011). Humans, however, show prosocial behaviours more readily and in a wider range of contexts (Warneken and Tomasello  2006a), and in particular prosociality in the context of sharing food seems to be easier for humans than for chimpanzees (Gilby  2006; Jensen et al.  2006; Silk et al.  2005), perhaps because chimpanzees are extremely competitive about food—which may be the evolutionary result of scarcity and the uncertainty of food supply in the future. For a fair comparison, we may think of how willing humans would be to share their brand new car with a stranger.

2.2  Development of prosocial behaviour In the first few years after birth, prosocial behaviour has been suggested to increase with age (for a meta-­analysis, see Eisenberg et al. 1998). Such age-­related

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

Dimensions, Determinants, and Development of Prosocial  49 changes in prosocial behaviour may result from children’s advances in two domains, their cognitive empathy or theory of mind (i.e. the ability to represent others’ needs and mental states), and emotional empathy (the capacity to share others’ emotions and feelings). Although one-­day-­old infants who are exposed to a newborn cry, a synthetic cry, and silence do cry most when they hear the cry of their conspecific (Sagi and Hoffman 1976), this clearly does not qualify for pro­ social behaviour. It has even be questioned whether it provides evidence for inborn empathic competence or merely reflects mimicry or contagious crying. Sympathetic or contagious distress nevertheless might constitute the basis for the emergence of real empathic concern. However, children do help other persons solve their problems from a very young age, even when the other is an unrelated stranger and when they are not rewarded for their help. Warneken and Tomasello (2006a) showed that 18-­month-­old children more often than not helped out in situations where a male experimenter was having trouble achieving a goal, for example reaching an out-­ of-­reach object that he accidentally dropped on the floor. Moreover, the toddlers more often picked up the object in this situation than in the control situation with the experimenter intentionally throwing the object on the floor. Apparently, the toddlers had an idea about when their action would be helpful and when not, and preferred to be helpful when needed. Three chimpanzees that were tested with the same experiments at age 36, 54, and 54 months, respectively, did offer their help in out-­of-­reach situations, but failed to do so in some other situations that did trigger prosociality in human children, such as removing obstacles. This pattern of results may suggest that both infants and chimpanzees were willing to help, but that some scenarios were maybe too difficult to understand for the chimpanzees. Selecting the six tasks in which 18-­month-­old toddlers showed robust helping behaviour, Warneken and Tomasello (2007) then administered these tasks to 14-­month-­old infants. In the out-­of-­reach tasks, they were helpful in on average 39 per cent of the situations; in other tasks, including removing obstacles, they showed hardly any helping behaviour—similar to the chimpanzees’ behaviour in these tasks (Warneken and Tomasello  2006), and perhaps indicating that they lacked the cognitive capacities to understand the experimenter’s needs in these somewhat more complicated contexts. If the 14-­month-­olds helped, they did so very soon, before any verbal hint was given, even before the experimenter looked at them. Some studies included even younger children, but these focused on young infants’ preferences for helping versus non-­helping figures, and not helping or otherwise prosocial behaviour itself. In the helper–hinderer paradigm, for ex­ample, children are shown an animation of a character (the ‘climber’) who is at the bottom of a hill and repeatedly attempts to climb the hill. On the third attempt he is either aided up by a helper pushing from behind or pushed down by a hinderer (Hamlin, Wynn, and Bloom 2007). When offered the choice between the

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

50  M. J. Bakermans-Kranenburg and M. H. van IJzendoorn wooden forms of the helper and the hinderer, both 6- and 10-­month-­olds more often reached for the helper, showing that they do understand the need of the climbing character and prefer helping behaviour. In an extension of the study, Hamlin and Wynn (2011) replicated the preference for a helper compared to a hinderer in somewhat younger infants with different stimuli. One of the paradigms consisted of a hand puppet trying to open the lid of a transparent plastic box with an attractive toy inside, and was aided by a prosocial ‘opener’ or hindered by an antisocial ‘closer’. Again, clear-­cut evidence for the prosocial preference was documented. However, in two independent exact replications of the study with the box paradigm, on the basis of detailed guidelines by the original authors, Salvadori et al. (2015) failed to find this preference, so the verdict on the early emergence of prosocial preference is still out. A very optimistic perspective on young children’s empathic prosocial behaviour in daily life is not warranted given the findings of Lamb and Zakhireh (1997), who observed 45 toddlers (18 months of age) in the natural setting of a day-­care centre. Twenty hours of video recordings presented 345 distress incidents and the responses of other children in the group. Counting only unambiguous instances of prosocial responses to distress (e.g. offering a toy or patting a crying child), and not simple approaches or concerned expressions, only eleven incidents were followed by a prosocial action from one of the peers. Whether or not a child ever responded prosocially was unrelated to age or gender. These observations do, of course, not deny that children of this age can be empathic, they only show that in the day-­care setting children display mostly indifferent ‘bystander’ behaviour. The children may have prosocial competence, but their prosocial performance does not emerge in this context. In our own studies on comforting, we saw from 18 to 24 months an increase of children’s empathic concern and prosocial comforting toward their mothers who pretended to have hurt themselves, but then a decrease from 24 to 89 months. Empathic concern toward the experimenter decreased from 18 to 24 to 89 months (Van der Mark, Van IJzendoorn, and Bakermans-­ Kranenburg 2002; Pannebakker 2007). Thus, contrary to the suggested developmental model, growing older did not necessarily mean becoming more empathic. Those who are nevertheless attached to the idea that children become more prosocial with age should be seriously troubled by the results of a replication of Milgram’s original test of obedience with children. Shanab and Yahya (1977) tested 192 Jordanian children (6–8, 10–12, and 14–16 years). Half of them received instructions to observe peers in a learning task, and to administer electric shocks each time they made a mistake, increasing the shock level with each additional mistake. The other half of the children were given a free choice of delivering or not delivering a shock each time the peer made a mistake. Pressed by the experimenter 73 per cent of the children continued to deliver shocks to  maximum voltage (only 16 per cent of the children with a free choice did so,  demonstrating the power of pressure). Neither age nor sex differences in obedi­ence rate were found. Moreover, the emotional responses (nervous laughter,

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

Dimensions, Determinants, and Development of Prosocial  51 lip biting, trembling) of the children in the Jordanian experiment were rated. Emotional responses significantly decreased with age. Of those who expressed intense tension, 44 per cent were in the youngest age group, 25 per cent in the middle group, and 16 per cent in the oldest age group. For donating, age-­related increases have been observed from 5 to 12 years when the currency was chocolate coins (Abramson, Danial, and Knafo-­Noam 2018), although the rate of donating overall remained lower than when they could give coins away that they could not have for themselves anyway. We were somewhat disappointed that 7-­year-­olds in our own studies, who had received money in return for their cooperation in our lab experiments, hardly donated any of the money to a charity (UNICEF). One explanation may be that donating to a charity is not linearly related to age. In their study of children aged 3–16 years, Grunberg et al. (1985) found that donating dropped around 7 years of age, perhaps because children of that age have become more aware of the importance of individual ownership. An alternative explanation may be the presence or absence of an experimenter, which makes an important difference: it may be especially spon­ tan­eous, unobserved donating that develops slowly. The mere presence or subtle hint of an experimenter may increase the willingness to share or donate substantially, as it did in our own studies (Van IJzendoorn et al. 2010). This should be taken into account when comparing donations across studies.

3.  Determinants of prosocial behaviour 3.1  Genes for generosity? An evolutionary basis for altruism does not imply that individual differences in prosocial behaviour are associated with differences in genetic make-­ up. To address the role of genetic factors in the explanation of individual differences in prosocial behaviour two approaches are available. The first approach aims to quantify the proportion of the variance in the phenotype under study that can be explained by genetic factors. For decades, behavioural genetic twin studies have played an important role in this approach; more recently, Genome-­Wide Complex Trait Analysis (GCTA) has emerged as informative with regard to the role of single nucleotide polymorphism (SNP) heritability. The second approach aims to identify the specific genes or SNPs involved in explaining differences in prosocial behaviour. For virtually all complex behavioural or psychiatric phenotypes there is a large gap between the (quantitative) estimate of the role of genetic factors and the identification of the actual genetic factors that play this role, and this is not different for studies on the genetics of prosocial behaviour. Studies with identical (or monozygotic) and fraternal (or dizygotic) twins on altruism have been conducted since the 1980s. Simply put, if correlations of concern for monozygotic twins (who are 100 per cent genetically similar) are

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

52  M. J. Bakermans-Kranenburg and M. H. van IJzendoorn substantially higher than correlations for dizygotic twins (on average 50 per cent genetically similar), there is evidence for a genetic basis of that behaviour. One of the first twin studies on altruism included 573 adult twin pairs who reported on the frequency with which they had engaged in 20 specific behaviours such as ‘I have donated blood’ (Rushton et al. 1986). More than half of the variance in altruism was accounted for by genetic factors, while only 2 per cent of the variance was due to common environmental factors that make individuals within a family similar to each other. An early study of observed empathic concern included twins at 14 and 20 months of age (Zahn-­Waxler, Robinson, and Emde 1992). Both mother and experimenter simulated distress according to specified scripts, such as closing a finger in the suitcase containing testing materials, and hurting a knee when getting up from the floor. They expressed pain vocalizations to attract the toddlers’ attention and assumed a pained facial expression over half a minute with a gradual subsiding of the distress for the rest of the minute, while direct eye contact was avoided. Empathic concern increased from 14 to 20 months of the infant’s age, and at both ages showed evidence of heritability. In the same age group, however, Volbrecht et al. (2007) observed more than 200 twin pairs and found no genetic influence for helping behaviour and empathic concern to distress. Both shared environmental influences, that make two children in a family similar to each other, and unique environmental factors (including measurement error) that make two children in a family different from each other explained the vari­ ation in these prosocial behaviours. In an even bigger study, Knafo et al. (2008) included more than 400 twin pairs longitudinally from 14 to 36 months of age. At 14 months, no genetic effect was found on a composite measure of empathy that was based on mother and experimenter simulations. By 24 months, a genetic effect emerged, accounting for about a quarter of the variance. The same genetic effect was found at 36 months. With age, genetic effects on prosocial behaviour and empathic concern appear to increase, and shared environmental effects decrease. In our own twin study 7-­year-­old twins received 10 coins of €0.20 for their cooperation after an hour of performing various sometimes boring tasks (Van IJzendoorn et al.  2010). They were then left alone and shown a two-­minute UNICEF promotional film of a child in a resource-­limited country. At the end of the fragment a voice-­over invited to donate money; a money box had been positioned next to the video screen. The money box was filled with several coins in order to enhance credibility. After one minute the experimenter came back into the room, and asked in a standardized way if the child would want to donate any money, and the children had another opportunity for donations. The first and somewhat disheartening observation was that the percentage of children who donated spontaneously after the promotional film was too small for genetic ­modelling. For donations after the probe (the experimenter’s question) 45 per cent

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

Dimensions, Determinants, and Development of Prosocial  53 of the variance was explained by shared environmental influences and 55 per cent by unique environmental factors. No heritability effects were found. An alternative approach to the quantification of heritability uses the genetic similarity among unrelated individuals as a predictor of their phenotypic similarity. When genome-­wide SNP data are available for a group of individuals, these can be used to obtain a measure of genetic similarity between all possible pairs. In a second step, this genetic relatedness is used as a predictor of the individuals’ similarity in prosocial behaviour. This is not fundamentally different from comparing monozygotic and dizygotic twin pairs who are divergent in genetic re­lated­ ness, but the variance in genetic similarity is measured (in contrast to twin studies, where the genetic similarity of dizygotic twins is simply assumed to be 50 per cent) and much larger. This also means that compensation for the large number of tests is necessary, and thus huge samples are a prerequisite. For the detection of SNP heritability estimates of 30 per cent with sufficient statistical power a sample size of at least 3,000 individuals is required (Visscher et al.  2014). In a sample of just over 3,000 children we found an SNP heritability for parent-­ reported and self-­reported prosocial behaviour in middle childhood of 21 per cent (Van IJzendoorn and Bakermans-­Kranenburg 2018). This is a finding in need of replication, but, if replicated, it implies that the heritability of prosocial behaviour is less pronounced than that of, for example, attention-­related or externalizing problems (Pappa et al. 2015). The search for specific genes and SNPs that are responsible for the genetic influence on prosocial behaviour has not been very successful. This is not specific to prosocial behaviour, indeed for virtually all complex traits the association with identified genes and pathways seems to be elusive. The oxytocin receptor gene has been suggested as a candidate gene for prosociality, and in one of the first studies it showed an association with parental sensitive caregiving, but the association between OXTR and (pro-)social behaviour did not survive a meta-­analytic combination of studies, and the early result must be considered an example of the so-­ called winner’s curse (Bakermans-­Kranenburg and Van IJzendoorn 2013); that is, the effect sizes of the promising pioneering studies are not replicated in a series of subsequent studies. Neither have other systematic associations of genes or genetic pathways with prosocial behaviour been documented, leading to the conclusion that there may be some genetic factors associated not only with the development of prosocial behaviour but also with individual differences in prosocial behaviour, but the identification of these factors is still in the future.

3.2  Benevolent brains? The brain is well equipped for prosociality: ancient subcortical circuits (brainstem, amygdala, hypothalamus, and basal ganglia) are involved in the perception and

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

54  M. J. Bakermans-Kranenburg and M. H. van IJzendoorn response to the distress of others (Decety et al. 2016). The exact neurobiological correlates of prosocial behaviour, however, have not been studied extensively. Two types of imaging studies are pertinent to the association between brain and prosocial behaviour: functional imaging studies, which are informative on the activity in brain areas during a specific task, and structural imaging studies providing information on the association between prosocial behaviour and brain morphology, including cortical surface, thickness, and volume. Cortical thickness is a brain morphometric measure that represents the combined thickness of the layers of the cerebral cortex. Cortical thickness has been related to psychiatric disorders (e.g. Fairchild et al. 2013), but also to aggression and conduct problems in the normal range in non-­clinical samples (Thijssen et al. 2015a; Walhovd et al. 2012). Cortical thickness increases during childhood, followed by thinning during and after puberty. Girls have an earlier peak of cor­ tical volume (Lenroot et al.  2007) and more advanced thinning during adolescence (Mutlu et al.  2013) than boys, implying a faster maturation of the social brain areas in females, which might be related (as cause or consequence or just correlate) to their better skills in empathizing and other prerequisites for pro­ social behaviour. In one large population-­based sample of 6- to 9-­year-­old children the as­so­ci­ ation between cortical thickness and prosocial behaviour was examined (Thijssen et al.  2015b), using the parent-­ report prosocial scale of the Strengths and Difficulties Questionnaire (SDQ; Goodman 1997; Paap et al. 2013). This scale consists of five items, such as ‘My child often spontaneously offers to help others’ and ‘My child is considerate of other people’s feelings’. More parent-­reported child prosocial behaviour was related to cortical thickness in regions related to theory of mind (superior frontal cortex, rostral middle frontal cortex cuneus, and precuneus) and inhibitory control (superior frontal and rostral middle frontal cortex). In a second study, involving a partially overlapping sample, brain structure and  functional connectivity were examined in relation to actual donating to a charity at 8 years of age (Wildeboer et al. 2018). Children received 20 coins of 20 eurocents because of their participation in a range of tasks, and were left alone watching a short UNICEF movie about a girl in Bangladesh who had to work in a stone pit and could not go to school. When the movie ended, a text on the computer screen accompanied by a voice-­over asked whether they wanted to donate money to the charity using a money box that was placed in front of them. A thicker cortex in the lateral orbitofrontal cortex/pars orbitalis and pre-/postcentral cortex was associated with higher donations. Interestingly, parental sensitivity during early childhood had also been found related to cortical thickness at age 8 years (Kok et al. 2015), nourishing the idea that parental sensitivity might affect structural brain development which in turn would be related to prosocial behaviour. Unfortunately, the cortical areas in the two studies were not overlapping;

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

Dimensions, Determinants, and Development of Prosocial  55 thus, although not impossible, the evidence for such a brain-­mediated pathway is still lacking. To measure the involvement of certain brain areas in acting prosocially, a specific version of the virtual ball tossing game named Cyberball has been used (Williams et al. 2000). In this game three players participate in a computerized game of ball tossing. During the first round all players participate and toss the ball to each other. In the next rounds one player no longer receives the ball, thus experiencing social exclusion. Many studies used Cyberball to examine the effects of being excluded (the participant no longer receives any tosses), and a few studies investigated how individuals respond when they observe someone else being ostracized. When, during Cyberball, individuals are excluded by others, they show increased neural activity in the dorsal anterior cingulate cortex (dACC) and bilateral insula (Eisenberger, Lieberman, and Williams 2003). Observing someone else being ostracized also activates brain regions involved in empathy (Masten et al. 2010; Van der Meulen et al. 2016). Masten et al. (2010) showed that activation in the empathy-­related brain regions was associated with later prosocial behaviour toward the victim, suggesting that individuals who felt more empathy for the person made greater efforts to comfort the victim afterwards. An adapted version of Cyberball is the four-­player Prosocial Cyberball game, which offers the participants the opportunity to toss the ball to the excluded player, thereby compensating for the exclusion by the other players (Williams and Jarvis  2006; Riem et al.  2013). The set-­up is similar: during the first round all players participate and toss the ball to each other. In the next rounds one player (not the participant) no longer receives the ball, thus experiencing social exclusion. Since the participant continues to receive tosses, he or she has the op­por­tun­ ity to either join the excluders and toss no more balls to the excluded player, or include the excluded player to the same extent as before, or compensate for the exclusion by tossing even more balls to the excluded player than before. In one study with young female adults (Van der Meulen et al.  2016) the participants indeed compensated by tossing the ball more often to the excluded player (50 per cent of tosses during exclusion compared to around 35 per cent during the inclusion round). Moreover, when they tossed the ball to the excluded player, they showed increased activation in the temporal parietal junction (TPJ) that plays a role in perspective taking and had previously been found related to adults’ altruism in the so-­called dictator game (with allocating money to anonymous partners, Morishima et al. 2012). Tossing the ball to the excluded player in the Cyberball game was also accompanied by increased activation in the nucleus accumbens (NAcc), which is involved in experiencing rewards. In a similar vein, the NAcc was found to be more active during voluntary money transfers to a charity than during mandatory transfers (Harbaugh, Mayr, and Burghart 2007); thus, the ­activation of reward areas may be part of the brain’s machinery of prosociality.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

56  M. J. Bakermans-Kranenburg and M. H. van IJzendoorn It may also be the case that individuals differ in the processes involved in their prosocial decisions or behaviour. Variation in motivation may be related to vari­ ation in neural activation. In one study, participants were invited to donate to a charity and then asked to rate how much they relied on their empathic responses (‘how much empathy did you experience for the beneficiaries of the charity’) or perspective taking (‘how much did you try to take the perspective of the beneficiaries of the charity’) during their donation decisions. Participants whose donating behaviour was mostly influenced by affective empathy exhibited more activity in the anterior insula during donations, while those who relied on cognitive perspective taking showed more activity in TPJ (Tusche et al. 2016). Different mo­tiv­ ations for prosocial behaviour may thus be reflected in or accompanied by differential neural activity—validating the idea that the same behaviour may result from multiple underlying motivations and neurobiological processes.

3.3  Enduring environmental influence: parenting Hoffman (1984) suggested that parenting may promote or hamper the development of children’s prosocial behaviour. If parents create a warm, sensitive atmosphere but also consistently discipline child behaviour that is damaging to others, feelings of empathy would be fostered in their children. Non-­empathic, authoritarian control combined with frequent threats and love withdrawal would however lead to compulsive compliance and a lack of moral internalization (Van IJzendoorn 1997; Eisenberger and Fabes 1998). The type of sensitive parenting that is recommended here is also the best predictor of a secure infant–parent attachment relationship. The role of sensitive caregiving may thus be twofold: first, sensitive parents model empathic, helping, and caregiving behaviour, and second, by promoting secure attachment they provide children with an emotional secure base, from which they can safely explore the world and are not overwhelmed by others’ emotions, which would interfere with effective prosocial actions despite good intentions. Numerous studies have examined the as­so­ci­ ations between sensitive parenting, attachment, and child prosocial behaviour (see Van IJzendoorn and Bakermans-­Kranenburg 2014). Results have not always been as straightforward as one might have expected. Brooker and Poulin-­Dubois (2013) found that for 2-­year-­old boys parental sensitivity did not predict the level of helping their mother reach a goal that she could not reach on her own, whereas for girls higher levels of parental sensitivity (observed in a different context) were associated with more helping behaviour. In our Leiden Longitudinal Empathy Study no associations between empathic concern toward the mother or the experimenter with attachment or parental sensitivity were found. Moreover, levels of observed empathy to mother and experimenter

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

Dimensions, Determinants, and Development of Prosocial  57 were not related, refuting the idea of an empathic personality independent of the target and the situation. The association between attachment and prosocial behaviour may be clearer in adulthood. Mikulincer and Shaver (2008a) primed young adults with security-­ enhancing figures, such as their best friend, and observed more compassion with a woman in distress and more willingness to take over her tasks after priming than in the control condition. In our own study we found that adults with a secure representation of attachment, as assessed with the Adult Attachment Interview (Hesse 2016), felt less irritated when they were exposed to crying sounds of an unknown baby, and used less excessive force as indicated by grip strength using a handgrip dynamometer (Riem, Bakermans-­Kranenburg, Van IJzendoorn, Out, and Rombouts 2012). Excessive force in this paradigm has been interpreted as an indicator of low empathy, and an analogy to the use of exaggerated punitive force in response to children’s disobedience.

3.4  Concurrent environmental influence: nudges It is well known that even tiny changes in the environment influence behaviour and—fortunately—can change behaviour for the better. Painting eyes on the wall increases the amount of money collected in an honesty box next to a coffee machine (Bateson, Nettle, and Roberts 2006). Telling 8-­year-­old children that we could see whether they were being honest during an fMRI task in which lying led to monetary rewards resulted in an increase of honest children from 26 per cent (when they thought they were unmonitored) to 66 per cent (when told that we could see whether they were honest) (Thijssen et al. 2017). Installing a security camera promotes students’ helping behaviour collecting and sorting a pile of questionnaires accidentally fallen on the floor (Van Rompay et al. 2009). Students who watched a video clip that documented moral excellence or who read a story about an extraordinary moral act or person donated more money to a charity that was somewhat antithetic to their own political views (Freeman et al. 2009). We mentioned above that in our studies on 7-­year-­old children only a very small minority were inclined to donate any money spontaneously (Van IJzendoorn et al. 2010). After being briefly prompted by an experimenter, the percentage of children who donated increased from one-­tenth to about two-­thirds. In one of our subsequent studies, we showed the same promotional clip for UNICEF including a call for donation, and for a random half of the 8-­year-­old children we added a fragment of a same-­sex peer donating money to the charity. Seeing a peer donate was associated with on average 27 per cent higher donations (Wildeboer et al. 2018). In behavioural economics, situational manipulations that effectively change human behaviour or nudges have become a central topic (see Thaler and Sunstein

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

58  M. J. Bakermans-Kranenburg and M. H. van IJzendoorn 2008). Some prime examples are putting healthy food on eye level and junk food on lower or higher levels in shops, or in preventive efforts emphasizing that the majority of students do not binge drink. But are all individuals similarly sus­cep­ tible to these nudges? Apparently not: as evident from the honesty task described above 34 per cent of the children were not affected by the idea that their lying could be monitored. What factors make individuals more or less susceptible to nudges and enduring environmental influences such as parenting? On the level of personality evidence is scarce. In a recent meta-­analysis of research on the ef­fect­ iv­ity of nudges, almost no studies on the influence of personality or character traits were found (Hummel and Maedche  2019) and the only pertinent study failed to find a moderating effect of personality, in particular conscientiousness, on openness for nudges (Stutzer, Goette, and Zehnder 2011).

3.5  Differential susceptibility Eight-­ year-­ olds donated more after seeing a peer donate (Wildeboer  2016/ Wildeboer et al. 2017). But not all children were similarly affected by the altruistic model, and two groups of children stood out from the majority. On the one hand, children in the upper quartile of anxiety problems donated significantly more after seeing a probe, but they donated less than average when not seeing this probe. They were thus highly susceptible to the nudge of seeing a peer donate. On the other hand, children with lower-­than-­average scores on a social responsiveness scale developed to screen for autistic symptoms (Constantino 2002) did not donate more when they saw a peer donate. Thus, children with more autistic-­like tendencies appeared to remain indifferent to the social pressure of a child exemplifying prosocial donating. In a somewhat similar way, Brooker and Poulin-­Dubois (2013) found that boys and girls differed in the influence of their mothers’ sensitive caregiving on their helping behaviour. In girls helping was related to experienced parental sensitivity, but in boys such an association was not found. Considering genetic influences, dopaminergic system genes have been suggested to be involved in openness to environmental influences (Bakermans-­Kranenburg and Van IJzendoorn 2010; 2015). Indeed, we found that for children carrying the DRD4 7-­repeat allele, security of attachment was related to higher donations to UNICEF and insecure attachment to lower donations, whereas for children without the DRD4 7-­repeat allele attachment security did not make a difference for their donations. In 3.5-­year-­old children Knafo and colleagues (Knafo, Israel, and Ebstein 2011) found that parenting style was related to self-­initiated prosocial behaviour in children carrying the DRD4 7-­repeat allele, but not in children without the 7-­repeat allele. In sum, genetic and environmental determinants of prosocial donating

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

Dimensions, Determinants, and Development of Prosocial  59 behaviour may be both important and need to be considered in interaction. Going beyond single candidate genes to genetic pathways and genome-­wide genetic ‘risk’ or ‘susceptibility’ scores, taking into account epigenetically regulated gene expression (Bakermans-­Kranenburg and Van IJzendoorn 2015; Belsky and Van IJzendoorn  2017) may open up new avenues of studying the interplay between environmental factors and personal characteristics.

4.  Stimulating prosociality Can prosocial behaviour be learned? Can we stimulate prosocial personalities? So far, we may only conclude that the idea or ideal of a moral ‘character’ or set of dispositions that would lead to predictable moral choices and prosocial behaviour in a preponderance of contexts is not supported by empirical research. The evidence showing contextual influences on prosocial behaviour seems to suggest that we could better spend our money and efforts on environmental nudges than on parent support or educational programmes. We may need to adjust the intervention aims: not aiming at individuals who act prosocially in more situations, but aiming at situations in which more individuals act prosocially. We have seen that not all individuals are affected to the same extent by nudge-­ like situational factors. In a similar vein, situational factors or nudges may vary in their breadth and effectiveness. An open question is whether habituation to nudges (e.g. seeing the pair of eyes painted on the honesty box for months) diminishes their effect or promotes prosocial behaviour as a habit. A link to the Aristotelian concept of ‘habituation’ can be made here: Virtuous affective dispositions are strengthened by conditioning behaviour (Steutel and Spiecker  2004). Repeating the same behaviour (paying for your tea when confronted with the painted eyes on the honesty box) over and over again, may over time strengthen the disposition to behave that way, with a corresponding decrease of the influence of the specific situational nudge. It is also tempting to think that some environmental influences could be so strong that their effects generalize over time and contexts. This would create a potential bridge to the idea of prosociality as an enduring personality trait. However, an example of such an encompassing environmental factor is the quality of the caregiving environment, and so far its effects on child prosocial behaviour have been disappointingly small. Another example along these lines is the just-­community approach to moral education (Kohlberg 1985; Power 1979; Oser et al. 2008), an intervention aimed at the immersion of students in an environment with just and fair role models, rules, and interactions, and with well-­defined behavioural norms. The peers in the just community embodied and sanctioned the socio-­ moral norms regulating individual students’ behaviour. Essential

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

60  M. J. Bakermans-Kranenburg and M. H. van IJzendoorn ingredients of the just-­ community approach were the community meeting for  democratic decision-­making and the discipline committee for confronting ­individuals with their misbehaviour, punishing, and forgiving. Moreover, moral dilemma discussions to enhance the level of moral reasoning were part of the curriculum (Oser et al. 2008). The just-­community approach was a short-­term success elevating the moral judgement level of the students, but one year after the conclusion of the experiment none of the teachers had stuck to the moral discussions. Effectiveness of the intervention has not been proved in a randomized control trial. Two other, totally different types of interventions can be mentioned here. They are characterized by a neurobiological approach, and were tested in the past decade. The first is a neurobiological intervention using transcutaneous Vagus Nerve Stimulation (tVNS) to stimulate prosocial behaviour. This is a non-­invasive method to stimulate the vagus nerve that in turn activates the insula and PFC. The method was used during the four-­player prosocial Cyberball game (Williams and Jarvis 2006). As described before, in this virtual tossing game participants observe one of the players being excluded by the others, and they can prosocially compensate by tossing the ball more often to the ostracized person or join the computer-­ controlled other players in excluding this person. The idea was that tVNS, by stimulating the prefrontal cortex-­insula network, would promote prosocial helping behaviour toward the excluded person. In the placebo condition a different placement of the device on the earlobe ensured no stimulation of the vagus nerve. No effect of tVNS was found: participants compensated for exclusion to the same extent with and without tVNS (Sellaro et al. 2015). The second neurobiological approach focuses on oxytocin, the neuropeptide called the ‘love hormone’ related to interpersonal trust and empathic concern. The past fifteen years have seen an explosion of experimental studies with nasal administration of oxytocin, both in clinical and non-­clinical samples (for reviews, see Van IJzendoorn and Bakermans-­ Kranenburg  2012,  2016; Bakermans-­ Kranenburg and Van IJzendoorn  2013). These studies have revealed that the effects of oxytocin depend on context, personal characteristics, and the perception of in-­groups and out-­groups. For example, in a competitive game that triggered decisions with financial consequences to the subjects themselves, their in-­group, and a competing out-­group (De Dreu et al. 2010), oxytocin enhanced a ‘tend and defend’ response in that it promoted in-­group trust and cooperation, but at the same time enhanced defensive aggression toward competing out-­ groups. The same effect can be observed in rodents with pups: oxytocin induces maternal behaviour such as nest building and adopting an optimal nursing posture, but higher oxytocin levels in the dam are also related to more protective aggression to intruders (see Bakermans-­Kranenburg and Van IJzendoorn 2017) In our own lab we found that a sniff of oxytocin increased donations to UNICEF in undergraduate students, but only for a specific subgroup of participants. In this study we also measured frontal brain activity, which is thought to

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

Dimensions, Determinants, and Development of Prosocial  61 reflect individual differences in approach–withdrawal motivation. We expected that relative left frontal activity (related to approach tendencies) would be related to larger donations, which was indeed the case. However, in participants with greater relative right frontal activity, who in general donated less, oxytocin administration resulted in larger donations, in particular when the participants had experienced supportive parenting during their childhood (Huffmeijer et al. 2012). In another sample, we found that oxytocin decreased the use of excessive force on the handgrip dynamometer during exposure to an unknown infant’s crying, but only in those participants who had not experienced harsh caregiving (Bakermans-­ Kranenburg, Van IJzendoorn, Riem, Tops, and Alink  2011). Lastly, in the pro­ social Cyberball game participants compensated for other players’ ostracism by throwing the ball even more often toward the excluded player after oxytocin administration, but only when they had experienced low levels of maternal love withdrawal (Riem et al. 2013). Although the evidence is certainly not yet conclusive, these findings all point into the same direction: oxytocin does promote prosocial behaviour but it is not effective in individuals with untoward childhood experiences. Only one study showed effects in the opposite direction: adults with an insecure representation of attachment felt less irritation and exerted less excessive power with the hand grip dynamometer during infant crying (of an unknown baby) after oxytocin administration (Riem, Bakermans-­Kranenburg, and Van IJzendoorn 2016). So far this is the exception, and oxytocin administration does not seem a viable way to promote prosocial behaviour at large scale. One of the reasons for the divergent effects of oxytocin may be that oxytocin affects social behaviour by increasing social salience. As a result oxytocin administration may have undesirable effects in contexts in which the out-­group is prominent and for (insecure) individuals who consider most of the world as a competing out-­group.

5.  Conclusion We started out with the definition of prosocial behaviours including all actions that serve to benefit others. However, to avoid confusion with selfish interest in one’s own genes, we excluded caregiving of own kin from prosocial behaviour, and we also excluded cooperation, where joining forces is needed to reach a wished-­for goal. The choice to exclude these two types of behaviour from the definition of prosociality is not indisputable, for the rationale relates to mo­tiv­ ations for behaviours, and we argued that motivations are difficult to pinpoint and less important than behaviours. Nevertheless, helping an unfamiliar person, volunteerism, donating to a charity, and favouring an excluded ‘underdog’ in a computer game are paramount examples of prosocial behaviours to out-­group members that are less, or less obviously, self-­serving.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

62  M. J. Bakermans-Kranenburg and M. H. van IJzendoorn If anything is clear from our review of empirical research of prosocial behaviour, it is that to find stability of prosocial behaviour across time and situations appears to be very difficult, as is pinning down the causes of prosocial behaviour in humans. Children who help a research confederate do not necessarily donate to a charity and are not necessarily helpful at a later point in their development. Prosociality does not appear to be a personality trait such that it can be easily measured with a task or questionnaire, with the resulting profile successfully ­predicted by caregiving experiences, neural signatures, or genetic make-­up. That being said, prosocial behaviours do belong to our innate repertoire and there are evolutionary reasons to expect humans to have evolved in such a way as to be generally disposed to behave prosocially in the context of the in-­group, since empathic concern and prosocial helping elevated the chances of survival (De Waal 2008; Hamlin 2014, Hrdy 2009). But apparently prosociality was not handed down over generations as a consistent set of prosocial behaviours, predictably applicable to diverse situations and embodied in individual’s neurobiology. The evidence for a stable mechanism that the individual carries within, explaining prosocial behaviour in different situations, is still elusive. What, then, was handed down over generations that makes us act more or less prosocially, not consistently so, but responding to nudges, situational characteristics such as eyes painted on the wall behind a donation box, or following the example of a stranger that we will never meet again? It may actually be this susceptibility to concurrent environmental factors that we inherited, for better and for worse. Given that in-­group solidarity and capacities for social evaluation were critical to successful breeding, we may have evolved with exactly that: a high susceptibility to environmental influences, particularly influences from conspecifics. It would be in line with this suggestion that prosocial behaviour increases under conditions of feeling monitored by or being involved with other human beings, and that prosociality decreases or becomes unpredictable in the absence of such concurrent environmental pressures or nudges. From this perspective the pattern of findings on prosocial behaviour and development, that seemed somewhat disappointing in the first round, might make sense. The influence of nudges, more specifically our behavioural adaptations in response to these nudges, may reflect an evolved dependence of individual behavioural tendencies on our en­vir­on­ ments, especially on other human beings. This would also be in line with variability in susceptibility to environmental influences, as we have found here (where anxious children were especially responsive to seeing a donating peer, while children with low social responsiveness were not affected) and in other contexts (Bakermans-­Kranenburg and Van IJzendoorn 2015). Indeed, even small between-­individual differences in social behaviour may have led to the coexistence of more and less susceptible individuals, because individual differences in social behaviour elicit flexible responses of conspecifics, and

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

Dimensions, Determinants, and Development of Prosocial  63 flexible responders offer opportunities to non-­flexible individuals (Dingemanse and Wolf  2013). This explains why both susceptible and non-­susceptible genotypes are maintained within a population—and possibly all gradients of susceptibility in between. If individuals have evolved as nodes in the social networks of their in-­groups, the search for a prosocial core in an individual person might focus on the wrong target, and should be redirected towards the interactions with the social context. Broadening the in-­group to include a wider social network than only biologically related conspecifics may be an effective way to promote genuine prosociality. The development of a secure network of attachments in childhood would help to explore this wider social environment, and to avoid or remediate the narrowed attention of insecure individuals to a few in-­group members. The in-­born (differential) susceptibility to social influences might be considered to be the evolutionary foundation upon which prosocial behaviours to out-­groups can be nurtured setting the stage for living a good life.

References Abramson L., Daniel E., & Knafo-Noam L., 2018. ‘The role of personal values in children's costly sharing and non-costly giving’. J Exp Child Psychol. 165, 117–134. Bakermans-Kranenburg, M.  J., and Van IJzendoorn, M.  H. 2013. ‘Sniffing around oxytocin: review and meta-analyses of trials in healthy and clinical groups with implications for pharmacotherapy’. Translational Psychiatry 3 e258. DOI:10.1038/ tp.2013.34 Bakermans-Kranenburg, M. J., Van IJzendoorn, M. H., Riem, M. M. E., Tops, M., and Alink, L. R. A. 2011. ‘Oxytocin decreases handgrip force in reaction to infant crying in females without harsh parenting experiences’. Social Cognitive and Affective Neuroscience. DOI: 10.1093/scan/nsr067. Bakermans-Kranenburg, M. J. & Van IJzendoorn, M. H. 2015. ‘The hidden efficacy of interventions: Gene x Environment experiments from a differential susceptibility perspective’. Annual Review of Psychology 66, 381–409. DOI:10.1146/ annurev-psych-010814-015407. Bakermans-Kranenburg  M. J. & Van IJzendoorn  M. H. 2017. ‘Protective parenting: Neurobiological and behavioural dimensions’, Current Opinion In Psychology 15, 45–49. DOI:10.1016/j.copsyc.2017.02.001. Bateson, M., Nettle, D., and Roberts, G. 2006. ‘Cues of being watched enhance co­oper­ ation in a real-world setting’. Biology Letters 2(3), 412–14. Belsky, J., and van IJzendoorn, M. H. 2017. ‘Genetic differential susceptibility to the effects of parenting’. Current Opinion in Psychology 15, 125–30.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

64  M. J. Bakermans-Kranenburg and M. H. van IJzendoorn Brooker, I., and Poulin-Dubois, D. 2013. ‘Is parental emotional reliability predictive of toddlers’ learning and helping?’ Infant Behaviour and Development 36, 403–18. Constantino, J.  N. 2002. Social Responsiveness Scale (SRS), Manual. Los Angeles: Western Psychological Services. Darwin, C. 1871/1981. The Descent of Man. Princeton: Princeton University Press. Decety, J., Bartal, I.  B.-A., Uzefovsky, F., and Knafo-Noam, A. 2016. ‘Empathy as a driver of prosocial behaviour: highly conserved neurobehavioural mechanisms across species’. Philosophical Transactions of the Royal Society B: Biological Sciences 371(1686), 20150077. http://DOI.org/10.1098/rstb.2015.0077 De Dreu, C. K. W., Greer, L. L., Handgraaf, M. J. J., Shalvi, S., Van Kleef, G. A., Baas, M. et  al. 2010. ‘The neuropeptide oxytocin regulates parochial altruism in intergroup conflict among humans’. Science 328. 1408–1411. De Waal, F.  B.  M. 2008. ‘Putting the altruism back into altruism: the evolution of empathy’. Annual Review of Psychology 59, 279–300. Dingemanse, N. J., and Wolf, M. 2013. ‘Between-individual differences in behavioral plasticity within populations: causes and consequences’. Anim. Behav. 85, 1031–9. Eisenberg, N., and Fabes, R. A. 1998. ‘Prosocial development’. In N. Eisenberg (ed.), Handbook of Child Psychology, vol. 3: Social, Emotional, and Personality Development (5th edn), 701–78. New York: Wiley. Eisenberg, N., Guthrie, I.  K., Murphy, B.  C., Shepard, S.  A., Cumberland, A., and Carlo, G. 1999. ‘Consistency and development of prosocial dispositions: a longitudinal study’. Child Development 70, 1360–72. Eisenberger, N. I., Lieberman M. D., and Williams, K. D. 2003. ‘Does rejection hurt? An fMRI study of social exclusion’. Science 302(5643), 290–2. Fairchild  G., Hagan  C., Walsh  N., Passamonti  L., Calder  A. J., Goodyer  I. M. 2013. ‘Brain structure abnormalities in adolescent girls with conduct disorder’. J Child Psychol. and Psychiatry 54(1), 86–95 Freeman, D., Aquino, K., and McFerran, B. 2009. ‘Overcoming beneficiary race as an impediment to charitable donations: social dominance orientation, the experience of moral elevation, and donation behaviour’. Pers. Soc. Psychol. Bull. 35(1), 72–84. Gilby, I.  C. 2006. ‘Meat sharing among the Gombe chimpanzees: harassment and reciprocal exchange’. Anim. Behav. 71, 953–63. DOI:10.1016/j.anbehav.2005.09.009. Goodman, R. 1997. ‘The strengths and difficulties questionnaire: A research note’. Journal of Child Psychology and Psychiatry, 38(5), 581–586. doi:10.1111/ jcpp.1997.38.issue-5 Grunberg, N. E., Maycock, V. A., and Anthony, B. J. 1985. ‘Material altruism in children’. Basic Appl. Soc. Psych. 6(1), 1–11. Hamlin, J.  K. 2014. ‘Context-dependent social evaluation in 4.5-month-old human infants: the role of domain-general versus domain-specific processes in the development of social evaluation’. Frontiers in Psychology 5, 614. DOI: 10.3389/ fpsyg.2014.00614.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

Dimensions, Determinants, and Development of Prosocial  65 Hamlin J. K. and Wynn K. 2011. ‘Young infants prefer prosocial to antisocial others’. Cognitive Development 26(1), 30–39. Hamlin, J.  K., Wynn, J., and Bloom, K. 2010. ‘Three-month-olds show a negativity bias in their social evaluations’. Developmental Science: A Journal of the European Society for Developmental Psychology 13(6), 923–9. Hamlin, J. K., Wynn, K., and Bloom, P. 2007. ‘Social evaluation by preverbal infants’. Nature 450(7169), 557–9. Harbaugh, W. T., Mayr, U., and Burghart, D. R. 2007. ‘Neural responses to taxation and voluntary giving reveal motives for charitable donations’. Science 316(5831), 1622–5. Hepach R., Vaish A. and Tomasello M. 2012. ‘Young children are intrinsically mo­tiv­ ated to see others helped’, Psychological Science 23(9), 967–72. Hepach R., Vaish A., Grossmann T. and Tomasello, M. 2016. ‘Young Children Want to See Others Get the Help They Need’, Child Development 87 (6), 1703–14. Hesse, E. 2016. ‘The Adult Attachment Interview. Protocol, method of analysis, and selected empirical studies: 1985–2015’. In J. Cassidy & P. R. Shaver (Eds.), Handbook of attachment: Theory, research, and clinical applications (3rd ed.), 553–597). New York: Guilford Press. Hoffman  M. 1984. ‘Parent Discipline, Moral Internalization, and Development of Prosocial Motivation’. In E.  Staub et al. (eds.), Development and Maintenance of Prosocial Behaviour. New York: Plenum, 117–137. Hrdy, Sarah. 2009. Mothers and others: The evolutionary origins of mutual understanding. Cambridge, MA: Belknap Press. Huffmeijer, R., Alink, L.  R.  A., Tops, M., Bakermans-Kranenburg, M.  J., and Van IJzendoorn, M. H. 2012. ‘Asymmetric frontal brain activity and parental rejection predict altruistic behaviour: moderation of oxytocin effects’. Cognitive Affective and Behavioural Neuroscience 12, 382–92. DOI: 10.3758/s13415-011-0082-6. Hummel, D., and Maedche, A. 2019. ‘How effective is nudging? A quantitative review on the effect sizes and limits of empirical nudging studies’. Journal of Behavioral and Experimental Economics 80, 47–58. Jensen, K., Hare, B., Call, J., and Tomasello, M. 2006 ‘What’s in it for me? Self-regard precludes altruism and spite in chimpanzee’s. Proc. R.  Soc. B 273, 1013–21. DOI:10.1098/rspb.2005.3417. Knafo A., Israel S. and Ebstein R. 2011. ‘Heritability of children's prosocial behavior and differential susceptibility to parenting by variation in the dopamine receptor D4 gene’, Developmental Psychopathology 23(1), 53–67. Knafo, A., Zahn-Waxler, C., Van Hulle, C., Robinson, J. L., & Rhee, S. H. 2008. ‘The developmental origins of a disposition toward empathy: Genetic and environmental contributions’. Emotion 8(6), 737–752. Kohlberg, L. 1985. ‘Resolving moral conflicts within the just community’. In C. Harding (ed.), Moral Dilemmas, 71–98. Chicago: Precedent Publishing.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

66  M. J. Bakermans-Kranenburg and M. H. van IJzendoorn Kok  R., Thijssen  S., Bakermans-Kranenburg  M. J., Jaddoe  V. et al. 2015. ‘Normal Variation in Early Parental Sensitivity Predicts Child Structural Brain Development’. J. American Academy of Child Adolescent Psychiatry 54(10), 824–831. Lamb, S. and Zakhireh, B. 1997. ‘Toddlers' Attention to the Distress of Peers in a Day Care Setting.’ Early Education and Development 8(2), 105–18. Lenroot, R.  K., Gogtay, N., Greenstein, D.  K., Wells, E.  M., Wallace, G.  L., Clasen, L. S., Blumenthal, J. D., Lerch, J., Zijdenbos, A. P., Evans, A. C., Thompson, P. M., and Giedd, J. N. 2007. ‘Sexual dimorphism of brain developmental trajectories during childhood and adolescence’. NeuroImage 36, 1065–73. Masten, C. L., Eisenberger, N. I., Pfeifer, J. H., and Dapretto M. 2010. ‘Witnessing peer rejection during early adolescence: neural correlates of empathy for experiences of social exclusion’. Social Neuroscience 5, 496–507. DOI: 10.1080/17470919.2010.490673 Melis  A. P., Warneken  F., Jensen  K., Schneider  A. C., Call  J., Tomasello  M. 2011. ‘Chimpanzees help conspecifics obtain food and non-food items’, Proceedings of the Royal Society B: Biological Sciences 278 (1710), 1405–1413. Merritt, M. W., Doris, J. M., and Harman, G. 2010. ‘Character’. In J. M. Doris and the Moral Psychology Research Group (eds), The Moral Psychology Handbook, 355–401. Oxford: Oxford University Press. Mikulincer, M., and Shaver, P. R. 2008a. ‘Can’t buy me love: an attachment perspective on social support and money as psychological buffers’. Psychological Inquiry 19, 167–73. Morishima, Y., Schunk, D., Bruhin, A., Ruff, C. C., and Fehr, E. 2012. ‘Linking brain structure and activation in temporoparietal junction to explain the neurobiology of human altruism’. Neuron 75, 73–9. DOI 10.1016/j.neuron.2012.05.021 Mutlu, A. K., Schneider, M., Debbané, M., Badoud, D., Eliez, S., and Schaer, M. 2013. ‘Sex differences in thickness, and folding developments throughout the cortex’. Neuroimage 82, 200–7. DOI:10.1016/j.neuroimage.2013.05.076 Ohtsuki, H., Hauert, C., Lieberman, E., and Nowak, M. A. 2006. ‘A simple rule for the evolution of cooperation on graphs and social networks’. Nature 441, 502–5. DOI:10.1038/nature04605. Oser, F.  K., Althof, W., and Higgins-D’Allessandro, A. 2008. ‘The Just Community approach to moral education: system change or individual change?’ Journal of Moral Education 37(3), 395–415. Paap, M. C., Haraldsen, I. R., Breivik, K., Butcher, P. R.,Hellem, F. M., & Stormark, K. M. 2013. ‘The link between peer relations, prosocial behavior, and ODD/ADHD symptoms in 7–9-year-old children’. Psychiatry Journal, 319874. Pannebakker, F. D. 2007. Morality from Infancy to Middle Childhood. Leiden: Mostert and Onderen. Pappa, I., Fedko, I.  O., Mileva-Seitz, V.  R., Hottenga, J.-J., Bakermans-Kranenburg, M. J., Bartels, M., and Boomsma, D. I. 2015. ‘Single nucleotide polymorphism heritability of behaviour problems in childhood: genome-wide complex trait analysis’.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

Dimensions, Determinants, and Development of Prosocial  67 J.  American Academy of Child and Adolescent Psychiatry 54(9), 737–44. DOI: 10.1016/j.jaac.2015.06.004. Power, F. C. 1979. The moral atmosphere of a Just Community high school: a fouryear longitudinal study. Unpublished doctoral dissertation, Harvard Graduate School of Education. Cambridge, Mass. Riem, M. M. E., Bakermans-Kranenburg, M. J., Van IJzendoorn, M. H., Out, D., and Rombouts, S. A. R. B. 2012.’ Attachment in the brain: adult attachment representations predict amygdala and behavioural responses to infant crying’. Attachment & Human Development 14, 533–551. Riem, M. M. E., Bakermans-Kranenburg, M. J., Huffmeijer, R., and Van IJzendoorn, M. H. 2013. ‘Does intranasal oxytocin promote prosocial behaviour to an excluded fellow player? A randomized controlled trial with Cyberball’. Psychoneuroendocrinology 38, 1418–25. Riem, M.  M.  E., Bakermans-Kranenburg, M.  J., and Van IJzendoorn, M.  H. 2016. ‘Intranasal administration of oxytocin modulates behavioural and amygdala responses to infant crying in females with insecure attachment representations’. Attachment & Human Development 18(3), 213–34. DOI:10.1080/14616734.201 6.1149872. Rilling, J., and Mascaro, J. S. 2017. ‘The neurobiology of fatherhood’. Current Opinion in Psychology 15, 26–32. Rushton, J. P., and Wheelwright, M. 1980. ‘Validation of donating to charity as a measure of children's altruism’. Psychological Reports, 47(3), 803–806. Rushton, J. P., Littlefield, C. H. and Lumsden, C. J. 1986. ‘Gene culture coevolution of complex social behaviour: human altruism and mate choice’. Proceedings of the National Academy of Sciences of the United States of America, 83(19), 7340–7343. Sagi, A., and Hoffman, M. L. 1976. ‘Empathic distress in the newborn’. Developmental Psychology 12(2), 175–176. Salvadori E, Blazsekova T, Volein A, Karap Z, Tatone D, Mascaro O, et al. 2015. ‘Probing the Strength of Infants' Preference for Helpers over Hinderers: Two Replication Attempts of Hamlin and Wynn (2011)’. PLoS ONE 10(11): e0140570. https://doi.org/10.1371/journal.pone.0140570 Sellaro, R., Steenbergen, L., Verkuil, B., Van Ijzendoorn, M. H., and Colzato, L. 2015. ‘Transcutaneous Vagus Nerve Stimulation (tVNS) does not increase prosocial behaviour in Cyberball’. Frontiers in Psychology 6. DOI.org/10.3389/ fpsyg.2015.00499 Shanab, M.  E. and Yahya, K.  A. 1977. ‘Behavioural study of obedience in children’. Journal of Personality and Social Psychology, 35, 530–6. Silk, J.  B., Brosnan, S.  F., Vonk, J., Henrich, J., Povinelli, D.  J., Richardson, A.  S., Lambeth, S. P., Mascaro, J., and Schapiro, S. J. 2005. ‘Chimpanzees are indifferent to the welfare of unrelated group members. Nature 437, 1357–9. DOI:10.1038/ nature04243.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

68  M. J. Bakermans-Kranenburg and M. H. van IJzendoorn Steutel, J., and Spiecker, B. 2004. ‘Cultivating sentimental dispositions through Aristotelian habituation’. Journal of Philosophy of Education 38, 531–49. Stutzer, A., Goette, L., and Zehnder, M. 2011. ‘Active decisions and prosocial behaviour: a field experiment on blood donation’. The Economic Journal 121(556), 476–93. Thaler, R.  H., and Sunstein, C.  R. 2008. Nudge: Improving decisions about health, wealth, and happiness. New Haven: Yale University Press. Thijssen, S., Ringoot, A. P., Wildeboer, A., Bakermans-Kranenburg, M. J., El Marroun, H., Hofman, A., and White, T. (2015a). ‘Brain morphology of childhood aggressive behaviour: a multi-informant study in school-age children’. Cognitive, Affective, & Behavioural Neuroscience 15(3), 564–77. DOI: 10.3758/s13415-015-0344-9 Thijssen, S., Wildeboer, A., Muetzel, R. L., Bakermans-Kranenburg, M. J., El Marroun, H., Hofman, A., Jaddoe, V. W. V., Van der Lugt, A., Verhulst, F. C., Tiemeier, H., Van IJzendoorn, M. H., and White, T. (2015b). ‘Cortical thickness and prosocial behaviour in school-age children: a population-based MRI study’. Social Neuroscience 10(6), 571–82. DOI: 10.1080/17470919.2015.1014063 Thijssen, S., Wildeboer, A.  Van IJzendoorn, M.  H., Muetzel, R., Langeslag, S.  J.  E., Jaddoe, V. W. V., Verhulst, F. C., Tiemeier, H., Bakermans-Kranenburg, M. J., and White, T. 2017. ‘The honest truth about deception: demographic, cognitive, and neural correlates of child repeated deceptive behaviour’. Journal of Experimental Child Psychology 162, 225–41. DOI: 10.1016/j.jecp.2017.05.009 Trivers, R. L. 1974. ‘Parent–offspring conflict’. American Zoologist 14(1), 249–64. Tucker, D. M., Luu, P., and Derryberry, D. 2005. ‘Love hurts: the evolution of empathic concern through the encephalization of nociceptive capacity’. Dev. Psychopathol. 17, 699–713. DOI:10.1017/S0954579405050339. Tusche, A., Böckler, A., Kanske, Ph., Trautwein, and F-M., Singer, T. 2016. ‘Decoding the charitable brain: empathy, perspective taking, and attention shifts differentially predict altruistic giving’. Journal of Neuroscience 36(17), 4719–32; DOI: 10.1523/ JNEUROSCI.3392–15. Van de Vondervoort, J.  W., and Hamlin, J.  K. 2017. ‘Preschoolers’ social and moral judgments of third-party helpers and hinderers align with infants’ social evaluations’. Journal of Experimental Child Psychology 164, 136–51. DOI: 10.1016/j. jecp.2017.07.004. Van der Mark, I. L., Van IJzendoorn, M. H., and Bakermans-Kranenburg, M. J. 2002. ‘Development of empathy in girls during the second year of life: associations with parenting, attachment and temperament’. Soc. Dev. 11(4), 451–68. Van der Meulen, M., van IJzendoorn, M.  H., and Crone, E.  A. 2016. ‘Neural ­correlates of prosocial behaviour: compensating social exclusion in a four-player Cyberball game’. PLOS ONE 11(7): e0159045. https://DOI.org/10.1371/journal. pone.0159045 Van IJzendoorn, M. H. 1997. ‘Attachment, emergent morality and aggression: toward a developmental socioemotional model of antisocial behaviour’. International Journal of Behavioural Development, 21(4), 703–27.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

Dimensions, Determinants, and Development of Prosocial  69 Van IJzendoorn, M.  H., and Bakermans-Kranenburg, M.  J. 2012. ‘A sniff of trust: meta-analysis of the effects of intranasal oxytocin administration on face recognition, trust to in-group, and trust to out-group’. Psychoneuroendocrinology 37, 438–43. DOI: 10.1016/j.psyneuen.2011.07.008 Van IJzendoorn, M. H., and Bakermans-Kranenburg, M. J. 2014. ‘Prosocial development and situational morality: neurobiological, parental, and contextual factors’. In J.  F.  Leckman, C.  Panter-Brick, and R.  Salah (eds), Pathways to Peace: The Transformative Power of Children and Families, 161–84. Cambridge, MA: MIT Press. Van IJzendoorn, M. H., and Bakermans-Kranenburg, M. J. 2016. ‘The role of oxytocin in parenting and as augmentative pharmacotherapy: critical issues and bold conjectures’. Journal of Neuroendocrinology 28. DOI:10.1111/jne.12355. Van IJzendoorn, M. H. & Bakermans-Kranenburg, M. J. 2018. ‘Prosocial behaviour in context: Influence of genetics, parenting, and context on children’s prosocial behaviour’. Invited contribution to the Moral Development Workshop at the Israeli Institute for Advanced Studies, Israel, Jerusalem July 9–13, 2018. Van IJzendoorn, M.  H., Bakermans-Kranenburg, M.  J., Pannebakker, F., and Out, D. 2010. ‘In defense of situational morality: genetic, dispositional and situational determinants of children’s donating to charity’. Journal of Moral Education 39, 1–20. Van Rompay, T. J. L., Vonk, D. J., and Fransen, M. L. 2009. ‘The eye of the camera: effects of security cameras on prosocial behaviour’. Environ. Behav. 41(1), 60–74. Visscher, P.  M., Hemani  G., Vinkhuyzen, A.  A.  E., et al. 2014. ‘Statistical power to detect genetic (co)variance of complex traits using SNP data in unrelated samples’. PLOS Genet. 10: e1004269. Volbrecht  M., Lemery  K., Aksan, N. Zahn-Waxler  C., Hill Goldsmith, H. 2007. ‘Examining the familial link between positive affect and empathy development in the second year’. Journal of Genetic Psychology 168(2), 105–130 Walhovd, K. B., Tamnes, C. K., Østby, Y., Due-Tønnessen, P., and Fjell, A. M. 2012. ‘Normal variation in behavioural adjustment relates to regional differences in cor­ tical thickness in children’. European Child & Adolescent Psychiatry 21, 133–40. DOI:10.1007/s00787-012-0241-5 Warneken, F., and Tomasello, M. 2006. ‘Altruistic helping in human infants and young chimpanzees’. Science 311(5765), 1301–3. DOI: 10.1126/science.1121448 Warneken, F., and Tomasello, M. 2007. ‘Helping and cooperation at 14 months of age’. Infancy 11, 271–94. Wesselmann, E.  D., Williams, K.  D., and Hales, A.  H. 2013. ‘Vicarious ostracism’. Frontiers in Human Neuroscience 7, 153. Wildeboer, A. 2016. ‘Nice traits or nasty states: dispositional and situational correlates of prosocial and antisocial behaviour in childhood’. Doctoral dissertation, Leiden University. Wildeboer, A., Thijssen, S., Bakermans-Kraneburg, M. J., Jaddoe, V. W. V., White, T., Tiemeier, H., & van IJzendoorn, M. H. 2017. ‘Anxiety and social responsiveness moderate the effect of situational demands in children’s donating behaviour’.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

70  M. J. Bakermans-Kranenburg and M. H. van IJzendoorn Merrill-Palmer Quarterly-Journal of Developmental Psychology 63(3), 240–366. DOI:10.13110/merrpalmquar1982.63.3.0340 Wildeboer, A., Thijssen, S., Muetzel, R. L., Bakermans-Kranenburg, M. J., Tiemeier, H., White, T. & van IJzendoorn, M. H. 2018.’ Neuroanatomical correlates of donating behavior in middle childhood’. Social Neuroscience 13(5), 541–52. Williams, K. D., Cheung, C. K. T., and Choi, W. 2000. ‘Cyberostracism: effects of being ignored over the Internet’. J. Pers. Soc. Psychology 79, 748–62. Williams, K. D., and Jarvis, B. 2006. ‘Cyberball: a program for use in research on interpersonal ostracism and acceptance’. Behav. Res. Methods 38, 174–80. DOI: 10.3758/ bf03192765 Witteman, J., Van IJzendoorn, M.  H., Rilling, J.  K., Bos, P.  A., Schiller, N.  O., and Bakermans-Kranenburg, M. J. 2019. ‘Towards a neural model of infant cry perception’. Neuroscience & Biobehavioral Reviews 99, 23–32. DOI: 10.1016/j. neubiorev.2019.01.026 Zahn-Waxler, C., Robinson, J. L., & Emde, R. N. 1992. ‘The development of empathy in twins’. Developmental psychology 28(6), 1038. Zimbardo, P. 2007. The Lucifer effect: Understanding how good people turn evil. New York: Random House.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

3

Attachment, the Virtues, and the Vices Some Developmental and Conceptual Considerations Jonathan Hill

1.  Introduction The topic of this volume crosses developmental, moral, psychiatric, and philosophical territories, and so poses interesting challenges of methodology and content. If we take the starting question to be, ‘can secure attachment contribute to our understanding of the virtues?’ with the implication that insecure attachment may inform the vices (or at least the absence of virtue), we have the very unusual situation of a developmental-­psychological predictor with an outcome identified (and scrutinized) as a philosophical concept. Paralleling this hybrid question we are also faced with the dilemma as to whether to try to answer it by reviewing the evidence or examining the argument. And if the evidence is against it, is that the end of the discussion? Or if the argument is strong is the (fallible and provisional) evidence peripheral? As we shall see, there is a third and related challenge, which is that attachment research has mainly been about a very specific, and circumscribed aspect of the child’s early development, whereas the virtues and the vices, discussed by Aristotle in terms of the man of good or bad character, cover almost every aspect of a person’s life and being. As it turns out the task is not as problematic as my opening may imply. This is in part because Aristotle provides a remarkable number of specifics in his characterization of the virtues and the vices, so that there is much that looks tractable, defining and focusing in ways that might bring the topic closer to empirical study. At the same time his concept of character ‘in the round’ provides a challenge to conceptualize its origins in a similarly comprehensive way. There is therefore no need to adopt a sternly empirical approach nor an insistent philosophical stance, and the chapter aims to avoid either. I start with an outline of aspects of Aristotle’s discussion of character, the virtues and the vices, drawing out some of the ways in which we can readily make links with concepts and findings in current developmental research. I then go on to discuss current concepts and evidence regarding the developmental origins of the virtues and the vices, before summarizing key aspects of what is known about Jonathan Hill, Attachment, the Virtues, and the Vices: Some Developmental and Conceptual Considerations In: Attachment and Character: Attachment Theory, Ethics, and the Developmental Psychology of Vice and Virtue. Edited by: Edward Harcourt, Oxford University Press. © Jonathan Hill 2021. DOI: 10.1093/oso/9780192898128.003.0004

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

72  Jonathan Hill early attachment processes. Having done this it will be evident that putting it all together cannot be straightforward because at almost every turn there are conditionals, exceptions, inconsistencies, and unknowns. Having stated largely eclectic aims for the chapter I am going to take as my starting point a developmental, psychiatric, and essentially empirical statement. The proposition that attachment processes may help us understand the virtues and the vices is a proposition about early developmental origins—for three main reasons. First attachment theory and the majority of the relevant research refers to early development. Second, by studying early development there is the possibility of identifying processes that operate prior to the establishment of character. Third, while measures of attachment in adolescence and adult life have been established, they are much less likely to illuminate origins because attachment status is as likely to be an outcome of character as the other way round. Thus the chapter centres on an enquiry into the childhood origins of the virtues and the vices in adults, and into the role of attachment and other processes in infancy and childhood.

2.  Aristotle, the Virtues and the Vices To summarize, according to Aristotle, the ethical virtues such as justice, courage, and temperance require, ‘a proper appreciation of the way in which such goods as friendship, pleasure, virtue, honour and wealth fit together as a whole’ (Kraut 2018). Coming from the empirical study of the early developmental origins of emotions and behaviours, this take on the virtues is so startling it is like walking into a sheet of glass mistaking it for an open door. The shock partly comes from the ambition and comprehensiveness of the concept, and the sense that this will never be embraced by the modest piecemeal measurement employed in the behavioural sciences. However the effect is not altogether painful because of the broader and interesting challenges that the ideas pose. One is the challenge to conceptualize a single quality covering multiple aspects of a person’s life. In the quotation above this quality is an appreciation of the way different experiences and activities fit together. However it is also clear that Aristotle conceives of the possibility that the ethical virtues also consist of the harmonious operation of these personal qualities and activities in the same individual. Given that human virtues seem often to be more patchy, with great leaders sometimes neglecting their families, and some wealth creators placing effectiveness over honour, this seems to set the bar very high (e.g. Gilbey 1994). Or might the idea be that the virtuous person achieves in all of these arenas but at a more modest level? Another is the conceptualization of the virtues as embedded in the social world, not only in terms of behaving well, but of making a contribution, and also participating as a sharer and a receiver with other people. Commonly developmental

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

Attachment, the Virtues, and the Vices  73 outcomes are assessed as individual competence or deviance, and rarely as a capacity for interdependence. While individuals may commonly possess patchy virtues, the vices are commonly embedded in shortcomings in many areas of life that Aristotle describes. Pervasive problems of this kind are characterized in psychiatry as ‘personality disorders’ and the pattern of behaviours associated with repeat offending as ‘Antisocial Personality Disorder’. The concept is much criticized, for good r­ easons, however it draws our attention to the fact that the problem for most repeat offenders is not confined to their antisocial behaviours, but also to failures in multiple aspects of their lives, in work, marriages and cohabitations, in friendships, and as parents. Around 80 per cent of violent offenders in UK prisons have Antisocial Personality Disorder (Nathan et al. 2003) and the majority when out of prison lack close relationships of any kind (Robins 1966). This characterization of the virtues will bring us back to the central dilemma for an attempt to find origins in attachment. How can we consider attachment status, measured in a circumscribed way, investigated as a predictor of discrete outcomes, as an explanation for this comprehensive set of human qualities? Do we broaden how we think of attachment, or in some way circumscribe or capture the virtues to bring them into the laboratory? If we do the first we will loosen the ties to the empirical evidence, and if we do the second we may remove the essence of the original concept. Aristotle himself brings the concept of the virtues closer to what we might be able to examine as an outcome from attachment, in at least two ways, by specifying behaviours, and by theorizing about their underlying processes. His statement that ‘The man who possesses character excellence does the right thing, at the right time, and in the right way’ provides us with specifics that we can readily translate into psychological constructs. Doing ‘the right thing’ implies that he appraises the demands of a situation accurately, acting ‘at the right time’ requires that action is neither premature nor too late, and achieving this ‘in the right way’ demands action that is effective in achieving a solution. We can also readily make links to further familiar ideas such as that doing the right thing in a social setting entails using the ‘Intentional Stance’ (Dennett  1989) and on the basis of an empathic response to another, that acting in the right way includes being assertive when rights are violated, that right action is not undermined by strong emotion, and that the person’s actions are consistent over situation and time. The meaning of ‘right’ here may seem vague, but can be readily translated to mean ‘according to laws or institutional rules’, ‘according to values’, or ‘according to the effect of actions on others’. Many of these constructs have been examined in relation to disruptive and aggressive behaviours in children as I shall outline later. Aristotle’s theorizing about internal processes is remarkably complex and psychological. The man of virtuous character not only acts in the right way, but does so as a result of harmonious internal processes, whereby acting in the right way is

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

74  Jonathan Hill a source of pleasure or satisfaction, or perhaps even something deeper, and even acts that seem generous, or self-­sacrificial, or likely to frustrate desires, are not conducted in the face of contrary processes. ‘Like anyone who has developed a skill in performing a complex and difficult activity, the virtuous person takes pleasure in exercising his intellectual skills. Furthermore, when he has decided what to do, he does not have to contend with internal pressures to act otherwise. He does not long to do something that he regards as shameful; and he is not greatly distressed at having to give up a pleasure that he realizes he should forego’(Kraut 2018). This leads to a paradox, which is that the person who acts in the right way, but who can only do that by controlling impulses or desires to act badly, is not considered virtuous, rather they are, according to Aristotle, ‘continent’. By contrast those who act badly because they are unable to regulate or control impulses or desires are ‘incontinent’. Others who are similarly prone to acting badly, refuse even to try to do what an ethically virtuous agent would do, because they have become convinced that justice, temperance, generosity, and the like are of little or no value. Such people Aristotle calls evil. He assumes that evil people are driven by desires for domination and luxury, and although they are single-­minded in their pursuit of these goals, he portrays them as deeply divided, because their desire for more and more leaves them dissatisfied and full of self-­hatred. Seen against Aristotle’s description of the components of virtuous action, and of ways in which action may fall short of the virtuous, making the link with current evidence regarding the developmental origins of social and antisocial behaviours becomes more feasible. Whether this also creates a link to attachment is discussed later in the chapter.

3.  Attachment Theory and Research Bowlby trained as a doctor, psychiatrist, and psychoanalyst, and we find echoes of each in attachment theory’s focus on biology, psychopathology, and the consequences of early experiences (Bowlby 1982). Key features of the theory include that infants seek care selectively from caregivers not because they reward them, as contemporary reinforcement theorists argued, but because they have regular contact with them. The process, Bowlby argued, has similarities to imprinting found in many species, and like imprinting this is an evolved universal phenomenon associated with offspring safety. In many species there is a fixed critical period for imprinting, so that contact with caregivers once that period has ended no longer has the same effect. The theory is concerned not only with the conditions under which selective attachments form but also what influences their quality and how this lays the foundations for later development. According to the theory, sensitive caregiving promotes the development of a secure ‘internal working model’ which

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

Attachment, the Virtues, and the Vices  75 is internalized and becomes, if not a template, a guide in later close relationships. Here Bowlby draws on ideas from the object relations school that early objects (oddly enough object here refers to people!) are internalized and become part of the personality. Early experiences that do not support a secure internal working model are thought to create vulnerability to psychiatric disorders. Intuitively, links between attachment security and anxiety and depression seem plausible. Individuals who lack confidence in close relationships, or fear their breakdown, may be expected to suffer from anxiety. Those with difficulties in establishing relationships may lack emotional supports, and those who have ­trouble maintaining relationships are likely to experience repeated losses, both associated with depression (Brown and Harris  1978). It is not so obvious that secure internal working models of relationships will lead to the virtues, and insecure to the vices. Nevertheless one of Bowlby’s earliest publications was his ‘44 Juvenile Thieves’ (Bowlby  1944). In this paper he compared 44 children aged 5–16 from a child guidance clinic with a history of stealing, and 44 controls from the same clinic. He characterized 14 of the children as having ‘affectionless psychopathy’ and recorded that 11 of them had been separated from their mothers in early childhood, in contrast to none of the controls. Bowlby inferred from this association that they had suffered early deprivation as a consequence of a lack of a warm continued love of a mother. Interestingly Bowlby’s explanation of the link between deprivation of maternal love and stealing is thoroughly Freudian: this is early Bowlby! Subsequently there has been considerable controversy regarding the role of separation from a parent, loss of a parent, and lack of stable caregivers. Overall the evidence is that separation and loss in themselves have few long-­term effects, independently of the circumstances that lead to them and the quality of care received from other caregivers (Rutter  1972). Lack of stable caregivers in early life, by contrast, has major effects as evidenced in the follow up into adult life of children adopted following early institutional experiences in Romania. Compared to other adopted children, they have markedly increased levels of ADHD, emotional problems, and impaired relationship and work functioning. However they are not more likely to be antisocial (Golm et al. 2020). Attachment theory not only proposes a key role for parental, and especially maternal, care, but also, as we have noted, a mechanism whereby it has an effect. This was conceptualized by Bowlby in terms of internal working models, and operationalized by Mary Ainsworth for infants aged 12–18 months as an or­gan­ iza­tion of behaviours with a caregiver in response to threat. Ainsworth created a standard procedure, the ‘Strange Situation Test’ (SST) in which infants are sep­ar­ ated from the parent, and then reunited, and are classified as ‘Secure’ or ‘Insecure’ on the basis of their behaviour towards the parent on reunion.1 Ainsworth

1  See summary in Introduction to the volume.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

76  Jonathan Hill proposed two insecure categories, ‘avoidant’ and ‘resistant’, and later investigators added ‘disorganized’. The SST has been very extensively used in studies of early attachment, and it is the measure against which other measures of attachment are commonly validated (Ainsworth et al. 1978). Three points regarding the SST need to be made for this discussion. First, the attachment classification of the same infant with different parents, or other stable caregivers, is commonly different. Thus, although the terms ‘secure infant’ or ‘disorganized infant’ are commonly used, in early life attachment status is an at­tri­ bute of a relationship not a child. Furthermore, stability of attachment status in infancy and early childhood is modest with correlations typically between 0.3 and 0.4 which means that changes of attachment status, even over periods of only a few months, are very common (Fraley and Thompson 2002). In summary attachment security, assessed in the SST and similar separation–reunion procedures after infancy, is not a strong candidate for understanding the origins of character, conceptualized as a stable coherent network of qualities in an individual. Second, the terminology of attachment classification in the SST is evaluative, and unless we are careful can inhibit focus on what is actually being measured. It seems obvious that being ‘secure’ is likely to be beneficial, and similarly to be insecure unhelpful or even harmful. However this cannot be assumed. The characteristic behaviour of the ‘secure’ infant with a caregiver on reunion, following separation that has caused distress, is to seek physical comfort, openly and directly. This, as the theory would predict, is likely to promote confidence in the relationship and the sense that upset and sadness can be relieved by that care­ giver—but only if the caregiver is warm and responsive. But such an open and trusting approach to an unresponsive or negative caregiver could have the reverse effect. The assumption is often made that this is an unlikely circumstance because secure infants have responsive parents. However, the association between sensitivity and security is modest (McElwain and Booth-­LaForce 2006), and for many there will be a mismatch between attachment status and the parenting they ex­peri­ence. In our longitudinal study, the Wirral Child Health and Development Study, we found evidence that secure attachment may under such circumstances be associated with a poor outcome. Children with secure attachment with their mothers in the SST at 14 months, whose mothers were critical of them when tasked to talk about them, had more emotional and behavioural problems at age 2.5 years than insecure children with critical mothers (Hill et al. in preparation). Third, a consequence of the prominence of the SST in attachment research is that the majority of attachment research employs a narrow attachment construct, referring only to the readiness to turn to a caregiver when distressed. It does not refer to other processes that might be thought of in relation to security, such as the reliability of a caregiver to keep a child safe, the predictability of their dis­cip­ line, the survivability of their anger, or the trustworthiness of their judgements. This narrowing also led to a focus on the relational consequences of parenting

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

Attachment, the Virtues, and the Vices  77 quality rather than the broader implications for the child. However, it is plausible to suppose that the effects of sensitive warm parenting extend beyond the relationship. Two lines of evidence support this idea. Maternal sensitivity in playful settings with infants is associated with later cognitive development, probably because the parent’s support for and interest in the infant’s exploration of the world promotes a curiosity that increases learning (Malmberg et al.  2016). Similarly in the Wirral study, maternal responsiveness to distress and their warmth with infants was associated with lower indifference to other children’s distress—‘callous unemotional traits’—between ages 2.5 and 5 years (Wright et al. 2018). This may be because the experience of empathic concern from a parent promotes it in the infant. This finding is very relevant to the topic of the chapter because callous unemotional traits in children, like psychopathy in adults, is associated with enduring and severe aggression. Crucially neither the link between parenting and cognitive ability or between parenting and lower callous unemotional traits appears to be mediated via attachment security. Thus, we find evidence that links between early parenting and some broader capabilities likely to be important to character development bypass attachment security, perhaps because the dominant concept is confined to a narrow set of processes in particular relationships.

4.  Current Evidence Regarding the Origins of Social and Antisocial Behaviours There is little doubt that many of the prosocial and antisocial behaviours that we associate with the way the virtues and the vices are described in adults are seen in very young children. Exactly how young is a matter of some debate. There is some evidence that remarkable socially sympathetic behaviours are seen in the first year of life. For example an intriguing study published in Nature reported that infants aged 6 months favoured a figure that provided help rather than one who hindered another figure’s progress up a hill (Hamlin et al.  2007). The extent to which such behaviours are seen as young as 6 months remains controversial, but empathic responding during the second year is well established. Between 12 and 18 months of age, toddlers have been found to show empathic concern toward others’ negative emotions and sometimes make helpful advances toward a victim of distress (i.e. patting, touching, hugging) or signal someone else to help (Roth-­ Hanania, Davidov, and Zahn-­Waxler 2011, Kim and Kochanska 2017). Similarly human infants as young as 14 to 18 months of age help others attain their goals, for example, by helping them to fetch out-­of-­reach objects or opening cabinets for them. They do this irrespective of any reward from adults and indeed external rewards undermine the tendency (Warneken and Tomasello 2009).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

78  Jonathan Hill Antisocial behaviours appear similarly early. Most repeat offenders in adult life have been aggressive and oppositional as young children, with these behaviours commonly seen during the second year of life (Côté et al. 2007). Strikingly, and in accord with Aristotle’s account, children and adults who repeatedly fail ‘to do the right thing at the right time’ also have difficulties in multiple domains of their lives. They commonly fail in school, are unpopular with peers, perform poorly in work, and have unstable adult relationships (Hill 2002; Hill and Maughan 2015). While these are very strong associations looking back to the childhoods of anti­ social adults, looking forwards the continuity is somewhat weaker. Around a half of antisocial young children do not grow up to be repeat offenders. Why do some young children, around 5 per cent, develop serious persistent antisocial behaviours? We can be moderately confident both that several factors combine to create the problem in any particular child, and that this occurs in different ways across children. A summary of the known risks illustrates how complex the area is. These include family history of criminality, physical abuse of the child, arguments and violence between parents, parental hostility, lack of parental monitoring, inconstant house rules, intense negative emotions in the child, lack of child empathy, low verbal ability and self-­control in the child, and being male. Some of these risks reflect genetic influences and others environmental. I will highlight three areas as background to our consideration of attachment and the vices. First, it is noteworthy that my list of established risk factors does not include attachment status. That is because the evidence is inconsistent. A meta-­analysis in 2010 concluded that insecure and disorganized attachment were associated with externalizing behaviours, which include aggressive and disruptive behaviours and inattentiveness and hyperactivity, in boys but not in girls, but that the findings varied considerably across studies (Fearon et al. 2010). By far the largest study of attachment and development, the NICHD Study of Early Child Care, with 1,149 children assessed in the Strange Situation at age 15 months, found a complex association between disorganized attachment and externalizing symptoms that varied by age of child, and that consistent risk for externalizing behaviours was seen only in a small group of 13 boys who also had been exposed to high levels of family and economic risk (Fearon and Belsky 2011). In other words attachment status in itself may not predict problems, but instead create vulnerability in the presence of other risks. Secondly, it is likely that there are several routes, or pathways, into aggressive and disruptive behaviours and hence also different underlying processes contributing to them. This poses a challenge and at the same time offers new op­por­tun­ ities in relation to the possible role of attachment. The challenge is that if there are different pathways to antisocial behaviour, it is unlikely that attachment processes will be relevant to all of them and hence provide a general account of the vices.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

Attachment, the Virtues, and the Vices  79 The opportunity arises because attachment processes may contribute to some and not other pathways, so the inconsistent findings and weak effects may arise because stronger effects in particular pathways are diluted by the inclusion of pathways in which there is no effect. We know little about this because research into attachment has not so far examined these different pathways. One of the most promising dimensions for the examination of pathways is ‘callous unemotional’ (CU) traits, which is the way the indifference to others’ distress found in adult psychopathy has been characterized in children. The aggression of children with high CU traits has been characterized as lacking in emotion, in contrast to aggression in other children which is often accompanied by anger. There may be a stronger genetic contribution to CU traits than to other risks for child aggressive and disruptive behaviours (Viding et al.  2008), and positive parenting may be important to reducing CU traits (Wagner et al. 2015), whereas by contrast other forms of behaviour problems may be more likely to arise in infants who are readily angered and also exposed to harsh or physically violent parenting (Hill 2002). Thirdly, aggressive and disruptive behaviours in childhood, and violence and criminality in adult life, are much more prevalent in males than females. The ­reasons for this are poorly understood. There are at least three possible explanations: first that the same risks operating in the same way account for antisocial behaviours in males and females, but males are exposed to more of them; secondly that the same risks operate in different ways in males and females; and thirdly that there are different risks. Sex differences in selection processes in evolution may be relevant. Ever since Darwin wrote On The Origin of Species sexual selection has been central to the theory of evolution depending ‘not on a struggle for existence, but on a struggle between the males for possession of the females; the result is not death to the unsuccessful competitor, but few or no offspring’ (Glover and Hill 2012]. While there are numerous differences across species, Darwin’s generalization has been supported that males, in contrast to females, at certain times give priority to competition and aggression with other males to maximize reproductive opportunities. Females by contrast give priority to care of offspring (Zahn-­Waxler 1993). This would imply that males are primed for aggression, and that socialization depends on inhibitory processes, which may be emotional (such as fear of punishment), social (such as responsiveness to other’s expressions of fear), or physiological (such as increased autonomic or endocrine reactivity). In females by contrast aggression may arise not from failures of inhibitory processes but by heightened activation, particularly of anger. In our research we have identified several examples of sex differences that may be relevant to these processes, including that prenatal maternal anxiety and infant low birth weight are associated with increased vagal (autonomic) reactivity in girls and reduced in boys (Tibu et al. 2014), and that prenatal maternal cortisol is associated with increased emotionality in girls and decreased in boys (Braithwaite et al. 2017). We have also

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

80  Jonathan Hill found that increased vagal reactivity in girls, but decreased reactivity in boys, is associated with higher disruptive behaviour problems, and that decreased cortisol reactivity in boys, but not in girls, is associated with higher levels of aggression (Vidal-­Ribas et al.  2017; Wright et al.  2019). Recent findings on associations between vagal and cortisol reactivity and child disruptive behaviour problems are consistent with the idea that in boys poorly regulated behaviours arise from underactivity of inhibitory or regulatory processes, while in girls they are an outcome of emotional and physiological over-­reactivity.

5.  Why might Early Secure Attachment Lay the Foundations for the Virtues and Insecure or Disorganized Attachment Create Risk for the Vices? The answer to this question depends considerably on which concept of attachment, and which characterization of the virtues and the vices, we are talking about. If we take the concept behind the majority of attachment research with infants and young children, namely the regulation of distress or worry with a specific caregiver, and the concept of the virtues and vices as characteristic of the person in general and evident across multiple situations, there are immense distances to be covered. We have to explain why patterns of behaviours centred on one type of emotion, and in one specific relationship, may influence behaviours linked to multiple emotions, and shown not only across a diversity of relationships, but also with people with whom a person has no relationship. The argument would run perhaps that the child’s early capacity to trust a caregiver enough to open themselves to being soothed by them underpins their ability and willingness to respond empathically to others. No doubt other accounts can be given. But all will have to deal with the question of how attachment security assessed within a very demarcated arena of human behaviour may be linked with the very broad landscape of character. How do the specific context, of threat to the child; of the specific emotions of distress and worry; of the specific relationship, one caregiver and not another; and the specific interpersonal circumstance, close relationship and not social processes more generally, translate into the multiple cap­ abil­ities of the virtuous person and the multiple struggles of the antisocial person? The question is not only whether attachment security might achieve this but why it might be the best candidate when considered against others. Why not for ex­ample, rather than starting with the ability to turn to others to be soothed, go straight to the early capacity to respond empathically to others, or to be cooperative? We know that children vary considerably in these capabilities during the second year of life and that these variations are predictive of later empathy and cooperation (Kochanska et al. 2010).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

Attachment, the Virtues, and the Vices  81 Furthermore different early childhood processes probably contribute to the virtues in different ways. Attachment-­oriented research has focused mainly on the role of parental, often maternal, sensitivity, which is the extent to which the parent is attentive and responsive to the child’s cues. Greater sensitivity, especially to infant distress, is modestly associated with secure attachment, and might also be a good candidate for promoting children’s own sensitive responding to others. However there are other dimensions of parenting that might be expected to impact on a child’s emerging participation in social structures. For example children who are assigned more regular household chores are more likely to show spontaneous concern for others in the family, including helping, sharing, comforting, defending, and demonstrations of concern (Grusec et al.  1996). Furthermore different aspects of parenting may contribute in different ways. All positive aspects of parenting may expose the child to parental values by keeping the child in the parent’s vicinity and attentive to the parent’s behaviour, but each may also affect differently the child’s willingness to adhere to those values. Attachment, for example, may be particularly important in its impact on emotion regulation. Mutual compliance may foster acceptance under conditions in which compliance is more easily given, whereas warmth may lead to strong identification with all aspects of parental actions and values, regardless of the extent of self-­sacrifice required (Grusec et al. 2000). As we noted earlier we may find a way of bridging this divide, either by broadening what we refer to by attachment, or by narrowing the characterization of character. The broadening of attachment commonly occurs in two ways. First the narrow definition of attachment is retained together with the expectation that it provides an index of, or may influence, broader processes, for example of the child’s capacity to engage in playful interactions with, or to learn from, a parent. There is good evidence that they are associated but not that attachment status influences these capabilities. Second the concept is taken to be broader, so that secure attachment refers to a confidence in the parent, in most or all aspects of the relationship, and in relation to diverse emotions, anger as well as distress, sharing, joking, planning, learning, and so on. Specifically in relation to behaving in the right way, Bretherton et al. (1997) suggested that a secure relationship with a stronger, wiser attachment figure may make children more likely to regard parental control as legitimate: they trust the attachment figure to guide them in appropriate directions and they are confident that the behaviour expected by that figure is in their own self-­interest. While there may be concerns about narrowing the concept of character, links with specific processes implied in Aristotle’s description of character excellence can be envisaged. For example ‘doing the right thing’ implies an ability to appraise a challenging situation accurately as a basis for action. The secure infant gives the impression of being able to do this, appraising the separation as a threat, without

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

82  Jonathan Hill ignoring it as seems to be the case for the avoidant infant, and without overestimating it, as may be the case for the resistant infant. Consistent with this idea we found that children who had been insecure, mostly avoidant, at 12 months, scored lower on an ‘intentionality’ scale in responding to a threatening situation presented using dolls when they were 5 years old (Hill et al. 2008). The intentionality scale captured the extent to which portrayed doll behaviours were systematically linked to the threat and social processes between the characters. Put another way, the previously avoidant children were more likely to drop to the ‘physical stance’ (Dennett 1989) hence appraising the interpersonal situation less accurately. Whatever the concept of attachment that we employ, our account will have to do justice to the many influences on development, and the nature of the interplay between children’s individual characteristics and their environments. Twin studies of indices of the virtues and the vices, such as empathy, prosocial behaviours, aggression, and personality disorders, all find substantial genetic effects. The extent to which these operate independently of parental influences and other environmental factors remains unclear. Findings of gene by environment inter­ actions, still controversial, provide examples of ways in which genetic effects may be conditional on the environment. In the seminal paper by Caspi and colleagues (Kim-­Cohen et al. 2006), a variant of a gene associated with regulation of a brain transmitter associated with violence was associated with increased antisocial behaviours in adults, but only where they had been maltreated. In the absence of maltreatment there was no effect. Other studies have found not only that such gene variants are associated with a poorer outcome in the presence of an un­favour­able environment, but with a markedly good outcome in a supportive environment. Thus the genetic variant does not confer vulnerability, rather it is associated with an increased responsiveness to the environment. The implications of this ‘differential susceptibility’ phenomenon are that there may be some individuals who are highly influenced by the environment ‘for better or for worse’ while others, for whom the environment is much less important, don’t turn out badly in unfavourable environments, nor so well in good environments.

6.  Conclusions Attachment theory and findings initially provided an important counterweight to the prevailing behavioural theories of the 1950s and subsequently have performed a similar function in relation to powerful genetic and reductionist biological perspectives of recent years. It also has identified a key role for parents over development. Equally as articulated by behavioural theories, learning is crucial to the development of prosocial and antisocial behaviours, and there are substantial

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

Attachment, the Virtues, and the Vices  83 genetic contributions to all human behaviours, so attachment-­based hypotheses have to take account of these influences. Similarly the roles of parents extend well beyond those of attachment processes as envisaged in current attachment theory (Hill et al. 2003) also calling for integration of attachment and other perspectives. As we have seen, Aristotle’s characterization of the virtues and vices extends well beyond the presence or absence of good behaviour, to a comprehensive account of personality and of the individual in their social context. Such a characterization poses challenges to empirical study, yet perhaps also can provide clues to the way we may in the future develop an integrated model of the role of early childhood experiences in relation to the multifaceted task of moving from childhood and into adult life.

References Ainsworth, M.  D.  S., Blehar, M.  C., Waters, E., and Wall, S. 1978. Patterns of Attachment. New York: Erlbaum. Bowlby, J. 1944. ‘Forty-four juvenile thieves: their characters and home-life’. International Journal of Psycho-Analysis 25, 19–53. Bowlby, J. 1982. ‘Attachment and loss: retrospect and prospect’. American Journal of Orthopsychiatry 52(4), 664–78. Braithwaite, E. C., Pickles, A., Sharp, H., Glover, V., O’Donnell, K. J., Tibu, F., and Hill, J. 2017. ‘Maternal prenatal cortisol predicts infant negative emotionality in a sexdependent manner’. Physiology & Behavior 175, 31–6. Bretherton, I., Golby, B., and Cho, E. 1997. ‘Attachment and the transmission of values’. In J. E. Grusec and L. Kuczynski (eds), Parenting and Children’s Internalization of Values: A Handbook of Contemporary Theory, 103–34. New York: Wiley. Brown, E. W., and Harris, T. O. 1978. Social Origins of Depression. London: Tavistock. Côté, S. M., Vaillancourt, T., Barker, E. D., Nagin, D., and Tremblay, R. E. 2007. ‘The joint development of physical and indirect aggression: predictors of continuity and change during childhood’. Development and Psychopathology 19(1), 37–55. Dennett, D. C. 1989. The Intentional Stance. Cambridge, MA: MIT Press. Fearon, R. P., Bakermans-Kranenburg, M. J., Van IJzendoorn, M. H., Lapsley, A. M., and Roisman, G. I. 2010. ‘The significance of insecure attachment and disor­gan­iza­ tion in the development of children’s externalizing behavior: a meta-analytic study’. Child Development 81(2), 435–56. Fearon, R. P., and Belsky, J. 2011. ‘Infant–mother attachment and the growth of externalizing problems across the primary-school years’. Journal of Child Psychology and Psychiatry 52(7), 782–91.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

84  Jonathan Hill Fraley, C., and Thompson, R. 2002. ‘Attachment stability from infancy to adulthood: meta-analysis and dynamic modeling of developmental mechanisms’. Personality and Social Psychology Review 6(2), 123–51. Gilbey, E. 1994. The Lady: The Life and Times of Winnie Mandela. London: Vintage. Glover, V., and Hill, J. 2012. ‘Sex differences in the programming effects of prenatal stress on psychopathology and stress responses: an evolutionary perspective’. Physiology & Behavior, 106(5), 736–40. Golm, D., Maughan, B., Barker, E. D., Hill, J., Kennedy, M., Knights, N., Kreppner, J., Kumsta, R., Schlotz, W., Rutter, M., and Sonuga-Barke, E. J. 2020. ‘Why does early childhood deprivation increase the risk for depression and anxiety in adulthood? A developmental cascade model’. Journal of Child Psychology and Psychiatry 61(9), 1043–53. Grusec, J. E., Goodnow, J. J., and Cohen, L. 1996. ‘Household work and the development of concern for others’. Developmental Psychology, 32(6), 999. Grusec, J. E., Goodnow, J. J., and Kuczynski, L. 2000. ‘New directions in analyses of parenting contributions to children’s acquisition of values’. Child Development, 71(1), 205–11. Hamlin, J. K., Wynn, K., and Bloom, P. 2007. ‘Social evaluation by preverbal infants’. Nature 450(7169), 557–9. Hill, J. 2002. ‘Biological, psychological and social processes in the conduct disorders’. Journal of Child Psychology and Psychiatry 43(1), 133–64. Hill, J., Fonagy, P., Safier, E., and Sargent, J. 2003. ‘The ecology of attachment in the family’. Family Process 42(2), 205–21. Hill, J., and Maughan, B. 2015. ‘Conceptual issues and empirical challenges in the disruptive behavior disorders’. Rutter’s Child and Adolescent Psychiatry (6th edn), 41–52. Chichester: John Wiley and Sons. Hill, J., Murray, L., Leidecker, V., and Sharp, H. 2008. ‘The dynamics of threat, fear and intentionality in the conduct disorders: longitudinal findings in the children of women with post-natal depression’. Philosophical Transactions of the Royal Society of London B: Biological Sciences 363(1503), 2529–41. Hill, J., Sharp, H. M., Wright, N. J., and Pickles, A. In preparation. ‘Increased vul­ner­ abil­ity to behavioural problems at 2.5 years in secure infants exposed to maternal criticism at 14 months’. Kim, S., and Kochanska, G. 2017. ‘Relational antecedents and social implications of the emotion of empathy: evidence from three studies’. Emotion 17, 981–92. Kim-Cohen  J., Caspi, A., Taylor, A., Williams, B., Newcombe, R., Craig, I.  W., and Moffitt, T. E. 2006. ‘MAOA, maltreatment, and gene–environment interaction predicting children’s mental health: new evidence and a meta-analysis’. Molecular psych­iatry 11(10), 903–13.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

Attachment, the Virtues, and the Vices  85 Kochanska, G., Koenig, J. L., Barry, R. A., Kim, S., and Yoon, J. E. 2010. ‘Children’s conscience during toddler and preschool years, moral self, and a competent, adaptive developmental trajectory’. Developmental Psychology 46(5), 1320. Kraut, R. 2018. ‘Aristotle’s ethics’. In E.  N.  Zalta (ed.), The Stanford Encyclopedia of Philosophy (Summer 2018 Edition), McElwain, N. L., and Booth-LaForce, C. 2006. ‘Maternal sensitivity to infant distress and nondistress as predictors of infant–mother attachment security’. Journal of Family Psychology 20(2), 247. Malmberg, L. E., Lewis, S., West, A., Murray, E., Sylva, K., and Stein, A. 2016. ‘The influence of mothers’ and fathers’ sensitivity in the first year of life on children’s cognitive outcomes at 18 and 36 months’. Child: Care, Health and Development 42(1), 1–7. Nathan, R., Rollinson, L., Harvey, K., and Hill, J. 2003. The Liverpool Violence Assessment: an investigator-based measure of serious violence. Criminal Behaviour and Mental Health 13(2), 106–20. Robins, L. N. 1966. Deviant Children Grown Up: A Sociological and Psychiatric Study of Sociopathic Personality. Baltimore: Williams and Wilkins. Roth-Hanania, R., Davidov, M., and Zahn-Waxler, C. 2011. ‘Empathy development from 8 to 16 months: early signs of concern for others’. Infant Behavior and Development 34(3), 447–58. Rutter, M. 1972. Maternal Deprivation Reassessed. Harmondsworth: Penguin. Tibu, F., Hill, J., Sharp, H., Marshall, K., Glover, V., and Pickles, A. 2014. ‘Evidence for sex differences in fetal programming of physiological stress reactivity in infancy’. Development and Psychopathology 26(4:1), 879–88. Vidal-Ribas, P., Pickles, A., Tibu, F., Sharp, H., and Hill, J. 2017. ‘Sex differences in the associations between vagal reactivity and oppositional defiant disorder symptoms’. Journal of Child Psychology and Psychiatry 58(9), 988–97. Viding, E., Jones, A. P., Paul, J. F., Moffitt, T. E., and Plomin, R. 2008. ‘Heritability of antisocial behaviour at 9: do callous-unemotional traits matter?’. Developmental Science 11(1), 17–22. Wagner, N. J., Mills-Koonce, W. R., Willoughby, M. T., Zvara, B., and Cox, M. J. 2015. ‘Parenting and children’s representations of family predict disruptive and callousunemotional behaviors’. Developmental Psychology, 51(7), 935. Warneken, F., and Tomasello, M. 2009. ‘The roots of human altruism’. British Journal of Psychology 100(3), 455–71. Wright, N., Hill, J., Pickles, A., and Sharp, H. 2019. ‘Callous-unemotional traits, low cortisol reactivity and physical aggression in children: findings from the Wirral Child Health and Development Study’. Translational Psychiatry 9(1), 1–9.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

86  Jonathan Hill Wright, N., Hill, J., Sharp, H., and Pickles, A. 2018. ‘Maternal sensitivity to distress, attachment and the development of callous-unemotional traits in young children’. Journal of Child Psychology and Psychiatry 59(7), 790–800. Zahn-Waxler, C. 1993. ‘Warriors and worriers: gender and psychopathology’. Development and Psychopathology 5(1–2), 79–89. Zahn-Waxler, C., Robinson, J. L., and Emde, R. N. 1992. ‘The development of em­pathy in twins’. Developmental Psychology 28(6), 1038.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

4

The Evolved Nest, Virtue, and Vice Darcia Narvaez

1.  The Setting There are over 7.5 billion people on the earth. Some people take it as a sign of evolutionary success. But imagine if at the next birthday party you attend, the tallest person took everything—all the cake, punch, and presents for himself. It would not be much of a party. Similarly, when one species takes over a biocommunity for itself, it’s not much of a community. This is what the dominant industrialized capitalist culture has done on the earth. This dominant culture is behaving like a weed species. Weed species appear for a while but they disappear when a more cooperative species comes along that fosters the well-­being of the biocommunity (Naess and Rothenberg 1989). The dominant human culture has won the species race, its apparent aim with its propaganda of human separation from and superiority to Nature (Moore 2016). Only there is no race. And it has been lost. But isn’t the whole goal of evolution to proliferate? No. Evolution brings about greater and greater diversity of species, not more of one species. ‘Endless forms most beautiful’ (Darwin 1871). Diversity is critical for the flourishing of an ecological community, with each species having its niche within that community. Darwin’s (1859) theory of natural selection describes how new species evolve and diverge. The theory does not discuss how a plethora of species get along day to day. Getting along requires a balanced cooperation within narrow parameters. Too much imbalance is abnormal and problematic. For example, there are trillions of microorganisms in your body that keep you alive. When a particular bacterium proliferates in your body, you get sick and you can die due to that imbalance. Imbalance in a biocommunity is a problem. Isn’t evolution about competition? No. Evolution is largely about conservation of prior adaptations from one generation to the next. Most things we inherit do not change from generation to generation. Human bodies are made up of adaptations hundreds of millions of years old (like the spine) and each human body is a community of cooperation carrying 90–9 per cent of non-­human genes of those trillions of microorganisms keeping it alive (Dunn  2011). But today genetic competition seems like an inadequate marker of ‘success’, as humans are Darcia Narvaez, The Evolved Nest, Virtue, and Vice In: Attachment and Character: Attachment Theory, Ethics, and the Developmental Psychology of Vice and Virtue. Edited by: Edward Harcourt, Oxford University Press. © Darcia Narvaez 2021. DOI: 10.1093/oso/9780192898128.003.0005

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

88  Darcia Narvaez overwhelming the planet’s biocommunities, throwing everything out of balance, and destroying the diversity that has been the outcome of evolutionary processes. Biocommunity balance within certain parameters is ‘normal’. On a day-­to-­day basis, the natural world evolved to be deeply mutualistic and cooperative (Margulis 1998). Birds warn other species of predators. Animals share water holes with their predators (when the predators are not hungry). Forests have community-­ oriented trees who share nutrients (Wohlleben  2016). Older (‘mother’) trees feed the young of other plant species nearby through their root systems (Wohlleben 2016). Don’t genes predict human behaviour? No. Generally, the emphasis on genes is misleading for psychology (Joseph 2013). Important distinctions must be made. Functional adaptation (or ‘experience-­ adaptive programming’, Marshall and Kenney 2009) within a lifetime is not the same as evolutionary adaptation (in the genetic fitness sense) (Narvaez Gettler, et al. 2016). A circus elephant adapts to learning tricks to entertain the audience so that he gets fed and not further abused, but this is not evolutionary adaptation. A child who learns to be insecurely attached from growing up with emotionally distant caregivers is functionally adaptive, as a need to form some kind of attachment appears innate, but this is not evolutionarily adaptive. A child having a baby at 9 years old does not represent an evolutionary phenotype but shows signs of an environment gone awry— most notably an environment poisoned by heavy metals or endocrine disruptors like plastics (Özen and Darcan  2011). Evolutionary adaptation is visible only retro­spect­ive­ly from a distance because an individual must outcompete rivals over multiple generations (which takes the survival, thriving, and reproduction of multiple generations) (Lewontin 2010). One must distinguish between robustness and plasticity, though they are complementary and difficult to separate (Bateson and Gluckman 2011). Robustness refers to how members of a species grow and develop in much the same way regardless of the environment. For humans it means that, physically, a child learns to walk, talk, grow taller, and so forth. Plasticity refers to the malleability, especially during early development, as the environment shapes the direction of growth (i.e. children who are malnourished do not grow as tall). But the effects of undercare go much deeper, beyond the physical visible traits to neurobiological structures like neurotransmitter function and number, stress, and immune functions. Humans are particularly immature at birth and particularly plastic in early life (Gómez-­Robles et al. 2015), not only physiologically but in terms of psych­ ology and sociality which, I suggest in my work, ground morality and virtue. Insecure attachment places the individual in a different trajectory, canalizing less fitted social behaviour and diminished cooperation (Atzil et al. 2018), as sensitive periods pass (Knudsen  2004), based on unmet experience expectancies (Greenough, Black, and Wallace 1987). One other thing about the focus on genes. Few physical disorders and nearly no psychological disorders are predicted exclusively by single genes or even

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

The Evolved Nest, Virtue, and Vice  89 networks of genes (Carey 2011). The lack of a main effect of genes with the ­finding that all psychological outcomes are the result of interactions is called the ‘gloomy prospect’ by behaviour geneticists (Turkheimer 2000), a field whose credentials have been challenged on the grounds of lack of generalizability and replicability (Ho 2010). It turns out that epigenetics, the effect of experience on genetic expression, is the big story, especially for humans (Zhang and Meaney 2010). Why bring up evolution, genes, and gene-­centred theories? Because they are often used to fatalistically argue that the way things are now are as they should be, as they evolved to be (a naturalistic fallacy if there ever was one!). All sorts of rationales are created to justify things as they are. This is a sign of shifted baselines for what is considered normal. Why am I bringing these things up in the context of a discussion of virtue and vice? Because signs all around indicate that we have created vicious societies and a vicious human nature. Created, not inherited. We have fostered people unable to fit into the biocommunity as fellow members and then rationalized the dis­ ordered result with anthropocentric fatalistic theories like selfish-­gene theory. Here we are today with half of all species identified fifty years ago extinct in the intervening years due to human activity (World Wildlife Fund 2014). Earth systems are breaking down. Climate instability is the new normal as the polar ice caps melt (IPCC 2013). Massive ecological disruption from human activity occurs virtually everywhere on the planet (Díaz et al.  2019; Millennium Ecosystem Assessment  2005). Poisoning of soil, air, water, animal bodies increases by the day. Blood in baby umbilical cords has dozens of pollutants (Houlihan et al. 2014). While economic wealth has burgeoned, social and ecological health are sacrificed in exchange (Korten 2015). For example, the health and well-­being of citizens in the wealthiest nation on the earth (USA) is on the decline, with mental illness increasing (in real numbers) (USDHHS 1999), citizens under age 50 at a health disadvantage compared to sixteen other advanced nations (NRC 2013), and life­ spans shrinking (Xu et al. 2016). Moreover, it is often assumed that human nature is something we are born with. Gene-­centric evolutionary theorists typically assume that ‘selfish’ genes make humans ‘naturally’ selfish (Dawkins 1976). But there are other evolutionary theories. Evolutionary systems theory notes a rich set of inheritances beyond genes, such as culture, ecology, self-­organization, and the developmental system or nest (Oyama, Griffiths, and Gray 2001). These too influence the type of nature an individual and culture exhibit.

2.  The Diagnosis What has gone wrong? To begin to answer this question we must understand what kind of animal we are and what brings about our flourishing. Ethology shows us the importance of attachment and the types of neurobiological

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

90  Darcia Narvaez structures, along with secure attachment, that nurturing parenting promotes (Eibl-­Eibesfeldt 1989). Attachment represents an ‘internal working model’ that is not only conceptual but an engraving on the neurobiology of the body (van der Kolk 2014). It helps to understand how we are different from other hominids. We have greater immaturity at birth (only 25 per cent of the adult-­sized brain developed; we should be in the womb another eighteen months in terms of, for example, bone development) and we have the longest maturational schedule (Dettwyler  1997; Montagu 1968; Trevathan 2011). As a result, many physiological and psychosocial systems develop postnatally, dynamically in response to experience. As noted, when developmental psychology’s use of genetic theory fails to make the distinction between functional and evolutionary adaptation it also confuses species-typical developmental systems with those that are species atypical (based on an evolutionary timeline). The baseline for child development and the in­ herit­ ed developmental system have been lacking. So, whereas attachment ­security scores are signals of whether things are going well enough for the child’s psychobiosocial development, they don’t take into account species-­typical and atypical developmental systems. In our ancestral environment, insecure attachment would have been a death knell as its comorbidity is social impairment and distorted social and self-­cognition. Highly dependent on interpersonal cooperation, our ancestral context did not have the extra infrastructures for keeping ­uncooperative people alive that we have today.1 The range for developmental outcomes is narrowed in civilized societies that do not provide the evolved developmental system. Thus attachment researchers, assessing individuals who are not provided with our species-­typical developmental system, fail to measure optimal development. The securely attached person appears soothable, able to deactivate defensive emotions and behaviours as well as ‘seeking, doing, achieving, and acquiring’ (Gilbert  2005: 28). However, as Colwyn Trevarthen (2005) points out, warmth attachment—which is what attachment measures assess—is less adequate than companionship attachment, which aligns better with species-­typicality. Companionship attachment requires experiences of playing ‘together-­with’ multiple responsive caregivers, sharing intentions, interests, and affective appraisals. Many capacities are shaped preverbally (Stern  2010), as noted by anthropologists of the ‘preconquest consciousness’ (Sorenson  1998). Such care fosters a child’s active curiosity, a confident self-­ consciousness, and the ability to take on independent acting and thinking. Allan Schore (2019) concurs with the significance of these early relationships: ‘Regulated and synchronized affective interactions with familiar, predictable primary caregivers create not only a sense of safety but also a positively charged curiosity, wonder, and surprise that fuels the burgeoning self ’s exploration of 1  For contestation of the claim that secure attachment is species-­typical, see e.g. Schaubroeck (this volume).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

The Evolved Nest, Virtue, and Vice  91 novel socioemotional and physical environments. This ability is a marker of ­adaptive infant mental health’ (p. 10). Secure attachment assessments collapse this more optimal type of attachment with warmth attachment, perhaps assessing only the ground floor for development, not the capacities of optimal species-­ typical development. You might wonder how we tell exactly what young children need to optimize normal development. Actually, it is not difficult but takes some interdisciplinary investigation. As animals, humans need warmth, nourishment, and protection. As mammals, we have needs for extensive affection, breastfeeding, and self-­ directed play (Panksepp 1998). As social mammals, we need extensive bonding and community support (Hrdy  2009). As human beings, our development and functioning are optimized from intersubjectivity with multiple responsive others, immersion in communal rituals and cultural narratives, and apprenticeship in life activities (Trevarthen 2005; Shepard 1998). All these comprise an extensive needs list for an animal that matures slowly and is biosocially shaped. Like all animals, humans have a nest for their young that matches up with maturational needs. It was first named the ‘hunter-­gatherer childhood model’ from identifying the common characteristics of child raising among hunter-­ gatherers worldwide, most of which emerged over 30 million years ago with the social mammals (Konner 2005). Humanity’s evolved developmental niche (EDN) is much more intensive and lengthy than for any other animal. Humans are more sensitive to experience than other animals because their postnatal experience epigenetically influences neurobiological development more so than for their hom­ inid cousins (Gómez-­Robles et al.  2015). The EDN2 includes soothing birth, extensive breastfeeding, responsive care, plenty of affection (and no punishment), multiple adult responsive caregivers, free play, and positive social support. When these are not provided, it represents a broken continuum of support and we should not be surprised that various forms of dysregulation result that promote weakness of the will or what looks like vice (Niehoff 1999). Just as a dog’s temperament is influenced by the mothering received in early life (Foyer, Wilsson, and Jensen 2016), so too is human temperament and much more so because of vast immaturity and scheduled postnatal growth (Schore 2003). How can we say that there is a typical niche when there is so much variability today? Those who condone cultural variability for child well-­being are not looking at critical features. For example, Levine (Levine et al. 1994) assumed that talking to one’s baby was vital for child well-­being. But in comparing Gusii mothers with Boston mothers, he found little talking among the Gusii. He was surprised to find out that in adolescence the Gusii children were thriving. He failed to 2 AKA: hunter-­ gatherer childhood model (Konner 2005); evolved developmental niche (Narvaez  2014; Narvaez, Gleason, et al.  2013; Narvaez, Wang, et al. 2016; Narvaez, Woodbury, et al., 2019)

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

92  Darcia Narvaez notice the non-­verbal sensorially rich, responsive, affectionate care the Gusii ­children received throughout life. He failed to attend to the evolved nest of care that young children in particular need, which centres more on non-­verbal communication, affectionate carrying and movement, rather than on talking (Hewlett and Lamb  2005). Lancy (2015) too has glossed over the earliest experiences of young children, endorsing the view that any parenting will do. These researchers fail to notice the common practices that anthropological observations have made for baby care and that we in my lab are now beginning to study. Just because in the last 10,000 years or so (1 per cent of human genus existence) the evolved nest has become degraded in civilized, then industrialized, nations does not mean species typicality has shifted. The nest is still provided in societies around the world, as it was for 99 per cent of human genus history.3 Neurobiological studies demonstrate the effects of evolved nest components on human functioning and disposition. Here are a few examples of a growing literature. Skin-­to-­skin contact at birth facilitates the release of oxytocin in mother and infant, reducing childbirth stress (Bystrova, Widstrom, Matthiesen, et al. 2013). When caregivers are warm and responsive to needs, a baby’s vagus nerve will become myelinated, influencing health and social capacities (Porges 2011). Harry Harlow (1958) studied the effects of maternal touch deprivation on mammalian brains (monkeys), finding long-­lasting effects on self-­regulation and sociality. Further studies in humans indicate that maternal touch decreases cortisol release, which benefits the immune system as cortisol kills immune cells (Field and Hernandez-­Reif 2013). A lack of breastfeeding in the first week of life is related to greater depression and withdrawal as well as abnormal reflexes (Hart et al. 2003, 2006) and at three months, breastfed infants show greater myelination than formula-­fed infants (Deoni et al.  2013). Allomothers are critical supports for mother’s attention to her child (Hrdy 2009). Play fosters the growth in the orbitofrontal cortex and its linkages to other parts of the brain which take a large role in decision-­making (Pellis and Pellis 2009). Although animal studies show the importance of each EDN component for normal development and sociality, neuroscientific and developmental psychological studies of humans are also accruing evidence (Atzil et al. 2018). Most studies focus on maternal responsiveness, a combination of comforting behaviours (calming touch) and honest communicative behaviour that leads to secure attachment (Crittenden  1995; Easterbrooks and Goldberg  1990). For example, keeping mother and child together after birth leads to greater self-­regulation a year later (Bystrova et al.  2009). A mutually responsive relationship between 3  I find that when I give lectures about these things Western parents are almost always offended by hearing about the evolved developmental niche, whereas those from collectivist societies, as in Southeast Asia, take for granted the list of characteristics as the normal way to raise children and are astounded that a parent would not provide the nest to their young.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

The Evolved Nest, Virtue, and Vice  93 mother and child leads to secure attachment and greater capacities for prosocial and cooperative behaviour (Eisenberg 1995; Kochanska 2002; Zahn-­Waxler and Radke-­Yarrow 1990). Touch experience influences genetic expression and development of various systems such as oxytocin and vasopression which are related to social bonding (Ardiel and Rankin 2010; Carter and Porges 2013) and spanking increases social aggression over the long term (Gershoff 2013). In one of our lab’s publications, maternal touch attitudes and behaviour were examined in a longitudinal sample of at-­risk and middle-­class mother–child dyads (Narvaez, Wang, et al. 2019). Lack of negative touch was positively related to a child’s concurrent behavioural regulation at both 18 and 30 months. Maternal negative touch behaviour at 18 months was positively related to children’s externalizing problems at 24 and 36 months and negatively related to 36-­month ratings of social competence. By 30 months of age, maternal avoidance of punishing touch related significantly to all of the child outcomes except internalizing problems. Our lab’s studies look at multiple nest components always controlling for responsiveness to see if additional nest components matter. For example, breastfeeding length has been positively related to the development of 3-­year-­old children’s inhibitory control and conscience (guilt and concern after wrongdoing), even after controlling for maternal responsivity (Narvaez, Gleason, et al. 2013). This is not surprising when breastfeeding’s effects on self-­regulation and brain development have been shown to take place in a matter of weeks, as mentioned above (Deoni et al.  2013; Hart et al. 2003). In the aforementioned longitudinal sample of at-­risk and middle-­class mother–child dyads (Narvaez, Wang, et al. 2013), maternal social support positively correlated with child cooperation at 18 and 30 months, child social competence at 24 months, and reduced aggressive behaviour at 18 months. The aim of our lab is to assess the linkages between early experience—using the baseline of the EDN—well-­being and moral capacities, not only empathy and perspective-­taking, but species-­typical ethical mindsets seen in our ancestral contexts (Narvaez 2008, 2014): ethical engagement (flexible relational attunement to others) and communal imagination (rooted in ethical engagement, using abstracting capabilities for practically wise planning and action). In contrast, dispositional self-­protectionist ethics (social opposition, social withdrawal) are indicators of a stressed neurobiology. In a three-­nation study, EDN provision in the past week (self-­directed play, affection, no corporal punishment, family togetherness) was related to multiple child outcomes even after controlling for parental income, age, education, responsiveness, and child gender (Narvaez, Woodbury, Gleason, et al.  2019). When latent variables for child outcomes were created— moral socialization (self-­regulation, internalized conduct), social maladaptation (e.g. social distrust, social opposition, misbehaviour), and social thriving (e.g. social engagement, well-­being)—structural equation models indicated that EDN

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

94  Darcia Narvaez provision significantly predicted, beyond control variables, all three latent variables in the USA and social thriving in all three samples. In a study of adults, path analyses of higher levels of reported evolved nest (affection, play, family togetherness) experienced in childhood led to higher ­levels of secure attachment, lower internalizing (anxiety and depression), greater perspective-­taking (rather than personal distress), and greater commitment to ethical engagement rather than self-­protectionist morality (Narvaez, Wang, and Cheng 2016). In a study with over 1,500 adults, several measures of morality were examined along with secure and insecure attachment (Narvaez and Hardy 2016). Higher secure attachment was related to lower protectionism which was linked to higher integrity. The evolved nest provision can be described as love in action where the manner of treating the child is vital—not just responsiveness to conversational bids, but actual physical contact, comfort, play, and breast milk. These are embodied love in action. In early life, when neurobiological, social, and moral foundations are being established, the continuum of feeling synchronous with mother and others, bonded to the fabric of life and embedded in with natural processes, is critical. No other animal intentionally breaks that continuum. What happens with a broken continuum? Do this to other animals, even separating a mammalian newborn from its mother for an hour a day after birth, and you get abnormality in hormonal systems (e.g. Kalinichev et al. 2002; Kanitz et al. 2004). Yet, as noted, humans are more massively influenced by postnatal experience because of their greater immaturity and plasticity. Raise a wolf in a human family and you still have a wolf. Raise a human in a wolf family and you get a wolf-­child—a human being that fits into the wolf world but never the human—because sensitive periods have come and gone. We can observe the vast difference in personality and culture between societies that provide the evolved nest and those that do not. Studies and accounts of nomadic foragers, the type of society in which humanity spent 99 per cent of its genus history, indicate a more virtuous nature (see below) than the vicious natures we now think are normal, such as selfish calculation of economic utility (Derber  2013). Renowned anthropologist Marshall Sahlins (2008) pointed out how ‘the Western illusion of human nature’ vastly differs from accounts in non-­ Western cultures. Darwin (1871) identified what he called the moral sense, whose components evolved through the tree of life and culminate in humans. The moral sense includes a set of characteristics that other animals display: social pleasure, empathy, memory function, concern for the opinion of others, and habit control to behave in socially appropriate ways. Darwin noted the moral sense in primitive societies, but found it weak in his own (British) society. Elsewhere I (Narvaez 2017, 2018) point out how the moral sense seems to be diminishing in the USA, where the evolved nest is the most degraded. Humans are dynamic complex systems that self-­organize according to e­ xperience, and whose initial conditions shape subsequent development and function,

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

The Evolved Nest, Virtue, and Vice  95 ­ arring later intervention. The nest components provide the type of stimulation b and support at the right times and in the right ways, likely for any genotype, bringing about well-­ functioning psychosocial neurobiology (Overton  2013). Human functions have multiple sensitive periods—where it is genetically determined that only certain kinds of stimuli affect a particular circuit or system. If those stimuli are missing, then that system does not develop properly, affecting later-­developing systems upon which it relies (Knudsen 2004). ‘Higher levels in a hierarchy depend on precise and reliable information from lower levels in order to accomplish their functions . . . Experience-­dependent shaping of high-­level circuits cannot occur until the computations being carried out by lower-­level circuits have become reliable’ (Knudsen 2004: 1414). Reliable development can presumably be fostered within the species-­typical nest.

3.  And now? Extending back into civilization’s undermining of child development, the USA has become one of the worst places in the world to raise a child. There are many reasons for this that interrelate, from the lack of support of parents (e.g. lack of parental leave, need for both parents to work for adequate income, lack of childcare facilities at work, poor quality childcare centres), manipulation of parents by corporate profitmakers like those who sell infant formula milk (Braden and Narvaez, in press), to widespread misunderstanding about what children need to flourish. Parents respond to a lack of community support by offering less support to their own children (Hrdy 2009). Thus, contrary to millions of years of evolutionary adaptation, parenting babies for detachment (avoidant attachment) has become a dominant force in the USA (a ‘taboo on tenderness’, Suttie 1935). This means that all sorts of moves are taken to force babies to be independent of parents (or parents of babies)—from baby-unfriendly hospitals, isolation in carriers and cribs, separate sleeping spaces, use of infant formula, and hours spent each day in stranger daycare. The implicit assumption is that babies do not need much mothering. Caregivers focus on meeting animal needs (nourishment, safety, warmth), with a blindness toward mammalian needs (extensive affection, breastfeeding, and self-­directed play), social mammalian needs (extensive bonding and community support), and human needs (intersubjectivity with multiple responsive others, immersion in communal rituals and cultural narratives, apprenticeship in life activities). And there is no informed baseline provided to parents for making appropriate judgements.4 4  A recent book, Cribsheet (Oster 2019) by a health economist uses research studies as the baseline for judging best parenting practice. Failing to attend not only to the limitations of research studies, especially in medical journals, but lacking awareness of basic child needs, she tells parents not to worry about sleep training as the research studies show it is safe (contrary to anthropological studies; McKenna et al.  1994). She claims that breastfeeding has no long-­term effects, disregarding studies

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

96  Darcia Narvaez The evolved nest provides an evolved baseline for optimizing normal d ­ evelopment. Any shift away from the evolved nest should be considered a risk factor for poor neurobiology, self-­regulation, social fittedness, sociality, and morality. Any shift away should require decades-­long longitudinal evidence that multiple relevant ­outcomes are not adverse. Truly, the evolved nest should be considered a human right because its degradation has long-­term harmful effects not only on the child but on society itself. Communities who don’t provide what a child evolved to need foster several types of mistrust: (1) mistrust in emotions and self-­signals: mistrust of the child’s body (and with undercare the body does not develop well and can seem like an alien that must be controlled/ignored); (2) mistrust of the parents/caregivers who inconsistently provide basic needs; (3) mistrust of the world—it is not a benign place but uncaring and threatening as the child regulatory capacities are set to be threat-­reactive. Sandler (1960) suggested that the early sense of danger grows into cynicism or anxiety, minimally into an adult with little trust or confidence in the self and the world, and maximally an adult with personality disorders. One has to wonder whether the widespread mistrust that has spread throughout the twentieth century in the USA during the time period when the evolved nest has been particularly undermined is related to the nest’s demise—along with the lack of confidence plaguing young people of all ages, not to speak of the deteriorating health of everyone under 50 mentioned earlier. Lack of nest provision means the evolved trajectory for the development of human nature is broken. Instead the child becomes insecure and unconfident in self, parents, and the world. Such a child will look like they need adult guidance to grow. They will look and be dysregulated which again adults will interpret as ‘the way babies are’ and use coercion to shape them differently. The child will forever after be ruled by external forces because the development of an internal compass was broken early on by the ignorance of adult caregivers. The nest shapes

such as those of behaviour regulation differences and myelinization differences in short-­term comparisons of formula vs exclusive breastfeeding and makes no mention of the species-­normal length of breastfeeding and its effects (Prescott  1996). Oster uses a science-­ as-­ manipulation approach. Economist E. F. Schumacher pointed out how science shifted from a focus on wisdom to a focus on manipulation, playing a large role in developing the world crises we face. The science of manipulation assumes separable parts, fragments of reality, and attempts to find ways to control those parts. It treats all the parts as purposeless and only focuses on what can be measured and manipulated. And so it ignores the reality of dynamic, shapeable life, the interconnection, interpenetration of all of life. Schumacher said: ‘Western civilization is based on the philosophical error that manipulative science is the truth, and physics has caused and perpetuated this error. Physics got us into the mess we are in today . . . Science is concerned primarily with knowledge that is useful for manipulation, and the manipulation of nature almost invariably leads to the manipulation of people’ (Capra  1997: 35). Unfortunately, this manipulative form of science has been applied to parenting for the last century or more. Most explicitly in psychology, it began with John Watson’s 1928 book for parents that advocated treating babies like college students—ignoring them a lot so they would get used to such treatment from the beginning. He encouraged a cold unresponsiveness towards babies that unfortunately is still with us.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

The Evolved Nest, Virtue, and Vice  97 sub­ject­iv­ity—the nature of how one sees the world (Narvaez 2014). The child will learn to seek hierarchy because it provides a scaffolding for living life with some feeling of security. Self-­protective aggression or withdrawal will be at the ready if the script is challenged. The personal narrative the child develops is one of deep flaws and so the individual will forever seek relief through cultural narratives that justify current dysregulation and that assure safety (and dominance) (Narvaez 2011).

4.  Virtue Development in Communities that Provide the Evolved Nest Species-­typical human development is apparent in cultures that maintain our prehistoric ways of living (small-­band hunter-­gatherers, in which the human genus existed for 99 per cent of its presence on the earth). Of course, we cannot return to nomadic foraging, but we can shape institutions and incentives to support provision of the evolved nest, especially to young children. When the nest is provided, it offers grounding for virtue development, which used to be normal and part of survival, but in a wider sense of living sustainably in one’s landscape. When it is provided, societies show egalitarian relational attunement with others (including other-­than-­humans) and use their imaginations for communal ends that include the welfare of the biocommunity (plants, animals, rivers, forests) (Fry and Souillac  2017; Ingold  2005; Narvaez  2013). They live contentedly and sustainably (many for thousands if not tens of thousands of years). This is our human heritage. A successful species should not be assessed by their own generational success but by whether they help their species flourish in the future. This necessarily means cooperating with the biocommunity in which the species survives (Cajete 2000). Wiping out other species undermines the ecological balance that is required for the flourishing of individual species as well as the whole community. Of course, there are regular fluctuations in any dynamic system, but under normal conditions of local groundedness, this occurs within a narrow range of fluctuation that does not destroy a whole species. Kohák (2000) lists the characteristics of a species’ overpopulation: living off nature’s substance instead of its annual yield, the crowding out of other species, and the loss of ability to raise the young with necessary skills. Kohak noted that humanity was manifesting these indicators decades ago. As described here, the evolved nest is especially missing in ‘advanced’ nations, creating anxieties that are mitigated with work, addictions, and control. However, perhaps more im­port­ ant­ly in this era of planetary disaster, most children do not learn to live as partners of a local biocommunity. The Western worldview, deeply rooted in the metaphors used to guide life (e.g. progress, separation from and superiority to the rest of the natural world, Bowers  2003), perpetuates the disconnection from

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

98  Darcia Narvaez nature. Most members of Western societies today are missing attachment to the natural world, to the particularities of a place, which is a significant characteristic of sustainable communities such as the San Bushmen who have been extant for over 100,000 years (Suzman 2017). Surely attachment assessments should include bonding to nature as part of a full humanity and of a grounded virtue (Narvaez 2014).

5.  Conclusion If we are going to discuss virtue and vice, we must take into account the life world in which we exist. We must take into account the totality of flourishing. We live among a slippage of baselines: standards and expectations for child raising, child outcomes, adult health and well-­being, social support throughout life, living with other-­than-­humans. These shifts represent extensive social poverty leading to social and nature disconnection, a massive deterioration of relationships away from respect and responsibility. The undermining of the human nature that emerges when the evolved nest is provided and that was adaptive for our ancestors leads to a world increasingly filled with dysregulated, self-­centred people who are more vicious than virtuous, and who create theories of human nature to rationalize the slippage as normal and in need of coercive strategies (Hobbes 1651/2010; Wrangham and Peterson 1996), affecting how parents view and treat their children (Gerhardt  2010). Without baselines to guide judgements about species typicality, it is easy to think ‘there is no other way to be’ or that humanity is ‘progressing’ from a violent past (Pinker 2011). Moving forward, researchers need to take into account how nested their subjects are before drawing conclusions about human nature. Parents need to understand the short- and long-­ term effects of evolved nest provision and that investment in a young child’s nurturing is well worth it. But this can only happen with community and institutional support. Policymakers need to understand that if a child’s early nest is not supportive for proper psychosocial and neurobiological development, costs will go up for dealing with the resulting ­dysregulated people. The evolved nest is critical for restoring human virtue to its earth-­centric origins. Never before has this been more important.

References Ardiel, E.  L., and Rankin, C.  H. 2010. ‘The importance of touch in development’. Paediatrics & Child Health 15(3), 153–6. Atzil, S., Gao, W., Fradkin, I., and Feldman Barrett, L. 2018. ‘Growing a social brain’. Nature Human Behaviour 2, 624–36.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

The Evolved Nest, Virtue, and Vice  99 Bateson, P., and Gluckman, P. 2011. Plasticity, Robustness, Development and Evolution. New York: Cambridge University Press. Bowers, C. A. 2003. Mindful Conservatism: Rethinking the Ideological and Educational Basis of an Ecologically Sustainable Future. Lanham, MD: Rowman & Littlefield. Braden, A., & Narvaez, D. (in press). Primal parenting: Lasso the parent handlers and embrace your parenting gifts. New York: Oxford University Press. Bystrova, K., et al. 2009. ‘Early contact versus separation: effects on mother-infant interaction one year later’. Birth 36(2), 97–109. Bystrova K., Widström A.-M., A.-S. Matthiesen et al., 2013. ‘Skin-to-skin contact may reduce negative consequences of “the stress of being born”: a study on temperature in newborn infants, subjected to different ward routines in St. Petersburg’. Acta Paediatrica 92(3), 320–26. Cajete, G. 2000. Native Science: Natural Laws of Interdependence. Santa Fe, CA: Clear Light. Capra, F. 1997. ‘Encounter at the edge of the new paradigm: a dialogue with E. F. Schumacher’. What is Enlightenment? 11, Spring–Summer. Carey, N. 2011. The Epigenetic Revolution: How Modern Biology is Rewriting our Understanding of Genetics, Disease and Inheritance. London: Icon. Carter, C. S., and Porges, S. W. 2013. ‘Neurobiology and the evolution of mammalian social behavior’. In D.  Narvaez, J.  Panksepp, A.  N.  Schore, and T.  Gleason (eds), Evolution, Early Experience and Human Development: From Research to Practice and Policy, 132–51. New York: Oxford University Press. Crittenden, P. M. 1995. ‘Attachment and psychopathology’. In S. Goldberg, R. Muir, and J.  Kerr (eds), Attachment Theory: Social, Developmental, and Clinical Perspectives, 367–406. Hillsdale, NJ: The Analytic Press. Darwin, C. 1859/1962. The Origin of Species. New York: Collier Books. Darwin, C. (1871/1981). The Descent of Man. Princeton: Princeton University Press. Dawkins, R. 1976. The Selfish Gene. Oxford: Oxford University Press. Deoni, S. C. L., Dean III, D. C., Piryatinksy, I., O’Muircheartaigh, J., Waskiewicz, N., Lehman, K., Han, M., and Dirks, H. 2013. ‘Breastfeeding and early white matter development: a cross-sectional study’. Neuroimage 82, 77–86. Derber, C. 2013. Sociopathic Society: A People’s Sociology of the United States. Boulder, CO: Paradigm Press. Dettwyler, K. A. 1997. ‘When to wean: biological versus cultural perspectives’. Clinical Obstetrics and Gynecology 47(3), 712–23. Díaz, S., Settele, J., Brondizio, E., Ngo, H. T., Gueze, M., Agard, J., and Zayas, C. 2019. IPBES Summary for Policymakers of the Global Assessment Report on Biodiversity and Ecosystem Services. Bonn: The Intergovernmental Science-Policy Platform on Biodiversity and Ecosystem Services (IPBES). Dunn, R. 2011. The Wild Life of our Bodies: Predators, Parasites, and Partners that Shape who we are Today. New York: Harper.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

100  Darcia Narvaez Easterbrooks, M. A., and Goldberg, W. A. 1990. Parental attitudes toward childrearing (PACR). In J.  Touliatos, B.  Perlmutter, and M.  Strauss (eds), Handbook of Family Measurement Techniques, 341–2. Newbury Park, CA: Sage. Eibl-Eibesfeldt, I. 1989. Human Ethology. New York: Aldine de Gruyter. Eisenberg, L. 1995. ‘The social construction of the human brain’. American Journal of Psychiatry 152, 1563–75. Eisler, R., and Fry, D.  P. 2019. Nurturing our Humanity. New York: Oxford University Press. Field, T., Diego, M., and Hernandez-Reif, M. 2010. ‘Prenatal depression effects and interventions: a review’. Infant Behavior Development 33(4), 409–18. Field, T., and Hernandez-Reif, M. 2013. ‘Touch and pain perception in infants’. In D. Narvaez et al. (eds), Evolution, Early Experience and Human Development: From Research to Practice and Policy, 262–76. New York: Oxford University Press. Foyer, P., Wilsson, E., and Jensen, P. 2016. ‘Levels of maternal care in dogs affect adult offspring temperament’. Scientific Reports 6, 19253. DOI: 10.1038/srep19253. Fry, D. P., and Souillac, G. 2017. ‘The original partnership societies: evolved pro­pen­ sities for equality, prosociality, and peace’. Interdisciplinary Journal of Partnership Studies 4(1), Article 4. Gerhardt, S. 2010. The Selfish Society. London: Simon & Schuster. Gershoff, E. T. 2013. ‘Spanking and child development: we know enough now to stop hitting our children’. Child Development Perspectives 7(3), 133–7. Gilbert, P. 2005. Compassion and cruelty: a biopsychosocial approach. In P. Gilbert (ed.), Compassion: Conceptualisations, Research and Use in Psychotherapy, 9–74. New York: Routledge. Gómez-Robles, A., Hopkins, W. D., Schapiro, S. J., and Sherwood, C. C. 2015. ‘Relaxed genetic control of cortical organization in human brains compared with chimpanzees’. Proceedings of the National Academy of Sciences 112(48), 14799–804. DOI:10.1073/pnas.1512646112 Greenough, W. T., Black, J. E., and Wallace, C. S. 1987. ‘Experience and brain development’. Child Development 58, 539–59. Harlow, H. 1958. ‘The nature of love’. American Psychologist 13, 673–85. Hart, S., L. Boylan, S. Carroll, Y. Musick and R. Lampe. 2003. Breast-fed one-weekolds demonstrate superior neurobehavioral organization.’ J.  Pediatr. Psychol. 28(8), 529–34. Hart, S., Boylan, L. M., Carroll, S. R., Musick, Y. A., Kuratko, C., Border, B. G., and Lampe, R. L. 2006. ‘Newborn behavior differs with docosahexaenoic acid (DHA) levels in breast milk’. Journal of Pediatric Psychology 31, 221–26. Hewlett, B.  S., and Lamb, M.  E. 2005. Hunter-Gatherer Childhoods: Evolutionary, Developmental and Cultural Perspectives. New Brunswick, NJ: Aldine. Ho, M.  W. 2010. ‘Development and evolution revisited’. In K.  E.  Hood et al. (eds), Handbook of Developmental Science, Behavior, and Genetics, 61–109. Chichester: Wiley-Blackwell.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

The Evolved Nest, Virtue, and Vice  101 Hobbes, T. 1651/2010. Leviathan. Ed. A.  P.  Martinich and B.  Battiste (rev. edn). Peterborough, ONT: Broadview Press. Houlihan, J., Kropp, T., Wiles, R., Gray, S., Campbell, C. 2014. Body Burden: The Pollution in Newborns. Environment Working Group. Downloaded 11.8.2017 from

Hrdy, S. 2009. Mothers and Others: The Evolutionary Origins of Mutual Understanding. Cambridge, MA: Belknap Press. Ingold, T. 2005. ‘On the social relations of the hunter-gatherer band’. In R.  B.  Lee and R. Daly (eds), The Cambridge Encyclopedia of Hunters and Gatherers, 399–410. New York: Cambridge University Press. Intergovernmental Panel on Climate Change. 2013. Climate Change 2013: The Physical Science Basis. Working Group I Contribution to the IPCC 5th Assessment Report— Changes to the Underlying Scientific/Technical Assessment (IPCC-XXVI/Doc.4). Geneva: United Nations. Joseph, J. 2013. ‘The missing heritability of psychiatric disorders: elusive genes or non-existent genes?’ Applied Developmental Science 16, 65–83. Kalinichev, M., Easterling, K. W., Plotsky, P. M., et al. 2002. ‘Long-lasting changes in stress-induced corticosterone response and anxiety-like behaviors as a consequence of neonatal maternal separation in Long-Evans rats’. Pharmacol. Biochem. Behav. 73(1), 131–40. Kanitz, E. et al. 2004. ‘Consequences of repeated early isolation in domestic piglets (Sus scrofa) on their behavioural, neuroendocrine, and immunological responses’. Brain, Behavior & Immunity 18(1), 35. Knudsen, E. I. 2004. ‘Sensitive periods in the development of the brain and behavior’. Journal of Cognitive Neuroscience 16(8), 1412–25. Kochanska, G. 1994. ‘Beyond cognition: expanding the search for the early roots of internalization and conscience’. Developmental Psychology 30(1), 20–2. Kochanska, G. 2002. ‘Committed compliance, moral self, and internalization: a mediational model’. Developmental Psychology 38, 339–51. Kohák, E. 2000. The Green Halo: A Bird’s Eye View of Ecological Ethics. Chicago: Open Court. Konner, M. J. 2005. ‘Hunter-gatherer infancy and childhood: The !Kung and others’. In Hewlett, B.  S., Lamb, M.  E. (eds.), Hunter-gatherer childhoods: Evolutionary, developmental and cultural perspectives 19–64. New Brunswick, Canada: Transaction Publishers. Korten, D. 2015. Change the Story, Change the Future. Oakland, CA: Berrett-Koehler. Lancy, D.  F. 2015. The Anthropology of Childhood: Cherubs, Chattel, Changelings (2nd edn). New York: Cambridge University Press. Levine, R. A., Dixon, S., LeVine, S. E., Richman, A., Keefer, C., Liederman, P. H., and Brazelton, T.  B. 1994. Child Care and Culture: Lessons from Africa. New York: Cambridge University Press.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

102  Darcia Narvaez Lewontin, R. 2010. Response to Colin Wells’ comment on review called ‘Not so Natural Selection’ (of What Darwin got wrong, 27 May 2010). New York Review of Books. Downloaded on 11 Oct. 2018 McKenna, J., Mosko, S., Richard, C., Drummond, S., Hunt, L., Cetel, M. B., and Arpaia, J. 1994. ‘Experimental studies of infant-parent co-sleeping: mutual physi­ology and behavioral influences and their relevance to SIDS (sudden infant death syndrome)’. Early Human Development 38(3), 187–201. DOI:10.1016/0378-3782(94)90211-9. Main, M. 1995. ‘Recent studies in attachment: overview, with selected implications for clinical work’. In S. Goldberg, R. Muir, and J. Kerr (eds), Attachment Theory: Social, Developmental and Clinical Perspectives, 407–74. Hillsdale, NJ: The Analytic Press. Margulis, L. 1998. Symbiotic Planet: A New Look at Evolution. Amherst, MA: Sciencewriters. Marshall, P., and Kenney, J. 2009. ‘Biological perspectives on the effects of early psychosocial experience’. Developmental Review 29, 96–119. Millennium Ecosystem Assessment. 2005. Ecosystems and Human Well-being: Synthesis. Washington, DC: Island Press. Montagu, A. 1968. Culture: Man’s Adaptive Dimension. New York: Oxford University Press. Moore, J. (ed.). 2016. Anthropocene or Capitalocene: Nature, History and the Crisis of Capitalism. Oakland, CA: PM Press. Naess, A., and Rothenberg, D. 1989. Ecology, Community and Lifestyle. Cambridge: Cambridge University Press. Narvaez, D. 2008. ‘Triune ethics: The neurobiological roots of our multiple moralities’. New Ideas in Psychology 26, 95–119. Narvaez, D. 2011. ‘The ethics of neurobiological narratives’. Poetics Today 32(1), 81–106. Narvaez, D. 2013. ‘The 99%—development and socialization within an evolutionary context: growing up to become “A good and useful human being”’. In D. Fry (ed.), War, Peace and Human Nature: The Convergence of Evolutionary and Cultural Views, 643–72. New York: Oxford University Press. Narvaez, D. 2014. Neurobiology and the Development of Human Morality: Evolution, Culture and Wisdom. New York: Norton. Narvaez, D. 2017. ‘Evolution, early experience and Darwin’s moral sense’. In. R. Joyce (ed.), Routledge Handbook of Evolution and Philosophy, 322–32. London: Routledge. Narvaez, D. 2018. ‘Ethogenesis: evolution, early experience and moral becoming’. In J.  Graham and K.  Gray (eds), The Atlas of Moral Psychology, 451–64. New York: Guilford Press. Narvaez, D., Gettler, L., Braungart-Rieker, J., Miller-Graff, L., and Hastings, P. 2016. ‘The flourishing of young children: evolutionary baselines’. In D.  Narvaez, J. Braungart-Rieker, L. Miller, L. Gettler, and P. Harris (eds), Contexts for Young Child Flourishing: Evolution, Family and Society, 3–27. New York: Oxford University Press.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

The Evolved Nest, Virtue, and Vice  103 Narvaez, D., Gleason, T., Wang, L., Brooks, J., Lefever, J., Cheng, A., and Centers for the Prevention of Child Neglect. 2013. ‘The evolved development niche: lon­gi­tu­ din­al effects of caregiving practices on early childhood psychosocial development’. Early Childhood Research Quarterly 28(4), 759–73. DOI: 10.1016/j.ecresq.2013. 07.003. Narvaez, D., and Hardy, S. 2016. ‘Measuring triune ethics orientations’. In Narvaez, D. (Ed.), Embodied morality: Protectionism, engagement and imagination, 47–72. New York: Palgrave Macmillan. Narvaez, D., Wang, L., Cheng, A., Gleason, T., Woodbury, R., Kurth, A., and Lefever, J. B. 2019. ‘The importance of early life touch for psychosocial and moral development’. Psicologia: Reflexão e Crítica 32(16). doi.org/10.1186/s41155-019-0129-0. Narvaez, D., Wang, L., and Cheng, Y. 2016. ‘The evolved developmental niche in childhood: relation to adult psychopathology and morality’. Applied Developmental Science 20(4), 294–309. Narvaez, D., Woodbury, R., Gleason, T., Kurth, A., Cheng, A., Wang, L., Deng, L., Gutzwiller-Helfenfinger, E., Christen, M., and Näpflin, C. 2019. ‘Evolved development niche provision: moral socialization, social maladaptation and social thriving in three countries’. Sage Open 9(2). National Research Council. 2013. U.S.  Health in International Perspective: Shorter Lives, Poorer Health. Washington, DC: The National Academies Press. Niehoff, D. 1999. The Biology of Violence: How Understanding the Brain, Behavior, and Environment can Break the Vicious Circle of Aggression. New York: Free Press. Oster, Emily. 2019. Cribsheet: A data-driven guide to better, more relaxed parenting. London: Souvenir Press. Overton, W.  F. 2013. ‘A new paradigm for developmental science: relationism and relational-developmental-systems’. Applied Developmental Science 17(2), 94–107. Oyama, S., Griffiths, P. E., and Gray, R. D. 2001. Cycles of Contingency: Developmental Systems and Evolution. Cambridge, MA: MIT Press. Özen, S., and Darcan, S. 2011. ‘Effects of environmental endocrine disruptors on pubertal development’. Journal of Clinical Research in Pediatric Endocrinology 3(1), 1–6. Panksepp, J. 1998. Affective Neuroscience: The Foundations of Human and Animal Emotions. New York: Oxford University Press. Pellis, S.  M., and Pellis, V.  C. 2009. The Playful Brain: Venturing to the Limits of Neuroscience. Oxford: Oneworld. Pinker, S. 2011. The Better Angels of our Nature. New York: Viking. Porges, S. W. 2011. The Polyvagal Theory: Neurophysiological Foundations of Emotions, Attachment, Communication, Self-regulation. New York: Norton. Prescott  J.  W. 1996. ‘The origins of human love and violence’. Pre- and Perinatal Psychology Journal 10(3), 143–88. Sahlins, M. 2008. The Western Illusion of Human Nature. Chicago: Prickly Paradigm Press.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

104  Darcia Narvaez Sandler, J. 1960. ‘The Background of Safety’. International Journal of Psycho-Analysis, 41, 352–356. Schore, A. N. 2003. Affect Dysregulation and Disorders of the Self. New York: Norton. Schore, A. N. 2019. The Development of the Unconscious Mind. New York: Norton. Shepard, P. 1998. Coming Home to the Pleistocene. Washington DC: Island Press/ Shearwater Books. Sorenson, E.  R. 1998. ‘Preconquest consciousness’. In H.  Wautischer (ed.), Tribal Epistemologies, 79–115. Aldershot: Ashgate. Stern, D. 2010. Forms of Vitality: Exploring Dynamic Experience in Psychology, the Arts, Psychotherapy, and Development. New York: Oxford University Press. Suttie, I. 1935. The Origins of Love and Hate. New York: The Julian Press. Suzman, J. 2017. Affluence without Abundance: The Disappearing World of the Bushmen. New York: Bloomsbury. Trevarthen, C. 2005. ‘Stepping away from the mirror: pride and shame in adventures of companionship—reflections on the nature and emotional needs of infant intersubjectivity’. In C. S. Carter, L. Ahnert, K. E. Grossmann, S. B. Hrdy, M. E. Lamb, S.  W.  Porges, and N.  Sachser (eds), Attachment and Bonding: A New Synthesis, 55–84. Cambridge, MA: MIT Press. Trevathan, W. R. 2011. Human Birth: An Evolutionary Perspective (2nd edn). New York: Aldine de Gruyter. Turkheimer, E. 2000. ‘Three laws of behavior genetics and what they mean’. Current Direction in Psychological Science 9(5), 160–4. US Department of Health and Human Services, Substance Abuse and Mental Health Services Administration. 1999. Mental Health: A Report of the Surgeon General. Rockville, MD: Center for Mental Health Services, National Institutes of Health, National Institute of Mental Health. Van der Kolk B. 2014. The body keeps the score: Brain, mind, and body in the healing of trauma. London: Viking. Watson, J. B. 1928. Psychological Care of Infant and Child. New York: Norton. Wohlleben, P. 2016. The Hidden Life of Trees: What they Feel, how they Communicate. Trans. J. Billinghurst. Vancouver: Greystone Books. World Wildlife Fund. 2014. Living Planet Report 2014: Summary. Ed. R.  McLellan, L. Iyengar, B. Jeffries, and N. Oerlemans. Gland, Switzerland: WWF. Wrangham, R.  W., and Peterson, D. 1996. Demonic Males: Apes and the Origins of Human Violence. Boston: Houghton, Mifflin. Xu, J. Q., Murphy, S. L., Kochanek, K. D., and Arias, E. 2016. Mortality in the United States, 2015. NCHS data brief, n.o 267. Hyattsville, MD: National Center for Health Statistics. Zahn-Waxler, C., and Radke-Yarrow, M. 1990. ‘The origins of empathic concern’. Motivation and Emotion 14, 107–25. Zhang, T. Y., and Meaney, M. J. 2010. ‘Epigenetics and the environmental regulation of the genome and its function’. Annual Review of Psychology 61, 439–66.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

5

The Virtues of Attachment Keith S. Cox and Micah Lott

Neither by nature, then, nor contrary to nature do the virtues arise in us. Rather we are adapted by nature to receive them, and are made perfect by habit. Aristotle, Nicomachean Ethics II.1

1.  Introduction: A Tale of Two Mothers Ellen and Kelly are neighbours, and they have each just given birth to a baby girl. Each loves her daughter and wishes her to have a good life—a life that is worthwhile and well-­lived, and also a happy life that is satisfying and good for the one who lives it.1 As part of this wish, both Ellen and Kelly hope that their daughters will be physically and mentally healthy. They also hope that their daughters will understand what matters in human life: that they will not waste their lives on trivial or evil pursuits, that they will become morally virtuous, that they will learn to love and respect other people in the right ways, and that they will undertake worthwhile projects and find joy in those projects. Up to this point (and at this level of generality), Ellen and Kelly agree. However, they differ sharply over how best to raise their daughters. Ellen thinks that the most important thing is to offer consistently warm and attentive care, to make the child feel provided for and safe. Ellen believes that this kind of parenting will help her daughter grow into a caring and self-­confident person, someone who is generous and capable of loving relationships. Kelly is sceptical of this parenting approach. She sees it as coddling a child and hindering her from achieving her full potential. She takes a ‘tiger mom’ approach. Kelly is not shy about using coercive tactics such as shaming and yelling to shape a child’s unruly will, in order to guide her to complete the hard work necessary to excel in academics and the arts. Kelly is convinced that without such parenting methods children will never

1  What they wish for their daughters is what Neera Badhwar has called ‘the highest prudential good’—‘happiness in an objectively worthwhile life’. See Badhwar (2014: part I).

Keith S. Cox and Micah Lott, The Virtues of Attachment In: Attachment and Character: Attachment Theory, Ethics, and the Developmental Psychology of Vice and Virtue. Edited by: Edward Harcourt, Oxford University Press. © Keith S. Cox and Micah Lott 2021. DOI: 10.1093/oso/9780192898128.003.0006

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

106  Keith S. Cox and Micah Lott perform beyond a middling level in these areas, and she believes that achieving excellence in them is central to a good life. Ellen and Kelly care about each other, and each is open to listening to the other. Could they reasonably hope to settle their disagreement about parenting by appealing to empirical research in psychology or cognitive science? This particular question is an instance of a more general question: how (if at all) might the empirical sciences vindicate, undermine, or otherwise shape our understanding of living well? In this chapter, we consider the particular case of Ellen and Kelly. And we narrow our focus in two ways. First, we ask whether Ellen might reasonably hope to change Kelly’s mind by appealing to attachment theory.2 What sort of evidence could Ellen appeal to, and what sort of arguments might she make? Second, we approach the issue from the perspective of Aristotelian naturalism, a view that has been articulated by philosophers such as Philippa Foot, Rosalind Hursthouse, and Michael Thompson.3 As a position in (meta)ethics, a core claim of Aristotelian naturalism is that moral goodness—goodness with respect to the rational will—is a form of natural goodness. In the following two sections, we briefly sketch Aristotelian naturalism and attachment theory, respectively. In section  4, we consider the case of Ellen and Kelly in greater detail. In the language of Aristotelian naturalism, Ellen and Kelly disagree over how best to characterize human form. But the nature and depth of their disagreement depends on how one further spells out the example. We spell out the example in three different scenarios, representing progressively deeper forms of disagreement. In each case, we argue, Ellen can reasonably make some appeal to attachment theory, but her appeal must vary according to the nature of their disagreement. We offer a roadmap of this argumentative landscape.

2.  Aristotelian Naturalism: A Sketch As we understand it, Aristotelian naturalism centres on two claims: (1) There is a distinctive ‘grammar of goodness’ that applies to living things. (2) Moral evaluation belongs to this grammar, and thus moral goodness in a human being is a type of natural goodness and vice a type of natural defect. Let us start with the first claim. In Natural Goodness, Philippa Foot writes: ‘Judgments of goodness and badness can have, it seems, a special “grammar” 2  One might also ask whether Kelly could hope to challenge Ellen’s mind by appealing to attachment theory (or another research programme in psychology). And some of what we say will speak to that question. But for reasons of simplicity and manageability, we will focus on one side of the im­agined conversation. 3  For more on Aristotelian naturalism, see Hähnel (2020).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

The Virtues of Attachment  107 when the subject belongs to a living thing, whether plant, animal, or human being’ (Foot 2001: 26). By ‘grammar’ Foot means the logical connections among a certain class of judgements.4 The judgements at issue here are those that concern a special type of goodness, which Foot calls ‘natural goodness’. Following the work of Michael Thompson, Foot argues that in order to represent anything as living, we must represent it as a particular kind of organism—as the member of some species, or (equivalently) as the bearer of some ‘life form’. A given life form, such as the horseshoe bat or the African elephant, can be represented in an ordered system of ‘natural historical’ judgements. These judgements describe the characteristic features and activities of the life form, such as ‘horseshoe bats hunt insects using echolocation’ or ‘African elephants have four legs.’ A system of such judgements provides an ordered account of the organs and operations of a given kind of living thing. The system, or natural history, provides an answer to the question, ‘how do they live?’5 Moreover, a natural history establishes the good of creatures of this kind. To grasp what is true of ‘the horseshoe bat’ or ‘the African elephant’ is to understand what flourishing amounts to for such creatures. And the good of an organism provides the criterion for judgements of excellence or defect in individual members of the species. To be good as a particular kind of living thing is to possess those qualities that enable the organism to live its characteristic life, as articulated in the natural historical description of the life form. Thus from the fact that ‘the African elephant has four legs’, together with the fact that this African elephant has three legs, it follows that this one is missing a leg. It is naturally defective as an elephant, with respect to its legs. Furthermore, what is good for an organism, what benefits it, is that which enables it to lead its characteristic life—that is, to do the things that make it good as a member of its kind. Thus there is a distinctive grammar of goodness, a set of interlocking concepts, that includes: (1) the good of, (2) good(ness) as, and (3) good for.6 Aristotelian naturalists hold that this sort of normativity is built into our representation of anything as living. Now to the second central claim. Aristotelian naturalists hold that this grammar of goodness applies to our own life form, and that it provides the proper framework for thinking about virtue and vice. In short, the moral virtues are those dispositions of the will that fit one to live and act well qua human being. Thus to understand what counts as excellence or defect in human action, choice, and feeling, we must understand the good of human beings. That is, we must understand our characteristic, non-­defective form of life. As Foot says: 4  Foot takes this sense of ‘grammar’ from Wittgenstein. See Foot (2001). 5  As Thompson says, a natural history provides ‘one’s interpretation or understanding of the life form shared by the members of that class’ (2008: 73). 6  We might also include goodness in the parts or operations of an organism. But goodness in, we think, is simply a different way of referring to as. For further discussion of Foot’s grammar of goodness, see Lott (2018).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

108  Keith S. Cox and Micah Lott I believe that evaluations of human will and action share a conceptual structure with evaluations of characteristics and operations of other living things, and can only be understood in these terms. I want to show moral evil as ‘a kind of nat­ ural defect’. Life will be at the centre of my discussion, and the fact that a human action or disposition is good of its kind will be taken to be simply a fact about a given feature of a certain kind of living thing.7 At this point, it will be helpful to note three distinctions that are part of Aristotelian naturalism as we understand it. The first distinction is between first nature and second nature. First nature refers to those aspects of ‘the human’ that are prior to teaching and training, such as an infant’s capacity to learn a language. Second nature refers to developed capacities and dispositions that are acquired through teaching and training, such as a child’s ability to speak Hindi or English. In spelling out ‘how they live’ in the case of humans, a Thompsonian natural history includes both first and second nature, and a developmental account of the connections between them.

A second distinction concerns different dimensions of natural goodness in human beings. For example, someone might have poor kidney functioning, and in that respect lack natural goodness (he is a defective human, renally speaking), but excel in moral virtue. Or vice versa. When we speak about someone as ‘a good person’, we typically have in mind the person’s rational will. But the same grammar of goodness can be applied to different dimensions, or layers, of human functioning—for example physical, mental, moral.8 And it is not hard to imagine cases in which a lack of natural goodness in one dimension is a factor that leads to greater natural goodness in another dimension. For example, experiencing dis­ abil­ity or disease might be a factor that leads a person to greater virtue, prompting the development of courage, patience, and compassion. Finally, it is crucial to distinguish between the formal and substantive levels of Aristotelian naturalism. At the formal level, Aristotelian naturalism claims that virtue is natural goodness and vice natural defect, but it leaves open the question of what, in fact, human good consists in, and which traits are, in fact, virtues or vices. At a substantive level, various versions of Aristotelian naturalism take a stand on what is true of ‘the human’—what our good consists in, and what the virtues and vices are. People who accept Aristotelian naturalism at the formal level might differ sharply in their respective substantive accounts of ‘the human’. 7  Foot (2001). There are, of course, many ways to challenge this Aristotelian grammar of goodness, both as it applied to non-­human animals and in the specific case of human beings. Addressing those challenges is outside the scope of this chapter. However, one of us has addressed a number of these challenges in other work. See the following: Lott (2012a), Lott (2012b), Lott (2013), Lott (2014), Groll and Lott (2015), Lott (2015), Lott (2018), Lott (2020). 8  Hursthouse notes this point in On Virtue Ethics (1999). See also Edward Harcourt’s discussion of ‘different layers in what counts as characteristically human living’ (227) in Harcourt (2016). See especially 227–33.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

The Virtues of Attachment  109 Given this distinction between formal and substantive levels of Aristotelian naturalism, it is natural to ask: how do we know what is actually true of ‘the human’, and how might we decide between competing substantive conceptions of natural goodness? At this point, we can distinguish two broad strategies that Aristotelian naturalists might take. One strategy attempts to build up an account of human good, including a substantive conception of the moral virtues, solely from ‘morally neutral’ materials, such as evidence from the empirical sciences or an account of human needs that could be accepted by persons of any moral outlook. Another strategy insists that we should not restrict ourselves to such non-­ moral materials when fleshing out Aristotelian naturalism at the substantive level. Rather our substantial conception of ‘the human’ should draw upon concepts and convictions that are internal to our moral outlook. A clear endorsement of this second strategy is found in the work of Rosalind Hursthouse. As she says: ‘Everyone who is taking the Aristotelian naturalist line takes it as obvious that they are not pretending to derive ethical evaluations of human beings from an ethically neutral human biology, but are already thinking about human beings in an ethically structured way’ (Hursthouse 2012: 174). We follow Hursthouse in adopting this second strategy. However, embracing this strategy raises a set of worries. If we are allowed to draw upon our moral outlook in developing an account of natural goodness in human beings, then won’t our account simply express the moral outlook we already accept? How will we have achieved any greater insight into the virtues? And what space will be left for revision in our understanding of human good? Won’t we end up with a trivial restatement of our moral views? Moreover, if we start from an ‘ethically structured’ conception of human beings, then what role is there for the sciences in figuring out what is true of human beings? If there is no role for the sciences, then it seems substantive Aristotelian naturalism is disengaged from biology and psych­ ology—a strange result for a view that begins by emphasizing the continuity between moral judgements and the evaluation of living things in general. We aim to address these worries. But rather than consider them in the abstract, we will focus on the case of Ellen and Kelly described earlier. In the language of Aristotelian naturalism, these two neighbours disagree about how best to characterize ‘the human’.9 Granting that Ellen and Kelly are both working with an eth­ic­ al­ ly structured conception of human beings, how might the empirical sciences—and in particular, the psychological programme of attachment theory—contribute to their substantive conceptions of ‘the human’, or support one substantive conception over another? We turn to that question in section 4. To prepare for our discussion, we first offer a brief sketch of attachment theory.

9  How much they actually disagree, and how far their disagreement goes, depends upon how we further spell out the case—an issue we take up in section 4.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

110  Keith S. Cox and Micah Lott

3.  Attachment Theory: A Sketch Attachment theory addresses how defenceless and dependent young survive early maturation. How is this evolutionary feat accomplished? By attaching the infant to the caregiver and the caregiver to the infant. The attachment system includes neurobiological, behavioural, and cognitive-­ affective components that bind young children to caregivers, and caregivers to young children, so that each is oriented to the other. Attached infants and toddlers seek nourishment, protection, and emotional comfort from the attachment figure. The caregiver in turn is oriented to monitor, engage, and maintain proximity with the child through internal motivations (e.g. reduced anxiety when the child is present) and external motivations (e.g. child’s pleas for proximity maintenance). Child–caregiver attachment, while not unique to humans, is widely held to be a species-­typical adaptation for humans (see e.g. Simpson and Belsky 2016 or Marvin, Britner, and Russell 2016. It is statistically normative for human infants and toddlers to attach to one or more caregivers. Conversely, non-­attachment in infancy and toddlerhood is highly unusual and thought to be generative of psychological pathology, concurrently and prospectively (see e.g. DeKlyen and Greenberg 2016). Beyond the nearly universal presence of an attachment bond, there is also variability in the nature of that bond. At the most general level, infant and toddler attachment falls into two categories: secure vs insecure. Securely attached children engage the caregiver in a style that assumes the caregiver is consistently close, attentive, and responsive to the child’s physical and emotional needs. The securely attached child is distressed when the caregiver exits the en­vir­ on­ment, seeks the caregiver during an absence, and is readily soothed at reunion with the caregiver. The securely attached child uses the caregiver as a safe haven when challenged with threats of many kinds (physical, psychological, or emotional) and also as a secure base from which to explore, play, and achieve mastery in the proximal environment. Insecurely attached children are commonly described as falling into two sub-­ categories: anxious-­ambivalent and anxious-­avoidant.10 The child who has an anxious-­ambivalent attachment is characterized by anxious exploration when the caregiver is present, extreme distress when the caregiver leaves, and multi-­valent responses when the caregiver returns (e.g. alternating resistance and engagement). Anxious-­avoidant attachment is typified by limited engagement with the caregiver when the caregiver is present and lack of response or interest when the caregiver leaves or returns. Studies typically find that more than 60 per cent of infants are

10  Disorganized attachment is a third category that is characterized by incoherent, incomplete, or contradictory responses to caregiver presence and absence.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

The Virtues of Attachment  111 securely attached, and the remainder are insecurely attached (Van IJzendoorn, and Kroonenberg 1988; Mesman, van IJzendoorn, and Sagi-­Schwartz 2016).

4.  Attached to Virtue? Let us return to the case of Ellen and Kelly, interpreting their disagreement as a debate over the proper substantive account of ‘the human’. We have seen that Aristotelian naturalism claims that moral virtue is a type of natural goodness. We have also said that Ellen and Kelly both wish for their daughters to grow up to be morally virtuous. We can now distinguish two questions about moral virtue that one might hope to answer by appealing to attachment theory: (1) Given that trait V is a moral virtue, what sort of practices, dispositions, and environments foster the development of V or hinder its development? Here we assume some fairly worked out picture of what acting from a virtue involves. We aim to understand how this virtue is acquired in the course of the characteristic life-­cycle of human beings. (2) Is trait V a genuine human virtue or not? In this case, we do not assume that trait V is actually a virtue. The situation might involve someone looking at a purported virtue—say, a particular conception of modesty—and wondering whether this is a true virtue or a sham virtue. Or it might involve one party affirming, and another party denying, that V is a virtue. In this case, our aim is to say which traits are, in fact, virtues—to vindicate or undermine various substantive conceptions of the virtues.11 Your answer to question (1) might impact your answer to question (2). For instance, you might be interested in (1) because you are wondering how best to bring about a particular virtue, as you understand it. Your motivation is moral education. But if it turns out that no practices, dispositions, or environments could bring about this trait in human beings, you might take that as evidence that this was not a virtue after all—an answer to question (2). Which of these questions is most relevant to the disagreement between Ellen and Kelly, and how might attachment theory play a role in answering the question? The answer depends on how we further spell out their respective positions and

11  There is a third question about virtue one might ask: given that trait V is a moral virtue, what is its profile—i.e. what are the different components of V and what does V look like in different contexts? Here we again assume that some trait is a virtue. Our goal is to refine, expand, and deepen our understanding of this virtue. We aim to give a more fine-­grained and realistic picture of the thoughts/feelings/reactions that characterize the virtue, and to understand how the virtue fits into the overall good life for human beings. For the sake of simplicity, we here set aside this question.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

112  Keith S. Cox and Micah Lott the nature of their disagreement. We will consider three different ways of spelling out the situation.

4.1  Scenario One: a question of development In this scenario, Ellen and Kelly disagree about the sort of parenting practices and home environment that are best for a child. Kelly embraces many practices that Ellen rejects, and she sees Ellen’s approach as too permissive and lacking in dis­ cip­line. However, Ellen and Kelly agree on the kind of life that counts as a well-­ lived, flourishing life. They simply disagree about what parenting strategies are most likely to help a child grow up to live that kind of life.12 Spelled out this way, their disagreement seems fairly shallow. They disagree about an important aspect of how ‘the human’ lives—namely, the proper way to parent a child.13 But they agree about the moral virtues and about what matters most in life. This means that they agree in their hopes for their daughters. Thus they will experience joy and sorrow over the same types of things in their daughters’ lives. Because each wishes the same type of life for her daughter, it is natural to say that they agree about the end but they disagree about the means. For this scenario, the relevant question about virtue is question (1), concerning the development of traits that are agreed to be virtues. Could attachment theory contribute to answering this question and settling a dispute of this sort? We believe the answer is ‘Yes.’ To understand the contribution that attachment theory might make, let us consider the work of the psychologist Darcia Narvaez. Narvaez gives an account of human development that describes what is typical for human beings, not merely in the sense of what is statistically common, but developmentally appropriate (Narvaez  2014). That is, Narvaez aims to capture both how human development usually proceeds and how it ought to proceed. In other words, we take Narvaez to be making claims about ‘the human’ of the sort identified by Thompsonian natural-­historical judgement. Narvaez sees moral development as having critical periods concentrated in early life that are especially anchored in the caregiver–child/infant relationship. She argues that an intensified version of secure attachment, which we will call secure attachment+ (Narvaez 2014), fosters the development of a suite of psychological tendencies that lead to and promote particular moral virtues, such as empathy. While standard models of secure attachment emphasize mutually responsive communication and intersubjective experience between caregiver and child, secure attachment+ is 12  We assume that a diverse set of lifestyles and choices will qualify equally as ‘the kind of life’ that is well lived and flourishing. So one could also speak about the range of well-­lived and flourishing lives. 13  In fact, even this way of putting things is too strong and needs to be qualified. For surely they will agree about many aspects of parenting. It is only in a certain range of situations, and with respect to certain considerations, that they disagree.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

The Virtues of Attachment  113 further characterized by caregiver–child relationships filled with playful inter­action, spontaneous sharing, joint activities, shared meaning and purpose, and the co-­construction of narratives about daily life experiences. Moreover, within secure attachment+ the caregiver is oriented to the child as an intelligent and purposeful agent. Narvaez argues that the regular experience of secure attachment+ tunes the neurobiological and psychological systems of the child for the development of virtue. She thus sees secure attachment+ as an optimal developmental pathway for human life (Narvaez 2014). Narvaez traces a number of particular developmental paths to various virtues. One particular path involves early attachment, the vagus nerve, and empathy-­ related traits. The vagus nerve is a central pathway of the parasympathetic ner­ vous system, which modulates numerous biological and emotional processes. Poor functioning of the vagus nerve has been linked to diverse biological problems—digestive issues, seizures, and general inflammation. In contrast, healthy functioning of the vagus nerve has been associated with diverse biological and psychological markers of health, such as emotion regulation and lowered risk for mental health disorders. Narvaez describes evidence that secure attachment, and in particular emotional attunement of mother to infant, promotes healthy vagal tone functioning in the infant (and later the child). Narvaez then points to evidence that healthy vagal tone is associated with greater cooperation and giving among children, and, in adults, compassion and openheartedness towards an individual from a different background. In this way, virtues like generosity and empathy are fostered by secure attachment+ in infancy and childhood. Narvaez’s account of the development of adult empathy is an instance of a more general type of explanation that occurs frequently in her account: (A) some aspect of our psychobiological nature (e.g. the vagus nerve) is iteratively shaped or tuned by (B) some aspect of the developmental environment (i.e. secure attachment+), which promotes (C) a building block (e.g. perspective-­taking) of (D) moral virtue (e.g. empathy). Narvaez makes this form of argument with respect to a number of building blocks of the virtues, as she understands them. These building blocks include things like ready pleasure in social engagement, openness to another’s influence, responsiveness, and the ability to emotionally attune to another. Narvaez also describes a common but defective developmental pathway that does not lead to virtue: (A) some aspect of our psychobiological nature (e.g. the vagus nerve) is iteratively shaped or tuned by (B) some aspect of the developmental environment (i.e. insecure attachment) in a way that impedes the building blocks (e.g. mistrust instead of perspective-­taking) of (D) moral virtue (e.g. empathy). In the language of Aristotelian naturalism, Narvaez is presenting an account of the life-­cycle of ‘the human’ that identifies developmental connections between both (a) first nature and second nature, and (b) different dimensions of human functioning—including physical, emotional, mental, and moral dimensions of appropriate human development and activity.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

114  Keith S. Cox and Micah Lott If we assume that Ellen and Kelly agree with Narvaez about which traits are moral virtues, then we can see how Narvaez’s arguments are relevant to their disagreement over the proper way to parent. Since Ellen and Kelly are having a debate about means and not ends, Ellen might reasonably point to the work of Narvaez and say: ‘You and I agree that empathy, generosity, etc. are important virtues. Research shows that yelling, shaming, and harsh discipline will tend to impede secure attachment. And that, in turn, will tend to hinder the psychological building blocks of those virtues. In parenting that way, you are making it harder, rather than easier, for your daughter to develop the moral character that you hope she will have.’ That said, there are two important limits to the general type of explanation that Narvaez offers. First, such explanations are not meant to establish that any particular trait, such as empathy, is a moral virtue. For instance, the developmental account linking secure attachment+ to superior vagal tone, which in turn supports the tendency to greater empathy, is not intended to vindicate the status of empathy as a virtue. In particular, the argument is not trying to show that em­pathy is a virtue because it is associated with superior vagal tone. More generally, no particular trait is countenanced as a virtue solely because it follows from or promotes some aspect of non-­moral human functioning. Instead the explanation starts with an account of the virtues and then maps out a developmental pathway likely to result in virtue development. Second, the explanations that Narvaez provides are best understood, we believe, as showing something about the psychological building blocks of virtue and not the full realization of virtue—at least not Aristotelian moral virtue. Take empathy. The developmental trajectory she traces results in adults who have a greater tendency to display empathy. But if empathy is a virtue in the Aristotelian sense, then it must be displayed at the right times, for the right reasons, to the right degree, towards the right persons, etc. Constructing an empirical way to simultaneously measure all these aspects of a virtue is a demanding problem for study design and measurement techniques. To our mind, and in agreement with others (Wright, Warren, and Snow 2020, this high bar has not been met by empirical research in psychological science to date.14 Thus, the developmental tra­jec­ tor­ies that Narvaez describes end with the psychological building blocks of virtue and not virtue itself. To conclude this subsection, we note three questions that are left open by a Narvaez-­style explanation of how virtue (or vice) develops. First, what is the status of developmental pathways that do not include or rely on secure attachment+? If secure attachment+ is characteristic of human beings, are all non-­attachment-­ based developmental theories lacking in some important respect? Or are there 14 We are primarily thinking of quantitative empirical research. Qualitative research methods might be better suited to capture some of the complexities of virtue.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

The Virtues of Attachment  115 other equally characteristic (non-­defective) developmental pathways that do not include or rely on secure attachment+? Perhaps there are a range of characteristic paths to virtue for human beings, some of which do not include secure attachment+.15 Second, what do we say about the familiar situation in which an individual appears to have grown up with the benefit of secure attachment+ (and other psychobiological and environmental goods) but nevertheless becomes an adult with poor, maybe even vicious, character? Such cases appear to show that secure attachment+ is not sufficient for developing the interpersonal moral virtues. Indeed, Narvaez’s account of the development of virtue recognizes this, as she includes elements beyond secure attachment+ in the pathway to good character. Third, what should we make of individuals who do not experience secure attachment+ in infancy or childhood, yet who grow up to be virtuous? There are many anecdotes of individuals who grow up in harsh contexts, characterized by instability, scarce resources, and poor treatment, yet become adults with ad­mir­ able moral character. Presumably, such developmental contexts are not characteristic of ‘the human’ (recall that what is characteristic here is not a statistical notion). What should we say, then, about individuals who grow up in such circumstances and yet become morally virtuous? What do they tell us about moral virtue and its development?16

4.2  Scenario Two: the good life and achieved excellence Fleshed out in a different way, the disagreement between Ellen and Kelly runs deeper. In the second scenario, they disagree not only about parenting strategies, but about what type of life will count as a well-­lived, flourishing life. To make the disagreement clear, recall that Kelly believes that achieving excellence in academics and the arts is central to a good life. We can distinguish between two very broad aspects of human good: moral virtue and achieved excellence. By ‘achieved excellence’ we mean the development of one’s talents and capabilities, exercised in skilful and significant ways. We take the category of achieved excellence to include a great many activities, such as excellence in art, science, literature, sports, cooking, business, etc. In this scenario, Ellen and Kelly both take a good life to include both moral virtue and achieved excellence. They both hope that their daughters will grow up

15  It appears that Narvaez does not leave room for a range of species-­typical trajectories in this respect, but it is not clear if there is consensus on this question among developmental psychologists, especially with regard to an emphasis on secure attachment+. 16  For an interesting treatment of these questions within an Aristotelian framework, see Vogler (2013). Vogler draws on the work of Anselm Müller, and his account of the Aristotelian distinction between natural virtue and ethical/moral virtue. We agree that this is an important distinction— another valuable analytic tool in the Aristotelian conceptual toolbox.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

116  Keith S. Cox and Micah Lott to be (a) good—in the sense of morally virtuous, and (b) excellent—skilful at some activities that are important in human life, such as medicine or journalism or physics or the violin. However, while they agree in their substantive conceptions of moral virtue, they disagree in their substantive conceptions of achieved excellence. That is, they disagree about the range of activities in which excellence matters, and/or about the level of achievement that qualifies as excellent, and/or about the relative importance of achieved excellence—including the importance of achieved excellence relative to moral goodness. This is a deeper form of disagreement, because it concerns some elements of what counts as a good human life. And thus Ellen and Kelly partly diverge in their hopes for their daughters. Suppose, for example, that Kelly values a narrower range of activities in which one might attain excellence, and she values excellence in those domains more highly than Ellen. In that case, Ellen and Kelly will agree about what it means for their daughters to grow up to be ‘a good person’, in the sense of being morally virtuous. But they will disagree, in at least some cases, over the manner and extent to which their daughters have lived good, flourishing lives. In that respect, they disagree about the end, and not merely about the means. In this scenario, there are two main ways that Ellen might appeal to attachment theory. First, it is an open empirical question whether or not Kelly’s parenting practices actually tend to promote achieved excellence as she understands it. Even if we understand excellence in Kelly’s terms, the parenting practices that best promote such excellence might turn out to be those that are embraced by Ellen, not Kelly. Indeed, Western developmental psychologists appeared to have come to a consensus in recent decades that the kinds of harsh parenting techniques that Kelly favours were associated with inferior levels of academic success, where academic success is a possible indicator for achieved excellence (e.g. Spera 2005). In add­ ition, several studies have found longitudinal links between secure attachment and superior academic achievement (Jacobsen and Hofmann 1997; Moss and St-­ Laurent 2001). However, this question was reignited in 2011 with the high-­profile publication of Amy Chua’s The Battle Hymn of the Tiger Mother. Chua framed her parenting style as a cultural issue, one of East vs West. She argued that high expectations were crucial for a child to achieve excellence in academics and the arts. Given the unruly will of a child, Chua posited that the parent must impel the child to meet high expectations by using strict disciplinarian tactics. Chua illustrated what she meant by telling stories of how she used yelling, shaming, and sustained coercion in parenting her daughters. In one memorable sequence, Chua related how she forbade her daughter to use the bathroom for several hours until her daughter mastered a difficult section of a piano piece. Still, Chua’s account did not disregard attachment-­related notions such as emotional responsiveness and physical tenderness. Rather, Chua deployed these relational elements in a very

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

The Virtues of Attachment  117 particular way. When her daughters attained a degree of excellence, such as mastering the piano piece, Chua rewarded them with tenderness and cuddling. If Kelly agrees with Chua’s conception of achieved excellence, then Kelly might point to Chua’s ideas as a vindication of her forceful parenting practices. But from an empirical perspective, is Chua correct about what tends to promote achieved excellence? Her book spurred a flurry of studies. Researchers appeared to agree that the previous consensus that harsh parenting stunted the development of excellence was overly based on Western samples. With a small set of results reported to date, the picture is mixed. Some studies have not found a link between tiger parenting and achieved excellence (e.g. Liew et al. 2014), while others have suggested that such a link might exist for Asian Americans but not European Americans (e.g. Fu and Markus 2014). Moreover, the studies to date have primarily examined outcomes such as high school maths and English grades or results on a laboratory task. Such outcomes reflect, at most, a limited aspect of achieved excellence, which can be instantiated in diverse ways. Thus Ellen and Kelly might have to wait years (or even decades) for empirical research to come to a new consensus on what are the best parenting practices to promote achieved excellence. Moreover, the research might find that the answer depends on the cultural context. Rather than asking about what promotes excellence as Kelly understands it, might Ellen appeal to attachment theory to challenge Kelly’s conception of achieved excellence—to challenge that aspect of her understanding of the good life? It is hard to see how such an argument would work. Suppose that Kelly believes that scientific knowledge is so sublime, so uniquely marvellous, that only a life of scientific understanding counts as a fully worthwhile life. And suppose Kelly admits that she has no way to prove this to anyone who has not undertaken the life of science: it is only from within that activity that one can discern its value. In this case, we doubt that attachment theory per se (or any psychological science) could show that Kelly was wrong to value this sort of achieved excellence, or to value it as highly as she does.17 However, Ellen can appeal to psychology to challenge Kelly’s conception of the good life in a less direct way. In particular, she can point to the type of evidence we discussed in the first scenario, in our discussion of Narvaez’s theory of human development. Because Ellen and Kelly agree about the moral virtues, Ellen can say to Kelly: ‘Even if your parenting practices are successful at promoting achieved excellence as you understand it, those same practices will hinder the development of moral virtue—as we both understand virtue. If you try to help your daughter achieve excellence in that way, you do so at the risk of undermining her moral development.’ 17 The same point would hold, of course, for many other substantive conceptions of achieved excellence.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

118  Keith S. Cox and Micah Lott

4.3  Scenario Three: competing conceptions of the virtues In a third scenario, Ellen and Kelly disagree not only about achieved excellence, but also in their substantive conceptions of moral virtue. For example, Ellen affirms, but Kelly rejects, that humility is a virtue. And while they agree that generosity is a virtue, they disagree substantially over its proper expression—that is, over to whom, and when, and in what ways one ought to give to others. And so on.18 This is a very deep kind of disagreement. For it amounts to disagreeing over how one ought to think, feel, and respond to the kind of considerations to which the virtues are sensitive—a disagreement about what it means to be a good person. They disagree about the end, as well as the means, and their disagreement about the end goes even further than in the second scenario. Thus they diverge widely in their hopes for their daughters. And here the relevant question about virtue is type 2: is trait V (as understood by either Ellen or Kelly) a genuine human virtue or not? In this case, Ellen cannot use Narvaez’s account of virtue and secure attachment+ to argue directly against Kelly’s alternative conception of the virtues. For Narvaez’s type of explanation assumes a conception of virtue and then maps out a developmental pathway likely to result in virtue. But in this scenario, Ellen needs an argument for a substantive conception of virtue, not an argument that starts from a substantive conception of virtue. So even if we grant that a certain way of parenting promotes secure attachment+, which in turn promotes vagal tone, which in turn promotes empathy, all of that need not matter to Kelly. For she does not regard empathy as a virtue, and so she does not hope that her daughter will develop empathy—at least not as Narvaez understands it. However, at least two other argumentative strategies are open to Ellen. Both face challenges, and both have important limitations. The first strategy draws on the distinction between first nature and second nature. Ellen might appeal to attachment theory to argue that there is no plausible developmental route from what a human child is prior to teaching and training (her first nature, qua human being) to the type of developed moral character that Kelly considers to be virtuous (a substantive account of the second nature that is characteristic of ‘the human’). In effect, Ellen would be arguing that human beings do not characteristically have it in them to develop the disposition to think/feel/choose in the way that Kelly considers to be virtuous. The main problem for this strategy is that our first nature is, by all accounts, capable of being shaped in a tremendous variety of ways. From the perspective of

18  In general, a substantive disagreement over moral virtue might be a disagreement over which traits count as virtues (e.g. ‘is modesty really a virtue?’) or a disagreement over what counts as the proper expression of a given virtue (e.g. ‘does modesty require that sort of behaviour?’). There is no sharp line between these two kinds of disagreement.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

The Virtues of Attachment  119 the true natural-­historical account of human beings, many of these will be regarded as interruptions of the proper unfolding of a human life—as distortions, rather than as proper developments. Still, human beings seem to have it in them to acquire the disposition to think/feel/choose in many different ways. Moreover, Kelly’s own substantive conception of the virtues is likely to embody some sensitivity to human potential. She is not likely to propose a pathway for virtue acquisition that anyone can see is more-­or-­less impossible for human beings. At the very least, then, someone pursuing this first strategy has their work cut out for them. A second strategy draws upon the distinction between different dimensions, or levels, of natural goodness in human beings. Ellen might appeal to attachment theory to argue that—given the biological, emotional, and mental needs of human beings—there is no way to bring about moral virtue, as Kelly understands it, without sacrificing proper functioning at some other level. In effect, Ellen would be arguing that virtue as Kelly understands it requires some sacrifice of physical or mental or emotional health. For such an argument to get a grip with Kelly, it would be necessary to characterize those other levels of proper functioning in a way that did not presuppose the very moral concepts that are in dispute. That is, Ellen needs to be able to say what counts as physical, emotional, or mental health without drawing upon assumptions about the moral virtues that Kelly rejects. Otherwise, the argument will presuppose what it aims to establish. Preliminary research on the costs of tiger parenting suggests that Ellen might find some empirical support for this second strategy. A single (and smaller) study among Chinese Americans found that parents who endorsed higher levels of psychologically controlling parental tactics had adolescents with greater levels of negative emotionality and lower levels of anger control and goal-­directed self-­ control (Liew et al.  2014). Assuming Kelly wants her daughter to enjoy lower ­levels of negative emotionality and higher levels of goal-­directed self-­control, these study results could put pressure on Kelly’s conception of the virtues by showing that controlling tactics come with a cost. Still, we think that at this early date in the development of the research base, Kelly can rightly insist that she need not revise her approach based on one study.19 Moreover, Kelly need not abandon her conception of the virtues, even if she acknowledges that there is no way to bring about moral virtue, as she understands it, without some cost to biological, emotional, or mental health, as she

19  One hazard of giving empirical results a role in these kinds of debates is that the achievement of a substantive research base on a topic can take decades, and even with time a clear consensus is not guaranteed. Furthermore, empirical results are neither error-­free nor self-­interpreting. Thus different interpretative frames can arrive at meaningfully different conclusions about the same empirical data. Finally, even after a scientific consensus is achieved, there is no guarantee that a scientific revolution in a field will not result in a radical reinterpretation of a once firmly held conclusion.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

120  Keith S. Cox and Micah Lott understands those notions. Instead she might simply claim that it is a fact—perhaps a tragic one—that moral virtue requires some sacrifice of non-­moral functioning. Finally, even if both of these strategies could find empirical support, neither takes us all the way to a full understanding of why a given trait is a moral virtue, or to the view of virtue that is characteristic of the virtuous themselves. For traits are not moral virtues simply because they are possible ways of shaping our first nature. Cruelty and selfishness are just as much ways of shaping first nature as are kindness and generosity. Nor are traits moral virtues simply because they foster, or are consistent with, other dimensions of proper human functioning. The point of virtue is not to promote good digestion. Rather the virtues are the proper development of our first nature; they enable us to live and act well as human beings. And it is characteristic of the virtuous to regard virtuous activity as intrinsically valuable—as noble or beautiful or ‘its own reward’. Thus in facing question (2) about virtue, we must undertake the interpretative task of showing why a trait matters in a good life, and how that trait is valuable. This task may involve showing how a trait supports some non-­moral dimensions of proper functioning. But it goes beyond that. We need an understanding of the trait’s intrinsic value. To convey such an understanding to another person, we need to present virtuous activity in a way that displays the forms of thought/ action/living that virtue makes possible, and reveals them as worthwhile. For example, if someone doubts that compassion is a virtue, and we hope to convince them that it is, we need to paint a picture of the kind of life that compassion makes possible: what compassionate activity looks like among family and friends; how compassion figures into trusting relations among neighbours; how being treated with compassion fosters honest self-­acceptance, etc. In painting such a picture, we offer an interpretation of ‘the human’. And in such interpretation, we seek a form of self-­understanding. For we seek to understand what is true of human form, and we are bearers of that form. Indeed, pursuing such self-­understanding appears to be one of the most characteristic activities of ‘the human’. Moreover, such self-­understanding is practical, insofar as it belongs to human beings to live and act from such an understanding—that is, to live on the basis of one’s understanding of living well. And this activity of practical self-­ interpretation is not an activity that properly belongs to some human beings but not others. Nor is it an activity that any of us can delegate to another. It is qua human being (rather than qua natural scientist) that each of us must strive for an intelligible and compelling interpretation of ‘the human’.

5. Conclusion In this chapter, we’ve attempted to integrate the (meta)ethical framework of Aristotelian naturalism with the psychological programme of attachment theory.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

The Virtues of Attachment  121 Our focus has been on how attachment theory, as incorporated into the Aristotelian framework, might inform a specific disagreement over how best to parent. At the outset, we noted that this disagreement points to a larger question: how might the empirical sciences vindicate, undermine, or otherwise shape our understanding of living well—or, in Aristotelian language, how might the sciences inform our substantive conception of ‘the human’? While we have not attempted to answer this larger question directly, what we have said here amounts to a partial answer. For in working through the case of Ellen and Kelly, we have sketched a conceptual roadmap that applies more broadly, to many other cases in which we might turn to the empirical sciences when thinking about how best to live.

References Badhwar, N. 2014. Well-Being: Happiness in a Worthwhile Life. Oxford: Oxford University Press. DeKlyen, M., and Greenberg,  T. 2016. ‘Attachment and Psychopathology in Childhood.’ In Cassidy. J. & Shaver, P.  R. (eds.) Handbook of Attachment. Theory, Research and Clinical Applications. 3rd edn, 639-666. New York/London: The Guilford Press. Fu, A. S., and Markus, H. R. 2014. ‘My mother and me: why tiger mothers motivate Asian Americans but not European Americans’. Personality and Social Psychology Bulletin 40(6), 739–49. Foot, P. 2001. Natural Goodness. Oxford: Oxford University Press. Groll, D., and Lott, M. 2015. ‘Is there a role for “human nature” in debates about human enhancement?’ Philosophy 90(4), 623–51. Hähnel, M. (ed.). 2020. Aristotelian Naturalism: A Research Companion. Cham, Switzerland: Springer Nature Switzerland. Harcourt, E. 2016. ‘“Mental health” and human excellence’. Aristotelian Society Supplementary Volume XC, 217–35. Hursthouse, R. 1999. On Virtue Ethics. Oxford: Oxford University Press. Hursthouse, R. 2012. ‘Human nature and Aristotelian virtue ethics. Royal Institute of Philosophy Supplement 70, 169–88. Jacobsen, T., and Hofmann, V. 1997. ‘Children’s attachment representations: longitudinal relations to school behavior and academic competency in middle childhood and adolescence’. Developmental Psychology 33(4), 703–10. Liew, J., Kwok, O., Chang, Y. P., Chang, B. W., and Yeh, Y. C. 2014. ‘Parental autonomy support predicts academic achievement through emotion-related self-regulation and adaptive skills in Chinese American adolescents’. Asian American Journal of Psychology 5, 214.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

122  Keith S. Cox and Micah Lott Lott, M. 2012a. ‘Have elephant seals refuted Aristotle?’ Journal of Moral Philosophy 9(3), 353–75. Lott, M. 2012b. ‘Moral virtue as knowledge of human form’. Social Theory and Practice 38(3), 407–31. Lott, M. 2013. ‘Does human nature conflict with itself? Human form and the harmony of the virtues’. American Catholic Philosophical Quarterly 87(4), 657–83. Lott, M. 2014. ‘Why be a good human being? Natural goodness, reason, and the authority of human nature’. Philosophia 42(3), 761–77. Lott, M. 2015. ‘Justice, function, and human form’. In M. Hähnel and M. Rothhaar (eds), Normativität des Lebens—Normativität der Vernunft?, 75–92. Berlin: De Gruyter. Lott, M. 2018. ‘Foot’s grammar of goodness’. In J.  Hacker-Wright, Philippa Foot on Goodness and Virtue, 257–76. London: Palgrave-Macmillan. Lott, M. 2020. ‘Aristotelian naturalism and the autonomy of ethics’. In M.  Hähnel (ed.), Aristotelian Naturalism: A Research Companion, 283–94. Cham, Switzerland: Springer Nature Switzerland. Marvin, R., Britner, P., and Russell, B. 2016. ‘Normative Development: The Ontogeny of Attachment in Childhood’. In Cassidy. J. & Shaver, P.  R. (eds.) Handbook of Attachment. Theory, Research and Clinical Applications. 3rd edn, 273–290. New York/London: The Guilford Press. Mesman, J., Van IJzendoorn, M.  H., and Sagi-Schwartz, A. 2016. ‘Cross-cultural patterns of attachment: universal and contextual dimensions’. 1994) Moss, E., and St-Laurent, D. 2001. ‘Attachment at school age and academic performance’. Developmental Psychology 37(6), 863. Narvaez, D. 2014. Neurobiology and the Development of Human Morality: Evolution, Culture and Wisdom. New York: Norton. Simpson, J. A. & Belsky, Jay. 2016. ‘Attachment Theory within a Modern Evolutionary Framework’. In Cassidy. J. & Shaver, P. R. (eds.) Handbook of Attachment. Theory, Research and Clinical Applications. 3rd edn, 91–116. New York/London: The Guilford Press. Spera, C. 2005. ‘A review of the relationship among parenting practices, parenting styles, and adolescent school achievement’. Educational Psychology Review, 17(2), 125–46. Thompson, M. 2008. Life and Action. Cambridge, MA: Harvard University Press. Van IJzendoorn, M.  H., and Kroonenberg, P.  M. 1988. ‘Cross-cultural patterns of attachment: a meta-analysis of the strange situation’. Child Development, 59, 147–56. Vogler, C. 2013. ‘Natural virtue and proper upbringing’. In J. Peters (ed.), Aristotelian Ethics in Contemporary Perspective, 145–57. London: Routledge. Wright, J.  C., Warren, M, and Snow, N. 2020. Understanding Virtue: Theory and Measurement. Oxford: Oxford University Press.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 26/07/21, SPi

6

Beyond Virtue The Development of Reproductive Strategies Jay Belsky

Many scholars today who study the psychology of individuals or who philo­soph­ ic­al­ly reflect on what it means to be human presume that by doing so they gain insight into human nature. But in today’s world this is typically accomplished by investigating or thinking about WEIRD people (Henrich, Heine, and Norenzayan 2010), namely, those living in Western, educated, industrialized, rich, and demo­ cratic societies. They comprise only about 12 per cent of the current population of the world. As Joe Henrich (2020) makes clear in his brilliant analysis of the devel­ opment of Western culture since around ad 1000—in a volume that could be subtitled ‘Guns, Germs and Steel 2.0’ after Jared Diamond’s (1997) remarkable volume—WEIRD people are quite different from those 88 per cent living in the rest of the world or who inhabited our planet since the first Homo sapiens walked the earth. In other words, our ways of thinking, feeling, and behaving are rather unique, even strange, but most certainly modern and not representative of most of humanity currently or historically. What might this kind of critique imply for our understanding of human virtues? Rather than drawing on Henrich’s cultural-­ ­ evolutionary analysis, in  this chapter I will rely on evolutionary-­biological thinking to critique the classical Aristotelian view of virtue. Just to be clear, the two perspectives just referred to are by no means mutually exclusive; indeed, they nicely fit together, as I make clear in the next paragraph. I will begin by articulating central assumptions of a modern evolutionary perspective on human nature which will serve as a launching point for considering—and critiquing—several long-­ standing even if not always contemporary theories of human psychology and development. I will argue that these, like the Aristotelian view of virtue, reflect a naive and romanticized view of human nature. This will lead me to develop the notion of re­pro­duct­ive strategies, a perhaps evolutionary, developmental, and contextualized view of virtue. Some final comments will be offered in a concluding section.

Jay Belsky, Beyond Virtue: The Development of Reproductive Strategies In: Attachment and Character: Attachment Theory, Ethics, and the Developmental Psychology of Vice and Virtue. Edited by: Edward Harcourt, Oxford University Press. © Jay Belsky 2021. DOI: 10.1093/oso/9780192898128.003.0007

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 26/07/21, SPi

124  Jay Belsky

1.  A Modern, Evolutionary Perspective A central premise on which this chapter is based is that of the modern evolutionary-­biological—and some would say sociobiological—view that what is virtuous in the case of humans is that which all living things seek to do, directly or indirectly. And that is to pass on their genes to future generations, either via direct descendants or collateral kin. This is not to say that such Dawkinsian self­ ishness is antagonistic toward cooperative and agreeable behaviour, as it is widely recognized—via the process of reciprocal altruism (Trivers 1971)—that giving is often in the service of getting, with getting being defined as those things that increase the passing on of genes to future generations. Dawkins (1976), of course, was communicating to a popular audience ideas which were first advanced by William Hamilton (1964). Let me point out in passing here that, without labelling it this way, Henrich’s (2020) original and illuminating analysis of cultural evolu­ tion over the past 1,000 years resulting in the economic ascendance of WEIRD societies essentially underscores a move from kin selection or groups and so­ci­ eties organized principally around kin to ones principally based on reciprocal altruism. Let me reiterate, then, what has already been stated, just so as not to be misunderstood, namely, that both ‘strategies’ (i.e. kin selection and reciprocal altruism) serve the same interest: passing on one’s genes to future generations. So, what is virtue from this perspective? If virtue for humans is about being a good specimen of one’s human kind, as Aristotle asserted, it is those things that serve fitness goals. That may involve, as with Aristotle, phenomena like honour, courage, good temper, friendliness, truthfulness, justice, and friendship. But what that classic view seems not to appreciate is that these orientations are context dependent. Imagine an African-­American slave toiling on a plantation who is forced to steal and then lie about it to his master in order to increase the nutri­ tional resources of his family. Are these not virtuous actions in an unfair world? Or consider the Jew in Spain during the Inquisition who must convert to Catholicism—and humiliate himself in his own eyes—in order to ensure the sur­ vival of himself or his family. Is this not an ‘honourable’ thing to do? These kinds of questions would seem to challenge the value of Aristotelian vir­ tues. As a further example, consider courage rather than cowardice, at least as typically considered. In terms of inclusive fitness, may it not be better, at least at certain times, to be afraid and run rather than to stand one’s ground and fight? Of course it is. How else to explain the evolved fight-­or-­flight response that we humans and many other animals display? And what about generosity, especially when one has few life-­sustaining resources? In such situations, assets should be restricted to kin who share genes, as a strategy for fostering fitness—whether one knows it or not—according to the evolutionary view of life. But if reciprocal ­altruism is possible—you scratch my back and I scratch yours—then being generous to non-­kin can also be a fitness-­enhancing strategy. Once again, context matters immensely.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 26/07/21, SPi

Beyond Virtue  125 The central critique of Aristotetelian virtues being advanced herein is that it represents an implicit if not explicit naive, idealistic, and romantic view of human nature. And this is an inclination one can see in many views of human nature that have held sway in Western thinking, as evidenced by widely embraced psycho­ logical theories to which I now turn. Recall that the forthcoming critique will lay the groundwork for an alternative evolutionary, developmental, and context-­ dependent perspective on what is virtuous.

2.  Psychological Views of Human Nature and the Perfectibility of Man Long-­standing perspectives on human development have failed to appreciate the biological foundations of all living things on planet earth, as reflected in modern biological thought underscoring the fundamental significance of fitness. This latter point needs to be appreciated to highlight an important distinction in the biological study of human psychology. While there is great emphasis today on the role that many diverse biological processes play in human psychology and development— whether focused on stress-­response physiology, genetics, epigenetics, and even most recently the microbiome—little of this work illuminating the ‘how’ of ­psychology and development considers, much less addresses, the ‘why’ of ­psych­ology. By that I mean that most modern biopsychology fails to reflect on whether and how the psychobiological mechanisms under investigation directly or indirectly serve the ultimate goal of all living things, the passing on of genes to future ­generations. That is, why they evolved in the first place to be central features of our species. This approach to illuminating the ‘how’ without considering the ‘why’ of biopsychology risks, metaphorically, trying to understand aero­dynam­ics while ignoring gravity. The three psychological perspectives to be considered, perhaps informed, knowingly or unknowingly, by Aristotelian views of virtue and certainly by the Enlightenment, have, implicitly or explicitly, advanced idealized and romantic views of human nature, even if they do not disregard the importance of context. To my way of thinking, all are based on a claim that Homo sapiens is a perfectible species. Let’s begin with Abraham Maslow, before considering the thinking of Erik Erikson and John Bowlby and his followers. I will end by making some com­ ments of what might be regarded as more contemporary theoretical perspectives.

2.1  Abraham Maslow’s theory of self-­actualization Maslow (1943) advanced a theory that humans can, would, and do develop to realize their full human potential—self-­actualization he called it—if and when allowed or enabled to do so. In his work, like that of others, this becomes the

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 26/07/21, SPi

126  Jay Belsky fundamental nature of humans—when given the opportunity to express it. Healthy—or what would often be referred to today as ‘optimal’ development— requires first and foremost the meeting of basic biological and psychological needs. Once these are fulfilled, the question turns to safety (meaning protection), security, order, and stability. Fulfilment of these needs then means that the real­ iza­tion of human potential requires the meeting of needs for belonging and love, as provided via family, affection, and relationships. The meeting of these needs in turn affords the possibility of fulfilling esteem needs; these involve achievement, status, responsibility, and reputation. And, finally, only once all these levels of need are met does self-­actualization become possible, involving personal growth and fulfilment. But what exactly is self-­actualization? It is what many might think of as being psychologically mature and secure, including the following: — a keen sense of reality—being aware of real situations—and thus the ability to make objective and not just subjective judgements; — the capacity to see problems in terms of challenges and as situations requir­ ing solutions, rather than viewing them as personal complaints or excuses; — the desire for privacy and thus the ability to be comfortable while alone; — the proclivity to rely on one’s own experiences and judgements, to think independently and not rely on cultural and environmental factors to form opinions and views; — the embracing of a democratic, fair, and non-­discriminating worldview and thus being able to embrace and enjoy all cultures, races, and individ­ ual styles; — the tendency to be socially compassionate, adopting a humanistic orienta­ tion toward others; —  the readiness to accept others as they are rather than trying to change people; — the establishment of a limited number of close intimate friendships rather than collecting a wide array of superficial relationships; — the possession of a sense of humour directed at oneself or the human con­ dition, rather than at the expense of others; — possessing a spontaneous and natural character, being true to oneself, rather than being how others want one to be; —  the tendency to be excited and interested in everything, even or­ din­ ary things; — and the capacity to be creative, inventive, and original. One might compare the Maslow scheme to when and how a beautiful flower becomes realized. In the seed is the capacity for such beauty. But only if watered,

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 26/07/21, SPi

Beyond Virtue  127 fertilized, exposed to the sun, and weeded will this inherent ‘nature’ of the flower be realized. To Maslow, then, humans are like flowers; they have the capacity for beauty, but nature’s realization is not guaranteed. Failure to realize it does not deny the thing’s nature; it simply reflects its lived experience.

2.2  Erik Erikson’s eight stages of development Maslow was a psychologist, but not one that could be considered a developmen­ talist, even if clearly a contextualist. But as a contextualist, he viewed failure to have foundational needs met as resulting in what today might be referred to as ‘dysregulated’ development; that is, the inability to realize one’s inherent, full, and beautiful—and virtuous?—potential. Erik Erikson also had a contextual and virtuous view of human nature, but it was cast in developmental perspective while operationalized in terms of eight stages of growth and change. Like Maslow, Erikson recognized the foundational importance of psychological security; and thus in his (idealized) model of human development, ‘basic trust’ is the first stage, which would or would not be achieved during infancy. If the child’s lived experiences fostered security, then, in toddler­ hood, the possibility of developing a sense of autonomy—rather than shame and doubt—became possible. This second stage, also shaped by lived experiences (as indeed were all eight stages), afforded the prospect of establishing the capacity for initiative rather than guilt during the preschool years. The developmental chal­ lenge during middle childhood was to embrace industry rather than inferiority; and, if successful, this set the stage for the development of identity during adoles­ cence, a clear sense of self rather than role confusion. Successful ‘negotiation’ of all these prior stages afforded the possibility during young adulthood for develop­ ing the capacity for intimacy rather than isolation; and this set the stage for the ability to care for the next generation—generativity rather than stagnation. Finally, in the ageing years, a sense of integrity rather than despair became possible. Once again we see the implicit metaphor of a flower, or perhaps a better one is of the caterpillar-­butterfly. Success at an earlier stage affords but does not guarantee the possibility of success—that is, realizing the organism’s full potential—at a later stage. Clearly, though, we humans have more (psychological) metamorphoses than the butterfly. And, like Maslow, Erikson’s idealized human nature required nurture. As with self-­actualization, realizing the inherent possibilities of the human was by no means assured; and, as with Maslow, failure to realize the posi­ tive and inherent possibilities of the human condition resulted in compromised, dysregulated, or maladaptive development. It could take the form of basic mis­ trust, shame and doubt, guilt, inferiority, role confusion, isolation, stagnation, and/or despair.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 26/07/21, SPi

128  Jay Belsky

2.3  John Bowlby’s attachment theory John Bowlby was the second evolutionary psychologist, after Darwin himself. Bowlby sought to explain why children develop strong emotional ties to their pri­ mary caregiver, which he labelled an ‘attachment’. Rather than explaining such psychological orientation via a process of learning (actually classical condition­ ing) as Freud did—‘caregiver feeds me, which makes me feel good, so I associate pleasure with her and she becomes a rewarding stimulus’—he turned to human ancestral history. Bowlby’s core Darwinian argument (Bowlby  1969) was that attachment developed in humans and other primates because in ancestral times it aided and abetted survival. The simple argument was that children who did not develop attachments were less likely to have descendants because they were more likely to die in early childhood; this could be due to their physically distancing themselves from their parent, especially as they developed the capacity to loco­ mote, and be taken by a predator, get lost, or fall into a fire. In contrast, those children who established strong emotional ties to their caregivers were more likely to stay close to them, thereby increasing their safety, and thus live to become parents themselves. As such, the genetic inclination to establish attachments came to be foundational to the make-­up of modern-­day humans. It was, simply put, a product of Darwinian natural selection. Further development of attachment theory by theorists and researchers like Mary Ainsworth (Ainsworth et al. 1978) and Alan Sroufe (1979) called attention to individual differences in the quality—that is, security—of attachment and its origins and sequelae. While most children develop attachments to one or more caregivers, except under unusual circumstances (e.g. severely deprived Romanian orphanages in Communist times), the quality of care they received was viewed as influencing whether an attachment was secure and insecure; and this was regarded as developmentally important because it set the child on a ‘developmen­ tal trajectory’. That is, it influenced how the child would develop in the future. And once again we see here an idealized and romantic view of human nature. Children sensitively cared for by a caregiver who was attentive, responsive, and provided appropriate care—perhaps especially when it came to providing com­ fort at times of illness, fear, and/or distress—were more likely than their insecure counterparts who failed to experience such care to develop ‘optimally’ or ‘compe­ tently’. That is, they were more likely—and here recall Erikson’s developmental achievements—to become curious and autonomous; to have the capacity for emotion regulation and stress resilience, as well as, eventually, for intimacy. They were also more likely as a developmental result to provide security-­inducing care when they became parents themselves and to be industrious. Freud defined mental health as the capacity ‘to love and to work’, which these developmental capabilities of adults surely seem to reflect. And once again the flower metaphor seems apt. These ‘outcomes’ of a particu­ lar developmental trajectory are by no means assured. To be realized they depend

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 26/07/21, SPi

Beyond Virtue  129 on continued contextual support. Perhaps a big difference between Erikson’s ­theory and attachment theory is that the latter viewed the child as a ‘producer of its own development’ (Belsky and Tolan 1981) in that early-­life capabilities derived from sensitive, security-­inducing, early-­life care were themselves influential in shaping future development. In a trajectory model, development takes on a life of its own, increasing the likelihood of some future ways of functioning rather than others. These include, in the case of those with secure relational foundations, the capacity to face challenge and resist the dysregulating effects of later-­life adversity— at least up to a point. Today we call this ‘resilience’. Grand theories like those just summarized are more or less absent in con­tem­ por­ary psychology. Perhaps this is due to the desire to focus—carefully—on psy­ chological mechanisms and the vagueness and imprecision of grand theories. In any event, naive, romanticized views of human nature still pervade the study of human development. This was already highlighted in the choice of terms used to describe what happens when the perfectibility of man is stymied by contexts that fail to foster it. Recall the use of terms like ‘optimal’, ‘dysregulated’, ‘maladaptive’— and perhaps add to them those of ‘disorder’. These are terms reminiscent of health and disease, and thus reference to a medical model of human development. But, as I will argue in the next section, that which is glibly labelled so often as ‘optimal’ or ‘dysregulated’ in the study of human development today is based on a WEIRD view of the world. It is my central contention that we confuse what we like and prefer and thus value with what we deem to be, for cultural and contextual ­reasons, virtuous and the more or less intended goal of human development, namely, the perfectibility of man.

3.  Evolutionary Challenge: Reproductive Strategies As stated in the opening of this chapter, there is no optimal development from an evolutionary perspective, with the single exception of succeeding in passing on (more rather than fewer of) one’s genes to future generations (i.e. inclusive fit­ ness). Important to appreciate here is that this does not mean that having more offspring or collateral kin is always the best way to promote inclusive fitness, though it certainly might be. For example, having a stellar reputation that enhances one’s own status—and, thereby, that of one’s children and those of the family for many subsequent generations—could prove more beneficial than just having more offspring. This would seem so especially if such status enhancement attracts high-­quality mates and resources. Think perhaps of the descendants of famous people—who inherit more than just the famous person’s genes. In any event, the central point to be made is that if there is any sense to the idea of what we as a species are ‘supposed’ to do, then the only thing that is optimal is that which succeeds in fostering inclusive fitness. It is not hard to imagine that in a safe and secure world what yields this result is just the course of development

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 26/07/21, SPi

130  Jay Belsky and manner of functioning outlined in the three preceding theoretical f­ rameworks— Maslow’s, Erikson’s, and Bowlby’s. Perhaps the same is true when it comes to realizing Aristotle’s virtues. But what about in a world of danger, fear, and insecurity? Is it dysregulated or dysfunctional to be distrusting, advantage-­taking, and ­anti-­social rather than cooperative and caring? That would not necessarily seem to be the case. How, then, can we regard Aristotle’s virtues as virtues for all times and places? Hitting first and asking questions later might make a lot more sense when it comes to survival and reproduction than doing the opposite. Indeed, may this be why victims of child abuse prove to be more prone to anger in the face of uncertainty and more vigilant to threat, as well as more aggressive, than children not so treated (Teisl and Cicchetti 2008)? Central to the branch of evolutionary biology known as life-­history theory and focused on ‘reproductive strategies’ is appreciation that life and development involve trade-­offs; and this is because there is no always-­better way to develop, think, and behave that will foster survival and reproduction (Bogin, Silva, and Rois  2007). Three trade-­offs or decisions are central when it comes to how to organize a life, though these are by no means conscious considerations (Del Giudice, Gangestad, and Kaplan  2015). First, do I continue to grow, while em­body­ing nutritional, educational, and psychological capital that will enable me to attract a better mate and care for my children well, or do I reproduce? In other words, do I privilege current or future reproduction? Secondly, once I have children, do I invest my time and energies heavily in them or in further mating, perhaps with others than the parent of my existing children? And, finally, do I emphasize number of offspring or their quality, as influenced by my investment in parenting? The correct—meaning fitness-­enhancing—answers to these questions shape the structure of life according to evolutionary life-­history theory. If and when survival is threatened, the choice is for reproduction sooner rather than later, thus privileging it over further growth in the service of embodied capital (Kaplan, Lancaster, and Robson 2003). If my resources are limited and there is little I can do to invest in my children in a way that will enhance their survival and repro­ duction, then I should seek additional matings. Relatedly, under these conditions I should bear more children even if, and indeed because, of my limited ability to care for them well. All these decisions, under these conditions, increase the chance that I will enhance my inclusive fitness. But can we move beyond such theoretical and logical arguments to advance the notion that what may be virtuous—or optimal—from an evolutionary perspective is that which fosters fitness? This is a challenge I posed to myself as a developmental scholar many years ago now (Belsky, Steinberg, and Draper  1991)—basically because I was not wedded to the argument being advanced in this chapter. To my way of thinking, if an evolutionary view emphasizing survival and fitness should replace claims about optimal development and the realization of inherent potential,

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 26/07/21, SPi

Beyond Virtue  131 then it had to yield predictions—and evidence—that could not be accounted for by these more traditional ways of thinking about human development. My thinking was thus informed by Thomas Kuhn’s (1962) philosophy of science which stipulated that a good theory should do three things if it is to trump and replace existing ways of thinking: explain what is already understood, generate novel predictions, and, in consequence, generate discoveries that would not be expected or explicable on the basis of existing understanding. This ultimately led to what can be referred to as ‘the puberty hypothesis’.

3.1  The puberty hypothesis To my way of thinking, the idea that supportive rearing conditions would pro­ mote optimal development defined in terms of psychological phenomena such as emotional security and the capacity to love and to work was tolerably well estab­ lished, just like the idea that adverse and especially hostile rearing conditions would generate the opposite. So evidence consistent with this view might be  ne­ces­sary but not sufficient to lead me—or anyone else—to embrace what I  regarded at the time as a less rather than more parsimonious explanation of human development. After all, if existing theories like attachment theory could explain such variation in development, at least to some extent, and was only based on the child’s lived experience—what happened to him or her while growing up— then why consider the process of evolution by natural selection as an explanation? The same went for existing evidence that children growing up under adverse conditions often became sexually active at a younger age, bore more children, but cared for them less intensively than those growing up under more benign if not supportive conditions (Duncan and Brooks-­Gunn 1997). Was not this, after all, evidence that fitness-­defined reproductive outcomes varied as a function of developmental experience? Case closed, it would seem. But, once again, simple social learning processes could account for these phenomena, so why embrace a seemingly less parsimonious evolutionary view? After all, Occam’s Razor favoured less rather than more complex explanations. If girls saw their single mothers involved with a series of men while growing up, then they just learned to do the same, resulting in early and unwanted pregnancies. And if boys had a badly behaving father to model themselves on or perhaps worse, a series of such men moving through their childhoods who did not treat them well, why would they not also prove to be opportunistic advantage takers, just like the men they were exposed to? A focus on insecure attachment or social learning would seem already to explain what was widely documented. So, again, why embrace an evo­ lutionary view? In the only eureka experience I ever had as a scientist, I realized that adverse childhood experiences of the kind just referred to would not just affect

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 26/07/21, SPi

132  Jay Belsky psychological and behavioural development as long appreciated, but somatic development as well, by fostering earlier sexual maturation, that is, puberty (Belsky, Steinberg, and Draper  1991). Why? Because in the face of risks that would lead one to appreciate, again not necessarily consciously, that the world was a risky, uncaring place, and that others could not be counted on to provide support and nurturance, it made sense not only to be an advantage taker, but to mature earlier as well. This way, before the risk of developmental compromise (in mental health terms) was fully realized, conceivably undermining one’s mating prospects and the ability to forge stable pair bonds and raise children with care and con­sid­er­ation, it made sense to mature early. This would, derivatively and most critically, afford the opportunity to mate sooner, conceive children sooner, have more children, and, consequently, care for them less well than others reared under better conditions. These others, my evolutionary theory of ‘reproductive strategies’ stipu­lated, could afford to bide their time to mature, mate, and rear offspring. Indeed, because of their own enhanced capabilities, they would be bet­ ter positioned after doing so to attract a high-­quality mate, establish and maintain enduring pair bonds, and raise fewer offspring with more care and consideration. And, also critically, this would increase their offspring’s own chances of survival and reproduction, which would mean that bearing fewer children, cared for well, could be more fitness-­enhancing than rearing more, cared for poorly. In the time since I first advanced these ideas two things have occurred. First, evolutionary-­developmental thinkers more knowledgeable than myself made clear that what early maturation really afforded was the increased opportunity to have progeny before dying (Chisholm 1999)! After all, early-­life adversity brought with it the risk of premature death, so getting to the starting gate—sexual maturation— sooner rather than later would decrease the risk of incurring this most costly reproductive expense, dying before reproducing. This extension of my thinking was certainly consistent with evidence that early-­life adversity increased the risk of childhood death (Pritchard and Williams  2011), and greatly strengthened the theoretical case for what I referred to as an ‘uncanny prediction’ regarding pubertal development—because, consistently with Kuhn, it could not be explained, should it be empirically supported, by any existing theories of development. In the thirty years since the publication of my evolutionary theory of socializa­ tion and reproductive strategy, evidence has emerged consistent with it, especially in the case of females, but not exclusively so (Belsky 2012). Indeed, prospective longitudinal research stimulated by the theory links a variety of early-­life adversi­ ties—such as child maltreatment, father absence, and foster care—with not only earlier sexual maturation, but, thereby, earlier onset of sexual behaviour, an obvi­ ous and necessary condition for reproduction. Consider, as just one example, work showing that harsh parenting during the preschool years predicts earlier age of menarche (even after controlling for maternal age of menarche) and, thereby, sexual risk taking (i.e. frequency of oral sex, vaginal sex, pregnancy, diagnosed

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 26/07/21, SPi

Beyond Virtue  133 STD) (Belsky et al. 2010). These, to repeat, are not findings that theories that still shape the study of human development can explain. Even more compelling than results of observational studies, even if lon­gi­tu­ din­al in design—following lives over time so that early-­life experiences can serve as predictors of the timing of pubertal development—are the results of natural experiments. Consider in this regard the following investigations with findings in line with the puberty hypothesis. The first is a long-­term study of the effects of children who experienced traumatic separation from their Finnish parents dur­ ing the Second World War in hopes of protecting them from the Nazis (Pesonen et al. 2008). The second focuses on children exposed to an earthquake in China (Lian et al. 2018). What makes such research on effects of early-­life adversity on pubertal development more convincing than observational research in which nothing is manipulated and statistical associations are examined is that there is no real chance of what scientists refer to as ‘selection effects’. After all, in observa­ tional studies it could be that parents genetically disposed to mature early them­ selves not only pass on their influential genes to their children, but are also more likely, perhaps due to the very same genes, to expose their children to adverse childhoods. Such ‘selection’ into such environments would make study results spurious, undermining the case that early-­life adversity is promoting earlier sex­ ual maturation. What natural experiments do, then, is what regular experiments do, more or less—randomly assign individuals to ‘treatments’.

3.2  Beyond the puberty hypothesis In the years since the puberty hypothesis was advanced, ever more work indicates that early-­ life adversity accelerates biological development (Belsky  2019). Intriguingly, this work focuses on cellular processes, specifically telomeres and the epigenetic methylation of genes. Telomeres are the ends of chromosomes and can be likened to the bindings at the ends of shoelaces. Every time a cell divides, telomeres shorten, eventually disappearing, at which point a cell dies; think of how a shoelace unravels once one of its terminal bindings is gone. Thus, many consider telomeres as an indicator of biological ageing. Ever more research indi­ cates that early-­life adversity is associated with shorter telomeres (Belsky and Shalev  2016), including, for example, work linking prenatal stress and shorter telomere length at birth (Entringer et al. 2013) and violence exposure in middle childhood and the actual shortening or ‘erosion’ of telomeres from 5 to 10 years of age (Shalev et al. 2013). Epigenetic methylation is a process by which the expression of genes—and thus whether or not they are transcribed and affect the phenotype—is regulated, with methylation turning off gene expression (and other chemical processes ­turning them on). Because the methylation of a select set of genes is strongly

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 26/07/21, SPi

134  Jay Belsky associated with ageing, the methylation of the relevant genes is also regarded as an index of biological ageing, making it a kind of ‘biological clock’. And, as it turns out, adverse childhood experiences are associated with the accelerated methyla­ tion of the genes in question and thus accelerated biological ageing. Consider in this regard evidence indicating that exposure to neighbourhood violence between the ages of 6 and 13 years is associated with epigenetic age acceleration (Jovanovic et al.  2017), as is lifetime stress exposure among adults (Zanas et al.  2015). Notably, accelerated epigenetic ageing—in childhood—itself predicts accelerated pubertal development (Binder et al. 2018). Research on the effects of childhood adversity on cellular ageing is not only consistent with that targeting puberty, but both bodies of work also underscore the notion of evolutionary-­developmental trade-­offs. And this is because earlier puberty, shorter telomeres, and the methylation of select genes defining the ‘epi­ genetic clock’ are each related to increased morbidity and mortality risk. What seems to be occurring is that Darwinian natural selection has traded off the increased chance of reproduction—by accelerating biological development—for poorer health and premature death in later life. In other words, nature privileges reproduction over health, happiness, and longevity! Such an analysis, of course, is exactly what gave rise to the evolutionary theory of socialization and re­pro­duct­ ive strategy and, thereby, the puberty hypothesis.

4.  Conclusion Even if we can surely debate the meaning of life, its purpose and goal seem rather clear. And it is to pass genes on to future generations, a principle that is founda­ tional to life on this planet—almost by definition and, perhaps, tautologically so. As stated at the outset, there is no psychological, behavioural, or biological ‘opti­ mal development’, that is, ‘best’ way of developing—because what promotes inclu­ sive fitness varies depending on ecological context. Nor should we entertain a notion of the perfectibility of man, a view which seems implicitly if not explicitly foundational to an Aristotelian view of virtues and many classic and con­tem­por­ ary perspectives on human development. Humans are on this planet for the same reason that are all other life forms; and it is to make more life that is genetically related to individuals alive at any point in time. From this standpoint, we are not designed to realize some inherent—and admirable—nature, but to reproduce, again, just like all other living things. This does not mean that being virtuous in the Aristotelian sense is a waste of time and effort and that we should take advantage of others, perhaps by raping and pil­ laging. And this is because that is not always the most strategic way to promote inclusive fitness. When cooperation with and care for those whom we do not share genes with yields the reciprocation of mutual benefit, then this is a

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 26/07/21, SPi

Beyond Virtue  135 re­pro­duct­ively strategic way to behave, on average, even if not always. But such enlightened conditions are by no mean those that currently characterize, or ever have characterized, human experience. As a result, we develop differently and value different ways of functioning due to our developmental experiences and con­text­ual exposures. We may not like or prefer this theory of human nature, with its less romantic view of human virtues, but it is the one we have been bequeathed.

References Ainsworth, M.  D.  S., Blehar, M.  C., Waters, E., and Wall, S. 1978. Patterns of Attachment. New York: Erlbaum. Belsky, J. 2012. ‘The development of human reproductive strategies: progress and prospects’. Current Directions in Psychological Science 21, 310–16. Belsky, J. 2019. ‘Early-life adversity accelerates child-adolescent development’. Current Directions in Psychological Science 28, 241–6. Belsky, J., and Shalev, I. 2016. ‘Contextual adversity, telomere erosion, pubertal devel­ opment and health: two models of accelerated aging—or one?’ Development and Psychopathology 28, 1367–83. Belsky, J., Steinberg, L., and Draper, P. 1991. ‘Childhood experience, interpersonal development and reproductive strategy: an evolutionary theory of socialization’. Child Development 62, 647–70. Belsky, J., Steinberg, L., Houts, R. M., and Halpern-Felsher, B. L. 2010. ‘The develop­ ment of reproductive strategy in females’. Developmental Psychology 46(1), 120–8. Belsky, J., and Tolan, W. 1981. ‘Infants as producers of their own development: an ecological analysis’. In R.  Lerner and N.  Busch-Rossnagel (eds), Individuals as Producers of their Own Development: A Lifespan Perspective, 87–116. New York: Academic Press. Binder, A. M., Corvalan, C., Mericq, V., Pereira, A., Santos, J. L., Horvath, S., Shepherd, J., and Michels, K. B. 2018. ‘Faster ticking rate of epigenetic clock is associated with faster pubertal development in girls’. Epigenetics 13, 85–94. Bogin, B., Silva, M. I. V., and Rios, L. 2007. ‘Life history trade-offs in human growth: adaptation or pathology’. American Journal of Human Biology 19, 631–42. Bowlby, J. 1969. Attachment and Loss, vol. 1: Attachment. New York: Basic Books. Chisholm, J. S. 1999. Death, Hope and Sex. Cambridge: Cambridge University Press. Dawkins, R. 1976. The Selfish Gene. Oxford: Oxford University Press. Del Giudice, M., Gangestad, S. W., and Kaplan, H. 2015. ‘Life history theory and evo­ lutionary psychology’. In D. M. Buss (ed.), The Handbook of Evolutionary Psychology (2nd edn), 88–114. New York: Wiley. Diamond, J. 1997. Guns, Germs, and Steel. New York: Norton.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 26/07/21, SPi

136  Jay Belsky Duncan, G., and Brooks-Gunn, J. 1997. Consequences of Growing Up Poor. New York: Russell Sage. Entringer, S., Epel, E. S., Lin, J., Buss, C., Shahbaba, B., Blackburn, E. H., and Wadhwa, P. D. 2013. ‘Maternal psychosocial stress during pregnancy is associated with new­ born leukocyte telomere length’. American Journal of Obstetrics and Gynecology 208, 134.e1–7. Erikson, E. 1950. Childhood and Society. New York: W. W. Norton. Hamilton, W.  D. 1964. ‘The genetical evolution of social behaviour’, parts I and II. Journal of Theoretical Biology 7, 1–52. Henrich, J. 2020. The WEIRDest People in the Word: How the West Became Psychologically Peculiar and Particularly Prosperous. New York: Farrar, Straus and Giroux. Henrich, J., Heine, S. J., and Norenzayan, A. 2010. ‘The weirdest people in the world?’ Behavioral and Brain Sciences 33, 61–83. Jovanovic, T., et al. 2017. ‘Exposure to violence accelerates epigenetic aging in chil­ dren’. Scientific Reports 7, 8962–7. Kaplan, H., Lancaster, J., and Robson, A. 2003. ‘Embodied capital and the evolutionary economics of the human life span’. Population and Development Review 29, 152–82. Kuhn, T. 1962. The Structure of Scientific Revolutions. Chicago: University of Chicago Press. Lian, Q., et al. 2018. ‘The impact of the Wenchuan earthquake on early puberty: a natural experiment’. PeerJ 6:e5085. DOI: 10.7717/peerj.5085. Maslow, A. 1943. A Theory of Human Motivation. Morrisville, NC: Lulu Press. Pesonen, A., et al. 2008. ‘Reproductive traits following a parent–child separation trauma during childhood: a natural experiment during World War II’. American Journal of Human Biology 20, 345–51. Pritchard, C., and Williams, R. 2011. ‘Poverty and child (0–14 years) mortality in the USA and other Western countries as an indicator of “how well a country meets the needs of its children” (UNICEF)’. International Journal of Medical Health 23, 251–5. Shalev, I., Moffitt, T. E., Sugden, K., Williams, B., Houts, R. M., Danese, A., Mill, J., Arseneault, L., and Caspi, A. 2013. ‘Exposure to violence during childhood is asso­ ciated with telomere erosion from 5 to 10 years of age: a longitudinal study’. Molecular Psychiatry 18, 576–81. Sroufe, L. A. 1979. ‘The coherence of individual development’. American Psychologist 34, 834–41. Teisl, M., and Cicchetti, D. 2008. ‘Physical abuse, cognitive and emotional processes, and aggressive/disruptive behavior problems’. Social Development 17, 1–23. Trivers, R. L. 1971. ‘The evolution of reciprocal altruism’. Quarterly Review of Biology 46, 35–57. Zanas, A.  S., et al. 2015. ‘Lifetime stress accelerates epigenetic aging in an urban, African American cohort’. Genome Biology 16, 266–77.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 26/07/21, SPi

7

Two Routes from Secure Attachment to Virtue Edward Harcourt

1.  Introduction Why is it of interest to try to sketch any route from secure attachment to virtue, or indeed from attachment in any form either to virtue or to vice? I shall attempt an answer to that question before I introduce the two routes of my title. Among the many things attachment theory tells us, it tells us about the developmental antecedents of people’s attachment classifications—secure, insecure-­ ambivalent, and so on. Classically it has had a great deal to say about their developmental antecedents in infancy, but the classifications don’t only apply to infants nor is there any reason to think the developmental story ends there. Secondly, the conceptual vocabulary of attachment theory is surely a char­ac­ter­ ology. The attachment classifications themselves look like character traits, at least because they are relatively stable (which is not to say fixed over time: but they don’t come and go like moods), and they also look explanatory,1 in that they are predictive (both at a time, and to some extent over time) of a range of other traits. Thirdly, the range of traits which attachment classification predicts (either at a time or over time) are often of interest (subject to endless qualifications, of course, about context, motivation, etc.) in the ascription of virtues and vices. For example these traits are connected with concern for others (e.g. a person’s sociability and cooperativeness) or, alternatively, with their being antisocial in various ways; with the quality of their executive function (e.g. their ability to concentrate or to complete a task); and with how well or ill regulated they are in their emotions. Finally, the differences between being securely attached and other attachment classifications are not bare differences, like differences between colours, say. On the

1  I say ‘look’ explanatory, because there is a question whether attachment classifications are underlying characteristics that explain sociability, emotion regulation etc., or whether the latter traits are simply the criteria for ascribing attachment classifications and so not to be seen as related to them as effects to causes. The same question arises for the explanatoriness of virtues and vices, however, so the existence of the question doesn’t threaten the observation that the attachment classifications are a characterology. Edward Harcourt, Two Routes from Secure Attachment to Virtue In: Attachment and Character: Attachment Theory, Ethics, and the Developmental Psychology of Vice and Virtue. Edited by: Edward Harcourt, Oxford University Press. © Edward Harcourt 2021. DOI: 10.1093/oso/9780192898128.003.0008

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 26/07/21, SPi

138  Edward Harcourt contrary, attachment classifications are evaluatively differentiated: there are goods realized by being securely attached that variously go unrealized by people who belong in the other classifications. I don’t mean by this that being securely attached is always the more advantageous state to be in for the person in it: there may be environments sufficiently hostile that it is more advantageous to be (say) insecure-­avoidant (Belsky 2008). But I take it that there are environments sufficiently hostile that it is advantageous to be bad in all sorts of ways, so the point about hostile environments and secure attachment is consistent with maintaining that secure attachment is the best attachment classification to belong to. Now Aristotle’s virtues and vices are relatively stable traits, that explain features such as sociability and emotion-­regulation (or their absence), and—obviously— they are evaluatively differentiated (the virtues are the good ones). To have two different ways of carving up what looks like the same territory simply sitting side by side is a theoretical embarrassment—surely at least one must be at least partially mistaken, or else there must be a way of merging the two taxonomies to get a more complete picture. What’s more, Aristotle gave us only the barest outline of a developmental history of virtue, telling us that we acquire virtues by ha­bitu­ ation. So if we could somehow map one characterology onto the other, one of the pay-­offs might be that we get a better developmental story about virtue and vice than Aristotle’s. Thus routes from attachment to virtue or vice, or from secure attachment to virtue, are something anyone interested in virtue and vice should be interested in. (In fact, as my title implies, I’m only going to talk about secure attachment and virtue here—vice will have to wait.) Now to the ‘two routes’. It’s of course in part an empirical question what—if anything—the connection is between secure attachment (as measured at this or that age), its developmental antecedents, and the various virtues. That is, one shouldn’t believe a philosopher who says they are connected without following up with some empirical research. However, the point of my ‘two routes’ discussion is to head off an a priori objection to the effect that, whatever empirical research purports to show to be actually the case, no connection between secure attachment and virtue can possibly obtain. So, first of all, I rehearse and indeed reinforce the philosophical reasons why one route which might be thought to be open from secure attachment to virtue is—whatever the empirical research seems to tell us— in fact closed. But, secondly, the message is ‘don’t give up’: the first route was always going to be the wrong place to look for a connection, and there’s a second route, or so I shall argue, which is still open.

2.  The First Route from Secure Attachment to Virtue Now I turn to the first supposed route from secure attachment to virtue. The availability of this route rests on two thoughts, each of which has a fair amount to

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 26/07/21, SPi

Two Routes from Secure Attachment to Virtue  139 be said in its favour, and each of which I will simply state before unpacking it further. The first is that love, or the capacity to love, makes us good. The second is that secure attachment is closely related to love, or to the capacity to love. If both are true, it should follow that secure attachment makes us good—not of course in the sense that it necessitates goodness in its possessor, but that it is a favoured route to goodness. Why might one think that love makes us good? Jealous or possessive love, after all, is still love, and it would seem to make us bad rather than good. But this is a verbal problem. There are really two relevant senses of the word ‘love’. When someone walks out of a gallery protesting that what he’s seen ‘isn’t art’, he is not on the whole saying that what he has seen doesn’t fit a definition of art, but rather that it’s bad art. The same ambiguity between an honorific or idealizing usage on the one hand and, on the other, something like a definition, occurs with ‘love’. Possessive love is indeed a bad variety of love but insofar as that’s what it is a bad variety of, still love. But when people say that love makes us good, what they mean is not that anything definable as love makes us good, but rather that love in its idealizing sense—love at its best—makes us good. That’s the thought whose appeal we need to explore. This thought is perhaps best captured by a well-­known line—though it seems to be a misquotation—from Augustine: ama et fac quod vis, ‘love and do what you will’.2 Get the (good) love disposition—the thought goes—and everything else comes for free, so you only have to do what you feel like doing—that is, with the tug of duty against impulse nowhere to be felt—and you won’t go wrong. Or the thought could be put in a developmental way: to get good at loving is to become excellent as a human being. This might of course be taken to mean that to get good at loving is to acquire a specific excellence—the excellence of being capable of love in, for example, a non-­possessive way—and thereby to become excellent as a human being in just that way. That might be true, but it would be a far less interesting thought than the one I have in mind. The thought I have in mind is that to get good at loving is to become not just narrowly excellent at loving but excellent overall—to possess all the virtues (if there is such a thing as all of them). This thought seems to lie close to the heart of the more sophisticated love-­ comedies. Though the reader is prepared for the marriage between Emma and Mr Knightley almost from page one, as Emma slowly learns to love him—which in part involves acknowledging that she already does—she doesn’t just get good at loving him and being loved by him, but also becomes more self-­knowing, less 2  In fact he seems to have said ‘dilige et quod vis fac’ (Augustine 2000: 7.8 l 36), which for present purposes may mean something less interesting, because ‘diligo’—unlike ‘amo’—may express an attitude that’s already further from personal love and closer at least to benevolence if not to virtue more generally. But I do not want to go further than ‘may’. If even the misquotation cannot do the work I’m assigning to it here, never mind—it is the idea I want, not what Augustine really meant. Thanks to Terence Irwin for the quotation, and for discussion.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 26/07/21, SPi

140  Edward Harcourt proud, less interfering, less oblivious of others’ needs and wants, and (as Mr Knightley himself emphasizes) more rational (Austen 1816/n.d.: 418). That is, she seems to acquire a suite of virtues which go way beyond the capacity to love, but which her having acquired that capacity somehow explains. Examples could be multiplied indefinitely, from Shakespeare to When Harry Met Sally.3 John Bowlby—the founder of attachment theory—would have had no difficulty with the second thought, that secure attachment is closely related to love in its best form, indeed he is to be found using ‘love’ and ‘attachment’ more or less interchangeably (Bowlby 1976–7/2005). Moreover one might think the fact that attachment comes in evaluatively differentiated sub-­varieties—secure, insecure-­ ambivalent, and so on—creates a kind of isomorphism between the concept of attachment and the concept of love, since love too (as I’ve already said) comes in good and bad sub-­varieties. How that putative isomorphism plays out in detail I’ve already tried to deal with elsewhere (Harcourt  2017). Here I’m going to explore a more modest version of the connection between attachment and love which focuses solely on secure attachment and love at its best, which consists in the claim that there is a very close association between secure attachment and what I will call the ‘virtues of intimacy’. To see the link between secure attachment and the virtues of intimacy, let us recall that secure attachment is associated with at least the following: 1. Delight in the other’s presence, including (but not only) the capacity to enjoy physical warmth and proximity (touching, holding hands, etc.). Attachment theorists note the ‘striking similarity in the physical intimacy that typifies lovers and mother–infant pairs’: kissing, touching, mutual absorption, and so on (Zeifman and Hazan 2008: 440). 2. The capacity to ask for help or comfort from special others when it is needed. 3. The disposition to manifest anxiety or sadness to the attachment figure when those emotions are felt, rather than suppressing them (for fear of alienating the attachment figure) or exaggerating them (to capture their wandering attention). 4. The capacity to handle separation anxiety: the ability to keep separation anxiety to levels that are reasonable in view of, for example, the length of separation means one will not attempt unduly to restrict special others’ freedom (which could be described as controllingness, or clinginess). Here, incidentally, it’s instructive to compare Aristotle’s ‘marks’ of ‘friendship for its own sake’, which include ‘living with’ one’s friend (sc. for no further 3  Related lines of thought are to be found in (a) psychoanalytically minded philosophers who think the right kind of relation to special others underlies all virtue (Chazan 1998); and (b) philosophers who have sought a trait that underlies all virtue in self-love: see e.g. Homiak (1981).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 26/07/21, SPi

Two Routes from Secure Attachment to Virtue  141 purpose; cf. 1 above), and ‘sharing [one’s] friend’s joys and sorrows’ (cf. 3 above). (Aristotle 1992: 1166a1–1166a9). Some of these—like (1)—have to do with a person’s ability to take pleasure in intimacy for its own sake; that is, without it, at least some aspects of the good of intimacy seem not to be instantiated at all. (I don’t say all aspects of the good of intimacy: if attachment theory is correct, even avoidant attachment, where touching etc. are kept to a minimum, serves some of the needs of intimacy since all attachment serves to protect against threat; individual attachment styles come about as ways to preserve the attachment, and so to preserve at least the good of security (Edelstein and Shaver 2004).) The same could be said of (3). Others have to do with facilitating, or not frustrating, other goods of intimate relationships, such as striking a balance between involvement with another and independence, both the other’s and one’s own—the capacity to handle separation anxiety may be a good of close relations, but absence of clinginess, allowing the other independence of thought and action, seem to be further goods, so the relation of (4) to these goods looks like an explanatory relation rather than, as in (1), identity. In any case, emotional warmth, emotional openness, the capacity for cooperation, and the capacity for independent activity are goods, and these goods are more fully realized in secure attachment than in the various insecure attachment types. Secure attachment, that is, is closely associated with a range of capacities which realize goods in distinctively intimate relationships, and which also perhaps protect against bads distinctively associated with them such as possessiveness, in­toler­ance of difference of mind, or certain sorts of favouritism. Indeed in some cases this attachment type is exclusively so associated, a stronger claim than any I want to make for the connection between secure attachment and virtue. It should be clear, then, where route one from secure attachment to virtue is supposed to lie: secure attachment predicts love at its best, and the capacity for love at its best—so runs the powerful intuition—brings with it not just those virtues in which love at its best consists (what we might call the ‘virtues of intimacy’), but a whole suite of further virtues, which manifest themselves in behaviour (thoughts, emotions, etc.) towards non-­intimates as well. So the link between secure attachment and love cements the intuitive link between love and virtue. Is possessing the virtues of intimacy just the same as possessing the capacity for love at its best? It might be objected that, if there can be intimate friends whom we would not say love one another, the virtues of intimacy cover more than just love at its best. I’m inclined to say that given the role of touch and physical proximity in the virtues of intimacy as I have characterized them, we would be pushed to describe intimate friendships—at least as personal bound­ar­ies tend to be set in our culture—as instantiating more than a subset of these virtues. But I don’t think we need to worry, for present purposes, where love ends and ‘mere’ friendship begins. Certainly the virtues of intimacy include the cap­acity for love

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 26/07/21, SPi

142  Edward Harcourt at its best, even if they also include something else;4 and if they do include ­something else, sponsors of the intuition that love makes us good should have no difficulty in conceding that the same ought to be true for intimate friendship. It is the cut-­off between people with whom we have intimate relations and people with whom we have no special ties that is important, whether that cut-­off occurs on the near or on the far side of ‘mere’ friendship.

3.  Why the First Route isn’t Open Attractive though route one from secure attachment to virtue may seem, I shall now argue—for philosophical rather than empirical reasons—that it is ir­rep­ar­ ably blocked. I do not know what work has been done, if any, on the connections between the virtues of intimacy in secure attachment and the exemplification, within those same secure attachment relationships, of other virtues. For example, are the securely attached more likely to be kind towards their attachment figures, and to receive kindness from them? The same questions arise for generosity, fairness, honesty, and so on. I can imagine the following story being told: the same impulse that makes one spontaneously touch an intimate might also make one spon­tan­ eous­ly give them something (indeed perhaps in touching one is already giving them something); that would be the outline story for generosity. The story would be more complicated for fairness, but would still make good sense. For example, a securely attached person might be better able than an insecure-­ambivalent person to give their attachment figure only what is their due, or indeed to make clear-­ headed judgements about what is and is not due, because they would have no reason to fear that, if they don’t give more, the attachment figure would abandon them; and thus the secure person would be better able to be fair. Think of Cordelia—‘I love your majesty | According to my bond; nor more nor less’ (Shakespeare 1606/1952: I.i). But let the connection between the virtues of intimacy and the display of just­ ice, fairness, generosity, kindness, fidelity to promises (and so on) towards those same intimates be as close as you like. This closeness, surprisingly or not so surprisingly, is precisely what frustrates the connection, envisaged by route one, between secure attachment and these further virtues more generally. Possessing the virtues of kindness, generosity, fairness, fidelity to promises, and so on requires one to be able to respond to the relevant kinds of reasons—‘he has less 4  Remember that Aristotle also says that the attitudes of ‘friendship for its own sake’ are characteristic of mothers’ attitudes to their children, which would bring them closer into the orbit of what we call ‘love’ rather than ‘friendship’.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 26/07/21, SPi

Two Routes from Secure Attachment to Virtue  143 than me’ or ‘he would so enjoy this’ for generosity, ‘I promised’ for fidelity to promises, and so on. (More of this later.) But the motivations characteristic of the virtues of intimacy depend on certain others’ special relationship to me—somebody’s being my wife, my son, etc. feature crucially in the reasons for action characteristic of me, if I have those virtues. Given that their respective kinds of reasons for action are so different, why should we expect the virtues of intimacy and the virtues more generally to cluster at all? The outlook does not improve if one points to the link, supposing there is one, between the virtues of intimacy and these other virtues—or, as one should more carefully say, action in accordance with these other virtues—as manifested towards one’s intimates. For these other virtues require one to respond to the relevant kinds of reasons no matter who is in question: one isn’t fair (or generous, truthful etc.), or at least has only a very limit­ed version of the virtue of fairness, if one is only fair (or generous, truthful, etc.) towards one’s intimates. But if what’s helping these reasons to operate in the case of one’s intimates is the special quality of relationship one has towards them, then these causes will be the weaker as the agent’s connection to relevant others gets weaker. Precisely what explains their operation in the intimate case will render their operation puzzling in other cases. (Indeed the explanatory connection might even work in reverse: the insecure-­ambivalent might find it easier to be fair in relation to their non-­intimates.) On either argument, it looks as if the psych­ology of the virtues of intimacy cannot also be the psychology of virtue generally. To summarize, secure attachment does very well at explaining the virtues of intimacy. But the virtues of intimacy do not do well at explaining the further virtues—justice, honesty, generosity, and so on. That is because the virtues of in­tim­ acy depend motivationally on special relationships—and of course secure attachment is such a relationship. But to instantiate the further virtues, we need to be capable reliably of exhibiting them towards any others, to whom for the most part we are not specially related. This is of course very like a problem that is already familiar from the history of philosophy. The obstacle we have hit is very similar to the one identified by Sidgwick in complaining of the unsuitedness of sympathy to be the psychology of morality generally: [T]he utmost development of sympathy . . . would not cause a perfect coincidence between [moral] duty and self-­interest. . . . Suppose a regard for the general good—[moral] Duty—demands  .  .  .  a sacrifice, or extreme risk, of life. . . . [T]hat particular portion of the general happiness, for which one is called upon to sacrifice one's own, may easily be the happiness of persons not especially dear to one. . . . There are very few persons, however strongly and widely sympathetic, who are so constituted as to feel for the pleasures and pains

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 26/07/21, SPi

144  Edward Harcourt of  mankind generally a degree of sympathy at all commensurate with their concern for wife or children, or lover, or intimate friend. (Sidgwick 1907/1962: 501–3)

Freud noted a similar obstacle, when speculating about the prospects of appealing to Eros to explain our compliance with the demands of civilization: Eros, he says, ‘binds in love among a loving pair’, but civilization’s demands are impartial: ‘we derive the antithesis between civilization and sexuality from the circumstance that sexual love is a relation between two individuals . . . whereas civilization depends on relations between a considerable number of individuals’ (Freud 1953–64: XXI. 108). I am not arguing that the present obstacle is the same as either Sidgwick’s or Freud’s. For example the attachment theorist is not committed to claiming any special status for sympathy (itself presumably either an emotion or a relation between different people’s emotions) in explaining the connection between the virtues of intimacy and the display of those same virtues (if they can be called that) towards those to whom one is intimately related: the explanation lies rather in secure attachment, which is not itself an emotion or a relation between emotions. But the structure of all the obstacles is the same, namely this: how can something that relates one to certain special others explain one’s possession of a set of dispositions which, if we have it, relates us not only to those special others but to—potentially—everybody else as well?

4.  Would Love for Everybody Still be Love? Before giving up altogether on route one, I want to explore a reply to the objection I raised in the last section, based on some thoughts of Iris Murdoch’s. She speaks not in terms of secure attachment but in terms of love and I will stick to that presentation for the time being in order not to overcomplicate things, but it will be easy to relate the discussion back to secure attachment in the end. Murdoch’s well-­known writings on love stand out not least for claiming that love can relate us not, indeed, to a very small number of things but, at least in principle, to everything: an attitude that extended to everyone or everything could still be love. In a characteristic and familiar passage, Murdoch describes the ‘fat relentless ego’ as obscuring from us the reality of others, and thus as an obstacle to love or, as I choose to put it, love at its best (Murdoch 1970: 52). We so often seem to find in others what is in fact no more than the echo of our own thoughts, feelings, ambitions, and so on, and thereby fail to see—or to hear—them as the originators of their own such things, to which we may then be required to respond: the father, for example, who simply assumes that his daughter must want to follow him to his old Oxford college, at a cost to her both in terms of stifled ambition, and in

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 26/07/21, SPi

Two Routes from Secure Attachment to Virtue  145 terms of stifling any appearance of difference of mind, which she knows would be so costly to him. In a similar vein, she writes: All just vision, even in the strictest problems of the intellect . . . is a moral matter. The same virtues, in the end the same virtue (love), are required throughout, and fantasy (self) can prevent us from seeing a blade of grass just as it can prevent us from seeing another person.  (Murdoch 1970: 70) [The ability to direct attention] away from self . . . towards the great surprising variety of the world . . . is love.  (Murdoch 1970: 66)

The thought seems to be as follows. Love is required for us to see justly. But we owe it to everything to see it justly—a blade of grass as much as a person. So even if it is a psychological truth about people that, because they cannot love every­ thing, they cannot see everything justly, were they to extend that attitude to every­thing it would not thereby cease to be love—on the contrary, we become more perfect in love the greater the number of things to which we are capable of extending the attitude. Now some of this seems right: whatever we may or may not go on to do with our true thoughts—such as uttering them—we do indeed seem required to think of everything as it is, not just of some things. What is more, the fact that  I  can only love a limited number of people at once isn’t per se an ­objection to Murdoch’s view. I am very limited in the number of people towards whom I can be compassionate (or the number of occasions on which I can be compassionate to a given person). But I am not complacent enough to draw the limits of those to whom I owe compassion just where I begin to experience compassion fatigue. Here it is important to stress over again the point that makes the misquotation from Augustine (‘ama’ etc.) so thought-­provoking: Murdoch is not to be taken as using the word ‘love’—as it may perfectly properly be used—as a synonym for ‘benevolence’, ‘compassion’, or the like. If she were, then ‘deriving’ the virtues generally from love would be a different and less interesting project than the one inspired by our pseudo-­Augustine, because so much of virtue would already be built in to the starting point. But she surely doesn’t mean benevolence. The flaws—‘fantasy’ and so on—that, in her view, get in the way of love are flaws to which we are especially prone in intimate relationships, that is, relationships with non-­ substitutable particular others. Not only that: it’s essential to the love Murdoch is talking about that when these flaws are got rid of, what’s thereby perfected is still an attitude towards a non-­substitutable particular. This is of course something Murdochian love has importantly in common with attachment, and which sets it apart from compassion or benevolence. So could an attitude in principle extendable to every person—or even every thing (remember the blades of grass)—still be recognizably (personal) love?

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 26/07/21, SPi

146  Edward Harcourt How exactly is one to go about settling the question? One way to do so is to ask whether it is ever true of love that love for one person is improved by its being at the expense of another or, conversely, that love for one person is vitiated by the lover’s also loving others. With sexual love, certainly, most people have no difficulty in saying that it is vitiated by the lover’s bearing the same relation to others. But I hesitate to rest too much on this thought, as it is culturally local and perhaps—as recent discussions of polyamory have suggested—optional even around here (Brunning 2020). The case of love for children is interesting, and more complicated. A child ought not to think that a parent’s love for him or her should be at the expense of his or her siblings, and a parent’s love which did have this character would be worse than love for all his or her children. So far so good for the Murdoch-­inspired reply. But still, can’t a child think its parents love for him or her is made worse by the fact that the parent also loves perhaps a great many other children who are not that person’s own? It is hard, however, to disentangle love per se and various other things—such as consanguinity, or hands-­on care— which are usually matters of mutual significance to people who love one another in the way parents and children do. If you love your foster children, and your children’s classmates, and the neighbours’ children, etc. etc., each one of them will receive less hands-­on care from you than they would if they were fewer in number, but whether this means you love each one less I do not know. Again, someone who spread their love so widely could presumably not attach much significance to consanguinity. Does that mean you don’t really love your own children, or just that something they thought bound you together because it had meaning for both of you (‘you look so like grandma!’) doesn’t in fact have meaning? But how much of that can you subtract before there isn’t anything that binds you together, and if there isn’t anything, do you really love each other? This binding together by things of mutual significance does not have a place in a conception of love as—solely— just vision. This inclines me to the view that love for one person at least can be vitiated by the lover’s also loving others, which would make love to some extent an essentially exclusive attitude. So an attitude in principle extendable to everyone would not be love. But that is hardly a decisive ruling on the question. In search of a more decisive ruling, let us try a different approach. I cannot pretend to offer a description of God’s actual attitude to humankind. However, the point of this thought experiment is to think of a being capable of personal love, but free of our psychological limitations. Such a being might, surely, love every human being in the way many parents love all their (comparatively few) children.5 5  If we grant this much, might we not just as well grant that an infinite being could have indefinitely many attachments? I am inclined to say no: since it is essential to an attachment that it has a certain evolved function, an infinite being—who does not depend on anything for survival—could not have any attachments. Given the arguable link between attachment and love, this may ultimately render the idea of indefinitely many love objects incoherent too. If it is, it would be apparent at an early stage that

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 26/07/21, SPi

Two Routes from Secure Attachment to Virtue  147 If one is trying to open up a route to virtue from actual humans’ dispositions to love, of course the fact that one has to imagine a being with powers infinitely greater than our own is not much help: the route is still closed to us. But we were after a priori reasons for finding it open or closed, so that ought not to matter: if we imagine away our psychological limitations, there’s no reason in principle why an attitude that extends to everyone could not still be love. But does this help the Murdoch-­inspired reply? It might appear as if it does: the problem we encountered in the last section was the problem of how an attitude which relates one to certain special others could explain one’s possession of a set of dispositions which, if we have it, relates us not only to those special others but to—potentially—everybody else as well. Now it looks as if it’s in principle possible for there to be a being such that everybody is that being’s special other. So why shouldn’t there be a route clear from love, whose objects are essentially non-­ substitutable, to the virtues (which need to be capable of being displayed towards anybody)? Before pursuing a negative response in detail, an aside is needed about love and attachment. Remember that the possibility of an attitude whose objects include everybody involved imagining a being free of our empirical limitations. It’s doubtful whether the same thought experiment would even make sense if we switched ‘attachment’ for ‘love’, because attachment is an essentially empirical concept (a concept whose home is evolutionary biology, ethology, developmental psychology). If it isn’t an evolved disposition which engages caregiving dispositions in a small number of conspecifics to see one through to reproductive age, it isn’t attachment, and I take it that God’s dispositions have not evolved at all. So, since the strategic goal of the argument is to evaluate a route from attachment to virtue, one might say that this disanalogy between love and attachment disqualifies the reply from the start. I want to press home a different problem with the reply, however, not least in order to forestall the objector who—seizing on the differences between Freud or Sidgwick and attachment theory rather than the similarities—asks why an obstacle in the way of route one from attachment to virtue should bother us when, for all that’s so far been said, a structurally similar route from love to virtue still lies open. On the contrary, as I shall now argue, the deep source of the failure of the Murdoch-­inspired reply goes to a feature attachment and love have in common. It’s often said that a hallmark of morality is impartiality (Wolf  1992; Scheffler 1982). What’s meant by morality in these discussions, I take it, includes some of the virtues even if there are also virtues—not least the ‘virtues of

switching from talk of attachment to talk of love would not be able to keep route 1 open. Making play with the possibility that love could have indefinitely many objects is a way of making sure that route 1 is definitely shut.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 26/07/21, SPi

148  Edward Harcourt in­tim­acy’ which I have already discussed—which fall outside it. Would love for every­body be impartial? Here, I want to distinguish between impartiality and identity-­indifference. Certainly, where a being loves everybody no one is specially favoured, so one might say that that being’s love is impartial. But one might equally well say that this love is universal but partial, in the sense that what drives the attitude is, in each case, a special relationship: it’s just that God, unlike us, has a special relationship with everybody. Perhaps the word ‘(im)partial’ is not after all much help here. What God’s presumptive love of everybody really couldn’t be, however, is identity-­indifferent, and this is a conceptual point, not a point about psy­cho­ logic­al limitations. It is essential to any attitude’s counting as personal love that its objects are not interchangeable—a feature as we’ve noted that love shares with attachment. But arguably, what’s needed for at least some virtues (and perhaps for all of them beyond the ‘virtues of intimacy’) is the capacity for general thought, that is, not simply thought which extends an attitude from a small range of individuals to a very large range, but thought that is not about particular individuals at all. Thus being fair, for example, involves undertaking to (say) distribute equal shares to anybody who . . . and one may be no less required to be fair in that sense if there is never in fact any individual who satisfies the condition in question.6 It thus appears that the psychology underlying at least some virtues—those which require general thought—is different from the psychology of love. The reason is that love (whether for everybody, or just for a few) is non-­identity-­indifferent, whereas the thoughts required to possess at least some virtues (justice, fairness, truthfulness . . .) are identity-­indifferent thoughts. Route one from love to virtue— and a fortiori route one from attachment to virtue—cannot therefore be rehabilitated by showing that an attitude in principle extendable to everybody can still qualify as love.

5.  The Second Route For all the difficulties we have encountered with route one, however, I do not think this is the end of tracing a route from secure attachment to virtue. But if we are going to show such a route is possible, we need to make sure the theoretical role played by relations to special others is not the very direct role it plays in route one. 6  Compare Gaita (2000), in which he connects love as a relation to an individual and love as the source of morality—we (he says) owe certain sorts of behaviour to ‘anyone who is a possible object of love’. But thinking about who is a possible object of love involves the universalization move thought to be characteristic of morality, and so doesn’t rest moral thinking on special relationships. Love plays a part in Gaita’s account, but it’s the part played by (say) ‘respect’ in some Kantian accounts, not the part love plays on route one.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 26/07/21, SPi

Two Routes from Secure Attachment to Virtue  149 Without quarrelling with the idea that the securely attached are specially favoured in respect of their ability to develop the virtues of intimacy, route two turns its back on that feature of secure attachment and exploits a different one to make the connection to virtue. One of the most impressive marks of the securely attached—whether children or adults—and which I haven’t drawn attention to so far is that they are confident explorers of the world (Holmes 2011: 32). Except in situations of danger, the securely attached have the freedom to look outwards to the world because their base is secure (Bowlby 1988)—that is, they don’t need to attend to their attachment figures for fear that the latter will ignore them when they’re most needed, attack them, etc., and as a result they have plenty of mental space left to devote to things outside the attachment relationship (Grossmann et  al.  2008: 870–1). By contrast at least one subvariety of insecure child—the insecure-­ambivalent—will focus on the caregiver rather than the environment even when there is nothing in the environment (no threat, for instance) to prompt this, thus inhibiting exploration (Weinfeld et al. 2008: 80; Simpson and Belsky 2016: 138–9). Let’s hold on to the thought of the securely attached as confident explorers, not of other people only but of the world generally, and step sideways to think about virtue. To return to a remark I made en passant in section  3, there is a widely accepted high-­level account of virtue—associated in modern times with John McDowell (1979) and Bernard Williams (2006), among others—according to which virtue consists in responsiveness to reasons, with virtue carved up into the different virtues according to the types of reason to which the virtuous person is responsive. Thus what marks someone out as the possessor of the virtue of fidelity to promises is their distinctive responsiveness to the reason ‘I promised’—that they promised is all on its own a reason for them to do something (namely, the thing they promised). Similarly, just people are distinctively responsive to reasons such as ‘that’s not fair’, ‘he has no right to that’; kind people distinctively sensitive to reasons such as ‘he’s in trouble’ or ‘she would love that!’ and so on for action, judgement, feeling, and the rest—though this isn’t the whole story as to how to taxonomize virtues, and most virtues don’t go with a proprietary set of reasons, let alone a single characteristic reason. This account of virtue is sometimes ­supplemented by the thesis that reasons are facts. This would make what I’m about to say neater, but for present purposes I don’t think it matters whether one sees reasons as states of the world (facts), or as on the subject side of a subject/ world line—what I’m about to say works just as well if ‘I promised’ expresses a belief (which is the reason) as it does if it reports a fact (and that is the reason). Now let’s put our new thought about secure attachment and our thought about virtue together. If virtue is responsiveness to reasons and reasons are either themselves facts—that is, states of the world—or beliefs about facts, and it is the securely attached who are distinctively able to focus on the condition of the world around them, then the securely attached will be the best responders to reasons,

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 26/07/21, SPi

150  Edward Harcourt because they will be best placed to see what the reasons are (either, to see the facts or to form true beliefs about the facts) in any given situation. So they are especially well placed to acquire the virtues, or having acquired them to exercise them. That is all there is to route two from secure attachment to virtue. Route two is not, of course, supposed to assert a guarantee that the securely attached will be virtuous, or will develop the virtues. I take it that in developmental psychology there aren’t any guarantees and one can see well enough (a) why one might be duly sensitive to reasons and yet not act as virtue requires and (b) why the securely attached might not even get that far. The thesis is only meant to defend the possibility of a route from secure attachment to virtue as solid as its failed rival, route one, was supposed to be. (No advocate of route one, I take it, ever said that equipped with the motivations of the virtues of intimacy you have to acquire the full set of virtues.) Nor is the thesis supposed to be that without secure attachment, it’s impossible to acquire any virtues. The insecure-­ambivalent are not, after all, wholly inattentive to their environments. Nor should we overlook the relationship of the insecure-­avoidant to the virtues. This attachment group, remember, are weak on the virtues of intimacy. Are they thereby compromised in their ability to acquire the virtues? It might be argued that their absence of emotional warmth impairs their ability to acquire those virtues to which the experience of emotion is (arguably) essential. Some would argue that this includes kindness, perhaps, or compassion, but predict that they would do as well as their secure counterparts with truthfulness or fidelity to promises; others, who argue that all virtues require emotion (Stocker 1996), would expect them to do less well than their secure counterparts across the board. One might point to the avoidant’s reluctance to admit vulnerability or to rely on others, which might compromise their ability to acquire virtues involving cooperation or teamwork. All these thoughts would leave the securely attached distinctively favoured, to a greater or lesser extent, in their ability to acquire virtues beyond the virtues of intimacy. One might on the other hand grant that emotion is essential to all virtues, but build on the idea that for the insecure-­avoidant, it’s emotional warmth vis-­à-­vis their special others that’s distinctively threatening,7 so when no special other is involved—as in most cases where virtue is called for—there’s no disposition to inhibit the experience of emotion, however much or little of it is needed for the various virtues. If this last line of thought were made to stick, it would not show there was anything blocking route two from secure attachment to virtue. But it would suggest that it’s not the case that secure attachment distinctively favours virtue, but rather that insecure ambivalent-­attachment distinctively disfavours it. 7  Or, more narrowly, special others in situations of stress—cf. observations of avoidant couples in airports who are stiff and distant when saying goodbye to one another, but hold hands etc. when saying goodbye to a child or friend (Fraley and Shaver 1998).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 26/07/21, SPi

Two Routes from Secure Attachment to Virtue  151 But that theoretical result would in its turn need to be reconciled with the em­pir­ic­al work which seems to show secure attachment is specially favoured in some way. But I say ‘seems’ advisedly: empirical work which tests for connections between virtue and secure attachment while respecting the differences between the virtues— acknowledging that truthfulness, fidelity to promises, fairness, and generosity can’t all be captured simply by testing for prosocial behaviour—is, by and large, empirical work we just don’t have yet. Setting the considerations about insecure-­avoidant attachment on one side, let me close by noting one further theoretical advantage to route two. This is that it gives us a credible alternative to ‘internalized other’ theories of self-­regulation— theories, that is, that perhaps like Freud’s theory of the superego, explain self-­ regulation by appeal to our having internalized an external figure (cf. Harcourt 2016). There is one insurmountable problem with such theories which can be stated in the form of a dilemma: is the internalized figure me, or isn’t it? If it isn’t, then the theory is not a description of self-regulation, but of regulation by another (albeit another that is internal to my mind—but in this context, that is a detail: no one would describe the schizophrenic who hears a voice telling him to do something rash as an effective self-­regulator, and yet the voice is internal). If on the other hand the internal figure is me, that objection lapses, but we have another: the theory is now not an account of self-­regulation, but a redescription of the phenomenon to be accounted for. Even setting the dilemma aside, there’s in any case something else that’s objectionable about internalized other theories, and grasping this point opens a connection between Murdoch’s writing on love and route two—notwithstanding the fact that Murdoch turned out to be no help in rehabilitating route one. Suppose that my grandmother was a loved and revered figure in my early life, and that she told me it was important always to write thank-­you letters. Now, this is an excellent moral lesson. But however loving and authoritative and free from coercion she was as a guide in life, there is surely nothing odd about my grandmother featuring, in maturity, absolutely nowhere in my reasons for writing thank-­you letters—supposing that’s what I do. On the contrary, her teaching has been 100 per cent effective if I write thank-­you letters because I so appreciated the present, because you were so thoughtful, or for whatever reasons there ever are for expressing gratitude. Indeed one can go further, and say something has gone wrong with my grandmother’s teaching if she does feature, because the teaching has somehow failed to enable the learner’s gaze to pass beyond the teacher to the world and the reasons for action it contains: the teacher remains herself a source of reasons, whereas (I take it) the best teaching—like Murdochian love—focuses the learner’s gaze not on the teacher but on what the teaching is about (cf. Velleman 2013). Route two not only gives us a way to avoid positing internalized others in the  ­theory of self-­regulation, but also explains why our moral teachers—and our  practical teachers generally—can be so recessive in the mature moral

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 26/07/21, SPi

152  Edward Harcourt consciousness; indeed the more recessive as they are more successful, because if they are secure attachment figures for us, they give us precisely the ability to look not at them, but at the world beyond them.

References Aristotle 1992. Nicomachean Ethics, trans. W. D. Ross, rev. J. O. Urmson. In Aristotle: The Complete Works. Electronic edition. BOLLINGEN SERIES LXXI, Vol. 2. ISBN: 978-1-57085-003-5. Charlottesville, VA: InteLex Corporation. Augustine 2000. ‘In epistulam Iohannis ad Parthos tractatus decem’. Opera: Part 8. Opera Omnia CAG. Electronic Edition. Austen, J. 1816/n.d.. Emma. London: Thomas Nelson. Belsky, J. 2008. ‘War, trauma and children’s development: observations from a modern evolutionary perspective’. International Journal of Behavioral Development 32, 260–71. Bowlby, J. 1976–7/2005. ‘The making and breaking of affectional bonds’. In J. Bowlby, The Making and Breaking of Affectional Bonds, 150–88. Abingdon: Routledge. Bowlby, J. 1988. ‘The role of attachment in personality development’. In J. Bowlby, A Secure Base, 134–54. London: Routledge. Brunning, L. 2020. ‘Compersion: an alternative to jealousy?’ J.  Amer. Phil. Ass. 6(2), 225–45. Chazan, P. 1998. The Moral Self. London and New York: Routledge. Edelstein R., and Shaver, P. 2004. ‘Avoidant attachment: exploration of an oxymoron’. In D.  Mashek and A.  Aron (eds), Handbook of Closeness and Intimacy, 397–414. Mahwah, NJ: Erlbaum. Fraley, C., and Shaver, P. 1998. ‘Airport separations: a naturalistic study of adult attachment dynamics in separating couples’. Journal of Personality and Social Psychology 75, 1198–212. DOI: 10.1037/0022-3514.75.5.1198. Freud, S. 1953–64. The Standard Edition of the Complete Psychological Works of Sigmund Freud. 24 vols. Vol. 21. Ed. J.  Strachey et al. London: Hogarth Press/ Institute of Psychoanalysis. Gaita, R. 2000. A Common Humanity. London: Routledge. Grossmann, K., et al. 2008. ‘A wider view of attachment and exploration: the influence of mothers and fathers on the development of psychological security from infancy to young adulthood’. In J.  Cassidy and P.  Shaver (eds), Handbook of Attachment: Theory, Research and Clinical Applications (2nd edn), 857–79. New York: Guilford Press. Harcourt, E. 2016. ‘Internalization, joint attention and the moral education of the child’. In L. Gormally, D. A. Jones, and R. Teichmann (eds), The Moral Philosophy of Elizabeth Anscombe (St Andrews Studies in Philosophy and Public Affairs), 243–62. Exeter: Imprint Academic.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 26/07/21, SPi

Two Routes from Secure Attachment to Virtue  153 Harcourt, E. 2017. ‘Attachment, autonomy and the evaluative variety of love’. In E. Kroeker and K. Schaubroeck (eds), Love, Reason and Morality, 39–56. London: Routledge. Holmes, J. 2011. Exploring in Security. London: Routledge. Homiak, M. 1981. ‘Virtue and self-love in Aristotle’s Ethics’. Canadian Journal of Philosophy 11(4), 633–51. McDowell, J. 1979. ‘Virtue and reason’. The Monist 62(3), 331–50. Murdoch, I. 1970. The Sovereignty of Good. London: Routledge and Kegan Paul. Scheffler, S. 1982. The Rejection of Consequentialism. Oxford: Oxford University Press. Shakespeare, W. 1606/1952. King Lear. Ed. K. Muir. London: Methuen. Sidgwick, H. 1907/1962. The Methods of Ethics (7th edn). London: Macmillan. Simpson, J. A., and Belsky, J. 2016. ‘Attachment theory within a modern evolutionary framework’. In J. Cassidy and P. R. Shaver (eds), Handbook of Attachment: Theory, Research and Clinical Applications (3rd edn), 91–116. New York/London: The Guilford Press. Velleman, J. D. 2013. ‘Sociality and solitude’. Philosophical Explorations 16(3), 324–35. DOI:10.1080/13869795.2013.767931. Weinfeld, N. S., et al. 2008. ‘Individual differences in infant–caregiver attachment’. In J. Cassidy and P. Shaver (eds), Handbook of Attachment, 78–101. New York/London: The Guilford Press. Williams, B. 2006. ‘Acting as the virtuous person acts’. In B. Williams, A Sense of the Past, 189–97. Princeton: Princeton University Press. Wolf, S. 1992. ‘Morality and partiality’. Philosophical Perspectives 6, 243–59. Zeifman, D., and Hazan, C. 2008. ‘Pair bonds as attachments’. In J.  Cassidy and P. Shaver (eds), Handbook of Attachment: Theory, Research and Clinical Applications, 436–55. New York/London: Guilford Press.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

8

Education for Virtue Aquinas on Passions and Attention Terence Irwin

1.  Introduction This chapter explores one way in which attachment is important, by exploring some of the possible effects of attachment. If we take one effect of attachment to be the establishment of a ‘secure base’ (Bowlby 1988), we may ask why it matters whether we have a secure base. The metaphor of a secure base suggests that we need some security and assurance if we are to explore. If an army establishes a base, but the base is insecure, the army may have to devote all its effort to prevent the base being overrun, and it may be too hazardous to explore outside the base. Similarly, if I am always afraid—reasonably or unreasonably—that my house is about to burn down, I will be unwilling to leave the house and to go exploring. This is how insecure attachment inhibits any exploratory attitude. If we are uncertain about the reliability of a parent, or carer, or other basis of attachment, our constant concern about their attitude to us may occupy our attention to a degree that prevents curiosity and exploration in areas that are not evidently relevant to reassurance about the basis of attachment. If, however, we do not have to worry constantly about the reliability of parents and others, we are more capable of noticing other things. A secure base is useful, then, for exploring other aspects of the world, and therefore we need it for cognitive development. It is not so obvious that we need it for moral development, and in particular for moral education. If, however, it turns out that moral education also requires the sort of exploration for which a secure base is necessary, we have found something else that is enabled by secure attachment, and therefore another respect in which secure attachment is important. It may not seem obvious, however, that moral education requires the exploratory attitude that supports cognitive development. We may think that it is a matter of training, and, more specifically, of being trained to do what other people tell us to do, so that it becomes habitual for us to do what we are trained to do. If we think of soldiers being drilled, or of learning to recite from memory, exploration Terence Irwin, Education for Virtue: Aquinas on Passions and Attention In: Attachment and Character: Attachment Theory, Ethics, and the Developmental Psychology of Vice and Virtue. Edited by: Edward Harcourt, Oxford University Press. © Terence Irwin 2021. DOI: 10.1093/oso/9780192898128.003.0009

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

Education for Virtue  155 does not seem especially relevant. An exploratory attitude may even seem out of place, if we find ourselves more interested in alternatives to the patterns of behaviour that we are being taught to acquire. These simple thoughts about moral training are relevant to an Aristotelian approach to ethics.1 The moral virtues, as Aristotle understands them, have to be acquired through a process that begins early in life, before we can be expected to understand the instruction, formal or informal, that allows us to understand what morality expects of us, and why it expects it. Aristotle expresses this view of the moral virtues by saying that they belong in the non-­rational part of the soul—to desires, impulses, and emotions that are initially unresponsive to rational deliberation.2 Virtue originates in the non-­rational part of us not only because we want ­people to do the right thing while they are young, before they understand the reasons for doing it, but also because we still have a non-­rational element that is part of us even when we are capable of acting on reason. We never outgrow our need for a non-­rational element in virtue. If our non-­rational aspects do not favour virtue, they will oppose it, and will prevent us from acquiring and maintaining it.

2.  Habituation and the Passions Aristotle’s term for the process by which we acquire virtues of character is ‘habituation’ (Aristotle 1992: ii, 1–2). He speaks of habituation in contrast to learning by reading a book, or listening to a course of lectures, or observing what an expert does. In all these cases we learn without (in the course of learning) necessarily doing the appropriate actions. Habituation involves our getting used to the relevant actions by practice in doing them. This general description of habituation covers different sorts of habit-­forming processes. Some of these processes happen to us without anyone’s intention. If we happen to be immersed in cold water often when we are young (because there are no hot baths available), we will get used to cold water; if we grow up with hot baths, we will be used to hot water and not to cold water. Other things happen to us through other people’s intention. If young children are made to play in water so that they will not be afraid of it and will find it easier to learn to swim, they are habituated by being trained. If we learn to ride a bicycle, we are habituated by our own actions (initially under supervision); we might say that this is active habituation. 1  In many respects, this is no less a Platonic than an Aristotelian approach. 2  Some of the connections between attachment theory and Aristotelian ethics are explained in Harcourt (2013).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

156  Terence Irwin All of these types of habituation are relevant to acquiring a moral virtue. Our purely passive habituation makes a difference to the types of active habituation that are open to us. If we are not exposed to cold water for many years, it may be very difficult or impossible for us to get used to swimming in cold water. The effects of purely passive habituation make different kinds of training or active habituation more and less difficult for different people. Aristotle recognizes purely passive habituation without training. He mentions it in non-­human animals in cases where no training is involved. When he discusses habituation in human beings, it is not always clear whether he has purely passive habituation or deliberate training in mind. He shows that he recognizes some role for purely passive habituation. This is why he regards the family, and especially (being Aristotle) the father, as the right source of early moral education: For in cities the provisions of law and the types of character have the appropriate strength, and, similarly, in households the discourses of fathers and the habits have the same sort of strength, and even more because of kinship and the benefits —for the children already love them and are naturally ready to obey.  (Aristotle 1992: 1180b3–7)

The mutual love of parents and children is the appropriate starting point for moral education and for receptivity to it. At this point Aristotle alludes to questions that have been explored by developmental psychologists, including attachment theorists. The space that is indicated by the allusion to mutual love might be filled by an account of the importance of secure attachment for the possibility of further cognitive and emotional development. Such an account will specify some of the preconditions of virtue, and the ways in which different types and degrees of attachment make it easier or more difficult to acquire the virtues. Some understanding of the virtues can help us to answer a further question: what kinds of cognitive or affective development should be made possible if early attachments are to result in the virtues? If we thought that being virtuous consisted entirely or largely in willingness to do what other people ask us to do, the appropriate sort of attachment would develop the relevant sort of compliance. If we thought it also required willingness to refuse unreasonable requests, a different sort of trait would need to be developed. I want to explore the possibility that, contrary to initial appearances, secure attachment is important for moral education because it allows curiosity about the world and further exploration of it. This sort of cognitive development is necessary for moral education of our non-­rational desires and tendencies. We might suppose that moral training consists in positive and negative reinforcement that forms the habit of doing whatever is reinforced by the sanctions of pleasure and pain. But if this is all we can do to train the non-­rational part, it may seem rather

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

Education for Virtue  157 disappointing. The sorts of habits we can acquire in this way may be coarse-­ grained and undiscriminating. I may form the habit of enjoying sweet cakes, because I have eaten them at birthday parties. This is a habit that may need to be restrained, though not necessarily eliminated. I may have formed the habit of not talking to adults I do not know; this may be a good habit to acquire in early life, for the sake of safety, but it needs some modification. These coarse-­grained habits do not seem to give us an accurate picture of most of the non-­rational states (emotions, impulses, feelings of pleasure and pain) that belong to adults, or to children who are beyond the very first stages of their development. We would like to understand, then, what is different about emotionally mature people. Aquinas offers a distinctive answer to this question. He starts from Aristotelian assumptions about non-­rational states of the soul, which he calls ‘passions’, but he goes beyond Aristotle in his analysis of the passions and his discussion of how they can be trained, so that he fills some gaps that Aristotle leaves in his description of moral education.3 In particular, since he affirms that the training of the passions requires flexible attention, he shows how the guidance of the non-­ rational by the rational part of the soul rests on secure attachment and on the attendant capacity for exploration.

3.  How can the Passions be Educated? The aim of educating the passions may appear to face a dilemma. On the one hand, habituation by repetition and reinforcement seems inadequate to produce mature and appropriately responsive passions. On the other hand, the sort of education that relies on reasoning and explanation seems inappropriate for the non-­rational part of the soul. What third option is available? Aquinas offers a third option. He seeks to explain how the rational will governs the non-­rational passions by a comparison with the way in which one rational agent rules another. He often refers to Aristotle’s comparison of rule over the passions with political rule.4 Political rule—the rule of one group of citizens over others—does not rely simply on coercion or on giving orders for thoughtless acceptance. It offers reasons and considerations that present the rulers’

3  I have not tried to separate the contributions of Aquinas’ various sources, including Aristotle, from his contributions. 4  ‘The irascible and appetitive powers do not obey the reason unquestioningly, but have their own proper movements, by which, at times, they fight back against reason, so that the Philosopher says . . . that “reason rules the irascible and appetitive powers with political primacy”, namely that with which free people are ruled, who have their own will in some things. And for this reason also there must be some virtues in the irascible and appetitive powers, by which these powers are well disposed for action’ (Aquinas, Summa Theologiae 1-­2 q56 a4 ad3). Quotations from Aquinas are all taken from the Summa Theologiae, cited by part, question, and article in the usual way.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

158  Terence Irwin instructions as deserving assent. He offers this analogy to show how the passions can be guided by a rational will. The analogy may seem altogether useless, for two reasons: 1. Aquinas uses, as Aristotle does, inter-­personal terms to describe the relations between different motives and impulses within a single person. He recognizes parts of the soul, and treats these parts as quasi-­agents or reduced agents. This may seem to introduce a useless reduplication of agents, which appears to explain agency while simply reduplicating it. Perhaps the treatment of motives and impulses as reduced agents constitutes an objectionable reification of them, or an objectionable resort to ‘faculty psychology’. 2. In this particular case, the inter-­personal terms that Aquinas chooses may seem especially unpersuasive, since they do not seem to offer a distinct option that applies to non-­rational passions. Citizens in a state have wills, but passions have none. It is pointless to represent them as though they were rational agents, since they are neither agents nor rational. In reply to these objections, I will argue that Aquinas’ comparison of the relation between the will and the passions with political rule is not useless, but illuminating. It explains how the passions can be educated, without attributing rationality or agency to them. A closer look at the analogy with political rule suggests that persuasion may work in different ways. When some citizens are rulers and others are ruled, the reasons and considerations that are offered to the ruled need not be those that convince the rulers that a particular course of action is the right one. It might be a mistake to present these original reasons, because they may be too complex to be grasped by everyone whose acceptance is needed. Honest and effective rulers are those who know how to communicate reasons for their decision that are ac­cess­ible to the right people and represent as fully as possible the original reasons for the decision. This comparison does not apply exactly to the rational will and the non-­ rational passions; for, since the passions are non-­rational, they have no will of their own. But Aquinas believes that the comparison is nonetheless useful, because the passions have something close enough to a will of their own. To see whether he is right to suppose that this is a useful comparison with the rule of will over passions, we need to look more closely at the similarities and differences between will and passions.

4.  To what Extent are the Passions Rational? Aquinas describes the passions by contrast with the will. Will is rational desire, and, being rational, aims at the apparent good. We do not simply notice that

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

Education for Virtue  159 certain things are sources of pleasure, for instance, and desire them in the ex­pect­ ation that they will give us more pleasure. We also suppose that some feature of them gives us a reason to pursue them. If we cease to believe that they have this feature, or that this feature gives us a reason to pursue them, we lose our reason-­ based desire for them, and no longer desire them as good. When Aquinas says that the will aims at the good, this is part of what he means. It is not all of what he means, however. We do not act on our will simply because we act on some reason or other. When we act on the basis of will, we act on a comparison of the goodness of different possible objects, and on the conclusion that one of these possible objects is best all things considered, or best overall.5 Aquinas expresses this point by saying that the will regards different goods ‘under the common character of good’, and that its object is what is good without qualification. Our comparison may be careless, ill informed, or misguided in other ways, but as long as we act on some conclusion about what is best among the options we are aware of, we act on our will. Whereas the will relies on a comparison of the goodness of different objects, passions are indifferent to this sort of comparison, because they rely on sensory judgements that are immediate (‘sudden’) rather than reflective (1-­2 q17 a7; q74 a10; q88 a2; Knuuttila  2004: 178–95; Sorabji  2000: chs 22–4). If I am angry because I have been insulted, the knowledge that it would be better to restrain the anger does not change the fact that I have been insulted, and if the anger responds to the insult, knowledge of some greater good that I will lose by being angry does not change the fact that my anger responds to. In this respect the passions are non-­rational. We cannot set them aside by noticing that a particular action will secure a small good, but will cause a greater evil in the future.6 It is one-­sided and misleading, however, to describe the passions as simply non-­rational. Aquinas does not believe that they are wholly indifferent to reason or to considerations of goodness. They are concerned about goodness to a limited extent, and the limited concern distinguishes them from the will. The object of a passion is ‘. . . a sensible good . . . taken without qualification’ (1-­2 q23 a1, a2). Passions respond only to particular goods, and only insofar as they are ‘pleasant from the point of view of sense and suitable to nature’ (1a q82 a5). They take the point of view of ‘particular reason’ (1a q81 a3).7 It is this limited reason, rather than the absence of reason altogether, that distinguishes the passions from the will. Because of their limited reliance on reason, a passion pursues a particular good, in contrast to the will, which pursues the universal good (1a q82 a5; 1-­2 q1 a2 ad3). Insofar as we pursue particular goods, we are guided by particular reason, and insofar as we pursue the universal good, we are guided by universal reason.

5  The comparative aspect of will: 1a q83 a1. 6  Will v. passions: 1a q82 a5; 1-­2 q45 a4. 7  In q83 a3 ad2 Aquinas mentions the analogy with political rule. In other animals sensory desire is moved by the estimative capacity (aestimativa virtus) (1a q78 a4), which has the role that particular reason has in human beings.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

160  Terence Irwin To see what Aquinas means, and to decide whether he is right, we need to look more closely at the connection between the pursuit of different sorts of goods and guidance by different types of reason. If we understand this connection, we will have a clearer idea of what particular reason is, and of how it differs from universal reason.

5.  Particular Goods v. Universal Good Aiming at a particular good is intelligible by contrast with merely liking or disliking something. Liking the taste of one kind of tea does not necessarily give me a reason to like the taste of another kind of tea that tastes quite similar to the first kind. Perhaps most people agree that English Breakfast tea tastes very like Irish Breakfast, and most people who like the one like the other. It does not follow that I am making a mistake or behaving irrationally if I like the one and dislike the other. I do not like the tea I like on the assumption that it has some desirable characteristic that I look for in tea. If I like English Breakfast, but I also like Earl Grey, which tastes very different from it, and which is disliked by most of those who like English Breakfast, I am not making any mistake and I am doing nothing irrational. In such a case, we might say, reason has nothing to take hold of. Let us take this as an example of merely liking or disliking. Not all passions are cases of merely liking or disliking, however. To see this, it is useful to turn to some suggestions by social psychologists of the early twentieth century. A. F. Shand’s The Foundations of Character begins with a discussion of ‘the system of the emotions’ and ‘the system of the sentiments’ (Shand 1920: chs 3–4). He argues that love and hatred are sentiments rather than emotions because they do not consist in a single type of response, but they include a disposition to respond with different emotions to different conditions. Shand illustrates this feature of love with a quotation from Swift: Love why do we one passion call when ’‘tis a compound of them all? Where hot and cold, where sharp and sweet, in all their equipages meet; where pleasures mixed with pain appear, sorrow with joy, and hope with fear. (Swift 1983: 150)

The answer to Swift’s question is that love is not a simple liking, but an attitude that makes these different emotions—pleasure, pain, sorrow, joy, hope, fear— appropriate on different occasions. Someone who merely liked another person but

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

Education for Virtue  161 was not disposed to have these different feelings in the appropriate circumstances would not love the other person. These are contrary emotions, but they do not introduce any inconsistency into love. On the contrary, love would not be a single consistent sentiment if it did not move us to joy on some occasions, sorrow on others, and so on (cf. Stout 1929: 625–8; McDougall 1919: 105–7). If we ask why love disposes us to these different responses on different occasions, our answer reveals the ways in which love is directed to a particular good, and is therefore subject to particular reason, as Aquinas understands it. Given the attitudes that characterize love, it is only reasonable to be pleased in some circumstances, sorrowful in others, and angry in others. It would be unreasonable to react to disappointment with pleasure, or to success with sorrow. If we reacted in these ways, our emotions would not be organized by the sentiment of love; they would be organized by some other sentiment, or they would be disorganized reactions that result from some tendencies distinct from sentiments. The way in which a sentiment is organized makes it subject to some sort of reason. To have a sentiment rather than a series of unconnected likings and dislikings is to be responsive to some rational considerations. To see how this particular reason differs from universal reason, it is useful to consider a misguided passion. If A hates B because B is a friend of C, and A sees an opportunity to cause a quarrel between B and C, A has a sufficient reason, on the basis of A’s hatred for B, to take this opportunity. It does not follow that A has a sufficient reason, all things considered, to cause A and C to quarrel. A may have a better reason to avoid stirring up trouble between B and C, since it would be better if A got over this hatred for B. If A recognizes and acts on this better reason, A is guided by universal reason, as Aquinas conceives it. If A causes B to quarrel with C, A acts on the particular reason that is restricted by the passion of hatred, but A does not act on universal reason. This example illustrates the extent to which passions do and do not respond to comparative judgements about goodness. Some comparisons stay within the range of goods that are relevant to a particular passion. My anger might be re­dir­ ect­ed from A to B by the thought that B has treated me worse than A has; I might form a stronger desire to retaliate against B that weakens my previous desire to retaliate against A. To this extent anger may be flexible in response to an apparently greater good or evil, and may therefore respond to comparative judgements. But it is flexible only to a limited degree; it may not be flexible in response to my recognition that I need to avoid offending this person, and that it would be better not to express my anger. This comparison is not open to us, insofar as we take the point of view of anger, and our anger is not necessarily responsive to this sort of comparative judgement. Cases of this sort illustrate the difference between particular reason, to which the passions respond, and universal reason. This is the difference between passion and will.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

162  Terence Irwin

6.  How Passions are Open to Persuasion I said earlier that, according to Aquinas, the rational will exercises political rule— the sort of rule that citizens exercise over one another—over the passions. I suggested that, contrary to appearances, this is an illuminating analogy. We can now see what is illuminating about it. Some passions have the structure of sentiments, and hence are guided to some degree by particular reason. This feature of them makes the analogy of political rule appropriate. Citizens are responsive to reason and persuasion as well as to compulsion. The same is true of passions to the extent that they are guided by particular reason and aim at particular goods. If a passion aims at a particular good and is guided by particular reason, we can modify it by reference to the sorts of considerations that are proper to it. I have given some examples from love, hatred, and anger. A different application of particular reason allows us to extend their range by similarity and analogy. We pursue the objects of some passions as particular goods, insofar as we pursue them for some reason that also appears to justify pursuing other things of the same sort. Perhaps, for instance, I enjoy watching car racing because I find it exciting to see people taking risks and dealing with them skilfully. Or I might enjoy watching boxing because I find it exciting to see people exercise skill in fa­cing danger and injuring each other. I might, then, be expected to notice that these sports have something in common, and that someone who enjoys watching motor racing might also enjoy watching boxing or bull-­fighting. If this is what I enjoy about them, I do not simply enjoy the visual impression they create. If I learned that the racing drivers or the boxers or the bull-­fighters are really in no danger, I would learn that these spectacles are not good in the way I supposed they were, and I would no longer get the same sort of pleasure out of them. Moreover, if we aim at a particular good, we recognize degrees of goodness. If danger makes it exciting to watch bull-­fighting, a bull-­fight is a better spectacle than a contest between a rat-­catcher and a rat, if the rat-­catcher is in less danger than the bull-­fighter. In these respects passions aim at particular goods, but they are nonetheless distinct from the will. For though we desire these things as goods, and as having degrees of goodness, we do not desire them with reference to the universal good. To recognize a universal good is to be disposed to evaluate all goods with reference to their comparative goodness. The place we assign to different goods in this comparative evaluation expresses our conception of the final good, as Aquinas describes it. The outlook of the passions we have mentioned has no place for this sort of evaluation by reference to a universal good. A passion, therefore, is open to persuasion, and not merely to compulsion, insofar as it follows particular reason aiming at a particular good. If the appropriate sort of consideration appears to it, within the limits set by the object of the passion, it sees a reason to pursue something else. The presentation of such a

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

Education for Virtue  163 c­ onsideration is a means of persuading it. The persuasion is rational, insofar as the content of the consideration explains why the passion changes direction.

7.  The Influence of Universal Reason on Particular Reason Given these features of the passions and particular reason, Aquinas believes that particular reason can be moved by universal reason, and that therefore passions can be moved by the will (1a q81 a3). We have seen why particular reason is not moved by universal reason as such. The passions and the will differ because passions do not respond to considerations of goodness that are derived from the universal good. They do not respond to the universal good, because they are not guided by universal reason, but by particular reason, which does not think about the universal good. If particular reason could be moved by universal reason as such, it would be universal reason. Particular reason, therefore, has to be moved by universal reason indirectly, because universal reason presents considerations that particular reason appreciates. We might say that universal reason can guide a passion insofar as it ‘speaks the language’ of the passion; it needs to present situations in terms that evoke a response from the particular reason that directly guides the passion. My anger at one person may be diverted if someone else makes me angrier, so that the new insult absorbs the anger that would have concentrated on the old insult. To see how indirect control by universal reason is possible, we need to look more closely at how passions are redirected. We become aware of a further suit­ able object through a shift in attention. Even if I do not try or intend to think about something else, something else captures my attention simply because I encounter it, and it makes a stronger impression on me. I may recognize I have no good reason to be as angry as I am, but I may nonetheless stay angry until I encounter something else that holds my attention. My anger then fades, but not because I see that it is unjustified; seeing that might make no difference to my anger. It fades because I have something else to absorb me. Attention may shift as a result of experience, without anyone else’s intervention. In that case, the direction in which one’s passions are modified depends on what happens to catch one’s attention. In passions that have the structure of sentiments, a shift in attention may focus the passion on a new object. A sentiment tends to produce the emotion appropriate to it in specific situations, if we attend to the right features of these situations. If we hate our enemies, we tend to be pleased if they suffer the bad things we think they deserve. While sentiments are sometimes affected by the way our attention is attracted or distracted, they also influence the direction of our attention. If hatred for our enemies organizes our emotions, it makes us look for ways in which they are better off or worse off. If I know that my enemy has bet a large sum on a horse, I will

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

164  Terence Irwin hope that the horse loses the race, and I will be especially interested in the results. It is reasonable, given my hatred, to be interested in the results, and, since my sentiment is responsive to particular reason, it directs my attention in the appropriate way. This role of attention reveals one of the mental preconditions for moral education. If our attention is to be captured by other possible objects of a passion, we have to be curious enough about the world to survey it, so that different things may capture our attention. If I am so preoccupied by my anger at A’s having insulted me that I do not notice anything else around me, I may not notice that B has insulted me more than A has, or that I misunderstood A’s remarks and A has not really insulted me at all. Some degree of curiosity and responsiveness to the external world is needed to guide passions if they are to find new objects. This point about the preconditions for moral education does not distinguish it from other kinds of learning. Many kinds of learning require curiosity and the capacity to attend to unfamiliar features of the world. The distinctive precondition of moral education is the flexibility of the passions that results from the structure of sentiments. If we have passions that are guided by particular reason, and if we recognize other suitable objects for them, we are capable of being guided to objects that are more suitable for a specific passion. Attraction or distraction of attention does not require the intervention of will and universal reason. Different features of my experience may attract my attention to one object or another independently of my will. In such cases the resulting structure and direction of different people’s passions may simply reflect different people’s experience and their different degrees of curiosity. While a curious and exploratory outlook may result in more complex passions that result from a wider range of objects that have captured my attention, it is not obvious that this greater degree of complexity has anything to do with virtue. But attention can also shift because it is directed, by oneself or by someone else. According to Aquinas, universal reason and will affect particular reason and passion by directing one’s attention. At the earlier stages of education, one person’s will directs another person’s passions. At the later stages, one learns to direct one’s own passions by one’s own will. We can attend to things as a result of deciding to consider them; will and reason are the source of these decisions; hence will and reason can direct attention and modify passions. In this way the flexibility that results from a curious and exploratory outlook contributes to virtue, because it is suitable for direction by reason and will.

8.  The Effects of Attention and Inattention Directed attention turns our mind to greater absorption in one activity than in another. If one task absorbs our attention, this may be good or bad. It is good if

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

Education for Virtue  165 the task is important, and needs our full attention; in such cases we need to avoid diversion of attention. It is bad, however, if the focus of our attention is not all that important, and we are less likely to notice other things that we need to do. So far I have spoken of direction of attention within the scope of a particular passion such as anger, in cases where attention to one insult, for instance, causes us to stop thinking about another one. In these cases it is relatively easy to explain how the will directs attention. If we think it is better to be angry at a more serious insult, and to overlook a minor one that occupies us at the moment, we can draw out attention to the fact that one insult is greater than another, and the particular reason that belongs to anger will focus on the greater insult. Diverted attention, however, has a further effect on passions. Instead of turning us from one object of a particular passion to another (e.g. from a smaller to a greater insult), it may provoke one passion and reduce another. If I am brooding on my anger, but I face some immediate and life-­threatening danger, my sudden fear overshadows my anger.8 Alternatively, something pleasant may divert me from my anger by presenting me with something more immediate. In these cases, the change in my passion does not result from the recognition of a more suitable object for one passion, but from a new passion—as we might say—pushing out the old one. Aquinas explains different aspects of attention and distraction by reference to the division and the unity of the self.9 On the one hand, he recognizes distinct parts of the soul, acknowledging that a passion may operate outside the control of reason, and reason may operate without reference to any passion. Will and passion are distinct parts, because neither of them necessarily responds to the other. On the other hand, the fact that we have a unified consciousness explains why attention makes a difference. Even though simple feelings and complex sentiments do not necessarily change in the ways that we think would be better, they nonetheless appear to us in ways that constrain our responses to them. If we have a severe toothache while we are watching a play, we may prefer to think about the action we are watching on the stage, but we may not be able to think about it as much as we would prefer to. A particular passion affects the whole soul, because the soul is unified in ways that make it susceptible to the state of one or another part; our consciousness of each of these operations is part of a unified consciousness. In order to decide what to do or what to think about, we have to take account of all our mental life at a particular time. If some event or state is especially

8  Kant offers an example of overshadowing fear in his comment on the threatened execution (Kant 1788/1997, 30). 9  ‘The powers of the appetite and of apprehension are indeed distinct parts, but belonging to the one soul. Consequently when the soul is very intent on the action of one part, it is hindered from attending to a contrary act of the other part.’ (1-­2 a33 a3 ad2)

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

166  Terence Irwin prominent in our consciousness, it influences our tendency to take account of one thing rather than another, and to aim at one end rather than another.10 The example that I have just given reminds us that both pleasure and pain concentrate our attention on one thing and distract it from other things.11 The close connection between pleasure and attention is appropriately emphasized, but also exaggerated by Gilbert Ryle. In explaining how pleasure differs from passions that distract us from an activity, he observes that it actually increases our attention. If a person is perfectly collected in his deliberations and movements, he cannot, logically cannot, be described as furious, revolted or in a panic. Some degree of temporary craziness is, by implicit definition, an internal feature of passion, in this sense of ‘passion’. But no such connotations attach to pleasure—though they do, of course, to such conditions as thrills, transports, raptures and convulsions. If a participant in a discussion or a game greatly enjoys the discussion or game, he is not thereby estopped from having his wits about him. Else the keener a person was on golf or playing the fiddle the less would he be capable of doing these things intelligently. If to enjoy a thing in some degree were to be in that degree beside oneself, one would be distraught throughout the prosecution of all one’s favourite occupations. Complete absorption in something would entail complete inability to think what one was doing; and this is absurd. Complete calmness does not exclude great pleasure. The concept of enjoyment refuses to go through the same logical hoops as fury, despair, panic or glee. (Ryle 1954: 56)

This observation about pleasure and absorption leads Ryle to identify pleasure (in one sense) with absorption.

10  ‘Since all the powers of the soul are rooted in the one essence of the soul, it is necessary that, when the direction of the soul is strongly drawn towards the action of one power, it is withdrawn from the action of another power: because the soul, being one, can only have one intention. And for this reason if one thing draws upon itself the entire direction of the soul, or a great part of it, it does not undergo together with it any other thing that demands great attention.’ (1-­2 q37 a1) ‘[S]ince all the soul’s powers are rooted in the one essence of the soul, it follows of necessity that, when one power is directed towards its act, another power becomes slack, or is even altogether impeded, in its act, both because all power is weakened through being divided, so that, conversely, when it is directed on to one thing, it is less able to be dispersed to several; and because, in the operations of the soul, a certain direction is required, and if this is strongly fixed on one thing, it cannot attend as strongly to another.’ (1-­2 q77 a1) 11  ‘Delight that is attendant upon the operation of the intellect does not hinder it, but on the contrary it perfects it, as stated in Ethics x, 4, since what we do with delight, we do with greater care and perseverance. On the other hand, delight which is extraneous to the operation is a hindrance thereto: sometimes by distracting the attention because, as already observed, we are more attentive to those things that delight us; and when we are very attentive to one thing, we must needs be less attentive to another: sometimes on account of opposition; thus a sensual delight that is contrary to reason, hinders the estimate of prudence more than it hinders the estimate of the speculative intellect.’ (1-­2 q4 a1 ad 3)

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

Education for Virtue  167 There is the sense [of ‘pleasure’] in which it is commonly replaced by the verbs ‘enjoy’ and ‘like’. To say that a person has been enjoying digging is not to say that he has been both digging and doing or experiencing something else as a concomitant or effect of the digging; it is to say that he dug with his whole heart in his task, i.e. that he dug, wanting to dig and not wanting to do anything else (or nothing) instead. His digging was a propensity-­fulfilment. His digging was his pleasure, and not a vehicle of his pleasure.  (Ryle 1949: 93)

We may reasonably reject Ryle’s reduction of enjoyment to whole-­hearted absorption. His example does not even seem to provide sufficient conditions for enjoying digging. If I am digging a trench to protect myself from enemy artillery, I may indeed have my whole heart in my task, because I am anxious to get it done as soon as I can, but I may not enjoy it at all. The whole-­hearted absorption has to be the right sort of absorption, but it may be difficult to find the right sort of absorption without saying that it is the enjoyable sort. We may find, then, that to modify Ryle’s account so as to reach a true account of enjoyment, we have to make it circular. This result, however, does not make Ryle’s account useless. Even if we do not reduce pleasure to attention, we may reasonably treat pleasure in an activity as a certain way of attending to that activity.12 This feature of a pleasure is both a difficulty and an opportunity for the education of the passions. If our attention is drawn to the wrong actions, and we find ourselves enjoying them, our enjoyment increases our attention to them, and distracts us from other possible objects of attention. The fact that we are subject to distracting pleasures makes education more difficult. Passions attract our attention and distract us from features of a situation that we ought to attend to. If the object of the passion is pleasant, it captures our attention more firmly, so that we attend to, and act on, the wrong features of the situation. Aquinas uses this feature of passions and attention to explain the condition that is sometimes called weakness of will, or akrasia, or incontinence. This is the condition of agents who believe that it would be better to do x than to do y, but do y nonetheless, because of non-­rational desires that they follow contrary to their rational judgement. If we describe this condition as lack of self-­control or as weakness of will, we suggest that the primary error is not cognitive, but motivational. Aquinas agrees that a fault in the will is responsible for this sort of weakness, but he also believes that some cognitive fault contributes to it. If the worse course of action appears pleasant to us, it diverts our attention from our belief that it is the worse course of action, and causes us to choose the action we believe to be worse.13 12  Since we may take pleasure in other things besides activities, I am not offering this as a general description of pleasure. 13  ‘Now, that a man sometimes fails to consider in particular what he knows habitually, may happen through mere lack of attention: for instance, a man who knows geometry, may not attend to the consideration of geometrical conclusions, which he is ready to consider at any moment. Sometimes

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

168  Terence Irwin Even if this is not an adequate account of weakness of will, at least Aquinas is right to observe that passion and pleasure may combine to divert our attention. But the same features of passions and pleasure may also combine to provide an opportunity for moral education. If our attention is directed to the right things, and we find that we enjoy them, our attention to them is strengthened, and we can learn more about why they are the right things to attend to. In this case of diverted attention, as in the case of diversion from one object of anger to another, the diversion may happen without the intervention of reason and will. If I encounter frightening situations, my attention will be diverted from playing chess, whether or not we would prefer to keep our mind on chess. Once again, however, reason and will—our own or someone else’s—may intervene to direct attention. Since we have a unified consciousness (so that one object of attention pushes out another), and our attention is flexible enough to be sus­cep­ tible to being directed, the focus of our attention does not depend only on what happens to us, but also on what a rational will chooses.

9.  How to Direct Attention While misdirected passions can divert attention to the wrong objects, rightly directed passions can absorb us in the right objects with good results. We direct passions correctly by directing our attention in the right way.14 A rational will redirects passions by turning our attention to objects that are relevant to our passions, but might be overlooked if our attention is not turned in the right direction. One case in which redirection is needed is consideration of the future. Such passions as love, hatred, and anger are not indifferent to the future, since (e.g.) hatred often makes us plan to harm someone in the future. But the present situ­ ation often absorbs our attention and diverts us from the future effects of our own man fails to consider actually what he knows habitually, on account of some hindrance supervening, e.g. some external occupation, or some bodily infirmity; and, in this way, a man who is in a state of passion, fails to consider in particular what he knows in general, in so far as the passions hinder him from considering it. . . . That this takes place in the passions is evident from the fact that sometimes, when the passions are very intense, man loses the use of reason altogether: for many have gone out of their minds through excess of love or anger. It is in this way that passion draws the reason to judge in particular, against the knowledge which it has in general.’ (1-­2 q77 a2) 14  ‘Now it is evident that pleasure fixes a man’s attention on whatever he takes pleasure in: hence the Philosopher says . . . that we all do best that which we take pleasure in doing, while as to other things, we do them either not at all, or in a faint-­hearted fashion.’ (2-­2 q15 a3) ‘Quickness of movement results from a man being intent on many things which he is in a hurry to accomplish, whereas the magnanimous is intent only on great things; these are few and require great attention, wherefore they call for slow movement.’ (2-­2 q129 a3 ad3) ‘Fear, considered in itself, is always apt to hinder exterior action, on account of the outward members being deprived, through fear, of their heat. But on the part of the soul, if the fear be moderate, without much disturbance of the reason, it conduces to working well, in so far as it causes a certain solicitude, and makes a man take counsel and work with greater attention. If, however, fear increases so much as to disturb the reason, it hinders action even on the part of the soul.’ (1-­2 q44 a4)

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

Education for Virtue  169 and other people’s actions. Deliberate choice turns our attention to the future, so that we consider it in more detail than we normally would, and it becomes more prominent in our practical thought than it would be without any deliberate effort to focus on the longer term as well as the shorter term. As Hume puts it, avarice directed to the future may restrain avarice directed to the present, if we are trained to think about the future with more concentration. There is no passion, therefore, capable of controlling the interested affection, but the very affection itself, by an alteration of its direction. Now this alteration must necessarily take place upon the least reflection; since ’tis evident, that the passion is much better satisfy’d by its restraint, than by its liberty, and that in preserving society, we make much greater advances in the acquiring possessions, than in the solitary and forlorn condition, which must follow upon violence and an universal licence.  (Hume 1739/1978: iii 2.2.13) But here ’tis observable, that this infirmity of human nature becomes a remedy to itself, and that we provide against our negligence about remote objects, merely because we are naturally inclin’d to that negligence. When we consider any objects at a distance, all their minute distinctions vanish, and we always give the preference to whatever is in itself preferable, without considering its situ­ation and circumstances. This gives rise to what in an improper sense we call reason, which is a principle, that is often contradictory to those propensities that display themselves upon the approach of the object.  (Hume 1739/1978: iii 2.7.5)

In order to correct the passion of avarice, we do not need anything more than the particular reason that belongs to avarice itself. But we need something more than avarice to make us aware of these self-­destructive effects. Making us aware of them is a task for universal reason that directs our attention towards the future. A distinct task for universal reason is the redirection of attention from the object of one passion to the object of another. We may not want to face danger or discomfort, but if we attend to the fact that (e.g.) other people are grateful to us afterwards, we will act on passions that make us ready to face the danger or discomfort. In this case our attention does not focus on different objects of fear, but on objects of a different passion. Selection of the appropriate object is a task for a rational will. Redirection of attention is a task for reason and will because we need them to redirect correctly. We ought not to form the indiscriminate tendency to concentrate on the longer term rather than the shorter term. Nor ought we always to ignore fear so that other people will be grateful to us. We ought to redirect attention to the right objects on the right occasions. The selection of the right objects and the right occasions is the task of reason and will. These suggestions avoid the error that I mentioned earlier, of treating the passions as though they were rational agents who can respond to universal reason.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

170  Terence Irwin Passions are guided by particular reason, insofar as they have the structure of sentiments. As we saw earlier, such sentiments as love and hatred give us reasons to inquire further (into what will help those we love, or what will harm those we hate), and direct our attention to whatever seems appropriate (to help our friends or harm our enemies). In the same way pride and emulation support the relevant sort of reasoning (about a difficult task that will be an appropriate source of pride, or about some achievement of another person that we can emulate). Similarly, to have our attention turned from the object of one passion to the object of another, we do not need to see the reason for turning our attention; we simply need to be capable of having our attention turned, and capable of responding with a further passion to the new situation. Education of the passions for virtue is possible if the virtues and their characteristic expressions in action can become appropriate objects for the passions, by attracting our attention and holding it through enjoyment. We can see how this is possible in some simple cases. If children love their parents, they want to please them, and they tend to trust them. If we have friends and have the appropriate attitudes to them, we tend to share activities and pleasures with them. If parents or friends attract our attention to the right objects, our passions turn us towards virtuous actions. The attention that results from our passions also directs our passions. If A likes B, B’s interest in some activity may draw A’s attention to it. When A attends to it, A may discover that A enjoys it, or may see that B will be pleased if A takes an interest in it. But if our attention is once drawn to an activity, we may discover something that appeals to a different passion besides the one that initially drew our attention. If A goes to watch cricket with B, A may find cricket interesting, and form the habit of going to watch it even without B. The motives that we acquire by turning our attention and modifying our  passions in these ways are not necessarily the distinctive motives of the virtuous person. If the passions were capable of taking pleasure in virtuous action precisely because it is virtuous, they would also value the virtuous action irrespective of whether it appealed to any passion. Motives of this sort are characteristic of the virtuous person, but they are not the only motives that are characteristic of the virtuous person who listens to reason. Listening to reason causes agreement with universal reason, but not on all the grounds that persuade universal reason. The outlook of universal reason does not entirely reject the outlook of the different passions, but it directs them to some objects in some circumstances and away from others. This direction is characteristic of the virtuous person. The direction of attention is a task for will and deliberation. At some stages of moral development we need someone else’s will and deliberation to direct our attention so as to modify our passions. But if the development is complete, we direct ourselves through our own will and deliberation.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

Education for Virtue  171

10.  How the Passions Contribute to Virtue I have discussed the way in which passions can be educated in someone who has not yet acquired the virtues. If our will and practical reason are insufficiently trained, we can benefit if our passions are trained well by someone who deliberates and chooses better than we would if we had to rely on our own practical reason. Our passions may help us to borrow some of the practical conclusions of wiser people, even if we cannot grasp all the considerations that lead them to these conclusions. But why do we need well-­trained passions, as opposed to no passions, if our practical reason has been perfected as far as it can be perfected in a human agent? Once we have acquired a virtue, does it matter whether the passions remain in the right relation to will and reason or simply fade away? If we are not bothered by passions that lead us in the wrong direction, is the guidance of the will all we need? Aquinas replies that appropriate passions are constituents of a virtue even when the virtue has been fully developed, because they are subject to reason and moved by reason (q56 a4). They can be moved by reason not only to remove hindrances to the actions that we rationally choose, but also to make us act better than we otherwise would.15 Even well-­developed practical reason is still subject to the limitations of a human being. To form the right conclusions on the right occasions, we need to keep our mind on the relevant questions. To carry out the right conclusions we have to keep our mind on them when the time comes to act. In both of these cases we need our attention turned in the right direction. Since passions capture our attention, well-­directed passions turn our attention in the right direction, and thereby cooperate with universal reason.16 Universal reason, as Aquinas conceives it, begins from a conception of the universal and final good, and argues from it to more specific conclusions about the states of character one should acquire, and eventually to conclusions about what one should do in this or that type of case. The deliberative task that it faces is complex and difficult. This is a task for the

15  As Augustine argues, if we have a good will, we are better off if we also have the right passions than if we have no passions (1957–72: xiv 6). 16  ‘[I]t belongs to the perfection of man’s good that his passions be moderated by reason. For since man’s good is founded on reason as its root, that good will be all the more perfect, according as it extends to more things pertaining to man. . . . Hence, since sensory desire can obey reason . . . it belongs to the perfection of moral or human good, that the passions themselves also should be controlled by reason.’ (1-­2 q24 a3) ‘[Two ways in which the passions may be affected by reason.] First, by way of redundance: because, that is to say, when the higher part of the soul is intensely moved to anything, the lower part also follows that movement: and thus the passion that results in consequence, in sensory desire, is a sign of the intensity of the will, and so indicates greater moral goodness. Secondly, by way of choice; when, that is, a man, by the judgment of his reason, chooses to be affected by a passion in order to work more promptly with the co-­operation of sensory desire. And thus a passion of the soul increases the goodness of an action.’ (1-­2 q24 a3 ad1)

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

172  Terence Irwin reflective agent. Some deliberation, however, is the task of agents who are trying to decide what to do, and who would be foolish to undertake, on this or that particular occasion, the complex task that the reflective agent undertakes. Since deliberation has a large place in the formation and the expression of character, we need some non-­arbitrary means to limit the scope of deliberation on particular occasions. This restriction is the role of well-­trained passions. A passion absorbs some of our attention, and turns it away from other things that may appear to be as good as or better than the object of the passion; and it encourages a more favourable view of the object of the passion than one would otherwise take (Herman 1993: 193–5). We therefore need passions, insofar as they are subject to the will. Though the passions do not move us to voluntary action without the consent of the will, they can influence us without any intervention of the will. They do this partly by attracting our attention. If we know this, we will want to form—through our will and choice—passions that attract us to the right things independently of our will and choice. In this survey of the role of the passions in different stages of moral education, I have often appealed to their flexibility. We can align passions with rational will because we can find new objects for the same passions, and because our attention can be diverted from the object of one passion to the object of another. We would lack this flexibility if we were always dominated by particular passions with specific objects on different occasions that confront us independently of our choices. We would suffer from such domination if, for instance, we were always fainting with hunger until we were satisfied, or if we were always paralysed by fear until the object of fear went away, or if we were obsessively anxious about one thing until something else arose to produce the same obsessive anxiety. We need to be free from dominant passions of this sort, if we are to form the habits of curiosity and exploration that we need. Since we can find new objects for the same passions, and our attention can be redirected from the object of one passion to the object of another, we can align passions with rational will. Our capacity for re­dir­ ect­ed attention and the discovery of new objects of a passion results from the habits of curiosity and exploration which are associated with secure attachment. In this way secure attachment is important for moral education, no less than for other types of cognitive development.

References Aristotle. 1992. Nicomachean Ethics, trans. W.D.  Ross, revised J.O.  Urmson. In Aristotle: The Complete Works. Electronic edition. BOLLINGEN SERIES LXXI, Volume 2. ISBN: 978-1-57085-003-5. Charlottesville, VA: InteLex Corporation. Augustine. 1957–72. City of God. Loeb Classical Library, 7 vols. Various translators. London: Heinemann/Cambridge, MA: Harvard University Press, 1957–72.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

Education for Virtue  173 Bowlby, J. 1988. ‘The role of attachment in personality development’. In Bowlby, A Secure Base, 134–54. London: Routledge. Harcourt, E. 2013. ‘Attachment theory, character and naturalism’. In J.  Peters (ed.), Aristotelian Ethics in Contemporary Perspective, 114–19. London: Routledge. Herman, B. 1993. The Practice of Moral Judgment. Cambridge, MA: Harvard University Press. Hume, D. 1739/1978. A Treatise of Human Nature, L. A. Selby-Bigge (ed.), 2nd rev. ed., P.H. Nidditch, Oxford: Clarendon Press. Kant, I. 1788/1997. Critique of Practical Reason. Ed. and trans. Mary Gregor. Cambridge: Cambridge University Press. Knuuttila, S. 2004. Emotions in Ancient and Mediaeval Philosophy. Oxford: Oxford University Press. McDougall, W. 1919. Social Psychology (14th edn). London: Methuen. Ryle, G. 1949. The Concept of Mind. London: Hutchinson. Ryle, G. 1954. Dilemmas. Cambridge: Cambridge University Press. Shand, A. F. 1920. The Foundations of Character, Being a Study of the Tendencies of the Emotions and Sentiments (2nd edn). London: Macmillan. Sorabji, R. R. K. 2000. Emotion and Peace of Mind. Oxford: Oxford University Press. Stout, G. F. 1929. A Manual of Psychology (4th edn). London: University Tutorial Press. Swift, Jonathan. 1983. Complete Poems, Ed. P. Rogers. Harmondsworth: Penguin.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 26/07/21, SPi

9

Attachment, Virtue, and the Second Person Katrien Schaubroeck

There is increasing evidence that childhood experiences can have a substantial effect on adult well-­being, which involves (yet is not exhausted by) an adult’s social skills and relational capacities (Narvaez 2014).1 It does not seem far-­fetched to hypothesize further that there is a connection between childhood experiences and one’s moral sensitivities. The more specific hypothesis that I want to examine and defend in this chapter is that responsive parenting leads not only to secure attachment but also to the development of moral capacities, or virtues. There is an empirical and a conceptual dimension to this thesis. First, I’ll review the em­pir­ ic­al evidence: studies show a connection between style of attachment and moral behaviour and prosocial behaviour which, for the purposes of this chapter, I shall treat as a proxy for moral behaviour, while acknowledging the complicated relationship between the two concepts.2 Secondly, I’ll describe several explanations for the connection that have been suggested in the literature, concluding that none of them is fully satisfactory. I will therefore, thirdly, introduce an alternative explanation that draws on recent work by R. Jay Wallace and Stephen Darwall on the relational nature of morality. I will bolster this alternative explanation by examining conceptual connections between attachment, virtue, and the second person, and by anticipating two objections.

1.  Does Secure Attachment Lead to Virtue? Empirical Evidence Attachment theory is a theory within developmental psychology developed mid-­ twentieth century by John Bowlby and Mary Ainsworth and describing the relation between caregiver and infant. Since the 1980s the theory has been extended 1  I would like to thank audiences at the workshop ‘Natural Goodness and the Concept of Optimal Development’ in Oxford, and at the work-­in-­progress seminar in Tilburg, as well as two anonymous reviewers and Edward Harcourt for very helpful feedback. 2  This complicated relationship is discussed by Harcourt in the Introduction to this volume, as well as further on in this chapter. Katrien Schaubroeck, Attachment, Virtue, and the Second Person In: Attachment and Character: Attachment Theory, Ethics, and the Developmental Psychology of Vice and Virtue. Edited by: Edward Harcourt, Oxford University Press. © Katrien Schaubroeck 2021. DOI: 10.1093/oso/9780192898128.003.0010

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 26/07/21, SPi

Attachment, Virtue, and the Second Person  175 to relationships between adults. Attachment is a motivational and behavioural system which directs the child to seek proximity with a familiar caregiver when they are alarmed, with the expectation that they will receive protection and emotional support. The attachment figure serves as a source of comfort for the attached (safe haven) and as a base from which the attached explores the physical and social world (secure base). The attachment system has clear evolutionary underpinnings: its expressions to maintain proximity (smiling, crying, clinging) tie an attachment figure to an infant, maximizing its survival chances. All normally developing infants become attached if provided with a caregiver, but there are individual differences in the quality of the relationships. Across all populations approximately 65 per cent is securely attached. The others fall either in the category of insecure-­ambivalent (10 per cent), insecure-­avoidant (15 per cent), or disorganized (15 per cent) attachment. This distribution is fairly robust over all non-­clinical populations (see Morelli 2015). This is indicative of something that sociobiologists like Jay Belsky and Patricia Draper have pointed out to attachment theorists from the 1980s onwards: since all four attachment styles evolved via natural selection to facilitate the survival of infants, non-­secure attachment must also offer some goods, otherwise it would have been selected away. Simpson and Belsky formulate the point this way: ‘The environment of evolutionary adaptedness was not nearly as uniform, resource-­rich, or benign as many early attachment theorists envisioned, which means that no single attachment pattern should have been primary or species-­typical. In fact, . . . the adoption of different attachment patterns (in children) or orientations (in adults) most likely reflect evolved . . . tactics that probably improved reproductive fitness in response to the specific environments in which individuals grew and developed in ancestral times and perhaps still today’ (2016: 92). In some circumstances (personal or cultural) insecure attachment may be more adaptive. Imagine being born in a war zone where threats are constant: the unwillingness to allow any distance and the tendency to be clingy is surely a good choice (see Simpson and Belsky 2016; see also Harcourt 2013). But also in cases where caregivers tend to behave inconsistently or neglecting for personal reasons, the demanding nature of anxious-­ambivalent children may reflect a strategy designed to obtain or retain parental attention. Similarly, it makes perfect sense for a child whose parents are overwhelmed or distressed or unmotivated for whatever reason, to behave like an avoidant child does. Being self-­reliant and keeping a distance might enable the infant to maintain reasonably close proximity to parents without driving them away. This strategy might increase survival chances of infants who would have been completely abandoned had they been more demanding. Even if there is no reason to call secure attachment patterns species-­typical, it is still tempting to think of secure attachment as the best attachment style.3 After 3  Obviously the standard of evaluation here cannot be adaptive success. As Sarah Blaffer Hrdy puts it: ‘A certain type of sociopathic personality, forged out of desperate circumstances, though highly

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 26/07/21, SPi

176  Katrien Schaubroeck all, it is by definition connected to a lot of goods, such as the capacity for sharing emotions and collaboration, but also for independent activity and dealing with separation. Important for our purposes is that the securely attached are more likely to treat their attachment figure in a kind, trusting, and generous way by definition. To examine the connection between attachment and virtue it should be asked whether securely attached children are also kind and generous with regard to strangers. It is not part of the conceptual framework of attachment theory that that would be the case. After all, the attachment system aims at seeking nurturance and care. Yet from early on there was the suggestion that the attachment system also played a fundamental and causal role in prosocial behaviour and the capacity for empathy beyond kin. Attachment security was considered a pillar of  mental health and social adjustment, as illustrated by Bowlby’s early study on juvenile thieves and their home-­life. In a series of studies that set off with his ‘44 juvenile thieves: their characters and home-­life’ (1944), Bowlby observed that sociopaths were disproportionately drawn from the portion of the population that had been avoidantly attached to their mothers. And also Ainsworth expressed the idea that the more sensitive and responsive a caregiver is, the more helping, sharing, and caring will be rewarded and thus cultivated (Ainsworth et al. 1974: 119). So an interest in moral development is inherent to attachment theory, and the working hypothesis has always been that secure attachment naturally gives rise to virtue. The connection between attachment style and moral virtue has also been ­empirically tested, in many longitudinal studies exploring the breadth of later behaviour associated with attachment security. Correlations have been found between early security and later relations with parents, peers, and other social partners, as well as with self-­concept, personality development, social cognition, behaviour problems, and psychopathology. One of the largest longitudinal ­studies, The Minnesota Study of Risk and Adaptation from Birth to Adulthood (Sroufe  2005), tracked the psychosocial outcomes of children observed in the Strange Situation Procedure in infancy and followed into adolescence. The focus was on associations between attachment and personality, the latter being assessed through a variety of measurements: behavioural observation, interviews, self-­ reports. They found a strong association between early attachment security and high levels of social competence, peer confidence, ego control, low externalizing behaviour—in short: prosocial behaviour—but also high self-­esteem and self-­ confidence (summarized in Thompson 2016). Although many attachment theorists intuit a connection between secure attachment and virtue, the empirical evidence is troubled by inconsistency and non-­replication (Fearon et al. 2010 presents a helpful review). In their illuminating, encompassing meta-­analysis Van IJzendoorn and Bakermans-­Kranenburg (2014) conclude that overall the results do not show a strong and straightforward undesirable from society’s point of view, might from the developing child’s perspective be a way of making the best of an appalling situation’ (Hrdy 1999: 527).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 26/07/21, SPi

Attachment, Virtue, and the Second Person  177 causal relation between a happy childhood (self-­report of low love withdrawal) and moral agency (prosocial behaviour in the lab). They found many inconsistencies. In a ball game where children can choose whether to compensate for the unfair losses of another (computerized) participant, giving the child an oxytocin sniff makes the child with a happy childhood more generous while it has no effect on the other subgroup. So this result suggests that the way in which a child is loved determines how susceptible it is to other factors that will influence its moral behaviour. But another set of studies shows that situational factors have an effect across the board: regardless of how happy their childhood, setting a good ­example will always trigger prosocial behaviour in children. And this history of non-­ replication is characteristic of the topic overall. Methodological explanations suggest themselves when meta-­analysis for example shows that while associations are consistent across high- and low-­SES samples, there are gender differences: perhaps observational data might be coloured by gender stereotypes (Fearon et al. 2010). Other meta-­analyses show that the younger the children are the more mixed the results (Shaver, Mikulincer, et al. 2016), that contemporaneous associations are stronger than predictive ones, and that pre­dict­ive associations are weaker the longer the distance in time (Thompson 2016). Perhaps the poor predictive value of attachment at a young age is not due to age as such but to the tests used to measure attachment, which of course cannot be separated from the age at which the assessment is conducted. Studies using the Strange Situation Procedure show smaller effect sizes than interview methods. And with regard to the measured outcome, it is striking that studies that assess externalizing behaviour problems via direct observation identify larger effect sizes than those that rely on questionnaires from teachers or questionnaires from parents (Fearon et al. 2010). Apart from methodological problems, conceptual issues also muddle the results. The type of measurement of prosocial behaviour is a crucial determinant, and defining what counts as prosocial behaviour is at least as crucial and thorny. What looks externally like prosocial behaviour may be motivated by empathy but not necessarily: a child may share a toy with a sad peer out of compliance with a teacher’s expectations, or out of deference to the peer’s social dominance, rather than out of genuine concern for the peer’s well-­being. This has been corrected for in adult studies (there are ways to distinguish authentic from so-­called defensive moral choices) but in children it is more difficult to examine the underlying mo­tiv­ations for their prosocial behaviour (see Shaver, Mikulincer, et al.  2016). And the conceptual issue runs even deeper: is authentic prosocial behaviour always a sign of virtue? There are cases (I am thinking of situations of oppression or institutional injustice) where compliant behaviour may be called prosocial, and may even be motivated by authentic, non-­defensive reasons, but where it is unclear whether it is virtuous.4 4  Mikulincer and Shaver (2007) did a study that found that more secure Palestinians living in the territories occupied by Israeli soldiers were more, rather than less, hostile towards Israeli Jews and

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 26/07/21, SPi

178  Katrien Schaubroeck Regardless of the ambiguities in the results due to methodological and conceptual obstacles, there is more or less consensus in the literature that there is a strong association between secure attachment and prosocial behaviour. One of the most convincing and illuminating sets of studies is done by Mikulincer and Shaver, who found a neat way to operationalize the difference between social competence and moral virtue in their studies with adults. In a 2005 study they examined whether an experimentally enhanced sense of security (using security priming techniques, such as subliminally presenting the names of attachment figures or supraliminally asking the participants to visualize the face of an attachment figure) increases participants’ compassion and willingness to help a distressed person (Mikulincer, Shaver, et al. 2005). Participants watched a confederate performing a series of aversive tasks. The confederate became increasingly distressed, and the participant was given an opportunity to take her place, helping her out. Shortly before the scenario unfolded, some participants were primed with a representation of attachment security (a name of a security provider, based on information they had supplied) or with the name of a familiar person who was not an attachment figure. Security priming increased participants’ compassion and willingness to take the distressed person’s place. Attachment security thus seems associated with helping behaviour. Whether this also shows a positive connection between attachment security and moral sensitivity is unclear, because, as already said, prosocial behaviour need not prove an altruistic character or virtuous motive beyond the behaviour. To differentiate between moral and non-­moral motives for prosocial behaviour, Shaver and Mikulincer (2012) developed a study in which they test whether threats to the self increase the tendency to act in a prosocial way. Using a standard questionnaire, they determined which attachment type subjects belong to. Subjects were then presented with a moral dilemma and asked what they would do. Making a situation more threatening to the self stimulated prosocial behaviour in anxious individuals, not in secure ones, which suggests indeed that the motives to do what is morally good in anxiously attached individuals are bound up with self-­defensive motives such as fear of rejection and need for approval. Securely attached individuals make what the authors call authentic moral choices, out of concern for others’ welfare rather than as a defence against threats and narcissistic wounds. The results of the Mikulincer and Shaver studies pose an interesting challenge: why would there be a connection between secure attachment (first and foremost a property of the relationship with caregivers, after all) and non-­affectional bonds

more accepting of violence against them. Reviewing this result Shaver, Mikulincer, et al. conclude that ‘Achieving a world at peace requires humane ethics, a more tolerant cultural and educational climate, and good judgment and effective political will on the part of leaders, not just securely attached individual citizens’ (Shaver, Mikulincer, et al. 2016: 904). One could additionally draw the implication that, in this world, virtue can manifest itself in non-­compliant behaviour. For further discussion of the relationship between the concepts of prosocial behaviour and of virtue, see Harcourt, Introduction to this volume.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 26/07/21, SPi

Attachment, Virtue, and the Second Person  179 with strangers? Pioneers like Bowlby and Ainsworth hypothesized the connection but it is, in fact, not obvious why the way in which we treat strangers would resemble our interactions with a close circle of loved ones. As conceded, the empirical evidence is ambiguous, and as a philosopher I cannot aspire to solve that problem. I take it that the results are interesting enough to invite reflection on why the connection is so intuitive and why in some studies indeed a causal connection has been established. While psychology should tell us whether and under what conditions there is a causal link or not, philosophy could be useful in explaining why there might be one. So this is the task I set myself for the remainder of this chapter: to find out what conceptual connections there are between secure attachment and moral virtue.

2.  Why would Secure Attachment Lead to Virtue? Four Linking Mechanisms Attachment theorists have themselves attempted to explain the connection in various ways. I detect four different linking mechanisms in the literature, and I will add a fifth one. The first explanation turns on the mechanism of modelling: secure children copy their parents’ harmonious relationships (with their children, but typically also with other people). Intergenerational transmission of attachment styles is relevant here: parents of securely attached children are typically also themselves securely attached and entertain harmonious relationships with kin and strangers. Parents thus set the example, and the infants copy their pro­ social behaviour as they copy lots of other behaviours. There is nothing special about moral development on this account. Shaver, Mikulincer, et al. (2016: 882) describe this linking mechanism in their overview as follows: ‘It is also possible that children incorporate behavioural routines for care in the same way they model other kinds of behaviour, such as eating with a spoon, brushing teeth, dancing, throwing a ball.’ A second, more popular explanation internalizes the linking mechanism and posits the influence of so-­called internal working models (IWMs). Bowlby himself coined the term of ‘an internal working model’ and seems to have meant it as a rich conceptual metaphor, but it sometimes gets applied so widely that it ­threatens to become an uninformative ad hoc explanation (Thompson  2016). Despite the disagreement about how IWMs function exactly, they are often invoked to explain developmental continuities between the organization of the attachment relationship and functioning beyond it (in time and space). Thompson offers the following useful specification of the workings of IWMs (2016: 332): IWMs are based on infants’ expectations for the accessibility and responsiveness of their caregivers. These expectations develop into broader representations of their attachment figures, interpretations of their relational experiences,

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 26/07/21, SPi

180  Katrien Schaubroeck guidelines about how to interact with others, and even beliefs about themselves as relational partners. These mental representations . . . expand into broader interpretive filters through which children and adults reconstruct their experience of new relationships in ways that are consistent with past experiences and the expectations arising from secure or insecure attachments . . . As a consequence, children choose new partners and behave with them in ways that are consistent with, and thus help to confirm, the expectations created from earlier attachments. IWMs therefore constitute the bridge between an infant’s experience of sensitive or insensitive care and the development of beliefs and expectations that affect [later experiences].

Key in the explanation are expectations: experiences of, for example, relational security amount to a secure base script, which sets expectations for relationships and encounters with other people in general. When social engagement is mediated by representations of self (as worthy of care) and others (as trustworthy), it is understandable that securely attached children behave less suspiciously, more confidently, more collaboratively, and in a less self-­focused way. One line of evidence showing the impact of inner working models and the mediating role of expectations comes from a remarkable study of infants’ response to geometric representations of a caregiver (a large oval) and a child (a small oval). Securely attached infants look longer at visual displays in which the caregiver oval is unresponsive to the child’s oval distress, whereas insecure infants look longer at displays in which the caregiver oval is responsive (Johnson, Dweck, and Chen 2007). Plausible as this explanation may seem, there remains a conceptual knot to be untied. Secure attachment gives children a reason to represent the self as worthy of care, and others as caring, responsive, and trustworthy. But there is no rational pressure to also represent others as worthy of care. As Shaver, Mikulincer, et al. (2016) rightly observe in their review article: ‘The precise mechanism by which the secure model of others as caring becomes integrated into a model of the self as caring for others remains unclear.’ One suggestion is that ‘care leading to secure attachment shows children both sides of a responsive relationship and that children can draw upon both representations when responding to the needs of others’ (Shaver, Mikulincer, et al. 2016: 882). But this suggestion makes the difference between the first and the second linking mechanism (modelling versus IWM) rather small. A third explanation is suggested by Edward Harcourt (this volume) and emphasizes the secure-­base aspect of secure attachment over the safe-­haven aspect. Securely attached infants rely on their caregivers when they need comforting. But an equally important mark of the securely attached is their confidence to explore the world and all the reasons for action it offers. Even when staying neutral on the metaphysics of reasons, one could say that in our lived experience reasons present themselves as part of the world. Likewise, without going into the

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 26/07/21, SPi

Attachment, Virtue, and the Second Person  181 details of virtue ethics, one could say that being kind, honest, or otherwise virtuous is being responsive to reasons (because I promised, because he likes it . . .). If securely attached children have more practice in confronting the world and dealing with various reason-­giving situations, it is only to be expected that they become well-­versed moral agents. This explanation is a valuable addition to the literature, although I have some concerns that I will come back to after explaining the fourth, well-­established linking mechanism. A fourth possible explanation locates the link between attachment and virtue in cognitive development rather than affect regulation or personality development. Simply put, the hypothesis is that secure children are more virtuous because they are better at reading others’ minds. The connection between attachment styles and mind-­reading capacities has been documented extensively by developmental psychologist Peter Hobson and his team. In Hobson (2002) he invokes the model of a triangle to explain how a child learns about the independent existence of the world and about other persons’ independent perspectives simultaneously. The philosopher Donald Davidson called this a process of triangulation, and argued that there needs to be more than one person for there to be a world about which one can hold true beliefs. It is for this reason that having knowledge of the content of one’s own mind, having knowledge of other people’s mental states, and having knowledge of a common, objective world are interdependent. In line with Davidson and going against Piaget’s legacy, Hobson holds that direct engagement with the world is not enough for developing the mind, and that an intermediate person is needed. Adding a psychological layer on Davidson’s triangulation model, Hobson points out that a caregiver first has to note the state a child is in (e.g. its anxiety) to pull off the triangulation process. And that is why attachment style can make a difference to the child’s mental development. Not all caregivers are equally sensitive and responsive to the child, or interested in the child’s mind. Several experiments provide evidence compatible with Hobson’s hypothesis that a child’s mental development is affected by the difference between caregivers who are ‘mind-­minded’ and caregivers who treat an infant as a creature whose needs must be met. For example, maternal use of mind-­minded language at 5 months predicts better performance at the false belief test at 5 years. Poor attunement between infants and depressed mothers and mothers with borderline personality disorder who are respectively less sensitive and more intrusive impacts negatively on mentalizing capacities later (Hobson 2002). If the development of thinking is influenced by a caregiver’s emotional relations with the infant, it is no surprise that disturbances in attachment may lead to disturbances in mind-­reading capacity. And indeed there is evidence that secure attachment predicts good mentalizing capacities (Hobson 2002). It may seem only a small step to the further hypothesis that disturbances in attachment may also lead to disturbances in moral development. After all, only if one is capable of  reading another’s mind can one react to it in morally appropriate ways.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 26/07/21, SPi

182  Katrien Schaubroeck The phil­oso­pher Hugh LaFollette indeed relies on this conditional when he writes (LaFollette 1995: 208–9): We cannot develop knowledge necessary to act morally unless we have been in intimate relationships. . . . Someone reared by uncaring parents, who never established close personal ties with others, will simply not know how to look after or promote the interests of intimates or strangers. We cannot promote interests we cannot identify, and the way we learn to identify the interest of others is by interacting with them.

LaFollette hastens to add that having been in close relationships does not guarantee that someone will be virtuous. His point is that exposure to personal relationships is necessary, yet not sufficient, to acquire the knowledge and motivation to be moral. And here we arrive at a fundamental problem with mind reading as the linking mechanism between attachment and virtue. At most this mechanism connects virtuous behaviour to conditions that enable an agent to behave morally. But note that many things must be true of an agent for him or her to be able to do the right thing, going from having a body and bodily control to being able to read other minds. If the envisioned explanation consists of no more than the statement that being securely attached makes it possible that one acts virtuously, we end up with a rather weak linking mechanism. Moreover, this linking mechanism makes the connection between secure attachment and virtue rather external. Hobson’s hypothesis that secure attachment provides the necessary environment in which children learn to recognize and ascribe mental states is interesting and plausible enough but it entails predictions about moral capacities as well as about dozens of other capacities that rely on good mind-­reading capacities. Hobson’s linking mechanism connects secure attachment to morality in a way that it would also connect secure attachment to being good at playing poker, to name one example. The latter requires good mind-­reading capacities, and thus indirectly depends on a loving context in which human beings learn to read minds. But obviously there is no internal connection between being securely attached to a caregiver and playing poker. It seems that something more is going on in the case of morality: the secure attachment between caregiver and child seems to reflect something about the content of a virtuous attitude or about the point of morality. The same criticism applies to the other linking mechanisms: modelling, opportunities to explore, IWM. If these explain the connection, they are not specific for morality: through these linking mechanisms, secure attachment should predict a lot of successes. For example, if Harcourt’s explanation of the connection via exploration of the world is correct, that would entail that securely attached children have an advantage in reasons-­responsive behaviour overall: they would be less likely to make stupid financial investments, they would be less likely to bump

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 26/07/21, SPi

Attachment, Virtue, and the Second Person  183 into things or to have car accidents, in short, they would be more likely to become successful in life overall—or at least in these domains that require reason-­ responsiveness. Sure enough, secure attachment may be of instrumental significance to a whole lot of things, as evidence suggests.5 And further empirical research might even show that there is as strong a connection between secure attachment and moral outcomes as between secure attachment and non-­moral outcomes. But it seems to me that there is a special connection between responsiveness to moral reasons and secure attachment that is absent in the case of responsiveness to prudential or non-­moral reasons. For lack of a better word I call this special connection ‘internal’, and in the next section I defend a specific interpretation of the internal connection. As will become clear, the point is philosophical and my comments about the four available linking mechanisms would not be answered by further empirical research.

3.  An Alternative Explanation of the Link between Attachment and Virtue On the basis of the four linking mechanisms suggested in the literature we do not get insight into a distinctive link between secure attachment and virtue, apart from the other good things it is associated with. It might seem as if Aristotelian virtue ethics could offer the wanted explanation. After all, if secure attachment leads to success in life, making life better, it could be thought to be conducive to well-­being in the Aristotelian sense of ‘eudaimonia’. But evolutionary or survival success does not map neatly onto moral success. As was pointed out before, no attachment style is ‘naturally defective’, and in some circumstances insecure attachment offers the best chances of survival. And as Harcourt (2013) rightly underlines: to carry on our species life, ‘it is surely good . . . to have some people around who are risk-­takers (and so presumptively insecure-­ambivalent), and some who are precociously self-­reliant (presumptively insecure-­avoidant)’ (2013: 127). Both for the group and for individuals it is conducive to survival that not everyone is cooperative and compliant. Personality traits that come with insecure attachment can undeniably lead to well-­being and success in a non-­moral sense (if not to virtue). Pace neo-­Aristotelians like Philippa Foot, moral vices thus cannot be thought of as natural defects within the evolutionary framework that 5  Correlations have been sought and found between attachment security and correlates that went far beyond what Bowlby originally envisioned: evidence suggests that secure attachment predicts later cognitive and language development, curiosity, ego resilience, maths achievement, political ideology . . . (see Thompson  2016 for an overview). A telling example is a study that draws a connection between attachment security and exploratory consumer behaviour: the more securely attached someone is, the more flavours of ice cream he wants to try (Li, Bruyneel, and Warlop 2012).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 26/07/21, SPi

184  Katrien Schaubroeck attachment theory is part of. A naturalized reading of virtues as benefiting the human species does not fit attachment theory. This should prompt us to look beyond virtue ethics in our search for an ex­plan­ ation between secure attachment and morality. We need a bridge between secure attachment and virtue that recognizes that all types of attachments are evolved adaptations (and thus good for something), and also leads to a distinctive ex­plan­ ation of why being securely attached leads to being virtuous (rather than to a multitude of beneficial effects). On the moral theory that I will rely on, the awkward individualist and inconsiderate risk-­taker act immorally not because their goals are incongruent with the goals of the human species, but because they neglect legitimate claims imposed on them. As I will argue, attachment theory aligns well with the moral framework that grounds morality not in ideas of eudaimonia but in second-­personal interdependent connections between agents, and that describes moral development not in terms of personal growth but in terms of growing awareness of interpersonal norms.6 The linking mechanism I propose turns on the notion of interpersonal norms, or second-­personal relationships in which norms are established, claims are exchanged, demands are made. My hypothesis is that secure children have more (positive) experience of standing in a second-­personal relationship where partners make demands, expect co­oper­ ation, or hold one another accountable and that this experience provides them with the knowledge of and willingness to comply with moral norms. To make sense of this linking mechanism we need to think of moral obligations as relational obligations, and of the moral domain as shaped by second-­ personal relationships in which people demand and expect things from one another. I borrow these terms respectively from R.  Jay Wallace and Stephen Darwall, and while I recognize the important differences between their positions (see Wallace  2007 and footnote 8), they are not relevant for my purposes. Interestingly, while Wallace opens the door for an association between parental love and moral capacities, Darwall is opposed to the idea that care and love could be conducive to morality. Through a reconstruction of how both see the relation between love and morality I will offer a specific interpretation of the connection between secure attachment and virtue. In ‘Duties of love’ (2012) Wallace argues in favour of an internal connection between love and morality. He uses the terminology of love and moral obligations, rather than secure attachment and virtue, but his main idea is relevant regardless. Wallace claims that relationships of love provide a paradigmatic understanding of the very notion of an obligation, without which moral obligations cannot be understood. This case depends on the thought that the notion of being under an obligation requires a relational understanding: in order to accept 6  Perhaps this is not an exclusive choice: one could argue that respecting interpersonal norms is part of good character-­formation.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 26/07/21, SPi

Attachment, Virtue, and the Second Person  185 the idea that our agency can be (legitimately) constrained by obligations we must think of obligations as conceptually connected to (valid) claims that other people have on us. And since relationships of love are the most familiar and extensive contexts for the application of the general idea of a relational obligation, they provide the most compelling examples of obligations that are relational. Not only does love offer illustrations of obligations, in Wallace’s reasoning loving relationships also provide the soil for an ethical understanding of our relationships to strangers. This means that duties of love cannot be reduced to moral obligations. Wallace defends ‘the idea that there are sui generis duties of love: duties, that is, that we owe to people in virtue of standing in loving relationships with them’ (2012: 175). Duties of love thus do not derive from general moral principles (such as that one should be grateful, or not betray trust). In support of this non-­reductionist view of duties of love he first cites examples from daily life. We say things such as: ‘I must cancel the meeting, my sister needs me.’ We offer the fact that we stand in a loving relationship as a reason: she is my sister, I love her, she requires assistance, and that is all there is to it. No extra (moral) justifications are offered, or requested. As a methodological rule Wallace holds that we should take phenomenology at face value, and not reinterpret first appearances unless there is strong independent reason to do so. Our practices of ordinary justification treat relationships of love as direct sources of obligation. A second reason in favour of the non-­reductionist view is that reductionist accounts of duties of love (accounts that interpret duties of love as moral duties in disguise) may be on unstable grounds. Wallace writes: ‘It is not really clear . . . that we can understand the notion of obligation if we are prepared to deny that there can be obligations that arise directly from our relationships with those we love’ (2012: 192). This thought is further elaborated in his latest book on morality as a domain of directed obligations, that is obligations that arise from a relationship and that are issued ‘from an I to a you’ (see Wallace 2019). I find Wallace’s assumption that we need experience with love in order to know what moral obligations are (in their form of relational obligations) intuitively plausible. But I am not sure that Wallace’s explanation for why that is so goes to the heart of the matter. Wallace thinks that duty is the connecting factor between love and morality. But clearly love provides us with the experience of many things, and I am not sure duties are the most salient feature, nor the most relevant in connection to morality. Note for example that love also gives us paradigmatic examples of a specific type of reasons, so-­called reasons of love. The relationship between lovers in virtue of which they have certain responsibilities towards one another gives rise to reasons which we might not want to call duties. In a love relationship for example one may have a reason to cook the other’s favourite meal when he or she feels down. Or friends may expect each other to take an interest in their hobbies. But these reasons and expectations are not formulated as claims and demands. While Wallace says that relationships of love provide the paradigm

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 26/07/21, SPi

186  Katrien Schaubroeck example of a relational obligation, I want to say that they provide the first and primary experience of possessing the standing to ask, expect, hope, or want things from each other. Formulating the point in terms of ‘a standing to ask’ is reminiscent of Stephen Darwall’s account of morality as grounded in second-­ personal demands and second-­personal reasons to comply. In his (2006), Darwall offers an analysis of the concept of a moral obligation: what are the assumptions underlying the thought that someone is under a moral obligation to do X? In exposing these assumptions, Darwall describes a circle of concepts that presuppose each other: an agent’s standing to make demands, the addressed person’s second-­personal competence and accountability for complying, and the exchange of second-­personal reasons to comply. So for example when agent A demands B to get off his foot by way of moral demand, he will (1) take himself to have the standing to make valid claims on B and the standing to hold B accountable if he fails to comply without excuse, (2) take B to have the competence to recognize A as someone with standing to make claims on him and the mirroring competence to hold himself accountable, and (3) take his claim to be valid since recognizable from a shared, or second-­personal standpoint that B can take as well. According to Darwall’s analysis, if these assumptions are not implicit in A’s claim, A is not putting B under a moral obligation but rather bullying or threatening him. The reference to the second-­personal standpoint is essential. From this standpoint reasons are offered that cannot be reduced to so-­called first-­personal or third-­personal reasons. If A’s demand to B to get off his foot is a valid demand, the reason that B has to get off A’s foot is a second-­personal reason: it is grounded in the standing that A has to ask things from B, and not in third-­personal reasons such as detached facts about pain and pleasure, nor in first-­personal reasons grounded in B’s desires. Moral agents have the standing to ask things from one another that need no further verification or foundation. This idea is a deeply Kantian idea that also tracks a feature of daily moral discourse, as illustrated in the slogan ‘no means no’. Importantly, this second-­person standpoint to which the title of Darwall’s book refers is not the standpoint of a victim. For a person to make valid claims on someone else, it is not necessary that the latter affects the former directly (e.g. by standing on his toe). What matters is that the former can judge and blame the latter for his actions including those affecting third parties. Thus the second-­ person standpoint underlying and enabling moral practices, in Darwall’s view, is the standpoint from which we hold others and ourselves accountable and from which we address demands to one another. It is a standpoint that all of us, as members of the moral community, can take up all the time. Against the background of this general understanding of second-­personality, it is all the more striking that Darwall excludes parental love. The love of a parent for his child, according to Darwall, does not invoke second-­personal reasons or

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 26/07/21, SPi

Attachment, Virtue, and the Second Person  187 the authority to address them, and does not rely on the child’s second-­personal competence to recognize and comply with those reasons. With his criticism of care as the wrong kind of moral attitude, Darwall stands in a long debate between Kantian ethicists and care ethicists. On the correct picture morality is a matter of second-­personal address, says Darwall. Not care but respect is the primary moral attitude because the attitude of respect responds to second-­personal reasons (of the form ‘That A demands B to get off his foot is a reason for B to do so’) while care responds to third-­personal reasons (of the form ‘That A is hurt is a reason for B to get off his foot’). Respect focuses on the dignity of persons (their standing to make claims), rather than on their well-­being. To bring out the differences between benevolent care and respect Darwall reflects on the relations between parents and children at different stages in the children’s life (2006: 128). He writes: Parents may legitimately give relatively little intrinsic weight to a sufficiently young child’s protest against eating healthful food, although they should take account of its bearing on the child’s welfare, for example, the likelihood that eating it will be an unpleasant experience, the long-­term effects on the child’s well-­ being of insisting that she eat it, and so on. At this stage, the parents may properly be guided by the child’s welfare alone.

In other words, up to a certain age, parental love takes the form of benevolent concern, thinks Darwall. Only later situations arise in which the desire to take care of a child may conflict with and be subordinated to the demands of respect. Once a child becomes a person with second-­personal competence, it needs to be respected, not to be taken care of. Darwall writes (2006: 128): For parents not to take a middle-­aged daughter’s will as having intrinsic weight, indeed, as governing, would clearly be disrespectful, paternalism in the pejorative sense. Now she has a second-­personal standing she simply did not have near the beginning of her life. And were her parents to attempt to pressure her to eat ‘her broccoli’ at this point, they would rightly be subject to reproach.

I agree with Darwall that in general it makes sense to distinguish between care that has well-­being as its focus and respect that responds to dignity. But in the case of parental love Darwall makes too sharp and too simple a distinction between care and respect. A parent’s loving attitude to his children is never one of just looking after their welfare—as he does with the plants in the garden. Real care for an infant incorporates respect, and respect does not silence concerns about welfare. Consider the broccoli example again. Of course parents have to see to it that their children eat vegetables. But there are many ways of doing that, one way better than the other. Parents do not force the broccoli down a child’s throat, and this is

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 26/07/21, SPi

188  Katrien Schaubroeck not just because they are concerned that it will traumatize the child and have bad consequences for its welfare in the long term. The real reason is that a loving parent is not only concerned about bringing it about that the broccoli is eaten, but also about how he can teach the child to eat vegetables. If the child will not eat  broccoli, he’ll mix it in the soup or he’ll try another vegetable. He figures out together with the child how to eat vegetables. In this sense, a loving parent is not just focused on the child’s well-­being: the eating of the broccoli is only an instance of a broader concern, namely teaching the child how to do things in life, how to be responsive to reasons, including considerations of welfare but also reasons that refer to someone’s autonomous choices. Darwall thinks that the possession of a second-­personal standing marks a crucial transition in the parent–child relationship: before that moment ‘paternalism’ would be alright, after that moment forcing the child to do what is in its interests would manifest a failure of respect. But there is no such moment. The difference between luring a child into eating broccoli or reasoning the child into eating veget­ables is merely a difference on account of the fact that loving care can manifest itself in many ways and these manifestations evolve together with the capacities of the parties. The second-­personal relation between parent and child is not reserved for later, it is there from the start. Parent and infant interact from the very beginning, as has been amply documented by developmental psychology in the tradition of attachment theory.7 To enable the interaction it must be true that even very young children have a will and are capable of expressing their will. Also for very young children it is true that their no means no. Their protest can be seen as a statement of their own will as a limit to the parent’s actions. And conversely, their ability to comply presupposes recognition of the parent as a limit of their will. Harcourt (2017) makes sense of a kind of Kantianism for pre-­rational beings in this way, and he adds that in early love relations one comes to recognize (in a primitive way) oneself as a limit to another’s will, and the other as a limit to one’s own. Parent–child relationships are in this sense second-­personal from the start. Now what is gained by introducing the notions of a second-­personal reason and a relational obligation into the discussion of attachment theory and morality? The critical assessment of available explanations for the connection between secure attachment and virtue in section 2 made clear that a successful explanation should avoid two pitfalls. It ought not to assume naively that love for one’s mother automatically (magically) expands to love for mankind. And it should not portray virtue as one of the many goods that secure attachment brings about (on a par with exploratory consumer behaviour and good mathematical skills) but should be informative with regard to the distinctiveness of the connection. On the 7  For example Vasu Reddy (2013) studied how two- to four-­month-­old babies anticipate being picked up (by tightening their muscles) and thus show very early signs of active participation in joint action directed at them.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 26/07/21, SPi

Attachment, Virtue, and the Second Person  189 ex­ plan­ ation proposed in this section, attachment provides experience with relational obligations and second-­ ­ personal reasons, with being bounded by another’s will—and yet also, if securely attached, with enjoying the cooperation involved in those things. Securely attached infants are not afraid of asking for help, or afraid to admit their vulnerability, nor do they avoid chances to explore and develop their agential capacities. When requests for cooperation or constraints are addressed in this safe environment, it is likely that compliance will occur more often, because securely attached children are not annoyed by their own dependence on others nor by the dependence of others on them (which shows in their request to cooperate). The more secure an attachment, the more often a child will experience the address of a second-­personal reason, and the more motivated it will be to acknowledge this type of reason in encounters with strangers too. This explanation avoids two pitfalls, which I will explain in turn. First, there is no naive assumption that love for mother expands to love for humanity. By giving second-­personal reasons centre stage, this explanation emphasizes the inter­ action, perhaps even negotiation, that is at the heart of both attachments and moral relationships. I hereby distance myself from more rosy pictures of attachment such as the one used in Axel Honneth’s foundational account of morality. Honneth holds that human beings enter the moral domain, treat each other as moral agents, strive for recognition in social integration and legal inclusion (and more besides) because they have a memory of complete fusion. For Honneth intersubjective recognition and social integration are surrogates for the state of pure unmediated togetherness that we purportedly experience in infancy. Honneth calls the infant’s experience of fusion with the mother (or primary care­ giver) ‘the zero-­point of all experiences of recognition’ (2012: 229) by which he means both that it offers the paradigm case we strive to realize, and that its in­ev­it­able loss is the motor that drives individual and social struggle for recognition. Once we are separate individuals, we can never regain the state of fusion but we strive for mediated forms of being together such as manifested in mutual recognition. Honneth’s use of attachment theory differs from mine. First, Honneth understands attachment as symbiosis and fusion, while this is a misleading description of (secure) attachment. Indeed in my view the important message that we learn in infancy is that responsive love is not the same as fusion, and that all human relationships are interactions between what is me and what is not me. Second, Honneth locates the importance of attachment in its irretrievable loss and the resultant striving for any experience that resembles it. My suggestion is not that moral motivation is a constant striving for a paradise lost. I do not believe that we strive for fusion with strangers, and I locate the experience of attachment firmly within the private sphere. Nevertheless, it is from that experience that we gain the knowledge necessary to understand what second-­personal demands are, and how they need not threaten our confidence and security.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 26/07/21, SPi

190  Katrien Schaubroeck With regard to the second pitfall, my preferred explanation can incorporate some of the explanatory power of the other explanations without losing sight of the distinctiveness of moral development. Securely attached children go and explore the world, and discover reasons, including second-­ personal reasons (namely reasons offered from an I to a you, inviting dialogue and acknowledgement of the validity of these reasons, and subsequent cooperation). Recognizing this particular kind of reasons that define the moral realm presupposes a particular way of relating to other persons that children typically learn in secure attachment relationships: they do not feel threatened by second-­ personal reasons addressed at them, they do not feel ignored when they address a second-­personal reason. On the basis of this experience they might very well form expectations and an inner working model that enables them to navigate confidently through the world, but again the important thing to explain moral capacities is not just the general feeling of confidence built into the IWM but the specific familiarity with second-­personal reasons, and the specific view of others as sources of valid demands and reasonable expectations.

4.  A Defence of the Constitutive Connection between Secure Attachment and Virtue against Two Objections To conclude I want to anticipate two objections against the proposed conceptual link between secure attachment and virtue via second-­personal competence. First one might detect a logical mistake in the reasoning, namely that I am committing the genetic fallacy. Second one might think that the explanation is too harsh on insecurely attached children, and their parents. The genetic fallacy takes place when one confuses how someone typically comes to learn a behaviour with the principles validating the behaviour. Applied to morality, the fallacy points out that how we come to know what is morally good does not necessarily show anything about what it means to be morally good, or what makes a certain action right. So the fact that we learn about morality in relational contexts does not mean that morality is relational. Compare mathematical knowledge: many children learn that two and two equals four in a loving environment but that fact does not reveal any conceptual connection between love and the nature of mathematical truths. Though I recognize the importance of avoiding this fallacy, it is hard to know how to avoid it in the case of morality. Morality is different from mathematics. Quite obviously the topic of mathematics has nothing to do with love. But moral education is nothing like lessons in mathematics. It is an open question whether moral education’s goals could be determined separately from the question what makes an action right or wrong. If one thinks that moral education should teach empathy, or care, or second-­personal competence, this is probably because one

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 26/07/21, SPi

Attachment, Virtue, and the Second Person  191 thinks that the morally good action is grounded in or is an expression of empathy, care, second-­personal competence.8 A theory of moral development, or moral education for that matter, can hardly be completely neutral with regard to ethical theories, as it needs to make certain assumptions about what matters in morality and, derived from that, about which capacities are crucial. On the hypothesis I defend, the competences that need to be developed in moral education have to do with recognizing demands or second-­personal reasons, and not with, for ex­ample, calculating harmful results. (By the way, this does not need to rule out every form of utilitarianism but it means that to be included it will have to be a version that acknowledges the importance of rights and recognition thereof.) The second objection questions whether secure attachment really is a necessary condition for moral development. Usually people learn about morality through their loving parents, as with much of what they know. But it does not follow that being loved by parents is a necessary condition, only a very common circumstance in which moral knowledge is transmitted. Moreover, these critics say, there is a strong reason against calling love a necessary condition because some children grow up without love, yet we assume that they too can and need to understand, for example, that lying is wrong. Given that about 40 per cent of the population is insecurely attached, and given that this number is fairly stable across times and cultures, does my suggestion of a conceptual tie between secure attachment and virtue ban 40 per cent of a population from moral virtue? Note that I never called secure attachment a necessary condition for virtue. At most I would call the connection between secure attachment and virtue internal or constitutive. Surely I want to allow for the possibility that moral competence can be constituted in various ways. There might be many routes to morality. Human beings are resourceful, and they can compensate for lacks of experience or capacities in various ways. So by calling secure attachment constitutive for virtue, I do not commit myself to saying that the connection establishes a necessary condition. Neither do I want to say something about the normal development of moral agency, because it would be egregious to label a portion of 40 per cent of the people abnormal. Calling the connection between secure attachment and virtue ‘constitutive’ is only meant to invoke the contrast between enabling conditions and constitutive conditions (as sometimes used in debates e.g. about the extended mind, see Gallagher  2018). Being securely attached does more than enabling 8  While my argument does entail a commitment to internalism about moral knowledge, it can remain neutral about fundamental metaphysical questions about ‘the source’ of morality. Putting it in the terms used by Wallace (2019) I do not have to choose between the relational model and the voluntaristic model of moral obligation yet. The relational model, defended by Wallace, only points towards a moral nexus between people in which the moral obligations manifest themselves. It refers to a structure of the moral domain, not to a source of moral obligations. For Wallace the relational structure of morality is perfectly compatible with moral realism. Darwall, on the other hand, is on the constructivist side of the meta-­ethical spectrum, and regards an autonomous agent’s will as the source of valid moral claims.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 26/07/21, SPi

192  Katrien Schaubroeck moral development, as does having a body. Being securely attached directly ­constitutes moral virtue because being responsive to second-­personal reasons is inherent to and definitive of both. If there is an internal connection between secure attachment and virtue, it is not a contingent fact that being securely attached helps to be virtuous. At the same time becoming virtuous without being securely attached is not impossible, though it may be harder to achieve. Insecurely attached children are not doomed by my account. Nor are their parents to be blamed. Attachment theory is often perceived as being moralistic and women-­unfriendly. Bowlby’s message was not welcomed by feminists, for example, and when he received an Honorary Degree from Cambridge University in 1977, women demonstrated in the street, apparently to the complete surprise and disbelief of Cambridge professors (recounted in Hrdy  1999). But ‘blame-­the-­mother’ is too simple a message to take from attachment theory. For one, attachment theory is a theory about infants’ needs, not about precisely who should meet those needs. The primary attachment figure can be a father; there is no reason why it should be the mother in contemporary societies where there is formula milk and even when a child is breastfed, childcare can be taken up by both parents. Second, human psychology is complicated, it is very hard to establish a clear causal path from birth to adulthood. Attachment theory should not be read as a crusade against failing parents. To the contrary, it shows that circumstances beyond a parent’s control may have a huge effect on the style of attachment. Secure attachment is not something one can choose or force, there is no point in recommending it to parents. Circumstances (both social and personal) determine what attachment style is developed, and in some circumstances insecure attachment is the more advantageous attachment style for an infant. With this conceptual analysis of connections between attachment and virtue I did not aspire to give any advice on how we should morally educate our children.9 If there is any advice in here at all, it is at most a call for social programmes that support young parents and that help to create safe, stress-­free en­ vir­ on­ ments that facilitate secure attachments between children and their ­caregivers. While it is not realistic, certainly not natural, and arguably not desirable to have securely attached children only, it remains true that secure attachment is pre­dict­ive of many good things in life, including a propensity for virtue. In this chapter I have offered an explanation for why that might be so, identifying as the

9  In this regard this chapter only partly sides with the position defended in Narvaez (2014). There is an affinity with Narvaez’s work on affective neurobiology and morality, as she offers fascinating insights in how morality develops in children starting in early caregiving. However the position defended in this chapter does not call for a cultural reform as Narvaez (2014) does. Insofar as my position assumes that the distribution of attachment styles is fairly robust throughout all populations including non-­Western societies (Morelli 2015, Hrdy 1999), it does not imply the prescriptive conclusion that we should imitate the lifestyle of hunter-­gatherer societies in order to increase the number of securely attached children.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 26/07/21, SPi

Attachment, Virtue, and the Second Person  193 linking mechanism the competence to recognize valid second-­personal reasons or interpersonal obligations and someone’s standing to address them.

References Ainsworth, M. D. S., Bell, S. M., and Stayton, D. 1974. ‘Infant–mother attachment and social development: socialization as a product of reciprocal responsiveness to signals’. In M.  P.  M.  Richards (ed.), The Integration of a Child into a Social World, 99–135. London: Cambridge University Press. Bowlby, J. 1944. ‘Forty-four juvenile thieves: their characters and home-life’. International Journal of Psycho-Analysis 25, 19–53. Darwall, S. 2006. The Second Person Standpoint: Morality, Respect, and Accountability. Cambridge, MA: Harvard University Press. Fearon, R. P., Bakermans-Kranenburg, M. J., Van IJzendoorn, M. H., Lapsley, A. M., and Roisman, G. I. 2010. ‘The significance of insecure attachment and disorganization in the development of children’s externalizing behavior: a meta-analytic study’. Child Development 81(2), 435–56. Gallagher, S. 2018. ‘The extended mind: state of the question’. Southern Journal of Philosophy 56(4), 421–47. Harcourt, E. 2013. ‘Attachment theory, character and naturalism’. In J.  Peters (ed.), Aristotelian Ethics in Contemporary Perspective, 114–29. London: Routledge. Harcourt, E. 2017. ‘Attachment, autonomy and the evaluative variety of love’. In E. Kroeker and K. Schaubroeck (eds), Love, Reason and Morality, 39–56. London: Routledge. Hobson, P. 2002. The Cradle of Thought. London: Pan Macmillan. Honneth, A. 2012. ‘Facets of the presocial self: rejoinder to Joel Whitebook’. In A. Honneth, The I in We: Studies in the Theory of Recognition. Trans. J. Ganahl. Cambridge: Polity Press. Hrdy, S. 1999. Mother Nature: Maternal Instincts and How They Shape the Human Species. New York: Ballantine Books. Johnson, S.  C., Dweck, C.  S., and Chen, F.  S. 2007. ‘Evidence for infants’ internal working models of attachment’. Psychological Science 18(6), 501–2. Koleva, S., et al. 2014. ‘The moral compass of insecurity: anxious and avoidant ­attachment predict moral judgment’. Social Psychological and Personality Science 5(2), 185–94. LaFollette, H. 1995. Relationships: Love, Identity and Morality. Oxford: Blackwell. Li, Y., Bruyneel, S., and Warlop, L. 2012. ‘Growing with love: priming attachment security enhances exploratory consumer behaviors’. Advances in Consumer Research 40, 759–60. Mikulincer, M., and Shaver, P. R. 2007. ‘Reflections on security dynamics: core constructs, psychological mechanisms, relational contexts, and the need for an integrative theory’. Psychological Inquiry 18, 197–209.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 26/07/21, SPi

194  Katrien Schaubroeck Mikulincer, M., Shaver, P. R., et al. 2005. ‘Attachment, caregiving, and altruism: boosting attachment security increases compassion and helping’. Journal of Personality and Social Psychology 89, 817–39. Morelli, G. 2015. ‘The evolution of attachment theory and cultures of human attachment in infancy and early childhood’. In L. A. Jensen (ed.), The Oxford Handbook of Human Development and Culture: An Interdisciplinary Perspective, 149–74. Oxford: Oxford University Press. Narvaez, D. 2014. Neurobiology and the Development of Human Morality: Evolution, Culture and Wisdom. New York: Norton. Narvaez, D., Wang, L., and Cheng, Y. 2016. ‘The evolved developmental niche in childhood: relation to adult psychopathology and morality’. Applied Developmental Science 20(4), 294–309. Reddy, V., et al. 2013. ‘Anticipatory adjustments to being picked up in infancy’. PLOS One 8(6), 1–9. Shaver, P. R., and Mikulincer, Mario. 2012. ‘An attachment perspective on morality: strengthening authentic forms of moral decision making’. In  M.  Mikulincer and P. R. Shaver (eds), The Social Psychology of Morality: Exploring the Causes of Good and Evil, 257–74. Washington DC: American Psychological Association. Shaver, P. R., Mikulincer, M., Gross, J. T., Stern, J. A., and Cassidy, J. 2016. ‘A lifespan perspective on attachment and care for others: empathy, altruism, and prosocial behavior’. In J. Cassidy and P. R. Shaver (eds), Handbook of Attachment (3rd edn), 878–916. New York: Guilford. Simpson, J. A., and Belsky, J. 2016. ‘Attachment theory within a modern evolutionary framework’. In J. Cassidy and P. R. Shaver (eds) Handbook of Attachment: Theory, Research and Clinical Applications (3rd edn), 91–116. New York/London: The Guilford Press. Sroufe, L.  A. 2005. ‘Attachment and development: a prospective, longitudinal study from birth to adulthood’. Attachment and Human Development 7(4), 349–67. Thompson, R. A. 2016. ‘Early attachment and later development: reframing the questions’. In J. Cassidy and P. R. Shaver (eds), Handbook of Attachment: Theory, Research and Clinical Applications (3rd edn), 330–48. New York/London: The Guilford Press. Van IJzendoorn, M. H., and Bakermans-Kranenburg, M. J. 2014. ‘Prosocial development and situational morality: neurobiological, parental, and contextual factors’. In J.  F.  Leckman, C.  Panter-Brick, and R.  Salah (eds), Pathways to Peace: The Transformative Power of Children and Families, 161–84. Cambridge, MA: MIT Press. Wallace, R. J. 2007. ‘Reasons, relations, and commands: reflections on Darwall’. Ethics 118(1), 24–36. Wallace, R.  J. 2012. ‘Duties of love’, Proceedings of the Aristotelian Society Supplementary Volume 86, 175–98. Wallace, R. J. 2019. The Moral Nexus. Princeton: Princeton University Press.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

10

An Aetiology of Recognition Empathy, Attachment, and Moral Competence A. E. Denham

1.  Introduction No father had watched my infant days, no mother had blessed me with smiles and caresses; or if they had, all my past life was now a blot, a blind vacancy in which I distinguished nothing. From my earli­est remembrance I had been as I then was . . . What was I? The creature (unnamed), in Frankenstein; or the Modern Prometheus (Shelley 1818/1969: 128) Secure attachment is associated with many traditional moral virtues, and ­compromised or absent attachment with various moral vices. What explains the interactions of attachment and moral development? This chapter explores the suggestion that early attachment underpins the human capacity for empathy, and that empathy, in turn, is a condition of moral competence. I begin with a fictional allegory for attachment, or rather for certain consequences of its absence: Mary Shelley’s Frankenstein.1 Shelley’s novel relates the creation, abandonment, development, dissolution, and tragic death of a being who, by almost any philosophical standard, is a person. Dr Frankenstein’s unnamed, monstrous invention—I shall call him ‘Creature’—is far from the mindless predator of popular imagination: he is a thinking, feeling, self-­conscious being endowed with both reason and affection. Despite these qualities, Creature’s repugnant appearance and fearful physical power provoke his creator to flee in horror, abandoning his ‘newborn’—then still developmentally, if not physically, an infant. Creature flees into the surrounding woodlands, spending his early years foraging and fending for physical survival. Later, he finds a kind of home, a hovel that adjoins the cottage of a close-­knit, harmonious family. A crack in the wall allows Creature secretly to observe and learn the language, the literature, the 1  I am indebted to Connie Rosati for recognizing the relevance of Shelley’s novel to themes in moral development. C.  Rosati, ‘Autonomy & Personal Good: Lessons from Frankenstein’s Monster’, unpublished MS. A. E. Denham, An Aetiology of Recognition: Empathy, Attachment, and Moral Competence In: Attachment and Character: Attachment Theory, Ethics, and the Developmental Psychology of Vice and Virtue. Edited by: Edward Harcourt, Oxford University Press. © A. E. Denham 2021. DOI: 10.1093/oso/9780192898128.003.0011

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

196  A. E. Denham music, and the domestic intimacy of ordinary humanity. Eventually reading and poring over the classics, he develops an appreciation of human morality, and a susceptibility to moral emotions such as pride, shame, and indignation. Above all, he understands what it is for the cottagers to ‘love and sympathise with one another’ and yearns to be ‘known and loved’ in his turn (Shelley 1818/1969). In these ways, Creature comes to understand—from without—the value of love and human connection, and to feel the bitter sorrow of solitude. He admires and envies the cottagers, and eventually makes a bid for acceptance into their family. This bid is met, as was inevitable, with rejection and terror, provoking in Creature a series of impulsive and violent acts of vengeance. As he puts it, ‘Evil thenceforth became [his] good’, and he begins his course of revenge on Dr Frankenstein, murdering his creator’s own close attachments one by one: his brother, his adoptive sister, his closest friend, and his lifelong love and bride. These subsequent rages and rampages, however, are less the expression of an inherently hostile nature than of a desire for revenge against a world—and especially a rejecting ‘parent’— which brought him into existence, only to deny him the affection and intimacy which his nearly human nature, like that of any person, required. At the close of the tale, Creature’s victory in destroying Frankenstein is worse than hollow: he is not only crushed by grief but by regret, crying out that ‘a frightful selfishness hurried me on, while my heart was poisoned with remorse . . . My heart was fashioned to be susceptible of love and sympathy . . .’ (Shelley 1818/1969). Frankenstein is a work of phantasy and fiction. But it is also an insightful ­illustration of an important psychological fact: beings such as us are disposed by our natures to seek intimacy with our human conspecifics, and its absence in our early years threatens not only our happiness, but mastery of our impulses, our emotions, and our moral characters. Attachment theory tells us more specifically that, like Creature, human beings are disposed from birth to seek one particular kind of intimacy: the proximity of a protective caregiver, typically a parent. The proximity sought is more than physical nearness; it is the caregiver’s emotional and cognitive responsiveness, reliability, and comfort. In its wider sense, ‘attachment’ refers to an enduring, intimate emotional bond that develops between two or more persons, normally through sustained personal contact, yielding a felt need for personal contact and conditioning the attached person’s sense of security and safety. In the context of child development, ‘attachment’ refers more particularly to this bond as it holds between an infant or toddler and his primary care­ giver—a connection that is instrumental in the child’s cognitive, affective, and social development (Bowlby  1969; Ainsworth et al. 1978; Fonagy et al. 1991). Securely attached children manifest behaviours consistent with a trusting, affectionate intimacy with their caregiver; the world of the securely attached child—at least the world within the orbit of that intimate relationship—is a fundamentally safe one in which threats will be diverted, needs will be met, and experiences shared. It is thus unsurprising that secure attachment early in life is longitudinally

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

An Aetiology of Recognition  197 associated with a range of characteristics that favour the creation and main­ten­ ance of intimate, personal relationships: trust, confidence, optimism, receptivity, and openness. The securely attached child learns that, whatever perils the world may hold, his well-­being is shielded within the private sphere of personal intimacy. It is less obvious, however, why secure attachment should also favour recognition of moral obligations, particularly towards those with whom we have no special standing and share no personal destiny—recognition that the claims of persons as such merit our attention and regard. Yet it does: secure attachment is developmentally associated with a wide range of traditional moral virtues such as reciprocity, honesty, and benevolence, where these are extended impersonally to strangers as much as friends, and even granted to our enemies. In short, secure attachment confers a sensitivity not only to the imperatives of personal intimacy, but to the wider imperatives of morality requiring moral recognition of persons generally. Let us call these ‘person-­regarding’ requirements. I will assume without argument that responsiveness to such requirements is a central part (although by no means the whole) of basic moral competence. Why should secure attachment promote sensitivity to requirements of this kind? One answer to this question looks beyond the fact of secure attachment to a further psychological capacity: our capacity for empathy with our conspecifics. Empathy has been conceived in many ways by many theorists, but nearly all delineate it with reference to a spontaneous sharing of affect, perceptual focus, and motivational direction. These are likewise hallmarks of what attachment theorists sometimes call the ‘reflective function’ (or ‘mentalizing’) dynamic of sensitive caregiving, whereby a harmonious, interpersonal synchrony is manifested in the carer’s (typically the mother’s) verbal and non-­verbal interactions with her child. These interactions provide one important foundation for the development of empathy in later life. For instance, successful synchronization and secure attachment strongly predict mature empathic responsiveness (Kestenbaum, Farber and Sroufe 1989), with mother–infant synchrony measures in the first year of infancy being directly associated with empathy levels at ages 6 and 16 (Feldman 2007). Sensitive, caregiver mentalizing is also a powerful predictor of optimal development in respect of a range of other morally relevant capacities, including cooperativeness, self-­regulation (including gratification deferral), and the ability to reliably identify, predict, and render intelligible others’ cognitive and affective states (Fonagy and Target 1997; Feeney et al. 2008). These same capacities are, in turn, both causally and constitutively related to altruistic motivation, and moral motivation of other kinds. Indeed, in one study directly examining the development of moral conscience it was found that the degree of mutually responsive orientation between an infant and caregiver, especially of positive affective states, was directly correlated both with higher empathic resonance at 22  months and with greater guilt awareness at 45 months (Zahn-­Waxler et al. 1992; Knafo et al. 2008).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

198  A. E. Denham So one reason that secure attachment is associated with person-­regarding morality might be this: secure attachment promotes susceptibility to empathy, and an appropriate susceptibility to empathy is a condition of basic moral competence. Put differently, secure attachment may support moral virtue by way of promoting empathy as one of its necessary conditions (Baron-­Cohen 2009). In what follows, I assess this proposal. I turn first in section 2 to the idea that empathy is necessary for morality, noting some of its history, and narrowing its legitimate scope to one component of basic moral competence—person-­regarding (including altruistic) moral norms. Section 3 then explores and distinguishes the different dimensions of empathy. Section 4 presents a sceptical argument against one version of the claim, namely, that empathy plays an essential synchronic role in moral judgement. Section 5 examines a developmental, diachronic version of the claim, focusing first on the evidence from psychopathology. I then return to attachment theory, proposing that the contributions of empathy to moral competence have their origins in early attachment. It is the dynamic, reciprocal mirroring between caregiver and child, I argue, which initiates our recognition of the reality and value of other persons.

2.  Empathy: The Indispensability Thesis Consider again the misfortunes of Frankenstein’s Creature. In his early development, his intuitive grasp of moral claims is revealed, in part, by his susceptibility to moral emotions—he recognizes his impulsive wrongdoings as such, and is subject to remorse and shame on their account. At this stage Creature also enjoys a natural and spontaneous empathic responsiveness to others; he is even moved to act altruistically on several occasions—for instance, by gathering wood in the night to aid the adored, hard-­labouring family of cottagers. Following his failed bid for their acceptance and friendship, however, he becomes consumed by a reactive rage—an ‘insatiable thirst for vengeance’—and embarks on his murderous course (Shelley 1818/1969). Thereafter, he seldom pauses over others’ needs and interests; to the contrary, he seems almost wilfully to de-­sensitize himself, and is undeterred by the fear and distress of his victims. Insofar as he concerns himself with Dr Frankenstein’s desires, for instance, that is only to ensure their bitter disappointment. Creature seems committed to casting aside not only shame and remorse, but his aversion to others’ suffering. His susceptibility to empathy appears to be extinguished, and along with it all regulation by moral requirements: when ‘evil becomes his good’, regard for others’ ends (almost) evaporates. Shelley’s tale of Creature’s moral dissolution thus echoes a familiar platitude of folk psychology: that empathy plays an important role in moral virtue. This platitude has also enjoyed a long and distinguished history in philosophical theory.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

An Aetiology of Recognition  199 Early British sentimentalists accorded to empathy (or to ‘sympathy’, as it was then labelled) a central role; Hume’s premise that ‘the minds of men are mirrors to one another’s’ lay at the heart of his aetiology of the ‘moral distinctions’ and their ability to move us to action. (Hume 1739/1978: 365) Adam Smith followed Hume, locating the affective power of moral claims in our natural propensity to reflect one another’s behaviours and inner lives (Smith 1759/2002).2 In the second half of the twentieth century, however, mainstream analytic philosophers largely abandoned empathy and its cognates, notwithstanding its close association with prominent notions such as universalizability, interpersonal cognition/other minds, and internal reasons. With a few notable exceptions, post-­war analytic philosophy regarded empathy with suspicion, as an ill-­defined, psychological construct that had no place in reasoned moral justification and motivation. Recent decades have seen a resurgence of interest in the role of empathy in moral motivation. A principal catalyst for this has been Daniel Batson’s landmark studies of moral motivation in the 1980s and 1990s (Batson 2011; Batson 2012). These studies put to the test what Batson called the ‘egoistic hypothesis’—the claim that the ultimate goal of all human action is to promote the agent’s own welfare. The competing hypothesis was that, in certain facilitating conditions, agents’ choices and actions can be altruistically motivated—motivated directly by a non-­instrumental or ultimate desire to benefit another, even when doing so incurs personal costs. Many candidate conditions might facilitate such mo­tiv­ ation, but the one on which Batson chose to focus was empathic concern, which he understood as involving ‘vicarious other-­focused emotions, including feelings of sympathy, compassion, tenderness and the like’ (Batson  1991:113). Batson’s studies explored the effect of empathic induction on subjects’ preparedness to respond altruistically to others, both in attitude and in action choices, using experimental designs that controlled for egoistic motives of reward seeking, punishment avoidance, and relief from aversive arousal. While his findings have met with many challenges, they are widely regarded as lending support to the ‘empathy-­altruism’ hypothesis—the claim that as ‘empathic feeling for a person in need increases, altruistic motivation to have that person’s need relieved increases’ (Batson 1991: 72). Batson’s claim that empathy evokes altruistic motivation harmonizes well with the common assumption that empathy moves us to do the right thing, and is a force for the (moral) good. Batson’s studies also lend support to the familiar intuition that empathy competes against the two forces most hostile to morality: indifference and self-­interest. Empathy competes with indifference in its epistemic role, by alerting us to 2  Parallel themes were mooted in German moral philosophy and aesthetics in the 1700s, and versions of the empathy construct remained prominent in continental accounts of moral motivation through the nineteenth and early twentieth centuries (Schiller 1794/1967; Schopenhauer 1840/1995; Lipps 1903; Scheler 1923/1954; Husserl 1931/1988).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

200  A. E. Denham circumstances that demand moral attention; in its motivational role, it serves as a corrective to our default position of egocentrically pursuing our own ends, and only our own ends. The reasoning behind these claims is straightforward. Other-­ regarding moral requirements often enjoin actions that require an understanding of others’ interests, and those interests often conflict with our concern for our own welfare. If we are to act competently as moral agents, then, indifference must be counteracted by a sound epistemic source, and self-­interest must be counteracted by a powerful motivating force. In our species, empathy both offers the right sort of informational resource and delivers motivational force. Hence empathy is, in such cases, indispensable to moral competence. This reasoning is plausible so far as it goes. Nonetheless, any identification of empathy and moral competence tout court would clearly be a mistake: countless moral requirements do not directly concern personal welfare at all, and enjoy no direct connection with empathy. Among these empathy-­irrelevant norms are various sexual, dietary, and hygiene prohibitions, norms deriving from religious commandments, and norms based on conceptions of social honour and prestige. Perhaps an evolutionary story can be told according to which these require empathic concern for human welfare, but at the level of individual motivation it is neither here nor there. Empathic responsiveness to human weal and woe will not dissuade a man from acts of necrophilia, nor keep him Kosher, nor prompt him patriotically to fall on his sword to honour his nation’s flag. These exceptions acknowledged, considerations of other persons’ interests still justify a central and ubiquitous core of moral requirements. Person-­regarding norms prescribe actions that are ‘prosocial’ in the sense that they direct the agent to protect or promote the interests of another person or persons. Among these are certain harm norms (prohibitions against harming persons and their property) as well as norms reflecting Aristotelian and Humean natural virtues, such as friendship, kindness, generosity, compassion, and loyalty. There is good reason to suppose that a sine qua non of respecting such norms is a propensity to respect and be moved by other people, and that doing this requires susceptibility to empathy. I will refer to the idea that empathy is a necessary condition of moral competence as the Empathy Indispensability Thesis (EIT). The EIT is ambiguous as between two claims. First, it may be taken as a claim about the epistemic and motivational contributions of immediate, occurrent synchronic empathy to token moral judgements. This is the claim that, necessarily, whenever one essays a person-­ regarding judgement, that judgement is in some way informed by empathic responsiveness. Secondly, it may be read as making a diachronic claim about the developmental contributions of empathy to our basic competence to essay person-­regarding moral judgements, such that a susceptibility to empathy is one of its necessary conditions. Is the EIT true, in either version? Before pursuing that question, let us first step back and consider more carefully the nature and varieties of empathy itself.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

An Aetiology of Recognition  201

3.  Dimensions of Empathy: Mindreading, Resonance, Attunement, Distress, and Concern Empathy is not an emotion, but a way of identifying and representing emotions and other affective states. I will use the terms ‘empathy’ and ‘affective empathy’ interchangeably; when we empathize, affective states are our objects of thought. In attachment theory, the term ‘empathy’ is often used more broadly, and Peter Fonagy identifies it as one among several functions constituting the more general capacity of ‘mentalising’—the impulse to understand and imagine both our own and other people’s thoughts (Fonagy et al. 2014). Mentalizing, as Fonagy uses the term, includes exercises in cognitive mindreading or perspective-­taking as well as affective responsiveness (Fonagy et al. 2014: 36). However, it is generally recognized that mindreading and affective empathy are distinguishable capacities: a plethora of experimental evidence testifies to this at both the functional and neuro­ physio­ logic­ al levels (Decety et al.  2013; Blair 2006; Smith 2006).3 ‘Mindreading’ refers to a capacity reliably to identify others’ action-­explaining intentional states—typically their beliefs, desires, and intentions. It is an ability accurately to represent the propositional attitudes that render actions intelligible, and to exercise these representations in explaining and predicting others’ behaviour. Affective empathy can also represent propositional attitudes, but it does so by a different mechanism and in a different mode. Jean Decety refers to affective empathy as empathy ‘proper’, and defines it as ‘a construct broadly reflecting a natural capacity to share and understand the affective states of others, comprising emotional, cognitive, and motivational facets’ (Decety et al. 2013). This requires that the empathizer not only represent, but also share in another’s target states: affective empathy is an experiential as well as a representational capacity. When we empathize, we do not only identify and individuate another’s affective/mo­tiv­ ation­al states (emotions, sensations, aversions, etc.) but do so by instantiating some of their first-­personal experiential character. The distinction between first-­ personal and other-­personal representations of experiential states is key to empathy’s motivating force: a solely conceptual or propositional representation of, for instance, another’s pain or pleasure, however detailed and accurate, does not constitute affective empathy, and indeed requires no affective or motivational engagement whatever. An empathic representation, by contrast, is what I elsewhere have called a ‘subjective conception’—a conception as from the first-­ personal perspective of the experiencing subject (Denham 2000; Denham 2012). 3  This distinction is not only a conceptual one; it is underwritten by the reliance of each on distinct neurological bases. As Luyten and Fonagy note, ‘there is increasing evidence that distinct, albeit to some extent overlapping, neurocognitive systems are involved in these capacity. . . . While cognitively oriented mentalization depends on several areas in the prefrontal cortex, affectively oriented mentalizing seems to be particularly dependent on the ventromedial prefrontal cortex’. Luyten and Fonagy 2014: 102)

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

202  A. E. Denham If one represents another’s pain by way of affective empathy, one’s own experience must feature some of the target state’s phenomenology—its qualitative and mo­tiv­ ation­al characteristics. To some degree, it is itself painful. So described, affective empathy is not yet a capacity for the solicitous concern that matters to moral motivation. To get there from here, we must progress four different dimensions of affective empathy: empathic resonance, empathic attunement, empathic distress, and empathic concern. I will briefly sketch each in turn. Empathic Resonance Infants famously mimic the facial musculature of their caregiver’s expressions, probably from only a few hours after birth (Hoffman 2000). Such motor mimicry is (a) reflexive and (b) non-­referential: the mimicking subject does not exercise voluntary control over his motor state, and nor is he typically consciously aware of its occurrence. Nonetheless, motor empathy arguably plays an important role in the development of affective empathy and interpersonal emotion regulation in the first few months of life; at the neurological level, the causal pathways between motor and affective responses are bi-­ directional (Hoffman 2008). Motor mimicry persists throughout our lives, and is an early and basic form of what I call ‘empathic resonance’—an innate capacity to reflect some features of the behaviour (especially facial expressions) and experiential states (especially the affective states) of others.4 Empathic resonance is automatic, non-­verbal, and, in Fonagy’s terms, ‘external’ in that it is cued by observation of perceptible conditions such as posture and facial expression. Hoffman observes that resonance (or ‘emotional contagion’) is ‘passive, involuntary, and based on surface cues; it requires little cognitive processing or awareness that the source is [someone else]’ (Hoffman 2008: 441). Nonetheless, resonance is an important early component of mentalizing, and can be profoundly psy­cho­ logic­al­ly efficacious: it is a psychophysical process, its effects often are consciously experienced, and it possesses a first-­personal phenomenology. But it is not yet a representational state, save in the attenuated sense of representing the resonating subject’s own condition. It serves no interpersonal, referential function. Empathic attunement Most developmental psychologists, including contemporary attachment theorists, regard empathic resonance as a developmental precursor to a second, cognitively more complex dimension of affective empathy: empathic attunement. Attunement occurs when (a) a subject conceives of (represents in thought) another’s experiential state, the conception being typically elicit­ed by observing or remembering or imagining the other; (b) via resonance, the subject’s occurrent state reflects (some constituents of) the content and phenomenological character of the target experience (or what he takes that

4  Resonance is vividly illustrated by Hume’s analogy between our responses to one another’s sentiments and the sympathetic vibrating of strings on a violin: when one string is plucked or bowed, it directly causes a vibration in the others (Hume 1739/1978).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

An Aetiology of Recognition  203 experience to be), and (c) the subject regards his reflective states as referring to and informing him of the other’s experience (Vignemont and Singer 2006: 435). This last feature registers that attunement constitutes a first-­personal conception of another’s emotions and other affective/motivational states as belonging to another subject of experience: the agent regards it as representing the content and character of the other’s inner life. Attunement is what theorists often mean when they use the word ‘empathy’.5 Attunement is essentially referential, and where the referent states are aversive ones such as fear, sadness, or other kinds of distress, attunement presents the agent with a motive for two further responses: empathic distress or empathic concern. Empathic distress names a familiar development of empathic attunement. (Batson terms it ‘personal distress’ (2011)). When empathic attunement is per­sist­ ent and intense, the empathizer can become ‘empathically over-­ aroused’ (Hoffman  2008): his focus of attention and his dominant motivation is then to relieve his own distress. In empathic distress, a subject (a) encounters another’s aversive state, typically by directly perceiving or imaginatively engaging with him, (b) empathically attunes to that aversive state, recognizing the other as its source and referent, and (c) incurs a self-­focused motivation to remove the aversive stimu­lus (the target subject’s distress) from his perceptual and/or cognitive en­vir­ on­ment—for instance, to abandon the victim or to pursue attentional diversions (Hoffman  2008). Empathic distress is thus an ‘egocentric’ motivational state in Batson’s sense of that term, which sometimes conflicts with our moral convictions—as when we guiltily bin the charity circular with its images of starving children or change the television channel to avoid scenes of desperate refugees. Empathic concern When attunement is manifested as empathic distress, it is negatively correlated with moral motivation—the opposite of the prosocial influence with which empathy is typically associated. Attunement must develop via a different transformation, as empathic concern, if it is to be recruited into the service of morality. Empathic concern is closely allied with Hume’s notion of ben­ evo­ lence—a non-­ instrumental desire to promote the welfare of another. A benevolent desire may, of course, arise by way of various causal trajectories, and not all are empathic; I will discuss one alternative shortly (Nichols  2004). As I  (stipulatively) use ‘empathic concern’ here, it names a species of the genus of concern, namely, concern that is a development from and conceptual elaboration of (empathic, affective) attunement: the former occurs contiguously or concurrently with the latter, and its content is informed by it. Empathic concern is thus distinguished from other modes of concerned attention by having resonance and attunement as constituents as well as causal conditions. Resonance and attunement do not just precede empathic concern, but contribute to its content and felt 5  Jesse Prinz, for example, says that the ‘core idea of empathy’ is that it is ‘a kind of vicarious emotion: it’s feeling what one takes another person to be feeling’ (Prinz 2011a: 212).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

204  A. E. Denham character, in part determining its valence, intensity, tone, and motivational force. If John responds to Sally’s painful toothache with empathic concern, he then is already in an internal state that refers to Sally, is aversive and negatively valenced, relatively intense, and has a particular attentional focus and motivational direction.

4.  The Synchronic Empathy Indispensability Thesis One interpretation of the EIT is that synchronous empathic arousal is necessary for an agent’s susceptibility to other-­regarding considerations—that the contribution of empathy occurs synchronously in our other-­regarding moral judgements. Is the synchronic version of the EIT true? In particular, is synchronous empathy necessary for recognition of the moral claims of persons as such? Let us first consider the positive evidence delivered by Batson’s studies of other-­ regarding, altruistic judgements. Batson’s initial experiments showed that subjects who are primed to empathize with victims are more strongly motivated to help them; ‘high empathy’ subjects are altruistically motivated even when helping comes at a significant cost to personal interests (Batson  2011). In later studies using the same basic design, Batson found further that empathy priming led subjects to act more altruistically even when (a) the helping was anonymous and offered no personal credit (thus challenging reward incentives), (b) there were good reasons to avoid helping, making helping demanding and not helping justified (challenging anticipated guilt incentives), (c) subjects were advised that they would receive no feedback on their assistance (challenging incentives of praise/ victim’s gratitude), and (d) when refusing to help promised a positive experience on par with that of helping (challenging anticipated pleasure incentives) (Batson 2011: Appendices B, C, D, F, G). These results are not conclusive, but they strongly suggest that empathy can promote attitudes and behaviour that are better explained by altruistic rather than egoistic motivation, at least in a context of heightened, targeted empathy induction. This is encouraging news for the friend of the synchronic EIT, insofar as altruistic judgements may be considered a central instance of person-­regarding ones. Numerous other studies have found strong correlations between empathy and other-­regarding actions and attitudes, including ‘impersonal’ person-­regarding recognition. Konrath’s and Grynberg’s extensive survey of the literature identifies a number of results supporting the claim that empathy promotes prosocial mo­tiv­ ation. To mention only a few: • For both attunement and empathic concern, and regardless of how these were measured (i.e. observer-­reports, self-­reports, self-­reported vicarious emotion, or targeted situational induction), empathy is positively associated

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

An Aetiology of Recognition  205 with prosocial behaviours towards strangers (sharing, assisting, giving) (Eisenberg and Miller 1987) • Empathy induction increases interpersonal cooperation, even in Prisoner’s Dilemma games in which the subjects know that their game partner has defected. In one study, situational empathy induction increased cooperation rates from 5 per cent (control) to 45 per cent in a one-­time play. (Batson and Ahmad 2001). • Empathy induction has been shown to improve outcomes in negotiations between parties with competing goals, producing better outcomes on both sides relative to controls (Galinsky et al. 2008) • Parents who rank high in empathy (on both self-­reports and observer reports) have more positive and effective interactions with their children. As Konrath and Grynberg note, this is unsurprising if, as the aforementioned studies suggest, ‘empathizing makes people kinder and more cooperative’ (Konrath and Grynberg 2013: 2). • In professional settings, higher empathy ratings by those in helping, ‘prosocial’ vocations (doctors, nurses, teachers, and therapists) were correlated with better performance and better outcomes for patients and students (Coffman 1981; Waxman 1983). Surveying the evidence, it is difficult to resist the conclusion that at least some dimension of empathy is causally efficacious in driving our recognition of persons as a source of moral claims. Hoffman is even more confident, asserting that there is ‘overwhelming evidence that people who feel empathically distressed at another’s misfortune are more motivated to help, that empathic distress makes people help more quickly, and that people who are empathically responsive to another’s distress feel better when they help than when they don’t’ (Hoffman 2000: 441). What more could the friend of the EIT require? To conclude in favour of the EIT on these grounds, however, is too hasty. First, many of the associations noted are merely correlational, and do not establish that empathy is the horse rather than the cart. This is not, in fact, a serious worry in every case: sometimes other considerations such as the order in which stimuli are presented (as in Batson’s studies) make the causal claims compelling. But there are other methodological worries as well, including inconsistencies in how ‘em­pathy’ is conceptualized. Some conceptualizations, for instance, include personal distress as an indicator of empathy, whereas others exclude it; again, some take perspective-­taking or cognitive mindreading as constituents of empathy and ­others do not. Measurement procedures are also inconsistent. Some studies rely solely on self-­report, which is notoriously unreliable for subjects who are independently invested in an empathic self-­conception. Others use observer reports, and still others assess autonomic, physiological correlates of affective arousal. Why should we suppose that all of these are measuring the same conditions?

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

206  A. E. Denham To further complicate matters, some studies target dispositional or trait empathy, while others assess empathy as aroused in a particular situation (situational empathy). Finally, and most problematic of all for the synchronic EIT, even very high positive correlations between empathy and indicators of moral competence can only show, at best, that empathy facilitates moral recognition of other persons, not that it is an indispensable condition of it. Moreover, empathy might inform and motivate other-­regarding, helping behaviour in ‘up close and personal circumstances’, while doing nothing to promote our recognition of the moral claims of persons who are more remote in space or time. For these and other reasons, many have argued that, the experimental evidence notwithstanding, the EIT is a non-­starter (Goldie 2011; Maibom 2014; Prinz 2011a, 2011b). Jesse Prinz is perhaps its most vociferous critic. He holds that empathy makes no indispensable (or even desirable) contribution to moral competence, arguing that it is neither constitutively, causally, epistemically, developmentally, nor motivationally necessary. A first objection is that many norms fail even to be candidates for empathic motivation. Sometimes this is because empathic concern directly recommends against them. Empathy—pace our usual norms and intuitions– would most likely recommend, for instance, that we steal from the rich to give to the poor, and that we refuse to punish transgressors. Prinz’s own list of ‘empathy indifferent’ norms includes crimes against oneself, offences against groups, ­victimless transgressions (bestiality, consensual incest), and moral judgements at a high level of abstraction (‘Tax evasion violates the obligations of citizenship’). The objection is well taken so far as it goes, but it does not go very far if the EIT concerns only moral judgements directly justified by recognition of persons’ interests, expressive of person-­directed virtues such as kindness, generosity, compassion, pity, fidelity, and forgivingness. Could we really find such judgements compelling without any kind of empathic responsiveness to others’ wants and needs? Prinz insists that we could, arguing that empathic concern fails to provide the best explanation of moral motivation, even in this restricted class. His argument relies on his particular meta-­ethical commitments, which are both internalist and sentimentalist. In brief, Prinz holds that moral judgements are intrinsically mo­tiv­at­ing because they have an emotional basis or ‘contain’ emotions, as he sometimes puts it. The emotions they contain may be negatively valenced (disapprobative) responses such as anger, disgust, guilt, and shame, or positively valenced (approbative) ones such as gratitude, admiration, or pride. A token moral judgement is in part constituted by such emotions, and that is why it is intrinsically motivating (Prinz  2011a: 219). This fact, Prinz argues, already de­ livers everything we need to explain why other-­ regarding considerations mo­tiv­ate us; empathy is simply surplus to requirements. There is no explanatory gap in the motivational story for it to fill.6 6  Might not empathy nonetheless be our most effective and reliable source of motivation, even if it is not a necessary one? Against this suggestion, Prinz adduces findings indicating that the emotions

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

An Aetiology of Recognition  207 Whether or not we find Prinz’s alternative account convincing, the synchronic EIT will fare badly in the court of everyday moral experience if our person-­ regarding judgements are ever motivating in the absence of occurrent empathy. And, in fact, this often happens: there are many modes and manifestations of interpersonal concern apart from occurrent empathic concern. Consider my standing disposition to help my child—to relieve his distress and to confer benefits such as affection, a secure home, an education. Fortunately for him, this disposition does not depend on my occurrent, empathic attunement; he is fed and comforted even when I am distracted or weary. While I am not then motivated by empathy, I still act from the ultimate goal of promoting his welfare. Similarly, person-­regarding attitudes may arise derivatively out of a social role with which one is identified. An overworked nurse suffering from compassion fatigue and long past empathic attunement can continue to be motivated by her commitment to caring for her patients; a committed Humanist may serve the homeless, even when he ceases to be animated by empathy for them. In both cases, their (non-­ empathic) concern may even see them through occasional episodes of irritation or distaste or revulsion. That gives us no reason to deny that their ultimate goal is to relieve the plight of those in need. No doubt some occurrent regard for ­others—some standing disposition to promote their welfare and avert their suffering—plays a role in recognizing our moral obligations to them. But this need not be empathic concern. That is, it need not be concern that is either caused or informed by empathic attunement of any kind, because alternative routes to moral judgement are available. First, deontological intuitions can suffice to put even the most hard-­hearted and avaricious shopkeeper—say, Kant’s famous shopkeeper—off cheating his customers, and prevent the most cool-­headed surgeon from engaging in random organ donation from an innocent patient. No empathy required. Secondly, Golden Rule considerations such as those in play behind Rawls’s veil of ignorance will suffice to rule slavery and ethnic discrimination off constitutive of moral judgements are also more powerful motivators than empathy: anger, disgust, happiness, and shame all, he claims, yield stronger effects than empathy (Prinz 2011a: 218–20). For example, he cites one study as showing ‘no correlation in children between empathy and pro-­social behaviour’ (Underwood and Moore  1982), another indicating only a modest correlation in adults between prosocial behaviour and shared sadness (Eisenberg et al. 1989), a meta-­analysis showing that empathy is only ‘weakly correlated’ with prosocial behaviour (Neuberg et al. 1997), and claims that in studies using economic games ‘empathy does not motivate moral behavior when there are significant costs’ (Fehr and Gachter  2002). This is puzzling data for the punter immersed in the experimental evidence adduced at the beginning of this section. What of Hoffman’s ‘overwhelming evidence’ for the very correlations Prinz denies? The puzzle disappears when one inspects how Prinz conceptualizes empathy: he elucidates it solely in terms of ‘shared’ or ‘vicarious’ affect (effectively, just resonance and attunement), excluding empathic concern. That is, the handful of studies on which he relies attribute empathy only to subjects who directly evidence affect-­sharing independently of concern. The Eisenberg study, for instance, distinguished displays of ‘concerned attention’ (e.g. a child wrinkling her brow) from displays of ‘shared emotion’ (direct mimicry of the target’s sadness); only the latter counts as manifesting empathy. This makes all the difference, for it assumes that subjects’ responses of sympathy or concern are not empathically driven. But this is almost to assume what Prinz aims to prove. Indeed, as Prinz judiciously acknowledges in a footnote, Batson’s ‘notion of empathic concern may be immune to many of the worries raised here’ (Prinz 2011a: 229).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

208  A. E. Denham the moral pitch. Again, no empathy required. Thirdly, old-­fashioned emotional conditioning and habituation can be enough to elicit our approbation for all manner of moral imperatives—it can prevent a jealous sibling from pinching his baby sister, it can keep a soldier at his post under fire, and it can refocus the wandering eye of a husband trapped in a stale marriage. Token, person-­regarding judgements require no here-­and-­now empathic input; while recognition of and respect for others may be indispensable, these often move us independently of any present empathic engagement. While empathy may play a role in motivating recognition of persons’ moral claims, it often is surplus to requirements. Even more worryingly, in everyday experience our empathic responses can be capricious and double-­edged, undermining the very sort of impartiality on which recognition of persons depends. The workings of empathy are often difficult to control, unreliable, and fleeting. Primo Levi accuses our empathic impulses of being insensitive to reason, arguing that they ‘elude all logic’. There is no proportion, he points out, ‘between the pity we feel and the extent of the pain by which the pity is aroused: a single Anne Frank excites more emotion than the myriads who suffered as she did but whose image has remained in the shadows’ (Levi 1988: 56). Levi’s scepticism about the contributions of empathy to moral conduct is at least partly borne out by its role in countless everyday, moral failings. Empathy may not only fail to move us to the right judgements about the right targets in the right measure and at the right times; it can move us to knowingly transgress. How many of us have lied to a friend about a sensitive issue (when they really needed to hear the truth) because we could not bear to witness their discomfort? What parent has not been tempted to assist his child a bit too actively with a challenging homework problem, or arts competition or job application, temporarily relegating considerations of authenticity and fairness to a lower shelf? How often have any of us picked up the pieces and covered for a feckless colleague, who by rights should have been held properly to account for some misdemeanour? Even if one rejects the thought that partiality and morality are incompatible, it is clear that empathy can sometimes deform our moral thought. As Prinz remarks, ‘We are grotesquely partial to the near and dear. But that does not confirm the epistemic status of empathy. On the contrary, it shows that we use empathy as an epistemic guide at the risk of profound moral error’ (2011a: 224). There is, moreover, compelling experimental evidence that empathy’s force is fickle (ebbing and waning whimsically), irrational (unmodulated by the seriousness or size of its targets), and wildly prejudicial, being subject to ­in-­group biases, to proximity, salience, and cuteness-­effects (Konrath and Grynberg  2013). Perhaps worst of all, the allure of its verdicts often persists even when they contra­dict our considered moral judgements (Navarrete 2012; Batson et al. 1995; Batson et al. 2004). In sum, both everyday moral experience and a growing body of scientific ­evidence suggest that basic moral competence need not, and often should not, be

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

An Aetiology of Recognition  209 underwritten by occurrent empathy—and that we are often better off without it. The synchronic EIT is false.

5.  The Diachronic Empathy Indispensability Thesis Even if the synchronic EIT fails, empathy may yet connect with basic moral competence in other ways. Indeed, it is difficult to believe that empathy plays no significant role in shaping the norms that govern our moral regard for other persons. As Maibom observes, ‘. . . without the influence of empathy-­related affect, morality might be unrecognizable to us’ (Maibom 2014: 38). Recall the earlier observations that empathy underwrites moral competence by providing an epistemic resource and a motivational force that can compete with, and sometimes defeat, indifference and self-­interest. The truth of this observation does not require the truth of the synchronic EIT. Instead, empathy may be diachronically necessary for moral competence, playing an indispensable role in the development of our recognition of others’ moral claims. This would be plausible, for instance, if empathy were a developmental precondition for (a) concerned attention to others and (b) regulating (restraining or deferring) one’s concern for oneself, balancing others’ needs against our egocentric ends. Even if occurrent empathy has moved largely off-­stage by the time mature moral judgement makes its entrance, it may have played a leading role earlier on in the developmental drama. One principal source of evidence for the diachronic EIT has been developmental psychopathology; another is attachment theory. These are not entirely independent, for there is good evidence that failed attachment contributes causally to a range of moral disorders.

5.1  Empathic concern and moral psychopathology Over the last two decades, several psychologists and philosophers have argued that psychopathic personality disorder provides evidence favouring some version of the diachronic EIT (Deigh 1995; Blair 2005; Soderstrom 2003; Denham 2000, 2012). It is widely believed that psychopaths exhibit deficits in affective empathy; indeed, ‘lack of empathy’ is among the disorder’s diagnostic criteria (Hare 2003). This is supported by behavioural observations as well as autonomic measures such as skin-­conductance and startle-­blink responses. EEG and fMRI data have further indicated that psychopaths are hypo-­responsive to others’ distress, and especially to fear and sadness (Blair 1995; Blair et al.  2001; Decety et al. 2013; Patrick  1994). Psychopaths are aware of the moral rules holding sway in their communities, but they fail regularly and systematically to be guided by moral

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

210  A. E. Denham considerations in their practical judgements—they seem to know what morality requires, but are unmoved by it. Moreover, some studies (albeit not all) indicate that they are less sensitive than controls to the special authority of moral as opposed to conventional rules (Blair 1995; Blair 2006). This anomaly is unlikely to be owed to deficits in cognitive mindreading, for most psychopaths typically perform as well as neurotypicals in that respect (Blair 1995). These findings have suggested to some that the psychopath’s moral failings are caused developmentally by a deficit in affective empathy. In normal moral development, affective empathy is thought to generate negative emotions in response to actions yielding distress in others (e.g. physical abuse) and positive emotions in response to actions promoting their well-­being (e.g. helping, comforting). On one standard developmental narrative, these action types come regularly to be associated with the elicited emotions; stable patterns of response are thus acquired throughout childhood and early adolescence, later developing into settled dis­ posi­tions to respond with disapproval to negative elicitors and approval to positive ones. Once this habituation has taken place, synchronic empathic responses are no longer required to motivate token moral judgements; our settled dis­posi­ tions do the job. While affective empathy may continue to be activated on occasion, its contribution to moral development is largely completed by late adolescence. In the case of the psychopath (the hypothesis goes), this process goes awry: because of his empathic deficits, he fails to lay down the requisite as­so­ci­ ations in the first place, and this explains his failure later to respond to moral transgressions/observances with appropriately valenced motivations. The EIT finds further support from comparative data on people with autism. Autistic subjects suffer significant mindreading deficits, as well as deficits in emotion recognition. However, their empathic responsiveness—and particularly their responsiveness to others’ distress—is largely intact: the affective empathy of high-­ functioning people with autism is often (although not always) on par with that of neurotypicals, as assessed by variety of measures including expression mimicry, autonomic arousal, and fMRI (Baron-­Cohen 1995; Blair 1995; Vignemont 2009). In view of this profile, the diachronic EIT would predict that people with autism are not, on the whole, deficient in moral motivation, and this prediction is largely fulfilled: while they struggle with subtler rules of social interaction, and show developmental delay on false belief and other mindreading tasks (especially in early years), they are not systematically transgressive of other-­regarding norms. Taking the evidence from psychopathy and autism together, then, seems to recommend some version of the diachronic EIT: while cognitive mindreading is neither necessary (being impaired in morally compliant people with autism) nor sufficient (being intact in morally unmotivated psychopaths) for other-­regarding moral competence, affective empathy is indispensable. Unfortunately, consideration of the wider evidence delivers a less straightforward picture. For one thing, recent research focusing on the psychopath’s cognitive

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

An Aetiology of Recognition  211 abilities has suggested that their deficits may not be specific to affective responsiveness as such, but to a failure to integrate affective and cognitive information (Decety 2015). Several other cognitive deficits, too, have been identified, including impairments in semantic processing (Kiehl et al. 2004) and emotion recognition (Wilson, Juodis, and Porter  2011). Secondly—and more worrying for the diachronic EIT—recent studies have challenged the pivotal claim that psychopaths have profound affective empathy deficits at all. Some recent experimental evidence has challenged this long-­standing view, including one study indicating that psychopaths ‘resonate’ with others’ distress at a sensorimotor level on par with controls (Maibom  2014: 14–16; Domes et al.  2013; Lishner et al.  2012; Mullins-­Nelson, Salekin, and Leistico 2006: 139–40). Scepticism seems also to be justified by Jean Decety’s finding that the neural regions in which psychopaths differ from non-­psychopaths are not those associated with affective resonance (amygdala and anterior insular cortex) but rather those associated with concern (ventromedial prefrontal cortex and lateral orbitofrontal cortex) (Decety 2015). If Decety is correct, intact affective resonance can combine with an absence of the most basic regard for other persons’ interests, suggesting that moral decency is not borne out of resonance/attunement-­based empathic concern, but has some independent source. In that case, the psychopath’s particular toolbox of capacities may even offer evidence against the diachronic EIT. At the least, it appears to be under-­determined by the evidence, opening the door to the possibility of a third explanans—some third condition which might independently explain both the psychopath’s empathy deficits and his lack of moral motivation (Prinz  2011a, 2011b; Maibom 2014).7 A simple example of a ‘third condition’ explanation is Shaun Nichols’s account of altruistic motivation. Nichols proposes that altruism (and other-­regarding moral responsiveness more generally) in our species is best explained by a ‘Concern Mechanism’—a dedicated, independent mechanism motivating us to act in ways that will relieve or reduce conspecifics’ distress. Like empathy, the Concern Mechanism plays both an epistemic and motivational role: it alerts the agent to the other’s distress, identifying it is as the other’s distress, and then ‘triggers’ an independent motivation to act altruistically (Nichols  2001: 444). As Nichols describes the process, ‘Altruistic motivation depends on a mechanism that takes as input representations that attribute distress, e.g., John is experiencing painful shock, and produces as output affect that inter alia motivates altruistic behavior . . . I’ll . . . call this system the Concern Mechanism’ (Nichols  2001: 446). This much is compatible with the diachronic EIT. Nichols’s Concern Mechanism,

7  Maibom, for instance, observes that psychopaths’ general hypo-­responsiveness to fear and high pain thresholds might fill that role. Owing to these deficits, ‘their understanding of, and ability to feel with and for people who are afraid, would also be impaired . . . lack of fear may itself cause a number of the deficits associated with psychopathy, including the moral ones’ (Maibom 2014: 16).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

212  A. E. Denham however, relies on neither empathic attunement nor on more sophisticated perspective-­taking skills, such as the ability to imaginatively elaborate the detail of the other’s experience, or to grasp its causes and consequences for someone in his position. It operates independently of empathic responsiveness, both in signalling to the agent that a conspecific is in distress (in its epistemic role) and triggering his altruistic response (in its motivational role) (Nichols 2001: 245). On this view, responsiveness to persons’ interests can float free of empathic resonance and attunement, driven solely by an autonomous, dedicated mechanism, so that ‘the representation of the other’s distress produces a distinctive emotion of . . . concern for the other person and this emotion is not homologous to the emotion of the person in need’ (Nichols 2001: 444, emphasis added).8 If it can be demonstrated that person-­regarding judgements can be explained by an empathy-­independent mechanism of this kind, that would plainly put paid to the diachronic EIT. At most, empathy might then play a modest epistemic role, providing detail of some distress ‘inputs’, with a functionally and neuro-­ physiologically discrete concern mechanism producing the altruistic outputs.9 How plausible is Nichols’s hypothesis? An initial worry is that it does little more than put a label to a hypothetical ‘black box’, defined in very minimal, functional terms. That might not matter if the psychological and biological sciences offered no other resources to better explain concern-­activated behaviour. But surely they do. Consider, first, that evolutionary adaptations are typically economical: nature rarely replicates functions to no point. Empathic attunement and concern are already inherently motivating, with the same attentional focus (a human conspecific) and part of the same motivational direction (aversion to the conspecific’s distress or attraction to his/her well-­being). Why render empathy redundant with a functionally independent system? It would be more efficient for empathic concern to develop out of and exploit both the information and the motivation inherent in resonance and attunement, perhaps modulated by certain cognitive skills (De Waal 2006, 2008. 8  The idea is not a new one: Darwin, for instance, maintained that sympathetic concern for others’ welfare constituted a ‘separate and distinct emotion’, and more recent evidence in its favour derives from studies associating altruistic behaviour with a distinctive facial expression (Darwin 1871: 215; Roberts and Strayer 1996: 456; Eisenberg et al. 1989: 58; Miller et al. 1996: 213). 9  Nichols also argues that basic altruistic motivation does not require sophisticated mindreading of the kind to which Batson appeals (Batson et al.  1997; Batson  2011; Batson  2012). Nichols adduces three empirical considerations in support of this claim. First, as a matter of chronology, very young children exhibit altruistic behaviour (at between 12 to 18 months) before they have developed sophisticated perspective-­taking/mindreading abilities—for instance, the ability to pass False Belief tests and to make relatively fine-­grained predictions of beliefs, desires, intentions, and actions. These do not emerge until 32–48 months (Nichols 2001: 447). Secondly, autistics also have restricted mindreading abilities and yet exhibit spontaneous altruistic behaviours (Nichols 2001: 449). Finally, psychopaths provide some negative evidence: as noted above, they are typically competent mindreaders, but exhibit significant deficits in their abilities to feel empathic concern and to behave altruistically towards ­others (Nichols 2001: 449).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

An Aetiology of Recognition  213 Secondly, the developmental chronology of empathic resonance, attunement, and person-­regarding concern tells against their independence. Ontogenetically, resonance is followed by attunement, which is in turn followed by concern (Preston and de Waal 2002). Phylogenetically, too, the neurological states realizing resonance and attunement (such as the amygdala, anterior insula, and an­ter­ ior cingulate cortex) antedate those associated with concern (the ventromedial prefrontal cortex and lateral orbitofrontal cortex) (Decety and Cowell  2014; Parsons et al. 2013). Finally, the chronology of the ontogenetic development of other-­regarding judgements suggests that empathic resonance paves the way for higher-­level cognitive awareness of our own and others’ inner lives. As a child acquires a more sophisticated conceptual repertoire, his exposure to such resonance modulates his ability to mark a self–other distinction, his awareness of his powers as a discrete agent, and his recognition of others as independent loci of affective experience (Hoffman 2008; Decety and Svetlova 2012). Taken together, these considerations suggest that Nichols’s postulation of a discrete concern mechanism, functionally independent of the other natural dis­posi­ tions and abilities subserving interpersonal responsiveness, is ad hoc to the point of arbitrariness. They also suggest that any credible account of our capacity for recognizing the claims of persons as such—of our responsiveness to person-­ regarding moral requirements—would do well to take into account what we know of its genealogy. Specifically, a credible account will follow attachment theory in considering the ontogenetic, developmental origins of interpersonal responsiveness as it emerges in concert with the complex tapestry of other affective and cognitive characteristics.

6.  Mentalizing, Attachment, and Moral Recognition It would be gratifying if, at this point in my narrative, I could point the reader to a compelling, evidence-­based explanatory account elucidating the distinctive contributions of empathy to attachment and, in turn, the contributions of secure attachment to person-­regarding moral competence. I have no such account to offer. What attachment research (and especially mentalization-­based attachment theory) does offer, however, is a promising and appropriately nuanced road-­ map—a set of investigative directions, as it were—for tracing a route from our very first intimate and empathic engagements with our caregivers in infancy, through the increasingly complex dynamics of our interpersonal (and especially intersubjective) engagement with them in early childhood, to our eventual development as inherently social beings, disposed to recognize ourselves and others as a source of moral claims. I cannot lay out the detail of that road-­map here; I shall only call attention to some of the principal signposts illuminated by the findings

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

214  A. E. Denham of attachment research. In the case of secure attachment, these signposts point (perhaps by indefinitely many individual routes) to several candidate foundations of moral competence—including empathy. In the case of insecure or disorganized attachment, they signal turns at which an individual’s development can falter, and they help to explain the losses this incurs. Consider again the fictional tale with which I began: Frankenstein’s monstrous Creature, deprived from birth of sensitive caregiving, or indeed any caregiving at all, destined to a life bereft of love or respect, to a dysregulated character marked by violence, impulsiveness, self-­loathing, anger, and alienation, and culminating in a solitary and sorrowful death. Creature’s life was not a well-­lived one; on the contrary, it serves almost as a summary profile of the key liabilities associated, within attachment theory, with bad beginnings and their inexorable progress to bad ends. Recall that the attachment system has evolved (the theory says) to regulate an individual’s internal conditions (his stress and fear responses) in ways appropriate to his external conditions, especially those associated with threats to his survival; this is its adaptive function. Fonagy neatly describes what it is for this process to unfold successfully: Perceived sources of distress trigger an attachment signal from the infant, who seeks protection from that threat by ‘evoking proximity and a matching regulating protective response from a caretaker disposed to reciprocate, form emotional bonds, mentalise and teach’. These caretaker responses are, as Fonagy conceives them, different aspects of a global mirroring on the part of the caretaker, forming the foundation of mentalization as ‘one of humanity’s most pervasive and powerful characteristics—the impulse to understand and imagine both our own and other people’s thoughts’ (Fonagy et al. 2014: 36). When optimal, this mirroring will be ‘contingent’ (accurately targeting specific and variable threats), attuned (accurately guided by the child’s inner experience), and marked (signalled back to the child in recognizable ways). None of this, of course, was available to Creature following his rejection and abandonment by Dr Frankenstein. The deprivations subsequently borne by Creature in his early development illustrate many of the ways in which compromised attachment can compromise optimal development—several of which specifically involve failures of empathic mirroring. • At the level of empathic resonance, newborns are biologically pre-­wired to envision and encode human mental states, and caregivers to reflect those of the child; this is an automatic, pre-­verbal, reciprocal sensitivity to emotional signposts—a natural capacity for intersubjective reflection. When a care­ giver is unavailable—either through emotional or, as in Creature’s case, physical—absence, the child is deprived not only of protection, but of the emotional displays of human presence which he is naturally motivated to seek out, and which signal to him his key source of protection. This leads to high levels of physiological stress, correlated with both deactivation and

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

An Aetiology of Recognition  215 hyperactivation of the attachment system itself. (Bowlby, for instance, held that fear, and in particular fear of the loss of the attachment figure, was the primary activator of the attachment system (Bowlby 1969)). The extent to which a child’s subjective experience is adequately mirrored by attachment figures is thus crucial both to the quality of his attachment relationship and to a global sense of emotional security. • At the level of empathic attunement, the caregiver is the principal source by which a child learns to identify, individuate, and represent his own and ­others’ internal mental states. Deprived of the opportunity to observe the perceptible manifestations of a caregiver (e.g. in his/her facial expressions, modulated vocalizations, modulated tactile contact, physical posture, etc.) the child loses not only his signposts for the subjective presence (the psy­ cho­logic­al reality) of the caregiver, but an understanding of how to navigate the opacity of others’ mental states. Empathic, attuned mirroring also plays a role in developing the child’s ability to differentiate himself from others, allowing him to inhibit tendencies to conflate the experience of the others with his own. Deficits in such inhibition are a risk to the integrity and independence of the self, which requires a stable sense of the distinction between self and other; such deficits are common in borderline personality dis­ ordered patients, for instances, who often feel their agency overwhelmed when confronted with the wishes of others. • Again at the level of emotional attunement, caregiver mirroring plays a crit­ ic­al role in the child’s developing ability to meta-­cognitively represent his own internal states (in part through the caregiver’s perceptible reflective marking of these). Meta-­cognitive self-­awareness in turn is essential to ef­fect­ive self-­regulation of many kinds: emotion regulation, attentional focus, and behavioural impulse control. Notoriously, one cannot regulate and ­control what one is unable to detect, identify, and monitor: it is unsurprising that so many contemporary recognized disorders (including ADHD, BPD, ASPD, and ASD) are strongly associated with meta-­cognitive dysfunction. • The meta-­cognitive capacities developed via attuned caregiver mirroring are also essential to the child’s recognition of his own agency in relation to ­others—his appreciation that his own thoughts and feelings, behaviourally manifested, can influence not only the physical environment but the thoughts, feelings, and actions of others. The child’s developing sense of agency is, in turn, not only essential to his concept of himself as empowered to influence other persons, but as bearing causal responsibility for certain of their experiences. Failure to recognize personal agency of this kind is, again unsurprisingly, associated in maturity with a range of morally relevant pathologies, including psychopathy. • Finally, caregiver mirroring and mentalization manifested as empathic concern matters profoundly to the child’s perception of others as collaborative

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

216  A. E. Denham beings, with shared interests and a common agenda of promoting well-­being and averting distress and harm. The caregiver’s capacity to respond reliably with empathic concern provides the foundational data for the child’s ‘in­tern­al working model’ of other persons and, more generally, his connectedness to the human social world. ‘The first minds that small children are presented with’, observe Fonagy and his co-­authors, ‘to wonder about and interpret . . . [provide] the earliest formative lessons in other people’s thinking and also, through these people’s reactions, for learning about how our thoughts are perceived: who we are imagined to be by others’ (Fonagy et al. 2014: 36). This lesson is writ large in Shelley’s narrative of Creature: he sees himself (accurately) as a monster, deformed and repellent, and this was indeed the motive for his abandonment by Frankenstein. This image of who he is, what value (or disvalue) he possesses, and how he is perceived by others shapes every aspect of his later personal relationships. Shelley’s ­ account of Creature’s fictional woes in this respect fit with contemporary evidence: threatening or hostile attachment figures are well-­known pre­dict­ ors of a disorganized attachment style and, as learned through studies of emotionally neglected (although often physically nurtured) orphans, the simple absence of a candidate attachment figure can inflict more serious and persisting effects on cognitive development than an abusive one. I hope that this brief survey serves to indicate the merits of mentalization-­ based attachment theory. It not only offers a rich and nuanced approach to the function and significance of the empathic interactions between caregiver and child, but points to the plethora of ways in which empathic mirroring is in­dis­ pens­able to optimal psychological development.10 It also suggests, if not an answer to the question with which I began, then a direction in which to look for one. To rehearse: why and how does secure attachment promote our respect for person-­regarding moral requirements—our ability to recognize in our practical reasoning the moral claims of persons as such? Mentalization theory’s mirroring or ‘reflective functioning’, I have proposed, exploits our natural susceptibility to empathic resonance, attunement, and concern. In doing so, moreover, the child is not merely encouraged but compelled to experience and to track, first-­personally, his caregiver’s subjective states and to calibrate to them his own experience. This spontaneous, unbidden responsiveness to the caregiver’s reality is non-­optional, not only in the sense that the child’s survival depends on it, but because his nature does not permit him to resist it,

10  Interestingly, a series of studies found that sensitive caregiver (and specifically maternal) mentalizing—her engagement with and representation of the child’s mental states—is a better predictor of a secure mother–child relationship even than her global sensitivity (reliability and responsiveness to needs). (Meins et al. 2001).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

An Aetiology of Recognition  217 disposing him as it does to maintain not only physical proximity but psychological intimacy. As the child develops, this early responsiveness must—if intimacy is to  be maintained—give way to a recognition of the caregiver’s cognitive and af­fect­ive independence, whilst remaining attuned to its reality and significance. Early solipsistic engagement, as it were, must give way to recognition of an ­independent locus of experience, without losing sight of its significance and value. Again, this is a transformation which the child is powerless to resist. He does not choose to acknowledge the reality and importance of his caregiver’s feelings, thoughts, and intentions: the attachment system leaves him no alternative. In this way the child, and especially the securely attached child, gains a disposition to recognize the reality and significance of other persons which, in later maturity, he remains unable to deny. Other persons can’t but matter. Recognition of the claims of ­others is not only morally, but psychologically obligatory. The suggestion that we are obliged by our natural constitutions to take the reality and value of others into account is not new to moral theory, although the connection I am proposing to attachment theory may be. One of its most recent defenders is Christine Korsgaard. ‘What makes you take my reasons into account or bridges the gap between your reasons and mine?’, she asks. The answer, she observes, is that there is no gap to bridge: Suppose that we are strangers and that you are tormenting me, and suppose that I call upon you to stop. I say: ‘How would you like it if someone did that to you?’ And now you cannot proceed as you did before. Oh, you can proceed all right, but not just as you did before. For I have obligated you to stop . . . How does the obligation come about? . . . [T]he argument would not go through if you failed to see yourself, to identify yourself, as just someone, a person, one person among others who are equally real. The argument invites you to change places with the other, and you could not do that if you failed to see what you and the other have in common . . . In hearing your words as words, I acknowledge that you are someone. If I listen to the argument at all, I have already admitted that each of us is someone.  (Korsgaard 1996: 143).

Korsgaard’s key idea here is that our recognition of the reality and claims of other persons is non-­optional: we are constitutionally unable to resist seeing ourselves as bound to them, and they to us—independent, yet capable of permeating our thoughts, our emotions, and our motivations as we calibrate our inner lives with theirs. As attachment theory predicts, this calibration of self to other, and the expectation of others’ calibration to ourselves—first encountered in the empathic reflection of our caregivers—is not, in maturity, a disposition we are at liberty to discard. Our personal psychologies are imbued from the outset with the powerful presence of others’ subjectivities: of their affections, motivations, wants, and needs. Far from requiring a reason to take others into account, we need a reason

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

218  A. E. Denham not to take them into account. Such reasons can and do arise, of course, often as a malign intrusion into our natural development. That is what happened to Dr Frankenstein’s Creature, as he knew too well himself. Reading in books about the lives and relations of human beings, he laments, As I read . . . I applied much personally to my own feelings and condition. I found myself similar, yet at the same time strangely unlike to the being concerning whom I read, and to whose conversation I was a listener. I sympathized with, and partly understood them, but I was . . . dependent on none, and related to none . . . My person was hideous . . . What did this mean? Who was I? What was I? Whence did I come?  (Shelley 1969: 128).

These questions press Creature precisely because he recognizes that he fails to be attached and bound in some essential way to humanity. Of course, similar questions sometimes press us all, if for different reasons. Moreover, what happened to Creature, in failing to develop into a fully moral being, happens to many in varying degrees—sometimes to a degree which obliterates their humanity altogether: countless tyrants, tormentors, and madmen have populated human history, often with natures even more deformed than Shelley’s tragic monster. But what develops in the optimal trajectory—the developmental trajectory of either the securely (or not-­hopelessly insecurely) attached child—is that his or her first reality is one permeated by the intimate presence of others’ inner lives. We have evolved in such a way that we arrive prepared for empathically attuned intimacy, and (barring misfortune) this preparation leads us to naturally calibrate our experience with that of others in ways which we are powerless to resist. For this reason, recognition and regard for other persons is, in maturity, our default position: we are, as Korsgaard observes, unable not to hear others call out, make demands, laugh and weep. We begin with the reality of persons who can’t but be heard when they speak to us, and whose concerns we can’t but register. ‘You could say that it is because we want to be cooperative’, says Korsgaard, ‘but that is like saying that you understand my words because you want to be cooperative. It ignores the same essential point, which is that it is so hard not to’ (Korsgaard 1996: 141).

7.  Conclusion I began with the question of why secure attachment is positively associated with what I call person-­regarding moral concern—a capacity not merely to guide one’s actions by other-­regarding moral norms, but to recognize that such actions are owed to other persons as such, so that we are obliged to take their claims into account in our practical reasoning. This capacity is central to any credible account of basic moral competence. I then set out to assess the claims of empathy—a

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

An Aetiology of Recognition  219 capacity significantly correlated with secure attachment—as a condition of such competence. We saw that synchronous empathic concern was neither necessary nor sufficient to ensure it; if empathy contributes to our recognition of others, it does so diachronically, by a more complicated and less direct, developmental route. I surveyed the evidence from psychopathology favouring this thesis; while suggestive, it proved far from conclusive, and offered no coherent developmental narrative. Looking instead to mentalization-­based attachment theory, I proposed, allows us to better understand how empathic mirroring enters into our earliest intimate interactions with other persons, securing our default commitment, as it were, to recognizing their reality as bound up with our own. In this way, empathy constitutes one of the natural foundations on which the more complex architecture of moral experience is constructed. It is not, of course, the only foundation: other correlates of secure attachment—self-­regulation, a capacity for the moral emotions, theory of mind and perspective-­taking—all are part of the groundwork of mature moral development. Nonetheless, attachment theory helps us better to understand the indispensable role empathy plays at the beginning of the cir­cu­it­ ous road to virtue.

References Ainsworth, M. D. S., and Bell, S. 1970. ‘Attachment, exploration, and separation: illustrated by the behavior of one-year-olds in a strange situation’. Child Development 41(1), 49–67. Ainsworth, M. D. S., Blehar, M. C., Waters, E., & Wall, S. 1978. Patterns of attachment. New York: Erlbaum. Baron-Cohen, S. 1995. Mindblindness: An essay on autism and theory of mind. Cambridge MA: MIT Press. Baron-Cohen, S. 2009. ‘Autism: the empathizing-systemizing (E-S) theory’. Annals of the New York Academy of Sciences 1156(1), 68–80. Batson, C.  D. 1991. The Altruism Question: Toward a Social-Psychological Answer. Hove: Lawrence Erlbaum. Batson, C. D. 2011. Altruism in Humans. Oxford: Oxford University Press. Batson, C.  D. 2012. ‘The empathy-altruism hypothesis: issues and implications’. In J. Decety (ed.), Empathy: From Bench to Bedside. Cambridge, MA: MIT Press. Batson, C.  D., and Ahmad, N. 2001. ‘Empathy-induced altruism in a prisoner’s dilemma II: what if the target of empathy has defected?’ European Journal of Social Psychology 31(1), 25–36. Batson, C. D., Ahmad, N., and Stocks, E. 2004. ‘Benefits and liabilities of empathyinduced altruism’. In A.  G.  Miller (ed.), The Social Psychology of Good and Evil, 359–85. New York: Guilford Press.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

220  A. E. Denham Batson, C.  D., Early, S., and Salvarani, G. 1997. ‘Perspective taking: imagining how another feels versus imagining how you would feel’. Personality and Social Psychology Bulletin 23(7), 751–8. Batson, C.  D., Klein, T.  R., Highberger, L., and Shaw, L.  L. 1995. ‘Immorality from empathy-induced altruism: when compassion and justice conflict’. Journal of Personality and Social Psychology 68(6), 1042. Blair, R. J. R. 1995. ‘A cognitive developmental approach to morality: investigating the psychopath’. Cognition 57(1), 1–29. Blair, R.  J.  R. 2005. ‘Responding to the emotions of others: dissociating forms of empathy through the study of typical and psychiatric populations’. Consciousness and Cognition, 14(4), 698–718. Blair, R.  J.  R. 2006. ‘Fine cuts of empathy and the amygdala: dissociable deficits in psychopathy and autism’. Quarterly J. of Experimental Psychology 61(1), 157–70. Blair, R. J. R., et al. 2001. ‘A selective impairment in the processing of sad and fearful expressions in children with psychopathic tendencies’. Journal of Abnormal Child Psychology 29(6), 491–8. Blair, R. J. R., Peschardt, K. S., Budhani, S., Mitchell, D. G. V., and Pine, D. S. 2006. ‘The development of psychopathy’. Journal of Child Psychology and Psychiatry 47(3–4), 262–76. Bowlby, J. 1969. Attachment and loss, vol. 1: Attachment. New York: Basic Books. Coffman, S.  L. 1981. ‘Empathy as a relevant instructor variable in the experiential classroom’. Group & Organization Management 6(1), 114–20. Darwin, C. 1871/2004. The Descent of Man. New York: Barnes & Noble. Decety, J. 2015. ‘The neural pathways, development and functions of empathy’. Current Opinion in Behavioral Sciences 3, 1–6. Decety, J., Chen, C., Harenski, C. L., and Kiehl, K. A. 2013. ‘An fMRI study of affective perspective taking in individuals with psychopathy: imagining another in pain does not evoke empathy’. Frontiers in Human Neuroscience 7, 489. Decety, J., and Cowell, J. 2014. ‘The complex relation between morality and empathy’. Trends in Cognitive Sciences 18(7), 337–9. Decety, J., and Svetlova, M. 2012. ‘Putting together phylogenetic and ontogenetic perspectives on empathy’. Developmental Cognitive Neuroscience 2, 1–24. Deigh, J. 1995. ‘Empathy and universalizability’. Ethics 105(4), 743–63. Denham, A. E. 2000. Metaphor and Moral Experience. Oxford: Oxford University Press. Denham, A. E. 2012. ‘Psychopathy, empathy & moral motivation’. In J. Broackes (ed.), Iris Murdoch: Philosopher. Oxford: Oxford University Press. De Waal, F. B. M. 2006. Primates and Philosophers: How Morality Evolved. Princeton: Princeton University Press. De Waal, F.  B.  M. 2008. ‘Putting the altruism back into altruism: the evolution of empathy’. Annual Review of Psychology 59, 279–300.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

An Aetiology of Recognition  221 Domes, G., Hollerbach, P., Vohs, K., Mokros, A., and Habermeyer, E. 2013. ‘Emotional empathy and psychopathy in offenders: an experimental study’. Journal of Personality Disorders 27(1), 67. Eisenberg, N., and Miller, P. A. 1987. ‘The relation of empathy to prosocial and related behaviors’. Psychological Bulletin 101(1), 91. Eisenberg, N., et al. 1989. ‘Relation of sympathy and personal distress to prosocial behavior: a multimethod study’. Journal of Personality and Social Psychology 57, 55–66. Feeney, B.  C., et al. 2008. ‘The generalization of attachment representations to new social situations: predicting behavior during initial interactions with strangers’. Journal of Personality and Social Psychology 95(6), 1481. Fehr, E., and Gächter, S. 2002. ‘Altruistic punishment in humans’. Nature 415, 137–40. Feldman, J. B., 2007. ‘The effect of support expectations on prenatal attachment: an evidence-based approach for intervention in an adolescent population’. Child and Adolescent Social Work Journal 24(3), 209–34. Fonagy, P., and Allison, E. 2014. ‘The role of mentalizing and epistemic trust in the therapeutic relationship’. Psychotherapy 51(3), 372–80. Fonagy, P., Lorenzini, N., Campbell, C., and Luyten, P. 2014. ‘Why are we interested in Attachments?’. In P.  Holmes and S.  Farnfield (eds.), The Routledge Handbook of Attachment, 31-48. London: Routledge. Fonagy, P., Steele, M., Steele, H., Moran, G. S., and Higgitt, A. C. 1991. ‘The capacity for understanding mental states: the reflective self in parent and child and its significance for security of attachment’. Infant Mental Health Journal 12(3), 201–18. Fonagy, P., and Target, M. 1997. ‘Attachment and reflective function: their role in selforganization’. Development and Psychopathology 9(4), 679–700. Galinsky, A. D., et al. 2008. ‘Why it pays to get inside the head of your opponent: the differential effects of perspective taking and empathy in negotiations’. Psychological Science 19(4), 378–84. Goldie, P. 2011. ‘Anti-empathy’. In P. Goldie and A. Coplan (eds), Empathy: Philosophical and Psychological Perspectives, 302–17. Oxford: Oxford University Press. Hare, R. D. 2003. The Hare Psychopathy Checklist-Revised (2nd edn). Toronto: MultiHealth Systems. Hoffman, M. 2000. Empathy and Moral Development. New York: Cambridge University Press. Hoffman, M. 2008. ‘Empathy and prosocial behavior’. In M. Lewis, J. Haviland Jones, and L.  Feldman Barrett (eds), Handbook of Emotion (3rd edn). New York: Guildford Press. Hume, D. 1739/1978. A Treatise of Human Nature. Ed. L. A. Selby-Bigge; 2nd rev. edn P. H. Nidditch. Oxford: Clarendon Press. Husserl, E. 1931/1988. Cartesian Meditations. Trans. D. Cairns. Dordrecht: Kluwer.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

222  A. E. Denham Kestenbaum, R., Farber, E.  A., and Sroufe, L.  A. 1989. ‘Individual differences in em­pathy among preschoolers: relation to attachment history’. New Directions for Child and Adolescent Development 1989(44), 51–64. Kiehl, K.  A., et al. 2004. ‘Temporal lobe abnormalities in semantic processing by criminal psychopaths as revealed by functional magnetic resonance imaging’. Psychiatry Research: Neuroimaging 130(1), 27–42. Knafo, A., Zahn-Waxler, C., Van Hulle, C., Robinson, J. L., and Rhee, S. H. 2008. ‘The developmental origins of a disposition toward empathy: genetic and environmental contributions’. Emotion 8(6), 737. Konrath, S., and Grynberg, D. 2013. ‘The positive (and negative) psychology of em­pathy’. In D.  Watt and J.  Panksepp (eds), The Neurobiology and Psychology of Empathy. New York: Nova Science Publishers. Korsgaard, C. 1996. The Sources of Normativity. Cambridge: Cambridge University Press. Levi, P. 1988. The Drowned and the Saved. New York: Summit Books. Lipps, T. 1903. Ästhetik, Teil I. Leipzig: Leopold Voss Verlag. Lishner, D. A., Vitacco, M. J., Hong, P. Y., Mosley, J., Miska, K., and Stocks, E. L. 2012. ‘Evaluating the relation between psychopathy and affective empathy: two pre­lim­in­ ary studies’. International Journal of Offender Therapy and Comparative Criminology 56(8), 1161–81. Luyten, P., and Fonagy, P. 2014. ‘Mentalising in attachment contexts’. In P. Holmes and S.  Farnfield (eds.), The Routledge Handbook of Attachment: Theory, 109–126. London: Routledge. Maibom, H. L. 2014. ‘Introduction: (almost) everything you ever wanted to know about empathy’. In H. Maibom (ed.), Empathy and Morality. Oxford: Oxford University Press. Meins, E., Ferneyhough, C., Fradley, E. and Tuckey, M. 2001. ‘Rethinking Maternal Sensitivity; Mothers’ comments on infants’ mental processes predict security of attachment at 12 months’. J Child Psychology and Psychiatry 42, 637-648. Miller, P. A., et al. 1996. ‘Relations of moral reasoning and vicarious emotion to young children’s prosocial behavior toward peers and adults’. Developmental Psychology 32(2), 210. Mullins-Nelson, J.  L., Salekin, R.  T., and Leistico, A.-M.  R. 2006. ‘Psychopathy, em­pathy, and perspective-taking ability in a community sample: implications for the successful psychopathy concept’. International Journal of Forensic Mental Health 5(2), 133–49. Nagel, T. 1979. The Possibility of Altruism. Princeton: Princeton University Press. Navarrete, C. D., McDonald, M. M., Mott, M. L., and Asher, B. 2012. ‘Virtual morality: emotion and action in a simulated three-dimensional “trolley problem” ’. Emotion 12(2), 364. Neuberg, S. L., et al. 1997. ‘Does empathy lead to anything more than superficial helping? Comment on Batson et al. (1997)’. Journal of Personality and Social Psychology 73(3), 510–16. Nichols, S. 2001. ‘Mindreading and the cognitive architecture underlying altruistic motivation’. Mind & Language 16, 425–55.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

An Aetiology of Recognition  223 Nichols, S. 2004. Sentimental Rules: On the Natural Foundations of Moral Judgment. New York: Oxford University Press. Parsons, C., Stark, E., Young, K., Stein, A., and Kringelbach, M. 2013. ‘Understanding the human parental brain: A critical role of the orbitofrontal cortex’. Social Neuroscience 8(6), 525-543, Patrick, C. J. 1994. ‘Emotion and psychopathy: startling new insights’. Psychophysiology 31(4), 319–30. Preston, S.  D., and de Waal, F. 2002. ‘Empathy: its ultimate and proximate bases’. Behavioural and Brain Sciences 1, 515–26. Prinz, J. 2007. The Emotional Construction of Morals. Oxford: Oxford University Press. Prinz, J. 2011a. ‘Is empathy necessary for morality?’ In A. Coplan and P. Goldie (eds), Empathy: Philosophical and Psychological Approaches, 211–29. New York: Oxford University Press. Prinz, J. 2011b. ‘Against empathy’. The Southern Journal of Philosophy 49(s1), 214–33. Roberts, W., and Strayer, J. 1996. ‘Empathy, emotional expressiveness, and prosocial behavior’. Child Development 67, 449–70. Scheler, M. 1923/1954. The Nature of Sympathy. Trans. P. Heath. London: Routledge. Schiller, F. 1794/1967. On the Aesthetic Education of Man: In a Series of Letters. Trans. and ed. E. M. Wilkinson and L. A. Willoughby. Oxford: Clarendon Press. Schopenhauer, A. 1840/1995. On the Basis of Morality. Trans. E.  F.  J.  Payne. Providence, RI: Berghahn Books. Shelley, M. 1818/1969. Frankenstein; or a Modern Prometheus. Oxford: Oxford University Press. Smith, A. 1759/2002. The Theory of Moral Sentiments. Ed. K. Haakonsen. Cambridge: Cambridge University Press. Soderstrom, H. 2003. ‘Psychopathy as a disorder of empathy’. European Child & Adolescent Psychiatry 12(5), 249–52. Underwood, B., and Moore, B. 1982. ‘Perspective-taking and altruism’. Psychological Bulletin 91(1), 143–73. Vignemont, F.  de. 2009. ‘Drawing the boundary between low-level and high-level mindreading’. Philosophical Studies 144(3), 457–66. Vignemont, F.  de, and Singer, T. 2006. ‘The empathic brain: how, when, and why?’ Trends in Cognitive Sciences 10, 435–4. Waxman, H.  C. 1983. ‘Effect of teachers’ empathy on students’ motivation’. Psychological Reports 53(2), 889–90. Wilson, K., Juodis, M., and Porter, S. 2011. ‘Fear and loathing in psychopaths: a metaanalytic investigation of the facial affect recognition deficit’. Criminal Justice and Behavior 38(7), 659–68. Zahn-Waxler, C., Radke-Yarrow, M., Wagner, E., & Chapman, M. 1992. Development of concern for others. Developmental Psychology 28(1), 126–136.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

11

Attachment, Addiction, and Vices of Valuing Monique Wonderly

1. Introduction Addiction and certain varieties of interpersonal attachment share strikingly similar psycho-­behavioural structures.1 For example, both the addicted and the (interpersonally) attached often report a common pattern of cognition, affect, and motivation directed toward the relevant object. This pattern includes, inter alia, recurring and persistent thoughts about the object that captivate one’s attention, intense longing for the object, feelings of bliss upon obtaining the object, and feelings of dejection when one is deprived of it for too long. These thoughts, feelings, and desires tend to motivate the agent to seek out the object for various kinds of interaction. Neuroscientists, psychologists, and philosophers have adduced the above (and other) similarities between addiction and attachment to argue that many typical cases of romantic love represent addictions to one’s partner and thus might be appropriate candidates for medical treatment. In this chapter, I argue for the far more neglected thesis that some paradigmatic cases of addiction are aptly characterized as emotional attachments to their objects. This has implications for how we should understand the nature of addiction and the ethics of attachment more broadly. The chapter will proceed as follows. In section  2, I review three overlapping conceptions of the relationship between addiction and attachment that find support in the literature. I will defend (a version of) the third conception: namely, that some substance addictions represent attachments. In section 3, I lay out the theory of attachment that I will use to argue for this view. In sections  4 and  5, I construct a defence of the claim that some paradigmatic cases of addiction contain the key marks of the relevant brand of attachment. Finally, in section  6,

1 Many thanks to Hanna Pickard, Edward Harcourt, David Beglin, Coleen Macnamara, two an­onym­ous reviewers from Oxford University Press, and seminar participants in my 2018 graduate course on the Philosophy of Attachment at UC San Diego for helpful comments on earlier drafts of this chapter. Monique Wonderly, Attachment, Addiction, and Vices of Valuing In: Attachment and Character: Attachment Theory, Ethics, and the Developmental Psychology of Vice and Virtue. Edited by: Edward Harcourt, Oxford University Press. © Monique Wonderly 2021. DOI: 10.1093/oso/9780192898128.003.0012

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

Attachment, Addiction, and Vices of Valuing  225 I  suggest that exploring the ethics of addiction can illuminate the potential for (what I will call) certain ‘vices of valuing’ in attachments more broadly.2

2.  Attachment and Addiction: Three Links from the Literature Addiction is notoriously difficult to define, and defining attachment too may not be without its difficulties, notwithstanding the extensive literature (including some contributions to this volume). The same goes for specifying the relationship between these two phenomena.3 We can, however, glean at least three overlapping conceptions of the relationship between addiction and attachment from the philosophical and social science literatures. These include: (1) that many typical cases of interpersonal love represent addictions, (2) that addiction is an attachment disorder, and (3) that paradigmatic cases of addiction sometimes represent attachment orientations. The idea that romantic love is, or can be, an addiction is a familiar one. References to love’s ‘addictive nature’ abound in popular music and classical poetry.4 Perhaps more surprisingly, this trend has recently found a foothold in scholarly research. For example, Helen Fisher et al. explain that some psychologists regard romantic love as an addiction owing to love’s ‘addiction characteristics’, which include (among other things) intensely focused attention on the love object, mood swings, cravings, obsessive thoughts, emotional dependence, and loss of self-­control (2010: 51–2). James Burkett and Larry Young associate the following pattern with both substance addiction and (what they refer to as) ‘partner addiction’: euphoria in early encounters, eventually replaced by ‘a subdued sense of contentment’ in later encounters, recalcitrant desires for further contact, compulsive seeking behaviour, separation anxiety, and depressive symptoms arising from permanent (or prolonged) cessation of contact (2012: 16).5 Recent advances in neuroscience suggest that there is also significant overlap in the neural processes underlying both romantic love and substance addiction, and some theorists have adduced these findings in support of the view that love is (or can be) an addiction. The relevant studies indicate that substance addictions and romantic attachments implicate similar neural reward pathways in the brain, 2  I use the term ‘vice’ somewhat reluctantly and merely to capture a tendency toward a problematic orientation rather than immoral character or conduct. 3  As psychologist Jean Mercer recalled, one ‘highly skilled and educated clinical psychologist’ who was asked to define attachment responded, ‘I don’t know but I know it when I see it’ (2006: 2). The term addiction has been ‘omitted from the DSM-­5 substance use disorder diagnostic terminology’ in part ‘because of its uncertain definition’ (APA 2013: 485). 4  Rolling Stone Magazine recently featured an article entitled ‘Your Love is a Drug: 20 Great Narcotic Love Songs’ containing a list of songs with this theme (Epstein 2015). For a classic poetry reference, see Ovid (1977: Bk II, IXb). 5  See also Peele and Brodsky (1975: esp. ch. 4).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

226  Monique Wonderly including activation of the ventral tegmental area and nucleus accumbens (among other neural structures) and the release of specific neurochemicals such as dopamine (Insel  2003; Frascella et al.  2010; Burkett and Young  2012). Philosopher Brian Earp and his co-­authors have drawn on this research in order to support the view that ‘to be in love is in some sense to be addicted . . . to another person’ and that ‘in some instances, [medical] “treatment” of love, could be justified or even desirable’ (Earp et al. 2017: 78–9).6 Another view that we find in the literature is that addiction is, or at least can be, an attachment disorder. Psychologist Philip Flores, for example, takes this view, explaining, ‘. . . individuals who have difficulty establishing emotionally regu­lat­ing attachments are more inclined to substitute drugs and alcohol for their deficiency in intimacy’ (2004: 6–7).7 On his account, the addiction ‘serves as both an obstacle [to] and as a substitute for interpersonal relationships’ (2004: 4). At least two observations from the psychopathology literature lend support to this picture. First, many addictions represent attempts to alleviate emotional distress via self-­medication—distress that often arises from psychological disorders and/ or relationship troubles (Khantzian 1985; Pickard  2012; Pickard 2019). Second, one important function of interpersonal attachment is to aid in emotion regulation (Bowlby 1969; Mikulincer and Shaver 2016; Schore 1994/2016). Thus, those who have attachment difficulties would seem especially vulnerable to addiction. Their relationship problems can both cause emotional pain and hinder their abil­ ities to cope with it, thereby making frequent substance use an attractive (and in some cases, the only obviously available) option for pain management.8 This leads us to a third, overlapping conception of the relationship between attachment and addiction. If, as Flores contends, addiction often serves as a substitute for interpersonal attachment, then it would make sense for (some cases of) addictions not only to function as sources of pain management—or again, hindrances to healthy attachment relationships—but also as forms of attachment themselves. Flores sometimes uses language suggestive of this view, describing addicted individuals as ‘attached to chemicals’ or ‘attached to alcohol’ (2004: 4, 34–6).9 In a similar vein, philosopher Gary Watson insightfully remarks that at a 6  Burkett and Young (2012) also explicitly make this suggestion. 7  See also Cihan et al. (2014) and Unterrainer et al. (2017). The term ‘attachment disorder’ can refer either to mental disorders for which attachment difficulties are presumed to play a significant causal role and/or psychologically disordered forms of attachment formation or maintenance. Note that the Diagnostic and Statistical Manual of Mental Disorders does not list attachment disorders as its own category. There are, however, several disorders that explicitly include attachment difficulties, including (but not limited to): Reactive Attachment Disorder, Separation Anxiety Disorder, Disinhibited Social Engagement Disorder, and Autism Spectrum Disorder (APA 2013). 8  Research suggests that early attachment difficulties and occurrent insecure attachment styles correlate positively with chemical and behavioural addictions (Mikulincer and Shaver 2016: 433–5). 9  See also Jim Orford’s work on addiction as ‘excessive appetite’ (2001). On Orford’s view, addiction is helpfully construed as ‘an attachment to an appetitive activity so strong that a person finds it hard to moderate the activity despite the fact that it is causing harm’ (2001: 18). While Orford suggests that

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

Attachment, Addiction, and Vices of Valuing  227 certain stage of dependence, one can develop an ‘attachment to one’s addiction’ that ‘in extreme cases . . . might amount to an attachment disorder’ (1999: 17).10 Watson has in mind cases in which the agent sees life without her addiction as a ‘grave loss’, potentially inducing terror and panic at the thought of having to get by without it (1999: 17). Neither Flores nor Watson, however, go into much detail about what it means to be attached to an addiction (or to an addiction object, such as alcohol). And to my knowledge, little has been written about ‘attachments’ of this sort and how they compare to—and might inform our understanding of— attachments more broadly, especially among philosophers. In what follows, I address this lacuna. But first, we need a working view of attachment.

3.  To Be Attached In an earlier work, I put forth a philosophical view of what I refer to as ‘security-­ based attachment’—a view that draws heavily on research from developmental and clinical psychology (Wonderly  2016).11 I will briefly reprise some of the details of that account here, as it will be helpful for establishing the case that some addictions represent attachments to their objects. To begin, let’s consider how psychologists have traditionally conceived of attachment. According to attachment theory, human infants typically develop a special bond with their primary caregivers. This bond is characterized by a set of evolutionarily adaptive behaviours that provide the infant with a sense of security. The attached infant attempts to remain in close proximity to her primary care­ giver, treats her as a ‘secure base’ from which to safely explore unfamiliar surroundings, seeks her out for protection as a ‘safe haven’ when threatened, and protests separation from her (by clinging, crying, etc.) (Bowlby  1969). Psychologists have also recognized that adult long-­term romantic partnerships are usually marked by similar behaviours. We seek proximity to our romantic partners and protest prolonged separation from them. Our romantic partners also serve as secure bases and safe havens for us. When they are nearby, we feel more competent to explore unfamiliar environments and to take on new

emotion regulation and operant reward in the form of ‘powerful emotional change’ play important roles in addiction, he says fairly little about what constitutes the relevant affective attachment (2001: 22). 10  Watson, following Seeburger (1993), calls the relevant stage existential dependence, defined as ‘the development of an identity to which the addictive practices are crucial’ (1999: 16). By ‘attachment to one’s addiction’, Watson appears to refer to an attachment to a practice (as opposed to an attachment to an addiction object). Watson also does not appear to be referring to the kind of security-­ based attachment at issue in this chapter (as will become plain in section  3), but he connects attachment to a ‘volitional necessity’ to continue to use the drug (Watson 1999: 17). 11  There has been fairly little in the way of sustained analyses of attachment, at least insofar as it is distinct from caring or love. Harcourt (2017) offers one of the most thorough and enlightening treatments of attachment in his recent work on attachment and love. See also Wonderly (2016; 2017).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

228  Monique Wonderly challenges. Also, when distressed, we tend to turn specifically to our romantic partners for support and relief (Mikulincer and Shaver 2016; Collins et al. 2006).12 In my own work, I develop and employ an expanded conception of (this type of) attachment—one that includes attachments to objects, activities, and ideas, an emphasis on engagement with (rather than mere proximity to) the relevant object, and an expanded notion of security (Wonderly  2016). On this view, attached agents experience recurrent, persistent desires to engage with some person or object. Different attachment relationships admit of different kinds of engagement. For example, in addition to wanting to be near her primary caregiver, the infant might desire to play with her, whereas the romantic partner might seek conversation or sexual contact from her beloved. Likewise, an individual who is attached to her car can engage with her attachment object by driving it, whereas an individual who is attached to a particular idea (e.g. the concept of infinity) or activity (e.g. running) can engage with the relevant objects by contemplating it or ‘doing it’, respectively (Wonderly 2016: 229).13 The relevant desire for engagement has two further important features. First, the object of engagement—i.e. the attachment object—is non-­substitutable. This is a familiar point in psychological theories of attachment. As attachment the­or­ ist, Mary Ainsworth, explains, ‘an attachment figure is never wholly interchangeable with or replaceable by another’ (Ainsworth et al. 1978: 38). To be attached is to feel as though one needs this person or object and going without her or it would come at a significant cost. An individual who could easily substitute biking for running is not attached to running. Similarly, an individual who could simply ‘trade out’ her spouse for a newer, more attractive partner, is not attached to her spouse. Merely finding pleasure in one’s engagement with the relevant object does not suffice for attachment. And even while one might be attached to multiple

12  Attachment theorists also suggest that our attachments to our primary caregivers in infancy shape how we relate to one another in later attachment relationships, including romantic pair bonds. As infants, we develop ‘attachment styles’ that tend to follow us through adulthood. The system by which attachment styles are classified grew largely out of Mary Ainsworth et al.’s empirical study of infant–primary caregiver interaction (1978). Ainsworth et al. identified two patterns of insecure infant attachment: avoidant and anxious. In infants, these patterns track certain atypical infant responses to separation and reunion with their primary caregivers. Securely attached infants tend to show some distress upon separation from their primary caregivers, but recover quickly upon reunion, exhibiting joy and a desire to return to exploration and play. Avoidant infants tend to show little distress upon separation from their primary caregivers and to avoid them when they return. Anxious infants are highly distressed during separation from their primary givers and display conflicting behaviours upon the caregiver’s return. In adulthood, an avoidant attachment style indicates ‘discomfort with closeness and dependence’ and a strong ‘preference for emotional distance and self-­reliance’. An anxious attachment style indicates a strong desire for closeness and intense worries about being abandoned or undervalued by one’s partner (Mikulincer and Shaver 2016: 23–5; see also Wonderly 2019: 31 n. 19). 13  The notion that ideas and activities can play regulative functional roles similar to that of relationship partners also appears in the psychological literature on object relations theory. Morris Eagle, for example, suggested that interests and values can serve some of the psychological functions typ­ic­ al­ly served by other persons (1981). Many thanks to Edward Harcourt for introducing me to this work and prompting me to reference it here.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

Attachment, Addiction, and Vices of Valuing  229 objects, each presents itself as indispensable and uniquely important for oneself in its own way.14 A second feature of attachment’s desire for engagement is that the attached party’s sense of security, in some sense, depends on whether or not that desire is satisfied. Engaging with an attachment object tends to increase one’s sense of security, while being deprived of it typically results in a decreased sense of se­cur­ ity. As I have argued, the relevant sense of security is not merely a feeling of safety or comfort, but is better understood as a kind of confidence in one’s well-­being and agential competence.15 In colloquial terms, without our attachment objects, we often feel ‘out of sorts’, off-­kilter, ‘no longer all of a piece’, etc. Engagement with our attachment objects helps us feel ‘on solid ground’, more stable, and more competent (Wonderly 2016: 231). As these descriptions make plain, a reduced sense of security has an affective, or emotional, element that is intimately connected to one’s agency. Individuals whose felt security is compromised often feel emotionally fractured and as though they are unable to get along in the world as well—in short, like less competent agents. Summing up, we can think of one form of attachment as a particular type of security-­based felt need for its object. Roughly, to be attached in this sense is (1) to have a relatively enduring desire for engagement with a non-­substitutable object, where (2) such engagement typically increases one’s felt security and prolonged separation from the object typically reduces one’s felt security.

4.  The Affective Divergence Worry On the view I argue for here, many paradigmatic cases of addiction represent attachments to their objects. In other words, just as an individual might be attached to a particular person, so too might an addicted person be attached to a particular drug.16 In the latter case, the individual experiences the substance as a felt need, such that engagement with it (or deprivation thereof) affects her sense of security—that is, her sense of confidence in her well-­being and agential competence. But before defending this view, it will be useful to lay out a set of interrelated concerns that suggests against it. To start, consider that one might explain (much of) the core phenomenology and motivational architecture of addiction exclusively in terms of somatic cravings and bodily feelings. Here is one way the story might go. After repeatedly 14 For more on the non-­ substitutability of attachment objects, see Bowlby (1969: 308–9); Bretherton (1991: 19); Weiss (1991: 66; Cassidy (2008: 12–15), and Wonderly (2017: 244). 15  See Wonderly (2016: esp. 230–2) for a more detailed discussion and defence of this claim. See also Maslow (1942); Blatz (1966); and Ainsworth (1988). 16  For simplicity’s sake, I will restrict discussion to substance or alcohol addictions, but I suspect that much of what I will say here also applies to behavioural addictions.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

230  Monique Wonderly ingesting a substance that produces highly pleasurable bodily sensations, an agent develops an increasingly strong—and difficult to satiate—recurrent craving for the substance. Failure to obtain the desired substance causes frustration, and in some cases, physical pain. The relevant pleasures and pains motivate the agent to seek out the substance compulsively. I take it that this picture will strike many readers as a familiar, if somewhat crude, model of addiction. While the philosophical literature on addiction tends to favour more nuanced accounts, here, too, theorists often analogize addiction to ‘appetites’ and other phenomena typically associated with meeting bodily needs. These accounts often accommodate roles for emotions.17 Yet, unsurprisingly, most do not immediately lend themselves to construing addiction in terms of the rich affective ties that we typically associate with attachment.18 These con­sid­er­ ations, though not necessarily incompatible with construing some substance addictions as attachments, encourage a particular worry that suggests against such a construal. According to the relevant worry, to construe addictions as attachments is to obfuscate a crucial difference between these two phenomena: namely, their differing affective qualities. We tend to think of an attachment as a form of emotional connectedness (typ­ic­ al­ly) toward another person with whom the attached party wants to engage. Addictions, on the other hand, are generally viewed as bodily appetites for substances that the addicted party wants to consume. This makes sense of why the locutions, ‘being attached to a drug’ and ‘engaging with a substance’, tend to strike the ear as rather odd. We have appetites for food and drugs and emotional attachments to other persons. What’s more, the ‘specialness’ and non-­substitutability of an individual’s attachment object is typically tied to her emotional bond with that particular person. And this might explain why addiction objects seem to lack similar qualities of particularity and irreplaceability. It is not as though the heroin addict craves that particular heroin, as opposed to any other bit (ounce, vial, batch, etc.) of the substance, and she might be more than willing to give up heroin altogether should a more pleasing drug come along. Finally, and for similar reasons, one might baulk at the idea that a drug can affect one’s felt security in a sense that even roughly parallels the impact that an attachment figure typically has. A drug, for example, might help one to feel well and empowered because it temporarily induces pleasure—or again, removes 17  Importantly, on many accounts, the objects of ‘appetites’ can include non-­ingestible objects (and activities) that confer emotional pleasure. 18  As noted in section 2, the accounts offered by Watson and Orford are notable exceptions. For other views that characterize addiction in terms of appetites or other visceral phenomena, see Loewenstein (1999); Dill and Holton (2014); Earp et al. (2017); and Foddy and Savulescu (2007; 2010). There are more and less capacious views of appetite represented here. Importantly, advocates of appetitive models needn’t claim that this is all there is to addiction, or again, sharply dichotomize appetites and emotions.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

Attachment, Addiction, and Vices of Valuing  231 pain—and we naturally feel better and more competent when in these states. Attachment figures would seem to impact one’s security in a deeper and more direct sense, as engagement with them seems to shape our senses of who we are and how we are able to get along in the world in a more lasting and meaningful way. And here again, this feature seems integrally connected to the emotional bond between two persons. The worrier, then, might conclude that owing to their divergent affective qual­ ities, addictions and attachments are essentially and ineluctably dissimilar. The animating desires for their respective objects, the substitutability of those objects, and their relationships to the agent’s felt security differ enough to render addictions unsuitable candidates for attachments. I take it that any plausible conception of addiction as attachment must deal with these concerns, which I will refer to jointly as the Affective Divergence Worry, and I will proceed with this desideratum in mind. As I will show in the following section, while addiction attachments and interpersonal attachments are typically not affectively on a par, many paradigmatic cases of addiction nevertheless meet the requisite criteria (affective and otherwise) for attachments.

5.  Addiction as Attachment: A Defence Contra the Affective Divergence Worry discussed above, many paradigmatic addictions share central affective characteristics with typical interpersonal attachments. To be sure, not all addictions do, but I hope to show that an attachment theoretical framework is remarkably apt for many addictions. To see, let’s consider how the attachment criteria discussed in section 3 might feature in paradigmatic cases of substance addiction. Recall that on my view, attachment involves, inter alia, a relatively enduring desire for engagement with a non-­substitutable object. In a fairly straightforward sense, addicted persons do often experience persistent desires to ‘engage with’— to smoke, to swallow, to snort, to inject, etc.—a specific drug that does not easily admit of a substitute substance. A heroin addict, for example, might use a plethora of drugs while experiencing a special felt need for heroin in particular.19 To say this, though, is not yet to say enough. One might attempt to capture the relevant ‘need’ exclusively in terms of a chemically induced, pleasure-­oriented desire for a preferred substance, thus giving traction to the Affective Divergence Worry. On this approach, the need aims at the attainment of physical pleasure (or again, the avoidance of physical pain), and talk of ‘emotional attachment’ seems 19  One might also imagine an attachment to the activity of ‘drug use’, where using a variety of different substances can still be explained by a single attachment. Thanks to Hanna Pickard for raising this possibility.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

232  Monique Wonderly more than a little out of place. Notably, though, evidence suggests that this approach (on its own) is ill equipped to capture many cases of addiction. Sometimes, for example, the object of one’s addiction isn’t experienced as pleasurable, and withdrawals aren’t always significantly physically painful.20 And even where these factors are present, addictions are also often constituted, in part, by complex emotional attitudes. Addiction is associated with a number of powerful emotions that come in both negative and positive varieties.21 In addition to physical pleasure-­terms like ‘euphoria’ and ‘ecstasy’, addicted persons also often describe their addictions in terms of richer affects, such as ‘joy’ and even ‘love’. According to psychologist Gabor Maté, ‘Since their addictions offer biochemical substitutes for love, connection, vitality and joy, to ask [substance addicts] to desist from their habits is to demand that they give up on the emotional experiences that make life worth living for them’ (2008: 363). Of course, to say that one’s addiction is a source of joy, love, or connectedness is not yet to say that one feels emotionally connected to the object of her addiction. But here again, numerous examples from addiction studies and first-­person reports support such a view. While the locution ‘attached to a drug’, is rare in everyday discourse, it is far less so among self-­described addicted persons and the theorists and practitioners who engage with them. In section 2, I indicated that Flores variously described addicted individuals as ‘attached to chemicals’ or ‘attached to alcohol’. To this, we can add descriptions of addicted persons ‘bonding with’ methamphetamines or heroin (Lewis 2015: 74; Hari 2015: 176), being ‘in love’ with alcohol (Knapp 1996: 5), and entering ‘into a relationship with [one’s addiction object]’ (Seeburger 1993: 55). And this is just a small sample. Neuroscientist and self-­described former addict Marc Lewis provides one poignant example of the relationship-­oriented phenomenology that is internal to some addictions in his discussion of a methamphetamine addict named Brian. Brian confided to Lewis, ‘As my relationship with [his wife] was collapsing, my relationship with the substance increased, to fill the gap’ (Lewis  2015: 74). As Lewis characterizes him, Brian developed a ‘bond with meth’ and lamented that ‘quitting meth was like turning his back on a friend or lover’, where his resultant ‘heartbreak’ was akin to that which he experienced upon losing his wife (2015: 74, 167–8). While Brian’s choice of words might strike one as confused or dishonest, first-­person reports of addiction often echo similar sentiments. Many describe their addictions in terms of their ‘relationships’ with the substance. In the Big Book of Alcoholics Anonymous, for example, there are numerous entries in which addicted persons refer to alcohol as a friend, lover, or companion 20  See e.g. Kennett and McConnell (2013); Pickard and Pearce (2013); and Pickard (2015). 21  The idea that shame plays a central role in (many cases of) addiction is a recurring theme in the addiction literature (Maté 2008: 225; Flanagan 2013; Hari 2015: 103–15; Reid 2016: 259).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

Attachment, Addiction, and Vices of Valuing  233 (AA 2001: 310, 289, 447). Smokers often ‘report feeling a sense of bereavement during the early stages of stopping smoking: as if they have lost a cherished friend or family member’ (West and Ubhi 2019: 47). Consider also that on a National Public Radio programme, health department outreach worker and former heroin user Nathan Fields explained, ‘Heroin has always been a great companion for people that are dependent . . . Best friend. It can talk to you. It can reason with you’ (Cornish 2016). I suspect that addicted persons rarely regard substances as literal love or friendship objects. But once we take seriously the ideas that addictions are frequently (partly) constituted by rich emotional material and that individuals often represent their addiction objects as something like friends or companions, it becomes more palatable to think of some substance addictions as attachments. Even while substances and persons admit of different varieties of engagement and different levels of particularity (particular drugs versus discrete individuals), one might nevertheless be emotionally connected to a substance that one regards as non-­substitutable. More, however, needs to be said. The above has not yet shown how addictions might constitute security-­based attachments proper. Recall that to be attached (in the relevant sense) is, among other things, to experience a security-­based felt need of the attachment object. Engagement with the object tends to increase one’s sense of security, while prolonged separation from the object typically results in a decreased sense of security. This brand of security represents a kind of confidence in one’s well-­being and agential competence—roughly captured by colloquialisms such as feeling ‘more together’ or ‘empowered’, as opposed to felt insecurity which often manifests in feeling ‘off-­kilter’, ‘out of sorts’, etc. An adequate defence of my view must show how an addiction might meet this criterion as well. First, we should note that theorists and practitioners have long suspected that the need for felt security often both motivates and sustains addictive behaviours. In their discussion of addiction recovery narratives, for example, Vilma Hanninen and Anja Koski-­Jannes explain that the ‘love story’ narrative depicts the agent’s addiction as ‘a justified way of striving for the feeling of security of which she has been deprived’ (1999: 1845). It is also instructive to consider how self-­described addicts frequently employ security-­laden language to explain their addictions. According to Maté, one addicted person reported that using heroin ‘felt like a warm, soft hug’, while one morphine-­addicted person described his first experience as feeling ‘like a warm, wet blanket’ and a ‘place of safety’ (2008: 236). Flanagan imparted that his drugs of choice afforded him ‘a safe-­haven feeling’ (2011: 275). Addicted individuals also frequently emphasize how using makes them feel more confident and capable, identifying the objects of their addiction with ‘power’, ‘self-­confidence’, or a ‘rush of contentment and well-­being’.22

22  See Knapp (1996: 5); Seeburger (1993: 10); Lewis (2015: 144, 176, 178); AA (2001: 502).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

234  Monique Wonderly In addition, going without one’s addiction object also tends to significantly diminish one’s sense of security. As philosopher Frances Seeburger, helpfully puts the point, ‘Whenever addicts are practicing their addictions, then by definition, they feel “all right.” In contrast, whenever addicts are not practicing their addictions, it feels to them as if something is wrong. They feel “restless, irritable, and discontented,” without ever being able to put a name on just what it is that is making them feel that way. In some way that they cannot further specify, things just feel out of joint to them’ (1993: 55). At first glance, this language suggests a strong connection between the addiction object and the addicted agent’s sense of security.23 Of course, if these ‘security-­related’ feelings are reducible to the drug’s direct impact on the agent’s neurophysiology, then the Affective Divergence Worry discussed in section 4 still looms large. Many types of addictive drugs are known to reduce physical pain. Similarly, particular families of drugs tend to chemically induce feelings of warmth; others, feelings of empowerment. Agitation and confusion are common drug withdrawal effects. These experiences are not attachment related, but predictable effects of substance use (regardless of whether or not the user is emotionally connected to the substance). In other words, even if the addicted person’s engagement with her drug impacts her sense of security, it may not do so in the requisite sense for an emotional attachment orientation. An example should clarify this point. Suppose that I have a pounding headache. The pain might leave me feeling unwell and unable to function as well as I otherwise could do. And I might reasonably report that in my current state, I feel quite ‘off-­kilter’ as a result of the pain. Pain, after all, often makes one feel less secure. Suppose, further, that a standard over-­the-­counter analgesic, such as ibuprofen, would do the trick, empowering me to better take on my day. While the analgesic impacts my felt security in some sense, I am no more attached to it than I am to the water that I might seek to quell my uncomfortable (and potentially debilitating) thirst. The ibuprofen operates as a mere tool for pain relief, one toward which I feel no affective connection and that I would readily trade out for a comparable medication such as acetaminophen. If an individual’s drug use affects her security in this way alone then it would not warrant the attachment label. To be sure, many drug addictions seem to operate in this very way. Recall that addicted persons often use a specific drug in order to self-­medicate. Using drugs often relieves pain, and since pain is a barrier to felt security, pain removal in­dir­ ect­ly affects the user’s sense of security. Similarly, as noted above, some substances 23  One potential objection to this claim that I will not explore in detail here is that more often than not substance use tends actually to diminish (rather than enhance) one’s security. This strikes me as a true statement insofar as it goes, but it poses no threat to the view on offer here. Once we acknowledge that the same interactions can impact a person’s sense of security in complex ways, it is not hard to see how substance use might increase one’s felt security in some sense, while diminishing it in another. This is a familiar phenomenon in many interpersonal attachments as well: think here of the abused partner who acknowledges the harm and fear her partner causes her but confesses that she nevertheless just doesn’t feel ‘all right’ without him.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

Attachment, Addiction, and Vices of Valuing  235 directly induce positive security-­related effects. Use that is sustained by these effects alone is not attachment. Importantly, though, even if most (or all) addictions begin this way, many eventually come to take on an additional character. Oftentimes, the addicted individual’s sense of security becomes wedded to her use in an affectively richer sense. Her substance use does not merely impact her sense of security by removing pain—or again, by inducing an intense, but fleeting sense of empowerment—but that particular substance comes to represent the (sadly, often false) promise of a more global confidence in her well-­being and agential competence. To see this more clearly, consider how Lewis characterizes his own addiction in his Memoirs: ‘The drug (or other substance) stands for a cluster of needs: in my case, needs for warmth, safety, freedom, and self-­sufficiency’ (2013: 256). These sentiments suggest a connection between the addiction object and the addicted person’s felt security that transcends the boundaries of the ‘highs’ associated with intoxication and the ‘lows’ associated with withdrawal. This should come as no surprise, as addicted individuals often continue to seek out their preferred substance even when that substance is no longer capable of producing its associated physiological effects in the user—and even when more potent substitutes are available.24 In addition, addicted persons often associate their substance use, not merely with relief, but with ‘meaning’ and one’s sense of ‘identity’, while they associate being unable to use with ‘grief ’ and ‘grave loss’ (Seeburger 1993; Watson 1999: 17; Maté 2008: 363; Hari 2015: 175; Lewis 2015: 168; Pickard 2019: 18–19). These terms call to mind the intimacy of close interpersonal relationships, even if only in a penumbral sense. Thus, it seems rea­son­ able to think that in at least some cases, the addicted agent’s orientation toward her addiction object reflects a security-­based felt need of it, in a sense similar to that of more paradigmatic attachments.25 An attachment framework nicely captures central aspects of addiction on which many models of addiction are largely silent.26 For example, many have noted that contemporary views of addiction often not only fail to adequately capture addiction’s phenomenology, but also its motivational structure, and emotional and social significance, more broadly—particularly with regard to

24  Thanks to Hanna Pickard and David Beglin for helpful discussion on this point. 25  One might worry that first-­personal accounts of addiction are unreliable, as a person might offer a romanticized description of her addiction merely to produce attractive prose—or perhaps for more practical reasons, such as mitigating personal responsibility or securing more sympathetic social support. While we should take such factors into account when evaluating the evidential force of addiction testimony, the sheer pervasiveness of stories like those presented above suggests against unreflective dismissal. Since addiction is not simply a mechanistic process, but a lived experience, first-­person accounts are bound to be uniquely informative routes to understanding it. For an insightful and balanced treatment of the role of testimony in addiction, see Pickard (2012: 44). 26  Importantly, I do not take myself to have shown that the Common View is incorrect, but merely incomplete. I take it that many, if not most, attachment addictions also have features associated with the Common View.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

236  Monique Wonderly identifying causal factors and appropriate treatment options.27 An attachment-­ theoretical conception of addiction captures the security-­based emotional underpinnings of many addictions, favouring treatment options that attend to this feature—such as attachment therapies, communal support structures, and neuropharmaceuticals aimed specifically at addressing diminished felt security. In addition, this conception offers a more nuanced alternative to accounts that depict addictive motivation as merely a matter of pleasure-­oriented desires. Often, the need for felt security is a crucial aspect of addiction’s motivational architecture, and, importantly, security is neither reducible to physical pleasure nor the absence of physical pain but has its own special conative force.28 Thus, if the view I have presented here is correct, it can potentially aid our understanding of the nature of addiction, how best to treat addicted persons, and how addicted desires impact human agency. What’s more, not only can the nature of attachment inform our understanding of addiction, but as I will suggest in the final section, the ethics of addiction can inform our understanding of the ethics of attachment more broadly.

6.  Addiction and the Ethics of Attachment If, as I have argued, persons sometimes become attached to their addiction objects, one might think that such attachments are especially and necessarily pernicious. After all, addiction is often associated with vice.29 For my part, I doubt that matters are so simple. First, it is unclear whether and to what extent all addictions have the (purportedly) bad-­ making features often attributed to them. Second, it is far from obvious that those features suffice to make addictions that do have them pernicious. Nevertheless, by considering what makes addictions objectionable when they are, we can glean insights into a particular set of vices to which we are susceptible in virtue of being attached—whether to persons or non-­persons. To see, let’s start by considering what makes addictions ethically objectionable. Addictions are often thought problematic because (1) they render one vulnerable 27  See e.g. Pickard (2012; 2015);Kennett (2013); Kennett and McConnell (2013). 28  Psychologists have long suspected that the need to restore and maintain a sense of security typ­ic­ al­ly enjoys a kind of primacy over other human needs; consequently, agents are prone to experience the need for their attachment objects as especially compelling. Bowlby (1969); Collins et al. (2006); and Mikulincer and Shaver (2016). 29  The point here is not that addiction itself constitutes a vice, but that at least some addictions have features internal to them that may render an agent susceptible to certain dispositions associated with vice. There are different views about the relevant association. Immanuel Kant, for example, deemed ‘drunkenness’ and (certain forms of) ‘intemperance in social drinking’ moral vices (Kant 1784–5/1997: 152–3; 1798/2006: 63). But we needn’t subscribe to such a view here. As I indicated in note 3, I use the term ‘vice’ here rather loosely and to capture a kind of problematic orientation, in character or conduct, that is inimical to certain (not necessarily moral) values.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

Attachment, Addiction, and Vices of Valuing  237 to suffering and diminished self-­control, and (2) they involve improper valuing orientations toward their objects.30 With regards to (1), we can be somewhat brief. Doubtless, addictions often cause suffering and undermine self-­control. Many addicted persons suffer physical pain due to withdrawal syndromes and other addiction-­related health problems, as well as mental anguish owing to societal rejection and frustrated attempts to quit. Addictions also typically bear the marks of excess and agential constraint. Addicted persons tend to consume ‘too much’ of their preferred substance. And while theorists differ about whether and in what sense addictions are ‘compulsive’, few would deny that, minimally, they are usually very difficult to quit.31 Does susceptibility to suffering and diminished self-­control suffice to make practising an addiction objectionable? This seems unlikely. Consider that interpersonal romantic love, one of the most revered forms of attachment, typically has these features as well. To love is to be susceptible to pain. Love is, after all, an especially powerful species of caring—a defining feature of which is vulnerability to emotions such as fear and sadness when the object is threatened or otherwise doing poorly (Jaworska 2007; Helm 2009). In varying degrees, painful longings, worry, and heartbreak occur in many overall worthwhile romantic attachments. To love is also to be susceptible to diminished self-­control. Watson is clear on this point, explaining, ‘like addictions’, certain loving relationships that we ‘encourage and honor’ also render one ‘vulnerable to diminished control of certain kinds’ (1999: 18). Just as we have reason to doubt that the smoker who can simply drop his habit without a second thought is an addict, so, too, do we have reason to doubt that the partner who could abandon his ‘love’ at will ever really loved at all. Interestingly, Harry Frankfurt suggests that the absence of volitional control over one’s love is both a constitutive and desirable feature of it (Frankfurt 2004: 44–7). Engagement with our attachment figures often feels non-­ optional, and this sense of necessity may contribute to love’s value. Thus, if suffering and diminished self-­control are the measures, then not only addiction, but loving attachments too, would be problematic. Perhaps then addictions are ultimately objectionable, when they are, because they involve improper valuing orientations toward their objects. Jesse Summers argues that ‘what is wrong with addiction’ is that it involves a misvalue of the importance of its object (2015: 33). Summers advocates making this a defining feature of addiction. On his preferred view, to say that one is addicted is to say 30  Interestingly, theorists working in the Ancient Stoic, Buddhist, and Daoist traditions raised similar concerns against attachments more broadly. For further discussion on this point, see Wonderly (2016). 31  Some philosophers, for example, have theorized that addiction undermines autonomy and self-­ control in virtue of subjecting the addict to impulses that are wholly (or almost wholly) irresistible. For examples, see Charland (2002) and Elliott (2002). Others have rejected irresistible impulse models of addiction in favour of other approaches, such as the disordered appetite account and the ego-­ depletion model (Watson 1999 and Levy 2006). For a detailed, informative critique of the view that addictions are compulsive, see Pickard (2015).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

238  Monique Wonderly that based on a pattern of behaviours, she values the object of her addiction ‘more than she should’ (2015: 33). I don’t know that all addictions have this feature, but it seems reasonable to think that some do and are problematic on that account. There may be, for example, some addiction objects that merit little or no value, in which case it seems that one ought not to value them. Think here of substances that are actively harmful and provide only fleeting, superficial benefits to their users. Notice, too, that (arguably) unlike vulnerability to suffering and diminished self-­control, we have little reason to think that an improper valuing orientation is a constitutive feature of healthy, loving interpersonal attachments. It seems rea­ son­ able to think that any attachments—including interpersonal ones—that involve placing high value on an object of little worth are problematic.32 There is another vice of valuing, however, that might occur in both substance addiction-­attachments and interpersonal attachments. To see this, it will be helpful to consider an addiction that seems, on the surface, very positive. Watson’s remarks on ‘Substance S’ will be instructive here. Watson describes (the hypothetical) S as a severely addictive substance to which one can become dependent while still leading a healthy and productive life (1999: 18–19). He asks readers to suppose that, ‘in a certain culture, otherwise similar to ours, the use of S is not only tolerated, but respected as highly spiritually beneficial. This culture regards the dependency on this substance, which is to say, the vulnerability to various kinds of diminished self-­control, as a small price to pay for the enrichment of human life provided by S. Fortunately, S is easily obtainable, perhaps even sub­sid­ ized by the society for religious reasons’ (1999: 18–19). On Watson’s view, it is not immediately clear why such an addiction would be objectionable. If one’s addiction to S is objectionable, it is not because the substance lacks value. One might think that drug addictions—even to beneficial substances such as S—might be problematic not because of the (low) value of their objects, but because of the way in which addicted persons value them. Since S is easy to access and S use is legal and socially accepted, the S addict does not need to dedicate inordinate amounts of time, energy, and other resources to practising her addiction, but she would do so readily under slightly different conditions. It’s tempting, then, to say that despite its high value, the addicted agent still values her addiction object ‘too much’. 32  Of course, one might doubt that interpersonal attachments admit of the vice, ‘valuing that which is not worthy of value’. On such a view, even the most troublesome of people are valuable qua persons and might be fitting objects of attachments on that account. A related point is that it isn’t clear that attachments respond to the value of their objects. Theorists have argued that love, for example, typically bestows rather than responds to value (see Singer 1991: 273 and Frankfurt 2004: 39). There is, however, room for debate here. Even if one grants that attachments typically confer, rather than track (inherent) value, one might think that attachments to ‘unsuitable’ objects are valueless and thus problematic (see e.g. Raz 2001: 16–18). Thanks to Edward Harcourt for raising this issue.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

Attachment, Addiction, and Vices of Valuing  239 Sometimes, when we accuse an individual of valuing something too much, what we mean to say is that she values it in such a way that prevents her from engaging with other things worth valuing, or more precisely, other things that she should and does value. The idea that addictions often ‘crowd out’, or preclude our abilities to properly engage with, other values is a common one in the addiction literature (Elster  1999; Watson  1999). When an attachment object garners too much of one’s attention, it can leave other aspects of one’s evaluative life impoverished by comparison. Perhaps S is troubling for this reason. But more needs to be said. Merely valuing the object in such a way that crowds out engagement with other values would not suffice to render an attachment problematic. Consider, for example, that a father may value his child to such a degree, or in such a way, that he is able to spend little time on his manuscript, his garden, or his volunteer work—even while he values those things as well. Because typically many things matter to a single agent, she will have to compromise her engagement with some. Quite often, she will give priority to her strongest attachments, and this is not obviously wrong. Thus, if the S user prioritizes S over most other valued objects in her life, it is not yet clear that she does something objectionable. But this doesn’t settle the issue. When one makes the clear-­eyed choice to give or (to assent to giving) priority to an attachment object, this seems like a fine thing to do. Yet, in the case of addiction, the agent’s focused attention on the object sometimes not only causes her to compromise other values, but, in a sense, it blinds her to them. Her mode of engagement is bleary-­eyed and narrow, and she is no position to recognize properly what she forgoes on account of her addiction. I doubt that all addictions function this way, but there is reason to think that some common (and relatively benign) addictions negatively impact our capacity to value. Evidence suggests that (some) addictions not only impact reasons responsiveness mid-­craving, but that they can impair one’s general capacity to be moved by certain kinds of reasons associated with valuing. As Gideon Yaffe explains in his (2011), one thing that addicts may not be able to do is to sufficiently ‘update their algorithms for determining what reasons they have to pursue various acts in light of the factual information they have about those acts and outcomes’ (2011: 128). This capacity is linked to our capacity for learning what value to attach to particular acts and outcomes in light of experience (2011: 134). One hypothesis is that repeated, over-­activation of the dopamine system causes structural changes in areas of the brain associated with assigning values to predicted and actual outcomes. This may explain why, even when sated, addicts often have trouble recognizing the salience and force of certain kinds of reasons. Recent research on addicts’ responses to fictive errors supports this point. Decreased reason/value responsiveness is evident even in relatively controlled addictions like smoking (see e.g. Chiu et al. 2008). Just as this feature may not be readily apparent— but nonetheless active—in smoking addictions, it might also be active in addictions to beneficial substances, like Watson’s Substance S.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

240  Monique Wonderly If Substance S addictions have this feature, then I think we would have reason to count them as objectionable. Insofar as we value ourselves as valuers, engagement with an object in such a way that impairs our abilities to recognize values— or again, to form new values—would seem to constitute a significant harm. Since some addictive substances directly and artificially stimulate the dopamine system, we may be more susceptible to this kind of problematic valuing orientation in the case of chemical addictions than in interpersonal attachments. However, recall that interpersonal attachments involve dopamine system activation as well, and add to this that the phenomenology of attachment often has a ‘crowding out’ component.33 It seems at least possible that interpersonal attachments, as well as substance addictions, can render us vulnerable to orientations of this kind. This, then, could be a ‘vice of valuing’ to which we are susceptible in attachments more broadly, including interpersonal attachments—though probably not to the same degree. One crucial difference between an attachment to a drug and an attachment to a person is that in the latter case, one’s attachment object can actively respond to one’s need. For example, if an individual is attached to someone who cares about her, then the attachment figure will actively mitigate some of the risks associated with attachment. For example, by regularly attending to one’s attachment needs, an attachment figure can prevent the suffering that would result from rejection and separation and reduce the negative impact of diminished self-­control by increasing one’s self-­confidence and self-­reliance.34 Finally, a caring attachment figure might redirect the attached party’s inclination to focus on her to other things, as this would better conduce to the attached agent’s own flourishing. In this way, an interpersonal attachment figure, unlike a drug, can mitigate the likelihood of one succumbing to the problematic valuing orientations described above. While I take it that neither addictions, nor attachments more broadly, are bad as such, I have argued that both might render us vulnerable to certain vices of valuing. If an agent values her attachment object in such a way that she cannot properly recognize values, or form new ones, then her orientation seems problematic on that account. Of course, healthy, interpersonal attachments, unlike attachments to substances, often have internal safeguards against it. Though this would not 33  While some might be tempted to think the relevant neurobiological effects are limited to chem­ ical addictions, it is worth noting that recent neuroscience research suggests that gambling addiction is associated with some of the same structural and functional brain abnormalities observed in stimulant addictions. Some of these abnormalities include, for example, elevated dopamine levels, abnormal activity in the ventral striatum, reduced subcortical volume, and diminished neural responsiveness to cues relevant to evaluative decision-­making (Whiting et al. 2019: 176–7). 34  Supportive interactions with attachment figures imbue us with a sense of security that enables us to persevere through difficult circumstances even without the physical presence of our attachment figures. Mikulincer and Shaver explain, ‘In adulthood, an attachment strategy . . . can also include activation of mental representations of relationship partners who regularly provide care and protection. These representations can create a sense of safety and security, which helps a person deal successfully with threats . . .’ (2016: 12).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

Attachment, Addiction, and Vices of Valuing  241 preclude the need to be vigilant in attachments of all kinds, it does suggest an important respect in which attachments to non-­persons might be inferior to interpersonal attachments. ~ ~ ~ Though in terms of human flourishing, interpersonal attachments and addictions seem to reside at opposite ends of the pole, theorists have identified a number of commonalities that unite them. While there are many ways of conceptualizing the relationship between addiction and attachment, I have argued for the relatively neglected view that some addictions represent attachments to their objects. In the final section, I suggested that exploring the ethics of addiction could potentially illuminate a particular set of problematic valuing orientations to which we are vulnerable in attachments of all kinds, including ordinary interpersonal relationships.

References AA. 2001. Alcoholics Anonymous: The Big Book (4th edn). New York: Alcoholics Anonymous World Services, Inc. Ainsworth, M. D. S. 1988. ‘On security’. Unpublished discussion paper prepared for the Foundations of Attachment Theory Workshop. Convened for the New York Attachment Consortium by G.  Cox Steiner and E.  Waters. Port Jefferson, NY. . Ainsworth, M.  D.  S., Blehar, M.  C., Waters, E., and Wall, S. 1978. Patterns of Attachment. New York: Erlbaum. American Psychiatric Association (APA). 2013. Diagnostic and Statistical Manual of Mental Disorders (5th edn). Arlington, VA: American Psychiatric Association. Blatz, W. 1966. Human Security: Some Reflections. Toronto: University of Toronto Press. Bowlby, J. 1969. Attachment and loss, vol. 1: Attachment. New York: Basic Books. Bretherton, I. 1991. ‘The roots and growing points of attachment theory’. In C. M. Parkes, J. Stevenson-Hinde, and P. Marris (eds), Attachment Across the Life Cycle, 9–32. New York: Routledge. Burkett, J. P., and Young, L. J. 2012. ‘The behavioral, anatomical and pharmacological parallels between social attachment, love and addiction’. Psychopharmacology 224(1), 1–26. Cassidy, J. 2008. ‘The Nature of the Child’s Ties’. In Cassidy, J., and Shaver, P. R. 2008. Handbook of Attachment: Theory, Research, and Clinical Applications (2nd edn), New York: Guilford Press, 3-22. Cassidy, J., and Shaver, P.  R. 2008. Handbook of Attachment: Theory, Research, and Clinical Applications (2nd edn), New York: Guilford Press.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

242  Monique Wonderly Charland, L. 2002. ‘Cynthia’s Dilemma: Consenting to Heroin Prescription’. American Journal of Bioethics 2(2): 37-47. Chiu, P. H., Lohrenz, T. M., and Montague, P. R. 2008. ‘Smokers’ brains compute, but ignore, a fictive error signal in a sequential investment task’. Nat. Neurosci. 11(4), 514–20. Cihan, A., Winstead, D. W., and Laulis, J. 2014. ‘Attachment Theory and Substance Abuse: Etiological Links’. Journal of Human Behavior in the Social Environment 24: 531-37. Collins, N., Guichard, A., Ford, M., and Feeney, B. 2006. ‘Responding to need in in­ tim­ ate relationships: normative processes and individual differences’. In M. Mikulincer and G. S. Goodman (eds), Dynamics of Romantic Love: Attachment, Caregiving, Sex, 149–89. New York: Guilford Press. Cornish, A. 2016. ‘Can Baltimore provide addiction treatment on demand?’ NPR. 2/24/2016. . Dill, B., and Holton, R. 2014. ‘The addict in us all’. Frontiers in Psychiatry 5, 1–20. Eagle, M.  N. 1981. ‘Interests as object relations’. Psychoanalysis and Contemporary Thought 4(4), 527–65. Earp, B.  D., Wudarczyk, O.  A., Foddy, B., and Savulescu, J. 2017. ‘Addicted to love: what is love addiction and when should it be treated?’ Philosophy, Psychiatry, & Psychology 24(1), 77–92. Elliott, C. 2002. ‘Who Holds the Leash?’ American Journal of Bioethics 2(2): 48. Elster, J. 1999. Strong Feelings: Emotion, Addiction, and Human Behavior. Cambridge, MA: MIT Press. Epstein, D. 2015. ‘Your love is a drug: 20 great narcotic love songs’. Rolling Stone 23 October. . Fisher, H. E., Brown, L. L., Aron, A., Strong, G., and Mashek, D. 2010. ‘Reward, addiction, and emotion regulation systems associated with rejection in love’. Journal of Neurophysiology 104(1), 51–60. DOI:10.1152/jn.00784.2009. Flanagan, O. 2011. ‘What is it like to be an addict?’ In J. Poland and G. Graham (eds), Addiction and Responsibility. Cambridge, MA: MIT Press. Flanagan, O. 2013. ‘The shame of addiction’. Frontiers in Psychiatry 4(120), 1–11. Flores, P.  J. 2004. ‘Addiction as an attachment disorder’. Lanham, MD: Rowman & Littlefield. Foddy, B., and Savulescu, J. 2007. ‘Addiction is not an affliction: addictive desires are merely pleasure-oriented desires’. American Journal of Bioethics 7(1), 29–32. Foddy, B., and Savulescu, J. 2010. ‘A liberal account of addiction’. Philosophy, Psychology, and Psychiatry 17(1), 1–22. Frankfurt, H. 2004. The Reasons of Love. Princeton: Princeton University Press.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

Attachment, Addiction, and Vices of Valuing  243 Frascella, J., Potenza, M. N., Brown, L. L., and Childress, A. R. 2010. ‘Carving addiction at a new joint? Shared brain vulnerabilities open the way for non-substance addictions’. In G. Uhl (ed.), Addiction Reviews. New York: New York Academy of Sciences. Hanninen, V., and Koski-Jannes, A. 1999. ‘Narratives of recovery from addictive behaviors’. Addiction 94(12), 1837–48. Harcourt, E. 2017. ‘Attachment, autonomy and the evaluative variety of love’. In E. Kroeker and K. Schaubroeck (eds), Love, Reason and Morality, 39–56. London: Routledge. Hari, J. 2015. Chasing the Scream: The First and Last Days of the War on Drugs. New York: Bloomsbury. Helm, B. 2009. ‘Love, identification, and the emotions’. American Philosophical Quarterly 46, 39–59. Insel, T. 2003. ‘Is social attachment an addictive disorder?’ Physiology & Behavior 79(3), 351–7. Jaworska, A. 2007. ‘Caring and internality. Philosophy and Phenomenological Research 74(3), 529–68. Kant, I. 1784–5/1997. ‘Moral philosophy: Collin’s lecture notes’. In Lectures on Ethics. Cambridge Edition of the Works of Immanuel Kant. Ed. and trans P.  Heath and J. B. Schneewind, 37–222. Cambridge: Cambridge University Press. Kant, I. 1798/2006. Anthropology from a Pragmatic Point of View. Ed. and trans. R. Louden. Cambridge: Cambridge University Press. Kennett, J. 2013. ‘Just Say No? Addiction and the Elements of Self-control’. In N. Levy (ed.), Addiction and Self-Control: Perspectives from Philosophy, Psychology, and Neuroscience, 144-64. New York, NY: Oxford University Press. Kennett, J., and McConnell, D. 2013. ‘Explaining addiction: how far does the reward account of motivation take us?’ Inquiry 56(5), 470–89. Khantzian, E. J. 1985. ‘The self-medication hypothesis of addictive disorders: focus on heroin and cocaine dependence’. American Journal of Psychiatry 142, 1259–64. Knapp, C. 1996. Drinking: A Love Story. New York: The Dial Press. Levy, N. 2006. ‘Autonomy and Addiction.’ Canadian Journal of Philosophy 36(3): 427-47. Lewis, M. 2013. Memoirs of an Addicted Brain. New York: Public Affairs. Lewis, M. 2015. The Biology of Desire. New York: Perseus. Loewenstein, G. 1999. ‘The visceral account of addiction’. In J.  Elster and O.  Skog (eds), Getting Hooked: Rationality and the Addictions, 235–64. Cambridge: Cambridge University Press. Maslow, A. 1942. ‘The dynamics of psychological security-insecurity’. Journal of Personality 10(4), 331–44. Maté, G. 2008. In the Realm of Hungry Ghosts: Close Encounters with Addiction. Toronto: Vintage Canada Press.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

244  Monique Wonderly Mercer, J. 2006. Understanding Attachment: Parenting, Child Care, and Emotional Development. Westport, CT: Praeger. Mikulincer, M., and Shaver, P. 2016. Attachment in Adulthood: Structure, Dynamics, and Change (2nd edn). New York: Guilford Press. Orford, J. 2001. ‘Addiction as excessive appetite’. Addiction 96, 15–31. Ovid. 1977. Heroides and Amores. Trans. G. Showerman, rev. G. P. Goold. Cambridge, MA: Harvard University Press. Peele, S., and Brodsky, A. 1975. Love and Addiction. New York: Taplinger. Pickard, H. 2012. ‘The purpose in chronic addiction’. AJOB Neuroscience 3(2), 40–9. Pickard, H. 2015. ‘Psychopathology and the ability to do otherwise’. Philosophy and Phenomenological Research 90(1), 135–63. Pickard, H. 2019. ‘The puzzle of addiction’. In H. Pickard and S. H. Ahmed (eds), The Routledge Handbook of Philosophy and Science of Addiction, 9–22. New York: Routledge. Pickard, H., and Pearce, S. 2013. ‘Addiction in context: philosophical lessons from a personality disorder clinic’. In N. Levy (ed.), Addiction and Self-Control: Perspectives from Philosophy, Psychology, and Neuroscience, 165–89. New York: Oxford Uinversity Press. Raz, J. 2001. Value, Respect, and Attachment. Cambridge: Cambridge University Press. Reid, S. 2016. ‘Junkie’. In L. Crozier and P. Lane (eds), Addicted: Notes from the Belly of the Beast, 255–76. British Columbia: Greystone Books. Schore, A.  N. 1994/2016. Affect Regulation and the Origin of the Self. New York: Routledge. Seeburger, F. 1993. Addiction and Responsibility: An Inquiry into the Addictive Mind. New York: Crossroads. Singer, I. 1991. ‘From The Nature of Love’. In R. C. Solomon and K. M. Higgins (eds), The Philosophy of (Erotic) Love, 259–78. Lawrence: Kansas University Press. Summers, J. 2015. ‘What is wrong with addiction?’ Philosophy, Psychiatry, & Psychology 22(1), 25–40. Unterrainer, H., Hiebler-Ragger, M., Koschutnig, K., Fuchshuber, J., Tscheschner, S., Uri, M., Wagner-Skacel, J., Reininghaus, E., Papousek, I., Weiss, E., and Fink, A. 2017. ‘Addiction as an Attachment Disorder: White Matter Impairment Is Linked to Increased Negative Affective States in Poly-Drug Use’. Frontiers in Human Neuroscience 11: 1-11. Watson, G. 1999. ‘Disordered appetites: addiction, compulsion, and dependence’. In J. Elster (ed.), Addiction: Entries and Exits, 3–28. New York: Russell Sage Foundation. Weiss, R. S. 1991. ‘The attachment bond in childhood and adulthood’. In C. M. Parkes, J.  Stevenson-Hinde, and P.  Marris (eds), Attachment Across the Life Cycle, 66–76. New York: Routledge. West, R., and Ubhi, H. 2019. ‘Addictive behaviors’. In C.  D.  Llewellyn et al. (eds), Cambridge Handbook of Psychology, Health, and Medicine (3rd edn), 46–50. Cambridge: Cambridge University Press.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

Attachment, Addiction, and Vices of Valuing  245 Whiting, S., et al. 2019. ‘Gambling disorder’. In H. Pickard and S. H. Ahmed (eds), The Routledge Handbook of Philosophy and Science of Addiction, 173–81. New York: Routledge. Wonderly, M. 2016. ‘On being attached’. Philosophical Studies 173(1), 223–42. Wonderly, M. 2017. ‘Love and attachment’. American Philosophical Quarterly 54(3), 232–50. Wonderly, M. 2019. ‘Early relationships, pathologies of attachment, and the capacity to love’. In A. Martin (ed.), The Routledge Handbook of Love in Philosophy, 23–34. New York: Routledge. Yaffe, G. 2011. ‘Lowering the bar for addicts’. In J.  Poland and G.  Graham (eds), Addiction and Responsibility, 269–92. Cambridge, MA: MIT Press.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 27/07/21, SPi

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 26/07/21, SPi

Index 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. Achieved excellence  115–17 Adaptation 12–13 Adaptation, evolutionary vs. functional  88 Addiction Motivational architecture of  229–30, 235–6 And vice  236 ADHD  75, 215 Adult Attachment Interview (AAI)  57 Affectionless psychopathy  75 Agential competence  229, 233–5 Ainsworth, Mary  1–3, 75–6, 128, 174–6, 228–9 Altruism  199, 211–12 Antisocial behaviour  78–80 Sex differences in the explanation of  79–80 Anxiety 75 Aquinas, St Thomas  157–72 Aristotle  1, 4–7, 59, 71–4, 78, 81–3, 105, 124, 129–30, 138, 140–1, 155–7 Attachment Attachment status as the attribute of a relationship 76 Attachment theory Individual difference component  1–2 Generic component  1–2, 7–8 Attachment types Secure  2–5, 76–7, 94, 112–13 And academic achievement  116–17 Insecure-ambivalent  2–3, 110–11 Insecure-avoidant 2–3 Disorganized 213–16 And evolution  11–12, 128, 227–8 ‘Secure attachment +’  113 Warmth vs Companionship attachment  90–1 Love and Attachment see Love Attention  145, 157, 163, 168–9 Augustine, St  139–40, 145 Autism  210, 226n.7 Biocommunity 97–8 Biological predispositions  19 Blame 192–3 Bloom, Paul  19

Bowlby, John  1, 74–5, 125–6, 128, 140, 179, 183n.5, 192–3, 214–15 Breastfeeding  91–3, 95, 192–3 Callous unemotional traits  76–9 Caregiving  2–3, 45–6, 59–61, 74–5, 147, 197, 214 Character Character trait  7, 11–12, 45, 47, 57–8 Character, concept of  81–2 Character, establishment of  4, 72 Characterology 138 Choice, autonomous  187–8 Co-regulatory support  24 Conscience 26–7 Cooperation  5–7, 45–6, 51, 60, 137–8, 141, 150, 197, 205, 218 Cooperativeness see Cooperation Coping skills  29 Cortical thickness  53–5 Cortisol 92 Curiosity see Exploration Darwall, Stephen  184, 186–8 Darwin, Charles  87, 94, 128, 134 Davidson, Donald  181 Dawkins, Richard  89, 124 Depression  6, 75 Developmental psychopathology  209, 218–19 Differential susceptibility to social influences  58–9, 82 Dopamine system  225–6, 239–40 Dysregulation  91, 96–7, 128–9 Emotion regulation  3, 5, 25–6, 29–30, 44–5, 128, 138, 226, 226n.9 Emotion understanding  25–6, 29–30, 33–4 Empathy Empathy, general  23, 26, 30, 33–4, 93–4, 112–14, 118, 176, 190–1 Cognitive empathy  48–9 Diachronic Empathy Indispensability Thesis 209 Empathic attunement  202–3

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 26/07/21, SPi

248 Index Empathy (cont.) Empathic concern  203–5 Empathic distress  203 Empathic resonance  197, 202, 214–17 Empathic responsiveness  197–8, 200, 206, 210–12 Leiden Longitudinal Empathy Study  56–7 Synchronic Empathy Indispensability Thesis 204–9 Empirical and philosophical methods, relationship between  4–5, 71, 119n.19, 120–1, 138, 174 Epigenetics 134 Erikson, Erik  125, 127 Ethology 89–90 Eudaimonia see Human Flourishing Evolved developmental niche  91, 91n.2 Evolved nest  91–2, 94, 96–8 Executive function  137–8 Exploration  12–13, 76–7, 128, 149, 154, 156–7, 182–3 Fairness  4–5, 142–3 Fidelity to promises  149–50 First nature  7–8, 108 Flourishing life see Human Flourishing fMRI 209–10 Foot, Philippa  10, 106–8, 183–4 Formula milk  92, 95, 192–3 Frankenstein’s monster (‘Creature’)  195–6 Friendship 141–2 Freud, Sigmund  1, 75, 144, 147, 151 Furious clinging  2–3 Fusion in Infancy  2–3, 189 Gender  50, 93–4, 176–7 Genetic fallacy  190 Genome-wide Complex Trait Analysis (GCTA)  51 Good For an organism  9–10, 105, 107 As an organism of its kind  107, 124 Particular vs universal  159–61 Guilt  35, 93 Habit 156–7 Habituation  11–12, 59, 155–6 Active 156 Passive 156 Haidt, Jonathan  19 Honneth, Axel  189 Human development, idealized model of  125, 127 Human flourishing  107, 112, 115–16, 183–4, 240–1

Human form  106, 120 Human nature Evolutionary perspective on  123, 134–5 Freudian views of  1 Perfectibility of  125, 134 Hume, David  168–9, 198–9, 202n.4 Hunter-gatherers 91 Hursthouse, Rosalind  106, 108n.8, 109 Identity-indifference 148 Impartiality 147–8 Imprinting 74–5 Impulsiveness 214 Inclusive fitness  124, 129–30 Incontinence 167 Intentional stance  73 Internal working models  22–3, 25, 28–30, 32–3, 36–7, 74–6, 179, 190 Internalization as a theory of moral learning  18, 20, 32–3 Internalized others  151–2 Interpersonal norms  184, 192–3 Intimacy Intimacy, capacity for  14, 127–8, 195–7, 216–19 Intimacy, virtues of  140–2, 150 Just-community approach to moral education  59–60 Justice 143 Kant, Immanuel  18–19, 165n.8, 186–8, 236n.29 Kohlberg, Lawrence  18–19, 24 Korsgaard, Christine  217–18 Life form  106–7 Life-history theory  130 Loss  226–7, 235 Love Love and Attachment  144, 146n.5, 147 Love and Benevolence  145 Love comedies  139–40 Love, duties of  184–5 Love as just vision  145–6 Love, non-substitutability of objects of  145, 147, 228–9 Love as a passion  160–1 Love, romantic  224–8, 237 Love and Virtue  141 Main, Mary  1 Maslow, Abraham  127 Maternal deprivation  1, 75 Maternal sensitivity  24–5, 30–1, 75–6

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 26/07/21, SPi

Index  249 Mentalization 213–17 Mind-mindedness 181 Mind-reading see Psychological Understanding Minnesota Study of Risk and Adaptation from Birth to Adulthood  3, 176 Mirroring 214–17 Mistrust 96 Moral community 186 competence  195, 197–8, 200, 206, 208–9, 213–14 development, relationally oriented theory of  20 emotions  195–6, 198, 218–19 motivation  199–200, 202–4, 206 obligations  197, 207–8 reasoning 18–19 self in young children  27, 29 Mother-infant synchrony  197 Murdoch, Iris  144–5, 151–2 Narvaez, Darcia  112–15, 192n.9 Natural defects  183–4 Natural goodness, moral virtue as a type of  106–9, 111, 119 Naturalism, ethical  7–8, 106–9, 113, 120–1 Neurology see neuroscience Neurophysiology see neuroscience Neuroscience  225–6, 234, 240n.33 NICHD Study of Early Child Care  78 Non-rational part of the soul  157–8 Nudge 57–8 Optimal development  10–11, 90–1, 118–20, 128–30, 197, 213–14, 216, 218 Oxytocin  60–1, 92, 176–7 Parental responsiveness see Caregiver Responsiveness Parenting Parent-child conversation  36–7 Parental discipline practices  36, 56 Parental sensitivity  3, 9, 44–5, 48, 53, 56, 58–9, 76–7, 81 ‘Tiger mom’ approach to  105–6, 117, 119 Passions education of  157–8 misdirected 168 rationality of  158–60 well trained  171 Paternalism 187–8 Personality disorder Antisocial personality disorder  73

Borderline Personality Disorder  181, 215 Personality disorders in general  82 Personality trait  19–20, 22, 29–30, 44–6, 57–60, 62 Perspective-taking 93–4 Persuasion 162–3 Piaget, Jean  18–19, 24 Premoral sensibility  33–5 Proper development see Optimal development Prosocial behaviour Association with secure attachment  5, 28, 56–7, 178 Definition of  5–7, 27–8, 45–7, 150–1, 178 Genetic factors in explaining individual differences in  51–3 Motives for  178 Contextual influences on  59–60 Prosocial Cyberball game  47, 55 Varieties of Donating  51, 54–6, 58–9, 62–3 Helping  28, 34–5, 45–50, 56–62, 175–6, 178 Sharing 27–8 Volunteerism  45, 61 Prosociality see Prosocial behaviour Psychological understanding  32, 34–5, 48–9, 54, 181–2, 201–2, 205–6, 210, 212n.9 Psychopathy  75–9, 209–11 Puberty hypothesis  131–3 Rational will  106, 108, 157–9, 162 Reality of others  215–18 Reason, practical  168–9, 171–2 Reasoning, practical  216, 218–19 Reasons for action  5–6, 142–3, 151–2 responsiveness to  149–50, 180–3, 187–8 Reciprocal altruism  45–7, 124 Relational Attunement  93–4, 97 capacities 174 consequences of parenting quality  76–7 influences on early development of moral conduct  20, 31, 37–8, 190 nature of morality  174, 191n.8 obligations  184–6, 188–9 Reproductive strategies  123, 129–34 Respect  186–7, 214 Responsiveness Affective responsiveness  201–2, 210–11 Caregiver responsiveness  30–1, 92–4, 174–6, 179–81 Sensitive responsiveness  3, 7, 22, 24, 30, 32–3, 38, 175–6

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 26/07/21, SPi

250 Index Second nature  7–8, 108 Second-personal Relationships 184 Reasons 185–92 Standpoint 186 Secondary-drive theory  1 Secure base  154, 227–8 Self-actualization 125–7 Self-regulation  23–4, 93–4, 96, 151, 197, 215, 218–19 Self-reliance  12–13, 175–6 Separation anxiety  140–1 Sexual maturation  131–3 Shame  35, 73–4, 195–6, 198, 206, 232n.21 Shand, A.F.  160 Shelley, Mary  195–6, 198–9, 215–16 Sidgwick, Henry  143, 147 Simple Nucleotide Polymorphism (SNP)  51, 53 Situationism  7, 44–5, 62, 80 Smith, Adam  198–9 Socialization, as a theory of moral learning  18, 35–7 Species-typical human development  90–1, 93–5, 97 Stable dispositions  6 Strange Situation  2–3, 75–6, 78–9, 176–7

Strangers  175–6, 178–9, 185, 188–9 Substance addiction  224–6, 230–1, 233, 238 Telomeres 133–4 Theory of mind see Psychological Understanding Thompson, Michael  106–8, 112–13 Transcutaneous Vagus Nerve Stimulation (tVNS) 60 Triangulation 181 Utilitarianism  18, 191 Vagus nerve  113 Vices of valuing  224–5, 236–41 Virtue Virtue, Context-dependence of  124–5, 127–9, 134 Virtue and Love see Love Virtue ethics  1, 4, 38 Wallace, R. Jay  174, 184–6 Watson, Gary  226–7, 227n.10, 230n.18 WEIRD people  123, 129 Well-being  186–8, 229, 233 Will, the see Rational will Wirral Child Health and Development Study 76–7