Attachment Theory, Character and Naturalism

Attachment Theory, Character and Naturalism Edward Harcourt Faculty of Philosophy, University of Oxford, and Keble College, Oxford OX1 3PG, UK Email: [email protected] published in Julia Peters (ed.), Aristotelian Ethics in Contemporary Perspective (London: Routledge, 2013), pp. 114-129 1 25/07/2013 0. The subject-matter of this paper falls within the triangle marked out by three Aristotelian ideas - human nature, human excellence, and human well-being. The paper is a highly programmatic attempt to introduce some material from developmental psychology – specifically, from attachment theory – and to explain why philosophers working somewhere within the Aristotelian triangle have reason to take more of an interest in it than they do now. I am assuming of course that attachment theory is true: if it isn’t, or doesn’t make some reasonable claim to be believed, there’s no reason for anyone to take an interest in it. But if you grant that much – and it is one of the leading theoretical orientations in developmental psychology1 - there are lots of true theories: why should neo-Aristotelians be interested in this one? Attachment theory is a theory of child development. Indeed properly speaking it is a theory of human development, but - partly because children are easier to study than adults, partly because childhood experience may be especially important in making us the way we are - attachment theorists have taken a special interest in the early years. This is already one, rather general, reason why neo-Aristotelians should be interested in the theory. For Aristotle’s ethics is (in part) a developmental theory: it aspires to provide not only a theory about what human excellence is, but a theory about how we acquire or fail to acquire it. And Aristotle himself says that because good character is produced by habituation, 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.2 So 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. But there is a more specific reason than this. Ethics in the neo-Aristotelian mould sees itself as a naturalistic undertaking: that is, it seeks to locate ethical life in the world as made intelligible to us by natural science. Moreover there is reason to think that the proper form of such an undertaking is to display the continuity between our second and our first, or between our ethical and our psychobiological 1 See H. Rudolph Schaffer, Key Concepts in Developmental Psychology (London: Sage, 2006), 160. 2 Aristotle, Nicomachean Ethics, trans. W.D. Ross, revised J.O. Urmson. In Aristotle: The Complete Works. Electronic edition. BOLLINGEN SERIES LXXI · 2 Volume II, 1103b24, 1743. 2 25/07/2013 natures.3 Now attachment theorists claim that some human traits described by the theory arise through natural selection. Those traits, then, would belong to our first natures, or to ‘an account of human beings which is to the greatest extent possible prior to ideas of the ethical’.4 But attachment theory is also a taxonomy of psychological dispositions, plus a theory about why people (children, adults) have the disposition(s) of that sort which they do. Furthermore, these psychological dispositions appear to stand in an explanatory relation to some traditional traits of character, that is, to some virtues and vices. Attachment theory therefore looks as if it is well equipped to put some empirical flesh on the bones of Aristotelian naturalism, by making vivid the continuity I have said this variety of naturalism demands. So far it looks as if I envisage the flow of ideas as being entirely from attachment theory to neo-Aristotelian ethics. But that’s not so. As far as I know, attachment theorists are unaware of the ways in which different forms of naturalism are debated within ethics, and might be surprised at the thought that their work had anything to do with ethics. But perhaps precisely because they think that what they are up to is just psychology or evolutionary biology (i.e. some sort of natural science), some of them have apparently signed up unawares to a version of ethical naturalism which goes much further than the continuity thesis I mentioned earlier. They are in distinguished philosophical company here – the late Philippa Foot’s, for example.5 But I shall argue that Foot’s ambitious version of neo-Aristotelian naturalism and the versions of attachment theory that unwittingly subscribe to it both run into difficulties. Thinking through the connections between attachment theory and neo-Aristotelian ethics, then, should be a way both of reining in some of 3 For senses of ‘naturalism’, see John McDowell, “Two Sorts of Naturalism,” in John McDowell, Mind, Value and Reality (Cambridge MA & London: Harvard University Press, 1998), 167-197; and Bernard Williams, “Naturalism and Genealogy,” in Morality, Reflection and Ideology, ed. E. Harcourt (Oxford: Oxford University Press, 2000), esp. 148-150. For the idea of second nature, which but for the term itself McDowell ascribes to Aristotle, see McDowell, “Two Sorts of Naturalism,” 184, and Mind and World (Cambridge MA: Harvard University Press, 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 question’ (‘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, “Naturalism and Genealogy,” 154, 150). See also Bernard Williams, “Replies”. In World, Mind and Ethics, ed. J.E.J. Altham and R. Harrison (Cambridge: Cambridge University Press, 1995), 203 (‘a conception of ethics … continuous with our understanding of human beings in other respects’). 4 Williams, “Naturalism and Genealogy,” 154. 5 See P. Foot, Natural Goodness (Oxford: Oxford University Press, 2001). 3 25/07/2013 attachment theory’s own more extravagant theoretical ambitions, and of demonstrating both the prospects (the continuity thesis itself) and some of the limits of naturalism in ethics. 1. I shall return to the continuity thesis shortly, but let me focus for now on the ambitious version of neo-Aristotelian ethical naturalism, as exemplified by Foot. In this version of naturalism, claims about human virtues and vices, or excellences and defects, are identified with claims about the way human beings should be if they are to be good of their kind, or properly suited to lead our characteristic species life. Here - in contrast to the kind of naturalism which Moore is famous for opposing, a species of reductionism in which moral properties are, one way or another, identified with ‘natural’ ones - it isn’t a matter of identifying one kind of property with another: only one kind of property, the excellences and defects themselves, is ever under discussion. Nor is it a matter of deriving surprising claims about which properties are excellences or defects in humans: it’s assumed we will more or less agree on that at the start. Rather it’s a matter of arguing that these interesting properties are the virtues and vices because they have a further property, that of playing a certain kind of role, namely, that they are necessary (in the case of the virtues) for us to lead the kind of life which, as members of that species, we’re supposed to lead. Just as it is an ‘Aristotelian necessity’6 – a necessity which ‘determine[s] what it is for members of a particular species to be as they should be’ - ‘for plants to have water or birds to build nests,’7 so ‘for human beings the teaching and following of morality is something necessary. We can’t get on without it’; again, ‘getting one another to do things without the application of physical force [and which morality accomplishes] is a necessity for human life.’ Correspondingly, a species member who does not do what it is necessary for the species to do is ‘naturally defective’, so immoral human beings are defective in just the same sense as birds who fail to build nests, or owls who cannot see in the dark. And because it’s a matter of plain fact that virtues are necessary for us, it is also a matter of plain fact that a human being with a given virtue is excellent, or with a given vice defective. Thus Foot’s ambitious naturalism is designed to fulfill the cognitivist meta-ethical ambition common to many ethical naturalisms. Contrast the continuity thesis, whose relation to cognitivism is looser, and won’t be discussed further here. 6 7 Foot, Natural Goodness, 15. Foot, Natural Goodness, 15. 4 25/07/2013 This ambitious form of naturalism ought to interest empirical investigators who address questions about the way human beings should be or - transposing the same idea into a developmental idiom – about optimal development,8 attachment theorists included. But there’s an interesting difference. There’s no difficulty in getting moral philosophers to recognize that what they’re working on are virtues and vices, that is, ethical notions. The controversy arises when ambitious naturalists try to get people to agree that these ethical notions are also ‘natural’ notions – that is, that virtue consists is a perfection of our first nature, or of our nature as we can make it intelligible to ourselves ‘to the greatest possible extent prior to ideas of the ethical’. With developmental psychologists the sticking point is different: many of them have no difficulty at all in agreeing that the notion of optimal development – human beings turning out as they should – is a psychobiological one. The surprise to them is that this notion is also ethical. But suppose they get over their initial surprise, and suppose the concept of optimal development, as deployed by attachment theorists, really is an ethical one. Isn’t the very fact that these developmentalists have been studying it unawares, using empirical methods, for all that time, evidence for the truth of ambitious naturalism? I shall argue (in section 5 below) that despite appearances, there’s no support for ambitious naturalism to be derived from developmental psychology, at least in the form of attachment theory. Roughly, in so far as the dispositions in which attachment theory deals are ‘adaptive’ – even if some theorists have claimed otherwise - this is not in the sense that they are the result of natural selection (and so belong to our first natures), but rather in the sense that they are favourably related to human social life – and perhaps to a particular form of it rather than to human social life in general. In so far as more is claimed, attachment theory overreaches itself, in a way that I shall suggest ambitious naturalism in ethics does too. 2. Attachment theory was first formulated by a dissenting psychoanalyst, John Bowlby.9 Partly inspired by the effects of maternal deprivation as a result of the 8 I take the phrase from Jay Belsky, “War, trauma and children's development: Observations from a modern evolutionary perspective”, International Journal of Behavioral Development 32 (2008): 260-271. 9 For a very brief history, see Jeremy Holmes, Exploring in Security (London & New York: Routledge, 2011), 3-5; for Bowlby and psychoanalysis, see Jeremy Holmes, Attachment, Intimacy and Autonomy (Northvale NJ & London: Aronson, 1996). 5 25/07/2013 wartime evacuation of children,10 he criticized classical Freudian ideas about human nature – he rejected, for example, the idea that infants are just in it for what they can get out of it (the ‘secondary-drive theory’), that is, form attachments to particular others simply as a means to getting food or warmth.11 He also wanted to test ideas about infancy using some of the methods of science and social science (for example large sample sizes and multiple observers), rather than relying only on what goes on in the consulting room. The theory was developed by two North American women, Mary Ainsworth and Mary Main, and has developed further since. It now has its own journals and is part of the academic mainstream, though of course it has its enemies as well as its friends. As to its main claims, attachment theorists distinguish between the ‘normative’ and the ‘individual-difference’ components of the theory,12 the ‘normative’ component – i.e. that part of the theory which deals with what is common to all (or almost all) human beings – being the more fundamental. The normative component (which is designed to apply also to other primates and perhaps to other mammals) notes that newborn offspring 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 proximity to an attachment figure who is able to protect the infant from threats to its 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 system’s behavioural expressions are various and maintain proximity in correspondingly various ways, including smiling and vocalizing (enlisting the attachment figure’s interaction), clinging (the attachment figure can’t get away), crying, and approaching and following. Thus we attach ourselves in infancy to a special attachment figure and by doing so tie them to us, enlisting the attachment figure’s responses, and if the infant responds in turn to those responses, a virtuous cycle is set in train that will help see to it that it 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 10 J. Robertson and J. Bowlby, ‘Responses of young children to separation from their mothers’, Courrier of the International Childrens Center, Paris, 2 (1952): 131-140. 11 Sigmund Freud, ‘Five Lectures on Psycho-Analysis’, The Standard Edition of the Complete Psychological Works of Sigmund Freud, ed. J. Strachey, vol. 11 (London: Hogarth Press, 1957). 12 See Jeffrey A. Simpson and Jay Belsky, “Attachment Theory within a Modern Evolutionary Framework,” in Handbook of Attachment: Theory, Research and Clinical Applications, ed. J. Cassidy and P. Shaver (New York/London: The Guilford Press, 2008), 136. 6 25/07/2013 non-behavioural characteristics of infants such as endearing ‘babyish’ features, 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 it is plausible that each one of a cluster of traits that, working together, serve to maintain infants’ proximity to attachment figures has emerged under pressure of natural selection.13 If evolutionary considerations explain why humans are equipped with the attachment system, isn’t this back to the bit of Freud which Bowlby rejected, i.e. that human infants form bonds with their special others in order to survive? No: if there’s an instrumentality here, it’s at the level of the unit on which selection operates, i.e. the gene. What helps genetic replication is that individual species members are ‘inherently motivated’14 to form attachment bonds to certain other species members. Dropping the jargon, the formation of such bonds is, like drinking or sleeping, something we are disposed to do for no reason (let alone any further reason) – it is simply our nature to do it, so looking for an individual’s reasons for doing it (which gives rise to the Freudian thought that the reasons are instrumental) is a mistake.15 Once exercised, however, the disposition gives rise to many goods (intimacy, reassurance, the pleasures of touch), which in turn are things we do ‘for their own sakes’ (i.e. for a reason but for no further reason). Thus the behaviour characteristic of attachment bonds – exclusivity, but also physical contact for its own sake, staying close especially in times of threat or distress, heightened anxiety at separation, grieving when an attachment figure dies – shares or (better) shares the general outlines of what we recognize in humans as relatedness to another that is valued or pursued for its own sake. 3. But though almost all human infants form attachments, not all attachments are alike in quality, and this fact and its explanation forms the subject-matter of attachment theory’s ‘individual-difference component’. The first measure of 13 In this paragraph I am indebted to Jude Cassidy, “The Nature of the Child’s Ties,” in Cassidy and Shaver, Handbook of Attachment, 3-20. 14 Cassidy, “The Nature of the Child’s Ties,” 5. 15 For discussion of things we do without doing them for a reason (whether for further reasons, or for their own sake), see M. Alvarez, Kinds of Reasons (Oxford: Oxford University Press 2010), 111 ff. Thanks to Luke Brunning for drawing my attention to this. 7 25/07/2013 attachment quality was the ‘Strange Situation’, developed by Mary Ainsworth.16 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 mother (let’s say) 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 mother tells the infant she is leaving, leaves for three minutes, and then comes back. The separation and reunion is repeated, with the stranger absent.17 The infant’s behaviour is recorded throughout. Observed infant behaviour falls into three recognizable patterns. (There is also a fourth pattern that was theorized later, but I omit that, again for simplicity’s sake.) In pattern B, the infant is overtly distressed when the mother leaves, then seeks proximity with her when she comes back and is comforted by it, and then resumes playing. In pattern A, the infant does not express distress when the mother leaves, though it displays other signs of distress such as more rapid breathing and heart rate, suggesting it is suppressing the expression of distress rather than simply indifferent. The infant doesn’t show a preference in play as between the mother and the stranger, and is then also relatively indifferent when the mother returns (for example looks or turns away from her), and ‘if picked up … makes no effort to maintain the contact’. In pattern C, the infant may be ‘clingy’ towards the mother and uninterested in the toys even before she leaves, is immoderately distressed when she does leave, but when she comes 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 ‘insecure-ambivalent’ or ‘insecureresistant’ (type C; also ‘hyperactivating’).18 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 life-span. 16 See for example M. Ainsworth and B. Wittig, ‘Attachment and Exploratory Behaviour in oneyear-olds in a Strange Situation’, in Determinants of Infant Behaviour, vol. 4, ed. B.M. Foss (London: Methuen, 1969), 111-136. 17 J. Solomon and C. George, “The Measurement of Attachment Security”, in Handbook of Attachment, ed. Cassidy and Shaver, 386. 18 Further variant terminology is also used: see Holmes, Exploring in Security, 3. Quotations in this paragraph are from Nancy S. Weinfeld et al., “Individual Differences in Infant-Caregiver Attachment,” in Handbook of Attachment, ed. Cassidy and Shaver, 78-101, to which I am more generally indebted. See also Solomon and George, “The Measurement of Attachment Security,” 383-418. 8 25/07/2013 Accordingly, other tests, based both on observation and on interview data, have been developed for various later stages of life (e.g. the Cassidy-Marvin system and the Preschool Assessment of Attachment for preschool-age children,19 and the Adult Attachment Interview20), with roughly the same number of attachment classifications, often with similar names to those used in the Strange Situation. Though there is some debate over which age-specific test is the most reliable for a given age, and over the extent to which different age-specific tests keep track of the same characteristics, there is an evident family resemblance between what these age-specific tests test for. The second aspect of attachment theory’s individual-difference component that I want to draw attention to concerns the further characteristics with which secure and insecure attachment are associated. These associations have both a diachronic and a synchronic dimension, the diachronic dimension concerning the characteristics predicted, at a greater or lesser distance in time, by a secure infantile attachment history, the synchronic dimension concerning the characteristics contemporaneously associated with secure attachment as measured by the test(s) appropriate to the life-stage in question. On the whole the contemporaneous associations are stronger than the predictive ones,21 and the predictive associations are weaker the longer the distance in time and the further removed the ‘outcome domain’ is from quality of relations with the infantile attachment figure him- or herself. This is thanks to the fact that attachment classifications can shift quite early in life (for example, an infant who is securely attached to its mother aged 12 months may become insecure if the mother suffers from post-natal depression following the birth of a second child; and an insecurely attached child may become securely attached to someone if it is fortunate in its adoptive or foster parents).22 And if they don’t shift, this is likely to be not only thanks to the infantile attachment history, but thanks to the persistence of the factors – such as warm relations with parents – which also explain the infantile attachment classification. Thus if there is a predictive relation to later characteristics it is likely to be mediated by a variety of further factors.23 19 Solomon and George, “The Measurement of Attachment Security,” 297, 299. J. Feeney, “Adult Romantic Attachment,” in Handbook of Attachment, ed. Cassidy and Shaver, 456-481. 21 R. Thompson, “Early Attachments and Later Development,” in Handbook of Attachment, ed. Cassidy and Shaver, 361. 22 Holmes, Exploring in Security, 4. 23 See Thompson, “Early Attachments,” 343. 20 9 25/07/2013 To summarize some recent findings,24 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 attachment in adulthood, meanwhile (whether or not itself predicted by secure attachment in infancy), is correlated with greater sensitivity to one’s own children’s needs, and ‘more warmth and appropriate structuring of learning tasks’;25 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’.26 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,27 are less dependent on teachers in preschool,28 less dependent on counselors at summer camp aged 10, and more sociable with unfamiliar adults.29 By contrast insecurely attached 4 year-old boys exhibit more ‘aggressive, assertive, controlling and attention-seeking behavior than their securely attached counterparts’. Finally, attachment theory argues for a connection between attachment security and broader personality traits. The Minnesota Study of Risk and Adaptation from Birth to Adulthood30 argues for ‘significant associations between early attachment security and personality characteristics throughout childhood and adolescence … [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 – insecurely attached either are more negative about themselves or do not admit flaws. There is also an important contemporaneous association between secure attachment and the capacity for emotion-regulation, including in adulthood;31 and between secure attachment and psychological understanding (they are more 24 See Thompson, “Early Attachments,” 348ff. Simpson and Belsky, “Attachment Theory within a Modern Evolutionary Framework,” 145. 26 M. Mikulincer and P. Shaver, “Adult Attachment and Affect Regulation”, in Handbook of Attachment, ed. Cassidy and Shaver, 507. 27 Thompson, “Early Attachments,” 355. 28 L.A. Sroufe, “Infant-caregiver attachment and patterns of adaptation in preschool,” in Minnesota Symposia on Child Psychology vol. 16, ed. M. Perlmutter (Hillsdale NJ: Erlbaum, 1983), 41-83. 29 Thompson, “Early Attachments,” 355. 30 L.A. Sroufe et al., The Development of the Person: The Minnesota Study of Risk and Adaptation from Birth to Adulthood (New York: Guilford Press, 2005). 31 Mikulincer and Shaver, “Adult Attachment”, 503-531. 25 10 25/07/2013 ‘proficient at identifying emotions in others, … especially … negative emotions and mixed feelings’). Thompson concludes that ‘children with a secure attachment history are capable of developing and maintaining more successful close relationships, especially with parents and with peers, than are insecure children; they develop a variety of desirable personality qualities in childhood and adolescence [including ‘social problem-solving skills’]; they are more likely to exhibit constructive forms of emotionality and emotion self-regulation; and they exhibit more positive self-regard’.32 4. I now want to use my sketch of attachment theory to explain why the theory is well placed to flesh out what I earlier called the continuity thesis. Attachment dispositions face both backwards towards the psychobiological (thanks to attachment theory’s normative component) and (thanks to its individual-difference component) forwards towards the ethical. As a result, the more we get to understand attachment, including what explains it and what it explains, the better we should be able to bring the continuity idea to life, and thus to take it beyond a picture of how some naturalistically minded philosophers think things ought to turn out. As regards the forwards-facing connection, it is striking how generally the characteristics with which secure attachment is associated, whether predictively or contemporaneously, seem to be positive characteristics: the capacity to relate harmoniously to others, co-operativeness, ‘positive affect’, self-esteem, the capacity to be realistic about oneself and to tolerate one’s own imperfections, are all apparently more worth having than those typically possessed by the insecurely attached or by those with insecure attachment histories (dependence, attentionseeking, low self-esteem, limited capacity for symbolic play and so on). It’s thus unsurprising that a great deal of effort is expended, by parents, educators, therapists and others, on trying to get children into attachment category B and making sure they stay there. However, the characteristics in question are ill-assorted. Some seem either to be virtues or to imply virtues. Here I am thinking of the capacity for friendship, and the capacity to form a realistic appraisal of one’s own excellences and defects, which is surely a virtue, though perhaps not an Aristotelian one;33 and the capacity to offer 32 33 Thompson, “Early Attachments,” 357-361. Cp. Aristotle’s virtue of truthfulness (Nicomachean Ethics, 1779). 11 25/07/2013 help and to ask for help when you need it. Co-operativeness, meanwhile, if it is not itself a virtue surely implies the traditional virtues of trust and, unless the cooperation is very short-term, also honesty and fidelity to promises. Other characteristics I mentioned seem like more general character-traits that have sometimes been argued (emotional self-regulation34) to go with or (self-esteem35) to underlie the virtues, while some are traits that may be as it were adverbially related both to virtues and to vices (aggression – good in fighting the local to get your child a place at school, not so good in bullying a colleague into accepting an unfair workload; the same goes for ‘positive affect’, though one really needs to know more about what’s meant – if it means ‘the capacity to enjoy life’s goods to the full’, arguably it is itself a virtue). Empathy is another tricky one, depending (for example) on whether one thinks cruel people genuinely possess it. There is a lesson to be learned both from the (as we might put it) evaluative asymmetry between secure and insecure attachment, and from the heterogeneity of the positive traits. The evaluative asymmetry seems to show that secure attachment stands in a privileged relation to the virtues (and insecure attachment to the vices). But how close is the relation? At one extreme, the answer would be that secure attachment is virtue; or perhaps that it’s the disposition that underlies and unifies the virtues. That would be a highly ambitious direction for neo-Aristotelian ethics to try to go in.36 On this view, attachment theory would not be an intermediate level of theory that merely mediates between the biological and the ethical, because the individual-difference component of attachment theory already is a theory about the ethical, that is, a theory that stands to virtue and vice as (say) Plato’s tripartite moral psychology does. On a more modest view, secure attachment would belong at a level intermediate between the biological and the ethical, with secure attachment occupying the place occupied (roughly) by Aristotelian ‘natural virtue’,37 a disposition which is not yet virtue but may turn into it if properly cultivated. That certainly fits the fact that attachment dispositions are relatively unstable in the early years. One might also take this more modest thought in a 34 See Aristotle, Nicomachean Ethics, 1106b, 1747. See e.g. Marcia Homiak, “Aristotle on Self-Love,” Canadian Journal of Philosophy 11 (1981). Pauline Chazan, The Moral Self (London & New York: Routledge, 1998). 36 Compare Chazan’s attempt to show that the Aristotelian virtues are underpinned by (roughly) good object relations, The Moral Self, esp. 63-154. There are clearly affinities between Bowlby’s work in attachment theory and the Kohutian psychoanalysis Chazan draws on. 37 Aristotle, Nicomachean Ethics, 1144b, 1807. 35 12 25/07/2013 skeptical direction: if secure attachment is clearly in some sense privileged among the attachment dispositions, the fact that it’s not straightforwardly related to the virtues (after all the list above leaves a great many out) could be used to explain why people are not typically unified in respect of the virtues (honest without being generous, or honest to colleagues but dishonest to lovers), or to challenge the Aristotelian thought that we’re ‘made for virtue’.38 I can’t develop these lines of thought here, let alone adjudicate between them, but this doesn’t matter for my present purpose: whichever way these lines of thought are developed, we are going to end up with a richer and more realistic version of the continuity thesis than we have got so far. 5. I want to comment now on the backwards-facing connection, from attachment to biology, and here my comments are of a more cautionary sort. Attachment theorists can hardly fail to think that attachment, meaning the disposition simply to form attachments, is ‘adaptive’. But when they go on to specify 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’,39 it is clear that they don’t mean attachment simpliciter but rather secure attachment. Again, Londerville and Main write that it is ‘adaptive’ to form a secure attachment in year one, because it increases the likelihood that a second ‘positive adaptation’, e.g. ‘the capacity for cooperation with the mother to gain needed help in problem 38 It would thus answer Bernard Williams’s complaint against Aristotle’s and Plato’s own moral psychologies that they do not answer to the demands of the continuity thesis, because they build too much of what they are trying to explain into the psychology that supposedly explains our ethical lives. (‘Aristotle’s psychology, despite its richness and elaboration, can seem ethically superficial. … [We need] a psychology that is less moralized, less adapted already to the demands of the ethical,’ Bernard Williams, “Replies,” 202.) But I take it that Williams also has in mind the thought that telling the continuity story in a non-circular way would display the fact that moral considerations do not have the privileged position in our practical lives that has sometimes been claimed for them: ‘A non-moralized, or less moralized, psychology … leaves it open, or even problematical, in what way moral reasons and ethical values fit with other motives and desires, and how far they are in conflict with them’, Williams, “Replies,” 202. 39 N. Humber and E. Moss, “The Relationship of Preschool and Early School Age Attachment to Mother–Child Interaction,” American Journal of Orthopsychiatry 75 (2005), 128–141. 13 25/07/2013 solving’, will develop in the second year of life.40 Now of course it is unhelpful to have too many people around who don’t comply with social or moral rules; who can’t cooperate, who can’t ask for help when they need it, who are also poor at striking out on their own (all characteristics associated with secure attachment). But we must be careful: ‘adaptive’ is a word borrowed from evolutionary theory, and appears (in the above quotations) to be used to smuggle in the idea that secure attachment is selected for. Depending on how close one takes the connection between secure attachment and the virtues to be, it might indeed be smuggling in even more – virtue is selected for. Thus some attachment theorists appear to make a strong connection between virtue and our biological natures analogous to that aspired to by ambitious ethical naturalists: we can get straight to the privilege of secure attachment by reflection on our first natures. However, it’s surely a mistake to say secure attachment is selected for. For one thing, if it were, it would be hard to explain why so many people fall into the various insecure attachment categories (between 35 and 45% of the population41) – why would selective pressure not have been better at rendering them extinct? One might think that secure attachment is more effective in enabling people to reach reproductive age. But this seems only very marginally to be so. It’s true that category C – insecure-ambivalent – is associated with certain sorts of risk-taking and self-destructive behaviour. (The idea is that you can attract a reasonably attentive mother by saying ‘mummy’; with a distracted or neglectful mother you need to start climbing the bookshelves.) But almost all the children who display these traits make it to reproductive age anyway, presumably because their ‘maximizing’ (attentionseeking) strategy is effective. Indeed, one reason why categories A and C are perpetuated is that they are born to parents who occupy these categories themselves.42 The second point is that difference of attachment type is explained not by natural selection but by the interaction between the evolutionarily determined generalized attachment disposition and the parental environment – among other things. The difference between secure and insecure attachment is not like the 40 “Security of attachment, compliance, and maternal training methods in the second year of life,” Developmental Psychology 17 (1981), 290. 41 Carol Magai, “Attachment in Middle and Later Life”, in Handbook of Attachment, ed. Cassidy and Shaver, 533. 42 P. Fonagy et al., “Maternal representations of attachment during pregnancy predict the organization of infant-mother attachment at one year of age,” Child Development 62 (1991), 891-905. 14 25/07/2013 difference between a picture that’s securely fixed to the wall and one that’s precariously hanging off a nail – whatever the terminology may suggest, secure/insecure isn’t a distinction between degrees of attachment. If it were it would be very hard to explain why the parents of insecurely attached children have such a big effect on them. Suppose the postman calls every day, he is a reliable and pleasant figure and he always gives you a smile – but if he is no more than that, why should the character of the postman have any effect on your character? And the same would be true of insecurely attached children, if insecurely attached meant ‘not very attached’. But attachments do not work like this. If you have a bad accountant, you can sack them and get another one. But children cannot change their parents, so even a cold or inconsistent parent gives you a better chance of making it out of infancy than no parent. So, instead of changing parents, you change. Thus for example, the insecure-avoidant type is an adaptation to indifferent or cold parents because, the thought is, such parents would be annoyed, and thus more rejecting, of a child who expressed its needs more overtly; the insecureambivalent child on the other hand has experience of parental interest so when this is replaced by neglect it will go to extremes to get it back. But all attachment types are children’s adaptations to different parental environments43 - adaptations the more complexly mediated the older the child and thus the more capable of complex forms of learning, comparison of goods and so on – aiming at the creation and ‘maintenance of a degree of proximity with the caregiver over time’.44 So it is a mistake to single out secure attachment as a better strategy for making it to reproductive age. If there is a sense in which secure attachment is ‘adaptive’ (or ‘optimal’), it is that the cost of adapting to a rejecting parental environment will be a sacrifice in intimacy, for example, and enjoyable mutual interactions: adaptations to a good environment are likely to be better, in ways that evolutionary biology and attachment theory itself are thoroughly unsuited to describe, than adaptations to a poor one. 43 ‘Each attachment pattern reflects a different ‘strategy’ that could have solved adaptive problems presented by different kinds of rearing environments’, Simpson and Belsky, “Attachment Theory,” 138. 44 E. Waters et al., “Infant-Parent Attachment and the Origins of Prosocial and Antisocial Behavior.” In Development of Antisocial and Prosocial Behavior : Research, Theories, and Issues, edited by D. Olweus et al., (Orlando: Academic Press, 1986), 105. 15 25/07/2013 6. I want finally to apply the foregoing reflections to raise some questions for Philippa Foot’s claim that vice is a natural defect in humans, and for the familiar idea that the virtues stand in a privileged relation to well-being. Whatever its relation to virtue, secure attachment appears to be in some sense best for the person whose disposition it is. Perhaps this is most clearly argued in connection with infancy itself. The salient characteristics of secure attachment in infancy – the freedom to express distress when it is felt without fear of rejection and in the expectation of comfort, the capacity for the pleasures of ‘affective sharing’45 and warm physical contact, and the freedom to become absorbed in the environment – are real human goods. These both reflect the real goods of the kinds of relationship which give rise to secure attachment, and – especially in the case of the freedom to become absorbed – make available to the infant a great many other goods in their turn. They are also goods which are to varying degrees and in varying combinations unavailable to the insecurely attached infant. Thus quite independently of what, if anything, secure attachment predicts about characteristics later in life, to describe secure attachment in infancy is to describe a good infancy in the sense that parallels ‘a good childhood’ or ‘a good life’. But if the point is especially vivid in connection with infancy, the capacity for good close relations in later life (which e.g. balance intimacy and autonomy) is also associated with secure attachment. One need only remind oneself of the number of people who refer themselves for psychotherapy because they find themselves unable to enjoy them to gauge the privilege of secure attachment in relation to well-being.46 This privileg plays well – as far as it goes – for the Aristotelian association between virtue and well-being if secure attachment also has a privileged relation to the virtues. I have also suggested that secure attachment enjoys another kind of privilege, in the sense that the characteristics associated with it seem desirable in a way those associated with insecure attachment are not. But there is surely an element of cultural relativity here. One only has to switch context to, say, ancient Sparta – of legend if not of fact – for it to be quite probably better (for me) 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, where real threats are more or less constant, insecure-ambivalent unwillingness to allow distance from the 45 46 Weinfeld, “Individual Differences,” 72. See Holmes, Attachment, Intimacy and Autonomy. 16 25/07/2013 attachment figure is the more ‘adaptive’ characteristic.47 If secure attachment is privileged in respect of suiting us better to social life, the privilege is thus surely relative to the more or less stable circumstances in which we live. Note however that this is not to endorse cultural relativism: one of the awful things about ancient Sparta might be that it prized the reproduction of insecure-avoidant types, thus leading many of its citizens to miss out on the real goods of warm personal relations; the same goes mutatis mutandis for war zones. But if secure attachment bears a privileged relation to some virtues, the relativity raises a problem for the idea that these virtues are necessary for our species life, or that lacking them is a ‘natural defect’: for it looks as if, as long as circumstances are imagined to be appropriately different, some version of our species life could be carried on just as well if the distribution of insecurely attached people in the population were the same as the distribution of securely attached people here and now. The point about relativity can be pressed further. If insecure attachment would be optimal for a majority of the population in radically challenging or threatening circumstances, it is surely useful in some of the population even in our circumstances. For though our circumstances are more or less stable, our social world is sufficiently complex to make it likely that some division of labour – made possible by the variety of attachment dispositions in a given population – is necessary to the form in which we, locally, carry on our species life. Thus it is surely good in our own fortunate though imperfect circumstances to have some people around who are risk-takers (and so presumptively insecure-ambivalent), and some who are precociously self-reliant (presumptively insecure-avoidant): it is not obvious that the full range of goods that are available to humans would be realized in a society in which everyone was cooperative, affectionate, and compliant. Thus it might not be that the only human analogue of Foot’s naturally excellent wolf who hunts with the pack is the cooperative person who sits attentively round the committee table: if the naturally excellent are those who have the characteristics necessary to sustain our species life, then granted the point about the division of labour, this description might net not only the good committee person but the odd person who angrily storms out of meetings (or simply, never attends meetings because they are hatching a plan on their own). The real human equivalent of the lone wolf or the night-blind owl would rather be the rare human being who has no 47 See Belsky, “War, trauma and children's development”, 265. 17 25/07/2013 disposition to form attachments – a defect indeed, but whose absence seems to leave just about everything open as far as virtues and vices are concerned. At the very least, the conclusion to draw is that since our species life can be carried on in circumstances which vary greatly in respect of stability and the presence of threats, there is no single attachment disposition (or single distribution of different attachment dispositions among a population) which is necessary for us to do so, and so no attachment disposition48 which is per se a natural defect. So far that sceptical point says nothing about virtue either way: the extent to which virtue comes in depends on the strength of the association between virtue and secure attachment. If secure attachment does have a privileged relation to virtue, it looks as if vices can’t be natural defects, because insecure attachment too is an ingredient in the mix necessary for our species life – it would be as impossible to sustain if no one had that as if no one was securely attached. But perhaps secure attachment doesn’t have a privileged relation to virtue. If that is so, then granted the apparently privileged relation of secure attachment to well-being, that would make trouble for the association between virtue and well-being, though perhaps that connection is in trouble anyway.49 In any case, virtue could be underpinned psychologically by either a secure or an insecure attachment disposition (perhaps depending on the virtue, perhaps depending on the circumstances): the good committee person and the awkward individualist who doesn’t turn up for meetings might, though in different attachment categories, both be virtuous. Whether secure attachment does bear a privileged relation to virtue, however, awaits a proper investigation of what in the way of virtues secure attachment is and is not correlated with - a question which attachment theorists may not ask in so many words, but on which their data bear in a multitude of ways. 48 49 Or organized attachment disposition? In part for reasons well stated by Foot herself: Foot, Natural Goodness, 85. 18 25/07/2013 References Ainsworth, M., and B. Wittig. “Attachment and Exploratory Behaviour in one-yearolds in a Strange Situation.” In Determinants of Infant Behaviour, vol. 4, edited by B.M. Foss, 111-136. London: Methuen, 1969. Alvarez, M.. Kinds of Reasons. Oxford: Oxford University Press 2010. Aristotle, Nicomachean Ethics, trans. W.D. Ross, revised J.O. Urmson. In Aristotle: The Complete Works. Electronic edition. BOLLINGEN SERIES LXXI, Volume 2. Belsky, Jay. “War, trauma and children's development: Observations from a modern evolutionary perspective”. International Journal of Behavioral Development 32 (2008): 260-71. Cassidy, Jude. “The Nature of the Child’s Ties.” In Handbook of Attachment: Theory, Research and Clinical Applications, edited by J. Cassidy and P. Shaver, 3-20. New York/London: The Guilford Press, 2008. Chazan, Pauline. The Moral Self. London & New York: Routledge, 1998. Feeney, J.. “Adult Romantic Attachment.” In Handbook of Attachment: Theory, Research and Clinical Applications, edited by J. Cassidy and P. Shaver, 456481. New York/London: The Guilford Press, 2008. Fonagy, P. et al.. “Maternal representations of attachment during pregnancy predict the organization of infant-mother attachment at one year of age.” Child Development 62 (1991): 891-905. Foot, Philippa, Natural Goodness. Oxford: Oxford University Press, 2001. Freud, Sigmund. “Five Lectures on Psycho-Analysis.” The Standard Edition of the Complete Psychological Works of Sigmund Freud, ed. J. Strachey, vol. 11. London: Hogarth Press, 1957. Holmes, Jeremy. Attachment, Intimacy and Autonomy. Northvale NJ & London: Aronson, 1996. Holmes, Jeremy. Exploring in Security. London & New York: Routledge, 2011. Homiak, Marcia. “Virtue and Self-Love in Aristotle’s Ethics.” Canadian Journal of Philosophy 11 (1981). Humber, N., and E. Moss. “The Relationship of Preschool and Early School Age Attachment to Mother–Child Interaction.” American Journal of Orthopsychiatry 75 (2005): 128–141. Londerville, Susan, and Mary Main. “Security of attachment, compliance, and maternal training methods in the second year of life.” Developmental Psychology 17 (1981). McDowell, John. “Two Sorts of Naturalism”. In Mind, Value and Reality by John McDowell, 167-197. Cambridge MA & London: Harvard University Press, 1998. McDowell, John. Mind and World. Cambridge MA: Harvard University Press, 1994. Magai, Carol. “Attachment in Middle and Later Life,” in Handbook of Attachment: Theory, Research and Clinical Applications, edited by J. Cassidy and P. Shaver, 532-551. New York/London: The Guilford Press, 2008. Mikulincer, M., and P. Shaver. “Adult Attachment and Affect Regulation”, in Handbook of Attachment: Theory, Research and Clinical Applications, edited by J. Cassidy and P. Shaver, 503-531. New York/London: The Guilford Press, 2008. Robertson, J., and J. Bowlby. “Responses of young children to separation from their mothers.” Courrier of the International Childrens’ Center, Paris, 2 (1952): 13140. Schaffer, H. Rudolph. Key Concepts in Developmental Psychology. London: Sage, 2006. 19 25/07/2013 Simpson, Jeffrey A., and Jay Belsky. “Attachment Theory within a Modern Evolutionary Framework.” In Handbook of Attachment: Theory, Research and Clinical Applications, edited by J. Cassidy and P. Shaver, 131-57. New York/London: The Guilford Press, 2008. Solomon, J., and C. George. “The Measurement of Attachment Security.” In Handbook of Attachment, ed. Cassidy and Shaver, 383-418 Sroufe, L.A.. “Infant-caregiver attachment and patterns of adaptation in preschool.” In Minnesota Symposia on Child Psychology vol. 16, edited by M. Perlmutter, 41-83. Hillsdale NJ: Erlbaum, 1983. Sroufe, L.A.. The Development of the person: The Minnesota Study of Risk and Adaptation from Birth to Adulthood. New York: Guilford Press, 2005. Thompson, Ross. “Early Attachment and Later Development.” In Handbook of Attachment: Theory, Research and Clinical Applications, edited by J. Cassidy and P. Shaver, 348-65. New York/London: The Guilford Press, 2008. Waters, E. et al.. “Infant-Parent Attachment and the Origins of Prosocial and Antisocial Behavior.” In Development of Antisocial and Prosocial Behavior : Research, Theories, and Issues, edited by D. Olweus et al., 97-128. Orlando: Academic Press, 1986. Weinfeld, Nancy S. et al., “Individual Differences in Infant-Caregiver Attachment.” In Handbook of Attachment: Theory, Research and Clinical Applications, edited by J. Cassidy and P. Shaver, 78-101. New York/London: The Guilford Press, 2008. Williams, Bernard. “Naturalism and Genealogy”. In Morality, Reflection and Ideology, edited by Edward Harcourt, 148-161. Oxford: Oxford University Press, 2000. Williams, Bernard. “Replies”. In World, Mind and Ethics, edited by J.E.J. Altham and R. Harrison, 185-224. Cambridge: Cambridge University Press, 1995. 20 25/07/2013