Attachments within close relationships lie at the heart of our emotional life, and the propensity to form attachments is an integral part of human nature. In the words of
John Bowlby (1907–90), 'It is characteristic of human beings to make strong affectional relationships with each other and for some of their strongest emotions to depend on how these relationships are faring. Whereas stable relationships are a source of enjoyment and security, separation, loss, or threatened loss arouse
anxiety or anger, or else sadness and
depression.' (Bowlby, 1st edition of this Companion, 1987, p. 57).
Bowlby produced a framework for understanding the development and implications of such close relationships, as set out in his trilogy
Attachment and Loss (1969/82; 1973; 1980). His early clinical observations pointed to the adverse effects of separation or loss of a mother figure, leading him to ask, 'If the disruption of a child's relationship with mother-figure in the early years creates much distress and anxiety, what is so special about the relationship that has been disrupted?' (Bowlby 1991
a: 302–3). The prevailing analytical thinking was that a special bond with the mother stemmed from her association with the provision of food. Not convinced by this, Bowlby noted the phenomenon of imprinting in precocial birds such as geese, with a bond developing in its own right, independently of being fed. He therefore turned to the scientific discipline of ethology and the writings of Konrad Lorenz and Niko Tinbergen. Within this framework, the occurrence of species-characteristic behaviour such as following the mother suggests that it may have been selected for in its own right during the course of evolution. Bowlby defined attachment behaviour as any form of behaviour that attains or maintains proximity to a caregiver in times of need or stress. In the environments in which we evolved, individuals who exhibited attachment behaviour would have been more apt to survive and leave offspring, who in turn would reproduce (i.e. would increase their 'inclusive fitness'), compared with those who did not show attachment behaviour. The presumed biological function was protection from harm. Selection for attachment behaviour could not have happened without a similar pressure on its complement, caregiving behaviour. 'During the course of time, the biologically given strategy of attachment in the young has evolved in parallel with the complementary parental strategy of responsive caregiving — the one presumes the other' (Bowlby 1991
b: 293).
Bowlby realized that an evolutionary argument could provide insight into behaviour which otherwise appeared abnormal, including the 'irrational fears of childhood'. The tendency to fear unfamiliar situations, darkness, or separation is 'to be regarded as a natural disposition of man ... that stays with him in some degree from infancy to old age ... . Thus it is not the presence of this tendency in childhood or later life that is pathological; pathology is indicated either when the tendency is apparently absent or when fear is aroused with unusual readiness and intensity' (Bowlby 1973: 84). In one of his final contributions Bowlby wrote, 'Once we postulate the presence within the organism of an attachment behavioural system regarded as the product of evolution and having protection as its biological function, many of the puzzles that have perplexed students of human relationships are found to be soluble. ... an urge to keep proximity or accessibility to someone seen as stronger or wiser, and who if responsive is deeply loved, comes to be recognised as an integral part of human nature and as having a vital role to play in life. Not only does its effective operation bring with it a strong feeling of security and contentment, but its temporary or long-term frustration causes acute or chronic anxiety and discontent. When seen in this light, the urge to keep proximity is to be respected, valued, and nurtured as making for potential strength, instead of being looked down upon, as so often hitherto, as a sign of inherent weakness' (Bowlby 1991
b: 293).
1. The development of attachment 2. The quality of attachment 3. Implications1. The development of attachment
As attachment behaviour develops, it forms the basis for an inferred attachment bond. Bowlby described particular phases of its development: pre-attachment (from birth to about 2 months), involving signalling without discriminating one person from another; attachment-in-the-making (2–6 months), where signals become directed to particular persons; clear-cut attachment (0.5–4 years), with locomotion and goal-corrected behaviour; and finally a goal-corrected partnership (4 years onwards) with perspective taking, communication skills, and sharing mutual plans. Although additional attachments may develop throughout life, early attachments endure.
Furthermore, Bowlby (1973) postulated that attachment relationships must become internalized. Internal working models may be defined as ' "operable" models of self and attachment partner, based on their joint relationship history. They serve to regulate, interpret, and predict both the attachment figure's and the self's attachment-related behavior, thoughts, and feelings' (Bretherton and Munholland, in Cassidy and Shaver 1999: 89). This definition reflects Bowlby's view of the complementary nature of an internal working model, representing both sides of the relationship. 'A working model of self as valued and competent, according to this view, is constructed in the context of a working model of parents as emotionally available, but also as supportive of exploratory activities. Conversely, a working model of self as devalued and incompetent is the counterpart of a working model of parents as rejecting or ignoring of attachment behavior and/or interfering with exploration'.
2. The quality of attachment
A wealth of empirical research within Bowlby's attachment theory was enabled by Mary Ainsworth (1913–99). Their friendship and collaboration spanned 40 years, starting from when Ainsworth worked with Bowlby in the early 1950s. Appreciating that assessments of attachment must involve how a child uses the mother (or other caregiver) as a 'secure base' when the attachment behaviour system is activated, Ainsworth developed a laboratory Strange Situation Test for infants (Ainsworth et al. 1978). This is a series of short episodes involving mother and a stranger, in which the child's attachment behaviour system is activated by the unfamiliarity of the situation and by the mother leaving. The return of the mother allows one to see how the child organizes his or her attachment behaviour to her. Ainsworth identified three patterns: Secure, Avoidant, and Ambivalent. A Secure pattern has been associated with antecedent interactions with a 'sensitively responsive' mother, as found in Mary Ainsworth's pioneering Baltimore study and subsequently in other studies (see the meta-analysis by DeWolff and van IJzendoorn 1997). The insecure patterns have been associated with different maternal styles, including Avoidance with rejection, Ambivalence with inconsistency, and a more recently documented pattern — Disorganization with fear (see Cassidy and Shaver 1999).
As for which pattern of attachment is desirable, Bowlby was concerned with what might be called 'psychological desiderata' (Hinde and Stevenson-Hinde 1991). Making an analogy with 'physical well-being', Bowlby argued that 'psychological well-being' had an absolute meaning, involving security of attachment. Research has supported this view, with security associated with self-reliance and efficacy, as opposed to dependency, anxiety, or anger. Insecure patterns are not seen as pathological in themselves, but rather as risk factors for pathology, while security is viewed as a protective factor (reviewed in Weinfield et al. in Cassidy and Shaver 1999).
In addition to behavioural assessments of attachment quality, once children reach Bowlby's 'goal-corrected partnership' stage, their verbal behaviour may be used to index their 'internal working models' of attachment. Representational methods typically involve a description of separations or other distress-provoking situations, with the child providing narratives around them (see Solomon and George in Cassidy and Shaver 1999).
With adolescents and adults, representations of attachment may be accessed through the Adult Attachment Interview, developed by Mary Main and colleagues in Berkeley (see Hesse in Cassidy and Shaver 1999). This is a semi-structured interview consisting of eighteen basic questions, starting with one's childhood relationships with parents. After providing five adjectives to describe each parent, the interviewee is probed for specific episodic memories to illustrate why each descriptor was chosen. The protocol goes on to explore what happened around physical and emotional upsets, including separation and loss. The final section involves asking in specific ways how this early experience may relate to current feelings and relationships. A verbatim transcript of the interview provides the basis for ratings on continuous scales as well as classifications: Secure/autonomous, Dismissing, Preoccupied, or Unresolved/disorganized. These classifications are related to parenting style and to infants' attachment classification. 'What is most striking about this association is that it suggests that the form in which an individual presents his or her life narrative (regardless of its content) predicts caregiving behavior in highly specific and systematic ways' (Hesse in Cassidy and Shaver 1999: 398).
3. Implications
The fruits of Bowlby and Ainsworth's pioneering work are gathered together in the
Handbook of Attachment (Cassidy and Shaver 1999), which is suitably dedicated to them and which contains references supporting the following implications of attachment theory:Normative development and parenting. Attachment research has found patterns in the development and expression of emotions within close relationships, including how patterns may be transmitted across generations. The sense of security which must underlie any appreciation of emotions within one's self and their expression to others depends upon having attachment figures who are sensitively responsive to emotional needs. A number of studies have revealed what types of parental interactions promote security, thereby paving the way for methods of parental guidance and childcare policies in general.Psychopathology. With children who have already developed disorders, particular patterns of insecure attachment shed light on aspects of parenting that might be modified to promote security. For example, the Disorganized pattern, often seen in clinical samples, appears to be associated with fearfulness in either parent or child or both, thereby suggesting a window for intervention. Additionally, attachment theory has direct implications for disorders related to separation or loss, such as abnormal grief, depression, or anxiety.Social policy issues. When a child goes to hospital, we now take it for granted that parents may visit or even live in. However, before Bowlby's influence, hospital practice involved leaving caregiving to the 'experts' and keeping relatives away. Obstetric practice followed similar lines, and Bowlby did not hesitate to challenge this, asking how early separation could possibly promote a close mother–child relationship. His concern over separation from attachment figures remains relevant, including our present provision of care for adults, such as the elderly or mentally ill.
Regarding day care, Bowlby was keen to put right the misinterpretation of what he wrote. In stressing the importance of early caregiving, he did not mean that a parent must be with the child all the time. From the above it will be clear that attachment theory is about the
quality of a relationship rather than amount of time spent together. Similarly with adoption, attachment theory does not say that caregivers must be biologically related parents, nor that early adversities cannot be overcome. Indeed the implication is that adoptive parents, if sensitively responsive to the child's needs, may do a great deal to set interactions in a cycle that will promote security.
In conclusion, the framework provided by Bowlby has proved to be robust. At the same time, attachment theory continues to grow — not as a closed system, but one that is open to input from many other disciplines, including neuroscience, cognitive psychology, and linguistics.
(Published 2004)— Joan Stevenson-Hinde
Bibliography- Ainsworth, M. D. A., Blehar, M. C., Waters, E., and Wall, S. (1978). Patterns of Attachment.
- Bowlby, J. (1969/1982). Attachment and Loss, i: Attachment.
- — — (1973). Attachment and Loss, ii: Separation, Anxiety and Anger.
- — — (1980). Attachment and Loss, iii: Loss, Sadness and Depression.
- — — (1991a). 'Ethological light on psychoanalytical problems'. In Bateson, P. (ed.), Development and Integration of Behaviour.
- — — (1991b). 'Postscript'. In Parkes, C. M., Stevenson-Hinde, J., and Marris, P. (eds.), Attachment across the Life Cycle.
- Cassidy, J., and Shaver, P. R., (eds.) (1999). Handbook of Attachment: Theory, Research, and Clinical Applications.
- DeWolff, M. S., and van IJzendoorn, M. H. (1997). 'Sensitivity and attachment: a meta-analysis on parental antecedents of infant attachment'. Child Development, 68.
- Hinde, R. A., and Stevenson-Hinde, J. (1991). 'Perspectives on attachment'. In Parkes, C. M., Stevenson-Hinde, J., and Marris, P. (eds.), Attachment across the Life Cycle.