In this article, Sroufe and Waters consider three possible conceptualisations of attachment and argue for a view of attachment as an organisational construct. They make 3 main points:

• The trait view of attachment must be discarded as it has been demonstrated that attachment is not amenable to being investigated on the basis of correlating discrete behaviours

• The social learning view of attachment (which came about as a reaction to the trait view) suggests that attachment is that which occurs within a given child-caregiver dyad in a given situation. According to this view, individual differences which are independent of the interaction should not be present. This is shown to be false by evidence from longitudinal studies that indicate that later social and cognitive competencies are related to the organisation of the early attachment system.

• A view of attachment as an organisational construct is both theoretically and empirically sound. This view explains both the absence of a consistent set of behavioural traits and the presence of a relatively stable attachment system. Attachment should be conceptualised as an adaptive relationship, within which there are changing, though predictable, patterns of behaviour. Individuals with stable attachment patterns will display different behaviours as they develop. Furthermore, since each specific behaviour can have a different meaning in different situations, behaviours cannot be considered in isolation. According to this conceptualisation, attachment must be inferred from the meaning of attachment behaviours within their individual contexts.

The authors point to Ainsworth’s classification system (Ainsworth, Bell & Stayton, 1971) as one which which determines attachment by considering the purpose and meaning of different behaviours, and so demonstrates the organisational construct view. Attachment data obtained using this system has proven reliable and stable over time, even though discrete attachment behaviours have showed little stability.

Given the above, the authors propose that the original view (Bowlby, 1969) of the evolutionary origins of the attachment mechanism needs revision. Bowlby viewed protection from predation as the biological function of attachment, with proximity, obtained through a variety of attachment behaviours, as the set goal. However, a model that better explains many of Bowlby's observations would have felt security as the set goal, and affect as the mechanism, aimed at fulfilling the biological functions of protection from predation and also exploration (a valuable human adaptation). Crucial to this view, is the acknowledgement that an affective bond develops between the attachment figure and the infant. This bond fosters felt security through the experience of preferential treatment from attachment figures, and it is this felt security permits the infant to engage in exploratory behaviour and so become independently competent.

In a nutshell, the authors suggest that parental responsiveness and degree of sensitivity to the child’s affective communication provides the context within which the child learns to organize emotional experience and regulate felt security. Based on the caregivers response, the child learns a set of strategies which are generally extended for organising emotional experience and for dealing with negative feelings. It is this set of organising strategies that is relatively stable over time, and this is what is meant when we refer to an individual's attachment style.

Refer to Waters (1978) for further support for this conceptualisation of attachment.