Nearly six decades have passed since John Bowlby wrote Maternal Care and Mental Health and redefined the bonds of love.

Bowlby’s formulation of attachment, dependency and play was intended to help practitioners and researchers advance the young science of human development, and over the years his central concept of attachment has slipped in and out of focus – sometimes productive, sometimes a hindrance to new thinking.

Today at London’s Regent’s College Professor Sir Michael Rutter uses the occasion of his Emanuel Miller Lecture to take fresh stock of the evidence on the variety of concepts associated with attachment.

He recalls the contribution of Bowlby’s colleague, the American Mary Ainsworth, who in her investigations of infant attachment famously developed the Strange Situation Procedure. Classic recordings of the procedure show a mother alone in a room with her toddler. The child explores the room while the mother watches. Then a stranger enters who talks to the parent and approaches the child. The parent quietly leaves the room. Ordinarily, the child will become distressed at this point. The parent returns and comforts the child.

What fascinated Ainsworth was what it signified if a child did not react in a typical way to this strange situation. What did it indicate if the child did not cry, or stopped crying, or avoided the mother when she returned to the room?

Out of her analysis came the concept of secure and insecure attachments. Later, with the study of high risk groups, Main and Solomons added a coding for ‘disorganized attachment’ which showed a much stronger association with psychopathology. As a rather separate development, Rutter explains, the American Psychiatric Association added to its classification scheme a diagnosis of ‘reactive attachment disorder’.

The elegance of such concepts began to be a problem: many commentators became overly attached to attachment theory itself and began to press it into service to capture all aspects of children's social relations and behavior.

In fact, although attachment theory has proved immensely useful, the specific focus on attachment insecurity has more limitations. By Rutter’s calculation, the Strange Situation Procedure is a good guide to the components of early relations, but it is not helpful in clinical assessment and it has nothing useful to say about conditions such as autism that involve a different type of problem in social relationships.

Much research has highlighted the very high frequency of ‘disorganized attachment’ in children experiencing abuse or neglect or an institutional upbringing, as well as the strong association with mental disorder. But it is also found in some 15% of children living in low risk families, so what might it mean?

Also, it seems that disorganized attachment does not persist if rearing conditions improve. A Greek study undertaken by Panayiota Vorria even showed that it was associated with greater security after the infants left the institution and were adopted. Rutter concludes that the category unquestionably exists and has importance, but he queries what it means with respect to social relationships.

In relation to the study of reactive attachment disorders, Rutter refers to the work of Charlie Zeanah, Sellars-Polchow Professor of Psychiatry at Tulane University in the US. This diagnosis is included in both the major classification systems, DSM-IV and ICD-10. It indicates markedly disturbed and inappropriate relationships that occur across social contexts. It has its onset before the age of five and persists over time.

To be counted as an attachment disorder, the behavior must not be a consequence of social anxiety or autism and it must be a reflection of the quality of early care. Inhibited RAD involves a reluctance to initiate or respond appropriately to social interactions and is common among institutionalized children. Disinhibited RAD is reflected in an increased risk of the child going off with strangers, or not checking with the parent or failing to distinguish between people who may help or hinder.

Zeanah's research is particularly important in showing that, although grouped together in the classification schemes, the two forms of RAD are actually very different. Whereas inhibited RAD ceased to be a problem after the children were placed in foster families, disinhibited RAD proved to be much more persistent. In Rutter’s own research, it was still evident more than seven years after children from institutions were adopted into mostly well functioning families.

Several times in the course of the lecture, he urges caution, particularly in relation to the clinical significance of results.

There is an urgent need, he says, to understand how children are processing information when they are faced with the stress associated with attachment difficulties. Early research made the case for altering child-rearing practices, particularly with respect to reductions in the use of institutional care; new evidence should signpost ways of preventing the wider sequelae of poor social relationships.

People relate to each other in a multitude of ways. Impairments to the ability to relate are also very varied. Among them are several variants of attachment disorder – and also apparently unrelated problems such as autism and Williams syndrome.

Attachment theory was crucially important in alerting people to the significance of early social relationships. It also did much to clarify the meaning of those relationships. However, it is now apparent that security is not all and that attention needs to be paid to the heterogeneity of disturbances in social relationships. Michael Rutter’s lecture puts its value into perspective.