Knowledge base Updated 2 November 2026 2 views

Attachment Theory

An account of the biologically-based system that leads infants to seek proximity to a caregiver under threat, and of how repeated experience with that caregiver shapes expectations about closeness, distress and dependence.

Definition

The attachment behavioural system is activated by threat, fatigue, illness or separation, and terminated by proximity to an attachment figure. A caregiver who reliably terminates it becomes a secure base for exploration and a safe haven in distress.

The Strange Situation procedure classifies infant patterns as:

Secure: distressed by separation, comforted and settled on reunion.
Insecure-avoidant: little overt distress, reunion minimised.
Insecure-ambivalent or resistant: high distress, reunion does not settle.
Disorganised: contradictory or apprehensive behaviour with no coherent strategy, added by Main and Solomon in 1986.

The first three are organised strategies, not deficits. Avoidance is an adaptation to a caregiver who withdraws under displays of need.

History

John Bowlby developed the theory from the late 1940s, drawing on ethology and on his work with separated and institutionalised children. Mary Ainsworth's Baltimore and Uganda observational work produced the classification and the Strange Situation.

The theory replaced accounts in which the infant bond derived from feeding, and Harlow's rhesus work — infant monkeys preferring a cloth surrogate over a wire one that fed them — was decisive against the drive-reduction view.

Bowlby's early maternal deprivation writing was used politically to argue mothers should not work. That was not what the evidence supported and it has coloured the theory's reception ever since.

Theory

The mechanism is the internal working model: a representation of self, others and relationships built from repeated interaction, carried forward and applied to new relationships, and revisable rather than fixed.

Adult attachment research developed along two partly separate tracks that are often conflated:

The Adult Attachment Interview, which classifies the coherence of a person's narrative about their own childhood and predicts their infant's classification.
Self-report romantic attachment measures on anxiety and avoidance dimensions.

These two correlate weakly. They are not measures of the same thing, and treating a self-report questionnaire result as an AAI classification is a common and serious error.

Research

Meta-analytic work shows parental sensitivity predicts security but modestly: de Wolff and van IJzendoorn found a correlation around .24, far smaller than the popular account implies. Something other than sensitivity accounts for most of the variance.

The transmission gap names this: parents' own AAI classification predicts infant classification more strongly than observed sensitivity does, and sensitivity does not mediate the association.

Stability from infancy to adulthood is moderate at best, and lower in samples with high life stress, which is what the theory predicts if working models are revisable.

Disorganised attachment is the classification with the clearest links to later difficulties, and it is also the one most often misused, because it is a low base-rate classification requiring trained coding.

Applications

Attachment-based interventions with the best evidence are short, behavioural and focused on sensitivity — Video-feedback Intervention to promote Positive Parenting is the clearest example — rather than long and insight-oriented.

The theory also informs care planning for looked-after children, work with foster and adoptive families, and the therapeutic relationship itself.

It does not support attachment therapies involving holding, rebirthing or coerced eye contact. These have no evidential basis, are repudiated by attachment researchers, and have caused deaths.

Criticism

The sensitivity effect is modest, and parenting advice built on it routinely overstates what a caregiver controls.
Cross-cultural validity of the Strange Situation is contested; separation from a caregiver does not mean the same thing everywhere, and distributions differ across cultures in ways the classification treats as pathology.
The two adult literatures are confused constantly.
Disorganised attachment is over-applied in family court and safeguarding contexts, well beyond what the coding supports.
The theory has been used to blame mothers for outcomes with many causes, and that history is not incidental to how it is received.

References

Bowlby, J. (1969). Attachment and Loss, Vol. 1: Attachment. Basic Books.
Ainsworth, M. D. S., et al. (1978). Patterns of Attachment. Erlbaum.
De Wolff, M. S., & van IJzendoorn, M. H. (1997). Sensitivity and attachment: A meta-analysis. Child Development, 68(4), 571-591.
Bakermans-Kranenburg, M. J., van IJzendoorn, M. H., & Juffer, F. (2003). Less is more: Meta-analyses of sensitivity and attachment interventions. Psychological Bulletin, 129(2), 195-215.
Granqvist, P., et al. (2017). Disorganized attachment in infancy: A review of the phenomenon and its implications. Attachment & Human Development, 19(6), 534-558.

Written for PsychologyOrbit and reviewed before publishing. It is a summary for professionals and students, not clinical guidance, and it is not a substitute for the primary sources listed above.