If attachment patterns reflect relationship characteristics rather than traits in the child, one would expect that characteristics of dyadic interaction would be associated with patterns of attachment. The research cited by van IJzendoorn provides support for a causal role of parental sensitivity in the development of attachment security, though much less research has addressed the interactive patterns that precede avoidant and resistant attachment. Research reviewed by Hennighausen and Lyons-Ruth has also demonstrated that certain parental behaviours, such as withdrawal, negative-intrusive responses, role-confused responses, disoriented responses, frightened or frightening behaviours and affective communication errors, which include contradictory responses to infant signals, are likely to be more evident in the context of certain types of parental psychopathology, and have been documented to be associated with disorganized attachment.3,4
A central tenet of attachment theory has been that early experiences between young children and their caregivers provide a model for intimate relationships in later life. Although this model is believed to be modifiable by subsequent experiences, the theory has posited a conservative tendency to resist change. These propositions suggest that in a stable caregiving environment, one would expect to find stable patterns of attachment, but in environments characterized by significant changes, one would expect less stability. On balance, these assertions are supported by research, although results from four longitudinal studies of attachment from infancy to adulthood do not support a linear relationship,5-8 as these studies do not uniformly demonstrate stability of attachment classifications from infancy to adulthood. They do, however, provide support for a relationship between life events and changes in attachment classifications. In the Grossmanns’ work, negative life events and stresses were also found to compromise...