Parent - Infant Interaction

The mother-infant relationship consists of interaction sequences, each of which is in some measure unique. We must therefore study pooled data from interactions of the same ‘type’ occurring within a limited time period. Changes in the relationship must be assessed by comparing data from different time periods. To understand the dynamics of the interaction, it is essential to formulate sharply the questions being asked. In particular, the following questions must not be confused: ‘Which partner is primarily responsible for determining the course of this type of interaction in a particular time period?, ‘Changes in which partner are responsible for changes in the pattern of interaction between time periods?’ and ‘Are differences between mothers or between infants primarily responsible for inter-dyad differences in any one time period? The argument is illustrated with data from rhesus monkeys. That the mother-infant relationship depends on interactions in which the behaviour of each partner affects and is affected by that of the other no longer needs to be emphasized (e.g. Bell 1968). That the interactions themselves, even those involving very young babies, can be of considerable complexity, is becoming progressively more widely recognized. By analogy with interactional sequences that have been studied in non-human species (e.g. Baerends et al. 1955; Hinde 1970; Nelson 1964; Simpson 1968, 1973), we must expect each response of each individual to be determined not only by the preceding behaviour of the partner, but also by earlier behavioural events, the individual’s current state, the environmental context, and so on. Because the determinants are multiple, susceptible to mutual influence, and complex, each interaction sequence is in some measure unique. But science has no techniques for coping with unique events: we can ask questions only about populations of events judged to be, by some criterion, similar. We must therefore study categories of individually unique interaction sequences, regarding the component events as having probabilistic relations to each other. In practice