Modality effects and the relational dimension in the abstraction of memory schema : a thesis presented in partial fulfilment of the requirements for the degree of Master of Arts in Psychology at Massey University

Abstract

This study investigated the nature of memory representations constructed from explicit and implied information about two different types of relationships between the objects or actors described in a set of ordered propositions. The propositions described a situation which referred to either an action sequence or to a set of spatial relationships and were presented to subjects in two forms, verbal and pictorial. Ten year old children were given a recognition memory task based on the paradigm used by Bransford, Barclay and Franks (1972). The results showed that subjects had difficulty in distinguishing old recognition items from new situation preserving propositions based on inferences derived from the acquisition sequence, but readily rejected those new propositions that were not consistent with the relationships described in the original premises. This was the case when the original premises were presented in either verbal or pictorial form, and when the relational term used referred to either action sequences or to spatial relationships. An analysis of the results for specific combinations of modality and relational term showed some variation in the general pattern of responses. The construction and integration of inferences into memory representation was facilitated when action sequences were presented in the verbal modality, and when spatial relationships were portrayed in the pictorial modality. Action verbs were found to be particularly sensitive to modality effects. The concept of 'abstraction' is discussed in the light of the results. It is suggested that this concept is not a unitary one, and that research concerned with the abstraction of memory schema refers to two different types of abstractive process. The distinction between paradigmatic and syntagmatic relationships, originally outlined in Sassure's (1916) study of linguistics, is proposed as a useful way to characterise these processes.