ENCODING STRATEGY DYNAMICS - WHEN RELATIONSHIPS BETWEEN WORDS DETERMINE STRATEGY USE

A number of questions remain largely unanswered about when subjects ar e likely to employ imagery as opposed to verbal encoding and the subse quent effect this will have on recall. The present experiments examine d some of the factors determining such use. Subjects were asked to lea rn mixed lists of word pairs which were either spatially related (i.e. their referents normally appear in spatial contiguity) or categorical ly related (their referents are members of a shared superordinate cate gory) using either imagery or verbal mediators throughout. For half th e pairs the encoding instructions were appropriate to the relationship between words (e.g. the use of imagery for spatially related pairs) a nd for the other pairs the instructions were inappropriate (e.g. the u se of imagery for categorically related pairs). Subjects were not infl uenced by experimental instructions under these conditions but overwhe lmingly used encodings appropriate to the relationship between words. Furthermore, it was found that subjects were able to adjust encoding s trategies from one pair to the next in a flexible manner in accordance with the nature of the relationship between words. What determines st rategy use at a given instant appears to result from the interaction o f cues provided by the experimental materials with prior world knowled ge relating to individual words or combinations of words. Recall for c ategorically related pairs was better than for spatially related pairs despite the fact that imagery was used more frequently for spatially related pairs and both types of pair were rated equally for concretene ss and imageability. Alternative explanations for this finding are dis cussed.