Cross-modality semantic integration of sentence and picture memory.

The present studies assessed- the effect of intervening items on the accuracy of recognition memory for pictures and sentences. Subjects were presented a sequence of 24 pictures and sentences, later followed by the presentation of 24 intervening items. Each intervening item corresponded to, but was in the opposite modality from, one of the original items. These intervening items were either semantically relevant or irrelevant to the corresponding originals. Subjects then received a same-different recognition test that included original and changed items. Despite the difference in modality, the presence of a semantically relevant intervening item depressed the obtained values of d and the probability of a hit, relative to the effects of an irrelevant intervening item. The data are discussed in terms of support for the integration property of constructive memory. The interpretation is that subjects semantically integrated the original items with the relevant intervening items and made subsequent recognition responses on the basis of the integrated memory. A number of studies have demonstrated subjects impressive ability to recognize pictorial material (Nickerson, 1965; Shepard, 1967; Standing, Conezio, Haber, 1970) and recall it (Bousfield, Esterson, Whitmarsh, 1957). The question of what is stored in memory when a picture is remembered is, however, still unanswered. The evidence that picture memory is superior to memory for words (Paivio, Rogers, Smythe, 1968) and sentences (Shepard, 1967) might suggest that pictures are stored in memory in a qualitatively different and separate way from verbal materials. An alternative position is that when a picture is remembered, the information in I thank James M. Royer ior his advice throughout this research. I also appreciate critical readings