Effects of type of input and type of processing on recall hypermnesia.

Abstract

Abstract Twenty undergraduates participated in one of six conditions representing two types of input (words, pictures) and three types of initial processing: imagery (generate object imagery or real - world imagery), semantic (rate pleasantness or form a verbal association), and physical (rate complexity or judge the presence of a feature). After viewing 70 items, subjects attempted to recall them four times. Except for physical processing of words, the number of items correctly recalled improved over trials, demonstrating hypermnesia. Intrusions increased over trials in most conditions and appeared later within a trial for words than for pictures. Of five alternative accounts of hypermnesia, these results favour the sensory/semantic model of memory coding.Resume Vingt etudiants du premier cycle ont pris part a une de six situations representant deux types d'intrants (mots, images) et trois types de traitement initial: images mentales (creer une imagerie d'objet ou une imagerie reelle), semantique (evaluer le caractere agreable ou etablir une association verbale), et physique (evaluer la complexite ou juger de la presence d'un trait caracteristique). Apres avoir regarde 70 articles, les sujets ont essaye de se les rappeler quatre fois. Sauf pour le traitement physique des mots, le nombre d'articles correctement rememores s'est ameliore d'un essai a l'autre, ce qui demontre l'hypermnesie. Les intrusions ont augmente d'un essai a l'autre dans la plupart des situations et sont apparues plus tard dans les essais portant sur des mots que sur ceux portant sur des images. A partir de cinq comptes rendus possibles d'hypermnesie, il semble que ces resultats favorisent le modele sensoriel/semantique de codage de la memoire.Change in recall over repeated trials has been of interest throughout the century of experimental psychology (Payne, 1987), particularly since the work of Erdelyi and Becker (1974). To control for the possibility of a progressively relaxed response criterion, they required subjects to produce a high fixed number of nonrepeating responses on each recall attempt. Over three trials, recall remained constant for concrete words but improved for pictures. Erdelyi and Becker called this improvement hypermnesia.On the assumption that recall involves a two - step process in which candidate items are generated then judged as correct or incorrect, Erdelyi and Becker suggested that retrieved picture input was almost perfectly recognized, leading to clear marking in memory and, on later trials, to faster recall of previous responses, giving extra time to search memory for unrecovered items. However, recognition of retrieved word input was poorer than that of pictures, leading to more incorrect responses (intrusions), and to less accurate marking. Because some subjects reported forming visual images during presentation and showed improved word recall over trials, Erdelyi and Becker also speculated that coding modality, not input modality, might be the critical factor in hypermnesia.In an extensive review, Payne (1987) concluded that picture hypermnesia is robust, but that word hypermnesia only occurs with imagery or semantic processing. If pictures are usually processed semantically (Erdelyi, Buschke, & Finkelstein, 1977), this implies that the critical factor in hypermnesia may be type of processing. Imagery and semantic processing both involve stimulus elaboration or richer coding, creating multiple retrieval cues (Craik & Tulving, 1975) that may be accessed on repeated trials (Roediger, Payne, Gillespie, & Lean, 1982). It follows that hypermnesia should not differ between imagery and semantic processing of pictures or words, and that it should not occur at all with physical or nonsemantic processing.The finding that single - trial free recall is higher for pictures than for words (e.g., Paivio, Rogers, & Smythe, 1968) could thus be attributed to depth of processing. …