The False-Recognition Effect for Homophones of Items Subjects Are Cued to Forget

Abstract

In two experiments, homophones of low orthographic distinctiveness from study items subjects were explicitly cued to forget yielded a lower false-recognition effect than did homophones of low orthographic distinctiveness from study items subjects were explicitly cued to remember. This variation in the magnitude of the effect is seen as reflecting a variation in the depth of featural processing accompanying the two different types of cue. By contrast, homophones of high orthographic distinctiveness yielded a reversal of the false-recognition effect (i.e., yielded fewer false positives than the control items), regardless of the type of cue on the study items. In recognition learning, new test items may be either related or unrelated to prior study items. The greater rate of false positives for the related than unrelated items defines the false-recognition effect (Underwood, 1965). A feature-tagging theory of recognition learning (see Kausler, 1974, pp. 481-487) accounts for this effect in terms of overlapping features between the representations of old and new items in permanent memory. Briefly, the processing of the study items yields 'occurrence' tags for their sensory and semantic features stored in memory. A decision whether a test item is old or new is presumed to be mediated by a search of that item's features for the presence or absence of such tags. Since related new items have a greater probability of revealing tagged features than unrelated items, they also have a greater probability of being misidentified as old.