Phonetic: Semantic Mediated False Recognition: Does Activation Fail to Spread?

Abstract

Spoken word recognition involves brief activation of candidate words. Six experiments examined whether words semantically related to phonologically activated words would be falsely recognized. Experiments 1 and 2 involved homophones as test words; Experiment 3 used strong associates for the semantic mediation link. Experiment 4 approximated lists of "strong" converging associates. Experiment 5 expanded the real time needed for word identification by using a gating procedure during study. In Experiment 6, the goal was to create a more sensitive test by requiring participants to indicate which of two lures (mediated or control) was "most likely" to be new. Recognition errors were sensitive to separate phonetic and semantic stages in the mediated chain; however, there was little evidence of mediated false recognition, despite expectations derived from common models of spreading activation.

False recognition following study of semantically related lists presented in jumbled word form

“…A recent series of experiments failed to demonstrate phonetic-semantic mediation effects in false recognition. Based on phonetic/orthographic similarity, early in the processing of a studied word such as hamster, words similar in sound and appearance (e.g., ham, hamburger, hammer) may be activated; however, data indicated that activation of hammer during the initial stages of processing hamster did not lead to an increase in false recognitions to the critical lure ''nail'' (Wallace, Salapska-Gelleri, Belz, & Owen, 2006). In the case of nonword study items such as dearm, if ''dream'' is activated (either through misperception or recoding), its semantic associates should have relatively few semantic competitors that would be directly elicited by the nonword dearm.…”