Auditory Word Frequency Effect within Homophone Families and the Activation of Homophone Representations

To those lexicons with larger homophone families,e.g.Chinese monosyllabic words,it has remained unclear what influence homophonic lexical ambiguity has on the lexical access of auditory modality and the activation of word representations.This study of speech comprehension included two experiments of auditory modality.Experiment 1 adapted a dictation paradigm requiring participants to listen to isolated Chinese homophonic monosyllables and write down one word corresponding to each syllable.The participants performed their tasks well and without hesitation.Although each individual could make the lexical access,the population showed lexical ambiguity:the larger a homophone family was,the larger variety of homophones was chosen from that family by the participants.However,listeners tended to chose words of high frequencies within homophone families rather than choosing arbitrarily or by chance.Experiment 2 adapted a cross-modal paradigm of syllable-character homophonic judgment task in which a monosyllabic sound was presented to the ears followed by a visually displayed Chinese character.The participants were required to judge whether the visually displayed characters had the same pronunciation as the preceding auditory syllables.The response time for judging homophonic pairs was compared to response time for judging non-homophonic pairs.The difference,called homophonic facilitation,revealed that when a heard syllable activated its phonological representation,high frequency homophones of that syllable were more often activated automatically,while low frequency homophones were mostly suppressed.The results of the two experiments demonstrate that there exists an auditory word frequency effect within each large homophone family.The activations of homophone representations are different between high,middle,and low frequency words within the families,resulting in non-exhaustive access.The most frequent homophones within families were more likely to be accessed.These findings are difficult to be explained by using the current models of lexical access and activations of homophone representations.We present a new model which is able to explain the new findings from auditory modality.