Effects of talker voice on lexical neighborhood and speech perception

Abstract

Previous research has shown that the number of words that a syllable is similar to (lexical neighborhood) influences the perception of the syllable and its phonemes. For example, in ‘‘bowth’’ ‘‘powth’’ and ‘‘bows’’ ‘‘pows’’ series, ‘‘bowth’’ is similar to more words than ‘‘powth,’’ while ‘‘pows’’ is similar to more words than ‘‘bows.’’ Listeners report more syllables beginning with /b/ in ‘‘bowth’’ ‘‘powth’’ and more that begin with /p/ in ‘‘bows’’ ‘‘pows.’’ However, with one particular talker, no effects of lexical neighborhood were found. This research focused on two explanations for the lack of neighborhood effects for this talker. One possibility is that the previous method of determining lexical neighborhood is inaccurate for this talker. Alternatively, the time course for phonetic and lexical information processing may be different for this talker. In this research, listeners identified the final consonant in two nonword CVC series: ‘‘faesh’’ ‘‘faech’’ and ‘‘taesh’’ ‘‘taech.’’ With a syllable final target phoneme, lexical neighborhood influences should build up prior to the processing of the target phoneme. Implications of the data for models of word recognition and perception with different talkers will be discussed. [Work supported by NIDCD Grant No. R01DC00219 to SUNY at Buffalo.]