Philip J. Monahan (University of Toronto) will give a virtual colloquium on November 4th.
< b>Title: Our Neurophysiology Supports Abstract Phonological Feat ures: Before and During Speech
Abstract: The fundamenta l nature of speech sound representations remains intensely debated. This is despite over a half-century of detailed work. While generative theories ha ve long posited a role for abstract phonological features\, their support f rom psycholinguistic investigation has been sparse and sometimes equivocal. This talk presents the results from four experiments investigating how spe ech sounds are represented in the brain. All four experiments employ a many -to-one oddball paradigm. In such designs\, participants habituate to a rep eated standard stimulus which is intermittently interrupted by a deviant. H uman brain responses are recorded and the mismatch negativity (MMN) is meas ured. First\, I use magnetoencephalography (MEG) to investigate how mid vow els are represented. Across two experiments\, I report asymmetric responses for when the standard is a low or high vowel compared to when the standard is a mid vowel. These results are consistent with an underspecified featur al account for the place of articulation of mid vowels. Moreover\, evidence from the time-frequency domain suggests that this abstract featural knowle dge plays a role in predicting the upcoming stimulus. Next\, I present resu lts from two electroencephalography (EEG) experiments using a similar parad igm. In particular\, I test consonant voicing in English and retroflex cons onants in Mandarin Chinese. Both experiments again show asymmetric brain re sponses consistent with featural representations of voicing and retroflex. Time-frequency analyses of the EEG response highlight differences prior to the presentation of the stimulus\, suggesting that the brain utilizes such featural representations to predict upcoming speech sounds. Taken together\ , abstract phonological features are not only supported by the brain but ar e used to help predict the incoming signal.
Zoom information: