Participants took part in two speech tests. In both tests, a model speaker produced vowel-consonant-vowels (VCVs) in which the initial vowel varied unpredictably in duration. In the simple response task, participants shadowed the initial vowel; when the model shifted to production of any of three CVs (/pa/, /ta/ or /ka/), participants produced a CV that they were assigned to say (one of /pa/, /ta/ or /ka/). In the choice task, participants shadowed the initial vowel; when the model shifted to a CV, participants shadowed that too. We found that, measured from the model's onset of closure for the consonant to the participant's closure onset, response times in the choice task exceeded those in the simple task by just 26 ms. This is much shorter than the canonical difference between simple and choice latencies [100-150 ms according to Luce (1986)] and is near the fastest simple times that Luce reports. The findings imply rapid access to articulatory speech information in the choice task. A second experiment found much longer choice times when the perception-production link for speech could not be exploited. A third experiment and an acoustic analysis verified that our measurement from closure in Experiment 1 provided a valid marker of speakers' onsets of consonant production. A final experiment showed that shadowing responses are imitations of the model's speech. We interpret the findings as evidence that listeners rapidly extract information about speakers' articulatory gestures.

Three experiments were designed to investigate how listeners to coarticulated speech use the acoustic speech signal during a vowel to extract information about a forthcoming oral or nasal consonant. A first experiment showed that listeners use evidence of nasalization in a vowel as information for a forthcoming nasal consonant. A second and third experiment attempted to distinguish two accounts of their ability to do so. According to one account, listeners hear nasalization in the vowel as such and use it to predict that a forthcoming nasal consonant is nasal. According to a second, they perceive speech gestures and hear nasalization in the acoustic domain of a vowel as the onset of a nasal consonant. Therefore, they parse nasal information from a vowel and hear the vowel as oral. In Experiment 2, evidence in favor of the parsing hypothesis was found. Experiment 3 showed, however, that parsing is incomplete.Theories of speech perception differ in their characterization of phonological categories and of the stimulus information that listeners are presumed to use to detect the phonological properties ofspoken utterances. In some theoretical accounts, phonological categories are cognitive categories in the mind; listeners identify intended phonological properties of an utterance from associated acoustic and, sometimes, optical cues. The results of initial stages of auditory (visual) analysis of a speech stimulus are mapped onto the mental phonological categories. In other accounts, phonological categories are gestural in nature; listeners identify the phonological properties of utterances by extracting acoustic and, sometimes, optical information about them that is available because gestural causes have specifying acoustic and, sometimes, optical consequences.As distinct as these views are conceptually, they have proven difficult to distinguish experimentally. Generally, researchers agree on what serves as acoustic information for the phonological properties of words. They disagree on what the acoustic information serves as information for: abstract, mental categories or phonological actions of the vocal tract.One approach that we have taken to try to distinguish the views experimentally has been to look closely at listeners' modes of attention to the acoustic speech signal. Although researchers generally agree on what serves as acoustic information, there may be some cases in which listeners who are attempting to extract information about gestures should differ in their mode of attention to the

No abstract