The use of duplex perception to study silence as a cue for stop consonants

Removal of the silent closure interval from /spa/ or /sta/ typically results in the perception of /sa/. To determine whether this phenomenon is an instance of phonetic (as distinct from auditory) perception, we used the method of duplex perception. This has the advantage that a single stimulus pattern yields speech and nonspeech percepts simultaneously [D. Isenberg and A. M. Liberman, J. Acoust. Soc. Am. 64, S20(A) (1978)]. Applying this method to the synthetic syllables /spa/ and /sta/, we separated the F₂‐F₃ transitions (which alone distinguished the syllables} from the base (the pattern common to the two syllables). When the base is presented in isolation it sounds like (s‐a/ or /as/ depending on whether or not a silent interval is present. The transitions alone sound like nonspeech chirps. When the base and the transitions are presented dichotically, listeners simultaneously hear a syllable and a nonspeech chirp. In the ear which receives the base, they hear /spa/ or /sta/ when a silent interval is present, contingent on the transition in the other ear; when the silent interval is absent, /sa/ is heard with either transition. In contrast, the nonspeech chirps, heard in the ear which receives the transitions, do not change with the presence or absence of the silent interval. Given that the speech changes while the chirp does not, the effect of silence in cueing the stops can hardly be interpreted as an instance of auditory interaction: it can more plausibly be seen as a result of distinctively phonetic processes. [Work supported by NINCDS and NICHD.]