The Stop/Glide Boundary Shift: Modelling Perceptual Data

Abstract

Research in speech perception has suggested that listeners often adjust category boundary locations for differing contexts. The stop/glide (e.g., /ba/ vs. /wa/) boundary shift occurring with changes in syllable duration has been interpreted as reflecting ‘rate normalization’ by some, whereas others have suggested an auditory basis for the shift. Clarification of both the causes and the effects of the stop/glide shift is being sought through experimentation intended to expand our knowledge of variations in the shift across different experimental treatments. The present study extends research on the stop/ glide shift by (1) showing that the shift is not an artifact of a particular stimulus pattern that had been present in all the previous studies of the effect, (2) further demonstrating the nonlinear character of the identification shift, (3) providing additional evidence of the role of presentation format (mixed vs. blocked) in the identification shift, and (4) exploring boundary locations in various discrimination paradigms. The complex outcomes of the studies reported here as well as a variety of previous results are interpreted as compatible with expanded versions of the durational contrast hypothesis. Two versions are considered, one in which the shift is seen as due to a mechanism of contrastive perception that applies to events of comparable duration, and one in which the shift is seen as a perceptual error that occurs when the auditory system attempts to track frequency trajectories that change (i.e., pivot) over time.

Perceptual effects of preceding nonspeech rate on temporal properties of speech categories

“…Thus, for example, listeners may accept longer formant transitions for [b] at slower speaking rates not because they reveal something about typical articulation, but simply because they appear shorter than longer surrounding segments, a pattern previously documented for nonspeech tone durations (Goldstone, Boardman, & Lhamon, 1959;Goldstone, Lhamon, & Boardman, 1957;Walker & Irion, 1979;Walker, Irion, & Gordon, 1981). Oller, Eilers, Miskiel, Burns, and Urbano (1991) speculated further on the precise workings of a durational contrast effect. To account for the nonlinear relation of boundary shifts and apparent rate, they introduced a "comparable range model" whereby contrast effects only occur between two sounds (e.g., a transition and following steady state) with absolute durations falling within a similar range, perhaps a 1:1 to 1:2 ratio.…”

“…Within a durational contrast account of rate effects, there is at least one major reason to predict this null result (see also Gordon,1988, for more general reasons). Whereas the precise workings of durational contrast are as yet unclear, it seems that some degree of spectral continuity (Walker & Irion, 1979) and temporal similarity (Oller et al, 1991) are necessary between contrast-providing context and target stimuli. In the case of phonetic contrasts such as stop-continuant…”

Entrained theta oscillations guide perception of subsequent speech: behavioural evidence from rate normalisation

“…Thus, this study should be viewed as providing a neural implementation of cognitive principles proposed to underlie rate normalisation, such as the principle of durational contrast (Diehl & Walsh, 1989;Oller, Eilers, Miskiel, Burns, & Urbano, 1991;Wade & Holt, 2005). This principle was formulated to have the psycholinguistic function of biasing the perception of ambiguous speech segments towards longer (shorter) percepts when they occur in the context of other shorter (longer) surrounding segments.…”