Perception of voicing and place of articulation in labial and alveolar English stop consonants

Abstract

Distinctive features define a multidimensional structure that must be implemented in speech production and perception. A multilevel Gaussian Gen eral Recognition Theory model is presented as a model of multidimensional fea ture perception. The model is fit to data from three experiments probing iden tification of noisemasked, naturallyproduced labial and alveolar English stop and [d] in onset (syllableinitial) and coda (syllablefinal) position. The results indicate systematic perceptual deviations from simple place and voicing structure in individual subjects and at the group level. Comparing onset and coda positions shows that syllable position modulates the deviation patterns, and comparing speechshaped noise and multitalker babble indicates that deviations from simple feature structure are reasonably robust to variation in noise characteristics. Possible causes of the observed perceptual confusion pat terns are discussed, and extensions of this work to studies of feature structure in speech production and investigation of nonnative speech perception are briefly outlined.

On the Origin of Post-Aspirated Stops: Production and Perception of /s/ + Voiceless Stop Sequences in Andalusian Spanish

“…104-106). 7 As far as perception is concerned, there are experiments indicating that the voicing contrast in English is more salient for alveolar than for bilabial stops (Silbert, 2014). If it is assumed that, due to aerodynamic factors, post-aspiration in /st/-sequences is perceptually more prominent than in /sp/-and /sk/-sequences, then it is possible that listeners first start to imitate the long VOT in /st/-words, and only later generalize it to the velar and the bilabial context.…”

Identifiability and testability in GRT with individual differences

“…We propose that, in addition to the necessity of fixing the location and scale of GRT models, the dimensional orthogonality of GRT models must be fixed, as well. We argue, following Silbert and Thomas (2013), that this is typically best done by assuming decisional separability, though we also discuss other possible approaches, noting that the scope of this assumption in single-subject identification and concurrent ratings models is straightforward, whereas it is somewhat less so in multilevel models like GRTwIND and the models described by Silbert (2012Silbert ( , 2014. We conclude with a brief discussion of the relationship between universal perception and the concept of perceptual primacy.…”

“…GRT is a two-stage model of perception and response selection (Ashby and Townsend, 1986;Kadlec and Townsend, 1992;Thomas, 2001b;Silbert, 2014;Wickens, 1992). The first stage consists of noisy perception.…”

“…The most common use of GRT is to analyze identification-confusion data in a 2 × 2 factorial paradigm, wherein the stimuli consist of the factorial combination of two levels on each of two dimensions (e.g., Ashby and Townsend, 1986;Silbert, 2012Silbert, , 2014Thomas, 2001b,a). For example, a factorial combination of frequency and intensity could produce a set of stimulus tones that are low or high frequency and low or high intensity.…”

“…Two recent extensions of GRT have focused on the simultaneous analysis of multiple subjects' data. One of these extensions is a Bayesian model in which each subject's data is fit to a standard 2 × 2 model, while, simultaneously, the individual subjects' parameters are modeled as random variables governed by group-level parameters (Silbert, 2012(Silbert, , 2014. The other extension is GRTwIND, in which a group-level set of parameters are shared as well as partially modified and complemented by subject-level parameters .…”