Perceptual test of the error severity model of the recovery process for articulatory disorders in children

Abstract

According to the error severity model of the recovery process from articulatory disorders, children progress from errors of omission to substitution to distortion before correct sound production. Inherent in this model is the assumption that the distorted phoneme is perceived as a unique sound category. Although categorical perception studies have shown within category perception, by far the most robust effects have been observed across categories. Consequently, categorical perception theory would provide only weak support for the above assumption. Since the perception of a distorted phoneme category is fundamental to the error severity model, this study tested the perceptual reality of the distorted /r/. In order to determine if subjects could categorize acoustic properties of /Cej/‐type utterances varying in F2 and F3 onset frequencies as distorted /r/, speech pathologists identified two 25‐stimuli manifolds synthesized so as to approximate /r/, /w/ and defective /r/ sounds modeled after child and adult vocal tracts. Subjects scaled the stimuli between /r/ and /w/ in one condition and labeled them as /r/, /w/, or distorted /r/ in another condition. The results showed that subjects consistently identified a distorted /r/ category for the adult manifold but not for the child manifold. These results cast some doubt that the distorted /r/ is a viable perceptual category for speakers who hold the greatest interest among speech pathologists—children. [Work supported in part by NINCDS.]