The effects of tongue shape categories on tongue segmentation in English

Abstract

In our recent work, we have proposed that the tongue moves by compressing and expanding local functional segments. For any single gesture, functional segments may move in similar or opposite directions to compress and expand the tongue locally. High correlations between segments suggest biomechanical constraints. Low correlations suggest independent control of these segments. Our previous studies on English have shown a front-back division of the tongue, where adjacent segments correlate positively with each other (moving in the same direction) and distal segments correlate negatively with each other (moving in opposite directions). Individual segments aggregate with adjacent or distal segments. Furthermore, the phonemic content of the dataset influences these aggregations and the location of the front-back division of the tongue. This study will more deeply explore these phonemic effects. In particular, we will examine the effects of the transitions between consonants and vowels of the four basic tongue shape categories of English [front raising (e.g., /n,i/), back raising (e.g., /ng,o/), complete channel (e.g., /s,ae/) and two point displacement (e.g., /l/)] on tongue segmentation and the location of the pivot point for the front-back division. [Work supported in part by NIDCD/NIH Grant RO1-DC01758 and by NIH Grant T32-DE07309.]