Charles Redmon, Allard Jongman, and Jie Zhang (2017)
Poster presented at the 174th Meeting of the Acoustical Society of America
Telugu is one of a small set of languages described as exhibiting three or more place contrasts among sibilant fricatives (less than 4% of languages in UPSID; Maddieson and Precoda, 1990). In a pilot study of alveolar, palatal, and retroflex sibilant productions in VCV sequences, Telugu speakers were found to show consistent inter-speaker variation in the production of the palatal sibilant, with half showing evidence (in production and perception) of a complete merger with the alveolar, and the other half showing evidence of a merger with the retroflex, and no clear demographic differences underlying the delineation of the two groups. Following up on this result, we present acoustic data from native speakers of Telugu recorded in Hyderabad producing words with the three sibilants in multiple vowel contexts, positions (CV, VC), and lexical neighborhood structures (with vs. without sibilant competitors in a second order neighborhood), focusing in the acoustics in particular on formant transition cues to palatal and retroflex places. This data are presented both to clarify the present state of the Telugu sibilant system and to explore what effect the lexical distribution of these sounds has on the acoustic feature mapping of the contrast.