Year: 2026 | Month: July-September | Volume: 11 | Issue: 3 | Pages: 282-293
DOI: https://doi.org/10.52403/gijhsr.20260331
Voice Onset Time Transfer Patterns in Assamese-English Bilinguals: Evidence for Asymmetric Cross-Linguistic Influence
Debarshi Das1, Dr Anne Varghese2
1Assistant Professor, Department of Speech-Language Pathology, NAISH College, Bengaluru, India.
2Professor, Department of Speech-Language Pathology, NAISH College, Bengaluru, India.
Corresponding Author: Debarshi Das
ABSTRACT
The Voice Onset Time (VOT) was examined in word-initial stops produced by thirty Assamese–English bilinguals (15 male, 15 females; aged 18–25 years) to characterise the pattern of cross-linguistic influence (CLI) between a four-way laryngeal system (Assamese)(L1) and a two-way system (English)(L2). Each speaker produced 96 tokens, covering voiced unaspirated, voice aspirated, voiceless unaspirated, and voiceless aspirated stops in word and non-word contexts in both languages, embedded in controlled carrier phrases. VOT was measured in Praat software from synchronised waveform and spectrogram displays, following current conventions for defining release and voicing onset. Paired comparisons of by-speaker VOT means revealed a category-specific, asymmetric transfer pattern. Voiced unaspirated stops showed a marginal cross-language difference (Assamese M = 80.0 ms vs. English M = 88.4 ms; d = −0.48, p = .064), with English /b d ɡ/ carrying substantial pre-voicing that is rarely reported for monolingual English speakers. English /p t k/ were produced with short-lag VOT (M = 20.7 ms) rather than the long-lag values typical of native English, indicating systematic L1→L2 transfer of the Assamese voiceless unaspirated category. The aspirated categories (breathy /dʱ/ and voiceless aspirated /tʰ/) did not differ across languages, consistent with wholesale transfer of L1 categories onto the English dental fricatives /θ ð/. Lexicality (word vs. non-word) had no effect on VOT in any category, confirming the phonetic robustness of the contrasts. Significant Gender vs Language interactions emerged, most notably for voiced unaspirated stops in English (males 103.5 ms vs. females 73.3 ms, p < .001). The findings support the Revised Speech Learning Model in predicting that bilingual categories share a common phonetic space, and show that CLI in this population is not uniform but is shaped by the structural mismatch between the two inventories.
Keywords: voice onset time; cross-linguistic influence; Assamese–English bilinguals; L1–L2 transfer; SLM-r; Indian English phonetics; laryngeal contrast