A word prosodic structure study in Vietnamese read and spontaneous speech
Résumé
Vietnamese is traditionally described as a monosyllabic language (see e.g. Nguyễn, 1989). But a significant number of polysyllabic words can be found at the lexical level (51.74% of the 5000 most commonly used Vietnamese words) (Trần & Vallée, 2009). Additionally, Michaud (2004) remarks that the Vietnamese language only is phonologically monosyllabic, but partially polysyllabic at the lexical level. Simple words are monosyllabic and differ from compound words only in the number of syllables (Trương, 1970; Nguyễn, 1999). Vietnamese syllable structure is presented as C1(w)VC2 (the brackets indicating the optional constituents) (Đoàn, 1999). The language licenses only eight segments /p t k m n N/ in coda position in which three voiceless obstruents are unreleased (Cao, 1985; Đoàn, 1999; Kirby, 2011). As a result, final consonant C2 can be found at inter-word boundaries (CVC2#CVC) or, in disyllables, at intra-word boundaries (CVC2.C3VC).
In order to verify if there is any impact of boundary type (inter-word vs. intra-word) on the realization and perception of final consonants, a series of acoustic and perceptual experiments in read speech were conducted (Trần, 2011). We have found that C2 at inter-word (CVC2#CVC) and at intra-word boundaries (C2.C3) are perceptually and acoustically different. This finding leads us to question the monosyllabic status of Vietnamese language: dissyllables have their own prosodic status rather than being simple juxtapositions. What then happens in spontaneous speech where articulatory contexts cannot be controlled? Do C2# and C2.C3 behave the same way in this context as in read speech already studied? Could significant differences of phonetic characteristics of final consonants in monosyllables and dissyllables establish that Vietnamese is not a monosyllabic language?
I will present some preliminary results of a multi-speaker (6 male, 6 female) ongoing study in which a semi-structured interview technique is used. 12 word pairs (already used in previous study) were selected to the test where monosyllabic words are identical to the first syllable of the dissyllabic words (for example, pháp vs. pháp lý, bát vs. bát ngát, lác vs. lác đác, đám vs. đám cưới, bán vs. bán kết, sáng vs. sáng kiến). The interviewer ensures that each stimulus must be pronounced at least 3 times. Acoustic analysis of the final consonant duration and the vowel formant transition has been conducted. The first results could be interesting to discuss the monosyllabic status of Vietnamese.