This is the abstract of a TULCON14 presentation.
Haili Su (University of Toronto)
The phenomenon of syllable contraction has been a major source of regional and idiosyncratic variations among Chinese languages. It is especially salient in the daily use of Taiwan Mandarin (TM), the predominant Standard Chinese (SC) variety spoken in Taiwan. The term syllable contraction refers to the shortening of a word by omitting one or more of its segments, usually during casual, spontaneous speech. There has been rich scholarly literature on the phonetic qualities, phonological distribution, pragmatics and motivations for syllabic contraction in TM. Yet little analysis was done on how the contracted forms are produced, which is unlikely an arbitrary process. To provide an account for the mainly phonological process, I analyzed the fully contracted forms gathered from Kuo (2010), Tseng (2005), and Chung (2006). The term full contraction refers to the monosyllabic contracted forms of disyllabic words. Here are two examples:
(1) /tʂɭ.taw/ 🡪 [tʂaw] (知道 zhī dào) ‘to know’ (Kuo, 2010)
(2) /ni.mən/ 🡪 [nim] (你們 niˇ mén) 2PL(Chung, 2006, Tseng, 2005)
With the data, it is not difficult to see patterns behind the contraction. First, most of these contracted forms retain the original left and right edges, and one of the original nuclei as its nucleus. But the edge-alignment is not always the case. What factors dictates the choice of nucleus is also crucial to be accounted. More unusually, what is happening when the contracted forms like (2) that appear to have violated the basic SC phonotactics? What could be the phonological and non-phonological—specifically, morphological—factors behind it? Finally, the gradient nature of this variation means that not everyone contracts the same words in the same ways all the time. A model for the derivation mechanism needs to take all these questions into consideration.
In my paper, I attempt to account for these disyllabic contraction forms with analyses under the Optimality Theory (OT) framework. The analyses establish that the full contraction process of disyllabic words in TM likely starts with edge-association, but only with the left edge—in contrast to the Edge-in theory common to account for contraction in other Chinese varieties. The second most important set of constraints concerns the nucleus of the contracted form, which is revealed to be determined by a sonority hierarchy that is language-specific (a > e > {o, ɤ} > ə > {i, y} > u). There is also a constraint to motivate to maximize the output syllable, yet it is subject to the constraint to preserve the original linear order of the segments. It is also found when there are two possibilities of glides and/or nasals right edges, nasals are preferred. Lastly, when the two potential right edges are equally ranked, the alignment constraint prefers the right edge of the input. The analyses reveal the prominence of the sonority of a segment in the contraction process in Taiwan Mandarin, while suggesting the low priority of the right edge—likely due to lower number of syllable codas in SC than other Chinese varieties. Then, through examining the cases that violate the phonotactics of SC, I will also discuss the possible phonological, morphological, semantic, and sociolinguistic motivation behind these examples, including the influence of Taiwan Southern Min (Taiwanese). Though more data is needed for further analysis