Phonotactic Constraints on Zhangzhou Onsets

Abstract Zhangzhou Southern Min is theoretically assumed to have 7320 possible syllables but more than 70% of them are not attested in the empirical data, implying that substantial constraints have been governing the segmental sequencing and segmental-suprasegmental alignment. This study explores phonotactic constraints on syllable onsets. It addresses two important issues as to in what way syllable onsets are constrained, and what mechanisms have governed the alignment of Zhangzhou onsets with other syllables components and tones to generate syllables that are attestable. The exploration substantially stretches and advances our knowledge of phonological constraints in this Sinitic dialect, while contributing important linguistic data to the typology of phonotactics as an important language phenomenon in world’s natural languages. Keywords: phonotactics, onsets, tones, finals, mechanisms, Zhangzhou 1 Introduction Sounds do not randomly combine with each other to form a functional syllable and/or a word in utterances. Instead, there are constraints posing restrictions on what sounds can be sequenced and how they can be sequenced to form a larger linguistic unit (e.g., Celata & Calderone, 2015; Zec, 1995; Kirby & Yu, 2007; Algeo, 1975; Pearce, 2007). For example, the segmental string *lbick is practically prohibited, because the onset cluster *lb violates the sonority sequencing principle that prefers a rising sonority from syllable edge to nucleus (Giegerich, 1992; Zec, 1995). However, in this case, the liquid sound [l] is more sonorous than the voiced obstruent [b]. Such restrictions on the speech sounds when they are aligned with others are generally known as phonotactics in the literature (e.g., Celata & Calderone, 2015; Algeo, 1975; Kirby & Yu, 2007; Zhang 2006; Pearce 2007). Phonotactics not only can determine syllable structure and inventory of any given language, but also can categorize the sequences of phonemes into linguistically permissible and impermissible. For example, consonant clusters are permissible at both onset and coda positions in English but are disallowable in Maori and Chinese. In Cantonese, there theoretically should have 5130 possible syllables (with tones), but only 36% can be attested in empirical data (Kirby & Yu, 2007). In Zhangzhou Southern Min, a Sinitic dialect spoken in southern Fujian province of southeast China, an inventory of 15 onsets, 61 finals and 8 tones can be identified phonemically (Huang, 2019; 2021). Given such, there should have 7320 (=15*61*8) theoretically possible syllables that would be generated in the empirical data. However, based on the calculation result on Huang (2019)’s rhyme tables, only 2105 syllables are attested in the synchronic speech, implying about 5215 syllables, occupying as many as 71%, are blocked to occur and disable to constitute the permissible syllable inventory. This can be best demonstrated by sequencing constraints on those syllable onsets in this Sinitic dialect. Each onset is logically expected to produce 488 (=61*8) possible syllables; however, 127 syllables are obtained under the unaspirated labial stop /p/, while only 33 syllables under the voiced alveolar fricative /z/, by far fewer than the assumption of 488 syllables for each onset. The strong divergency between the theoretical assumption and the practically attested number …

Content truncated to 500 words