A Phonological Reanalysis of Eastern Lawa

Journal of the Southeast Asian Linguistics Society JSEALS Vol. 10.2 (2017): 23-65 ISSN: 1836-6821, DOI: http://hdl.handle.net/10524/52413 University of Hawaiʼi Press A PHONOLOGICAL REANALYSIS OF EASTERN LAWA Elizabeth Munn Payap University Abstract Phonological descriptions of Western and Eastern Lawa, two related but mutually unintelligible languages (Nahhas, 2006), differ greatly. Western Lawa is relatively well described (c.f. Mitani, 1972, Schlatter, 1976, Ratanakul and Daoratanahongse, 1985). For Eastern Lawa, three partially conflicting phonological descriptions exist, with consonantal inventories ranging from 19 (Mitani, 1978) to 30 (Lipsius, n.d.) to 33 consonants (Blok, 2013). The vowel systems vary, from 9 (Mitani, 1978) to 24 (Blok, 2013) to 26 vowels (Lipsius, n.d.). In order to investigate the discrepancies between previous phonological descriptions, this study offers a phonological reanalysis of Eastern Lawa vowels and consonants based on recordings from nine Eastern Lawa speakers in Bo Luang and Kiu Lom, Thailand. A comparison with previous research on Eastern Lawa phonology suggests that the different results provided in earlier descriptions are partially caused by differing interpretations and partially due to undocumented phonological processes, which will be presented in this paper. Both synchronic and diachronic issues are considered. Keywords: Eastern Lawa, phonology, Lawa, Waic ISO 639-3 codes: lwl, lcp 1 Introduction This study presents a reanalysis of the phonology of the Eastern Lawa language. It evaluates previous conflicting descriptions of Eastern Lawa and then presents a reanalysis based on newly-colllected data. Eastern Lawa and Western Lawa are the only members of the Lawa group within the Waic subgroup of the Austroasiatic Palaungic language group (Sidwell, 2015a). They are most closely related to the Wa languages (Sidwell, 2015a), as shown in Figure 1. Figure 1: Classification of Eastern Lawa (following Sidwell, 2015a) Copyright vested in the author; Creative Commons Attribution Licence Elizabeth MUNN | Phonological reanalysis of Eastern Lawa | JSEALS 10.2 (2017) Eastern Lawa has about 7,000 speakers (Nahhas, 2006), and is located in northern Thailand, within Chiang Mai Province (Blok, 2013), as shown in Figure 2. Within Eastern Lawa, two major varieties have been mentioned briefly by Lipsius (n.d.) and Blok (2013), labeled according to village names as Bo Luang and Bo Sangae (Blok, 2013). Figure 2: Location of Eastern Lawa villages and Western Lawa villages (Nahhas, 2006) To date there exist three conflicting phonological descriptions of Eastern Lawa. This paper will compare them and offer a new analysis, which will clarify those points on which the previous analyses disagree. The Eastern Lawa recordings used for this study were collected from nine mother-tongue speakers in the fall of 2016 in Kiu Lom and Bo Luang, Hot District, Chiang Mai Province. These two villages speak the same variety. The language consultants available for this study included four men (two from Kiu Lom and two from Bo Luang) and five women (three from Kiu Lom and two from Bo Luang). Their ages ranged from 26 to 76. A 436-item wordlist with basic vocabulary was recorded from each speaker, and an additional 510 words were recorded from a 54-year-old female speaker from Kiu Lom. The wordlist was elicited in Central Thai. Each speaker was asked to repeat each word twice in isolation, then construct an utterance including the word, and then repeat the word once more in isolation. Data from all speakers was transcribed and analyzed, supported by the software programs Speech Analyzer and Phonology Assistant. 2 Previous research Although Proto-Austroasiatic likely did not have register (Sidwell, 2015b), many Austroasiatic languages have since developed a register distinction, and register is now a characteristic feature of Austroasiatic languages (Huffman, 1976). The prototypical pathway by which register develops and may subsequently be lost in Austroasiatic languages is the so-called “Khmer model of registrogenesis” (Huffman, 1976). Registrogenesis is triggered by distinctive voiced and voiceless onsets (stage 1), with voiced onsets leading to breathy phonation of the following vowel and subsequent devoicing of the initial stops (stage 2). Former environmentally conditioned register phonologizes as as the voicing contrast in onsets is lost (stage 3). Finally, breathy or lax register gives way to diphthongization or tonality; phonemic register contrast is no longer present. Different branches of the Austroasiatic family present variations on this central theme, but many have gone through this process in some form or other. The Katuic languages, for example, can be identified at various stages of this process: the most conservative, Katu, is still at stage 1, Souei (a Kuay variety) is at stage 2, and phonemic register (step 3) is found in several languages (Gehrmann, 2015). Some Katuic languages confirm Gregerson’s (1976) observation of a general dispreference for tense close vowels and lax open vowels. Some varieties have only lax close vowels and tense open vowels, while others lowered tense close vowels and raised lax open vowels (Gehrmann, 2015). 24 Elizabeth MUNN | Phonological reanalysis of Eastern Lawa | JSEALS 10.2 (2017) According to Diffloth (1980), Proto-Waic was at step 1 in this registrogenetic process, with a series of voiced and voiceless initials. He suggests that after that, a Proto-Wa-Lawa group split off from the rest of the Waic languages as it merged the voiced and voiceless series, and developed a register contrast subsequently. From there, the Proto-Wa-Lawa group itself split up. Many Wa-Lawa varieties no longer have contrastive register today (Diffloth, 1980), and they vary widely in their present-day realizations of those proto-language register contrasts. For some varieties such as Wa, however, the register contrast has been retained (cf. Watkins, 2002). Proto-Lawa had its own particular set of conditions in which it developed diphthongization and replaced its register contrast with a vowel contrast between diphthongs and monophthongs: “The evolution of PW *a to PLw *ɨa looks very similar to what we saw in Drage’s Wa, but the conditions for the appearance of diphthongization in Lawa are partly different from those of the Wa languages. These conditions have been discovered by Mitani (personal communication) and also apply to the Northern dialect which was not included in this survey; they are characteristic of the whole Lawa branch and no other branch of Waic.” Diffloth (1980:46) Proto-Wa-Lawa, with its register contrast, was in the diachronically unstable situation of having tense close vowels and lax open vowels, in contrast with the more natural lax close vowels and tense open vowels. After Proto-Lawa split off from Proto-Wa, Proto-Lawa vowel qualities shifted. Lax or breathy open vowels – in the case of Proto-Lawa, this was on (...truncated)