Loading...
A Reexamination of Tone in Tujia
Feng, Angela
Feng, Angela
Citations
Altmetric:
Abstract
This project concerns the representation of tone in Northern Tujia (henceforth Tujia), a Tibeto-Burman isolate. Drawing on fieldwork data collected in 2018-2024 in Dianfang Township, Hunan Province, China, I argue that tone in Tujia should be considered at the word level, not the syllables. That is, I propose a revision of tonal representation in Tujia such that monomorphemic polysyllabic words may be notated with one tone, as in (1)Íž I use superscripts rather than IPA to avoid confusion with pinyin. Secondarily, I infer that grammatical particlesâbolded in (2)âare underlyingly toneless, as their tone is predictable from the preceding morpheme and general prosody.
(1) khatChiemaHL
frog
âfrogâ
(2) k7ts1HL ĂŁmMH-poH wuMH-nouH s1L-tiuH%
3PL 1PL.OBL-GOAL cow-CLF send-DIR:PRF
âThey sent us a cow.â
Data & discussion. Tone inventory in Tujia has long been a matter of debate, beginning with Tianâs (1986) observation that high-level and high-falling tones appear to freely vary. More recently, Xiong (2016) has made the case that up to 6 tones may be distinguished. I believe that these authorsâ focus on the monosyllabic word narrowed their analytic vision. For instance, consider the collection of monomorphemic disyllabic words in Chen (2006). Of these 82, five preliminary contours can be identified that account for the majority of the data: H-H (30.5%), H-L (17.1%), M-H (15.9%), L-H (12.2%), and L-L (9.8%). Following analysis by Tian et al. (2019) that low tones alternate with rising tones at the end of utterance, L-H may be considered an allophone of L when followed by boundary tone H%. This distribution is unexpected in a system where each syllable has a theoretically equal chance of being high (or high-falling), low, and rising. In (3), I show a sample of words with their entries in Chen (2006) and Zhang (2006) followed by their representation under my proposed systemÍž my final inventory consists of H, HL, MH, and L.
(3) a. pu55tshi55 (Chen 2006) / pu53tshi53 (Zhang 2006) > putsh1H âspiderâ
b. xa53lie21 (Chen 2006) / xa53lie21 (Zhang 2006) > xalieHL âdogâ
Under this analysis, processes previously labeled as tone sandhi can be understood as rightward tone spreading. H and HL in monosyllabic words freely vary, but they are distinguishable in polysyllabic words, as in (3), and in sequences of monosyllabic words followed by grammatical particles, as in (4). Thus, true rightward replacive tone sandhi must be teased apart from rightward tone spreadingÍž in fact, my fieldwork shows that many compound words do not productively undergo tone sandhi.
(4) a. tChuH-poH
home-GOAL
âtoward homeâ
b. zĹŠmH-poL
little.sister-GOAL
âto/for (my) little sisterâ
Conclusion. This analysis aligns with the data and is theoretically motivated, as tone is expected to be assigned to a morphological unit rather than a syllable. Moreover, it is typologically unsurprising, given that nearby Qiangic languages have been analyzed as having word-level tone characterized by culminativity (Chirkova and Michaud 2009Íž Evans 2008). Such reexamination of tone is much-needed for accurate transcription of the Northern Tujia language.
Description
These are the slides from a presentation given at the 58th International Conference on Sino-Tibetan Languages and Linguistics (ICSTLL58) held at the University of Bern on 09/04/2025.
Date
2025-09-04
Journal Title
Journal ISSN
Volume Title
Publisher
University of Kansas
Files
Loading...
FengA_2025.pdf
Adobe PDF, 2.25 MB
Research Projects
Organizational Units
Journal Issue
Keywords
Phonology, Tone, Tibeto-Burman, Qiangic, Linguistics