Abstract
From an articulatory perspective, "uvularization" refers to the retraction of the tongue dorsum toward the uvular region, an articulatory gesture primarily caused by the contraction of the styloglossus and related muscles (Evans et al. 2016). Although uvularization has traditionally been viewed as a feature of consonants in the world's languages (e.g., Al-Tamimi & Heselwood 2011), in recent years, scholars have also identified vowel uvularization in several Tibeto-Burman languages, such as Qiang (Sun & Yu 2013; Evans et al. 2016), Xinlong Muya (Van Way 2018), and Queyu (Zheng 2023; Guan 2024). This paper aims to explore the acoustic and phonological characteristics of uvularized vowels in the Rongba dialect of Queyu. Queyu (ISO 639: qvy) is an under-documented Sino-Tibetan language mainly distributed in Litang, Xinlong, and Yajiang counties of the Ganzi Tibetan Autonomous Prefecture in Sichuan Province, China (Lu 1985). The vowel system of Rongba Queyu is relatively complex, consisting of eight plain vowels, five of which have corresponding uvularized counterparts (Zheng 2023). The pairing of plain and uvularized vowels is supported by morphological evidence, such as the parallelism in verbal vowel alternation patterns. As shown in the table below, in a parallel alternation pattern, the verb vowel changes to [o] or [oʶ] when the subject is first-person singular; conversely, when the subject is second-person plural, the vowel changes to [e] or [eʶ], respectively. Furthermore, uvularized vowels participate in vowel harmony processes. For example, the negative prefix [mɛ-] is assimilated by the uvularized vowel [oʶ] of the following verb stem, surfacing as [maʶ]. This confirms that [aʶ] is the uvularized counterpart of [ɛ]: [mɛ-tʰó] → [mɛ́-tʰʊ̀] 'neg.npst-drink.1sg'; [mɛ-roʶ] → [màʶ-róʶ] 'neg.npst-laugh.1sg'.
Acoustically, compared to their plain counterparts, uvularized vowels exhibit a lower second formant (F2) and a larger gap between the third formant (F3) and the second formant (F2). Perceptually, there may be significant differences in quality between plain vowels and their paired uvularized vowels. For instance, the plain vowel /e/ is realized phonetically as [ɪ], while its uvularized counterpart /eʶ/ sounds closer to the central vowel [ə]. These acoustic and perceptual differences are related to the physiological articulatory mechanism of uvularization: namely, tongue root retraction. Additionally, the phonetic realization of uvularized vowels is influenced by the preceding consonant onset, and their quality may exhibit subtle differences that require careful attention to distinguish. Phonologically, by examining the distribution of uvularized vowels and related phonological processes, this study chooses to analyze uvularization as a vowel feature rather than a consonantal feature. First, in terms of distribution, uvularized vowels, like plain vowels, can co-occur with consonants at various places of articulation (i.e., labial, coronal, and dorsal). This distributional pattern differs from the partial contrast patterns found in other languages (such as the "coronal emphatic consonants" in Jordanian Arabic, Al-Tamimi & Heselwood (2011)) or non-phonemic secondary articulations (such as the velarized lateral [ɫ] in English, Ladefoged & Maddieson (1996: 360-316)). This suggests that analyzing uvularization as a vowel feature is more consistent with the principle of economy in phonological analysis; otherwise, the number of contrasting consonants would double. Second, vowel harmony provides further evidence. The uvularization feature of the vowel spreads to the left, extending from the verb stem to the prefix. That is, if the vowel of the verb stem is uvularized, the directional prefix attached to that stem will also be fully uvularized, whereas plain vowels do not trigger this change. For example: [kə-tɜ́] → [kə́-tɜ̀] 'pfv-pour(water).1sg'; [kə-xtʰɜ́ʶ] → [qə́ʶ-χtʰɜ̀ʶ] 'pfv-ask.1sg'. Such evidence indicates that vowels are divided into two natural classes: uvularized vowels can trigger vowel harmony, while plain vowels cannot.
However, within the same syllable, uvularization appears to act simultaneously on both the vowel and the consonant. For example, when the vowel of a syllable is uvularized, the initial consonants [x-] and [ɣ-] in a consonant cluster are assimilated into uvular fricatives [χ-] and [ʁ-]: [xpóʶ] → [χpóʶ] 'ice'; [ɣzɜ́ʶ] → [ʁzɜ́ʶ] 'cow'. This suggests that uvularization might be a property of the entire syllable. Therefore, it remains to be determined whether uvularization should be regarded as a suprasegmental feature. From an areal perspective, similar secondary articulation phenomena also appear in neighboring languages, such as "velarization" (Sun 2000, 2004; Lin et al. 2012; Gong 2018), "pharyngealization" (Evans 2006; Chiu & Sun 2020), "Retracted Tongue Root (RTR)" (Gao 2015), or "tense-lax" contrasts (Huang 1991; Huang 2023). However, it is not yet clear whether these terms truly represent distinct articulatory mechanisms. To further investigate these differences, it is necessary to study the articulatory mechanisms using methods such as ultrasound imaging. Nevertheless, while acoustic descriptions can provide objective evidence, the phonological status of these secondary articulation features must still be determined through phonological evidence such as phonotactics and related phonological processes. A detailed investigation and description of the vowel systems in these languages will contribute to understanding the origin and evolution of such systems within Tibeto-Burman.
Full Text
Preamble
Department of Chinese Language and Literature, Peking University
Linguistics Laboratory
6. 结论与展望
Uvularization refers to the retraction of the tongue dorsum toward the uvula. This articulatory gesture is primarily caused by the contraction of the styloglossus and its associated muscles \cite{Evans}. This phenomenon has been classified as a characteristic of emphatic consonants in Jordanian Arabic \cite{Tamimi, Heselwood}. Scholars have also identified uvularization in several Tibeto-Burman languages, such as the Mawo and Yunlinsi dialects of Qiang \cite{Evans}, as well as in Queyu (Choyul), which is actually a dialect of the Horpa language \cite{Zheng}.
Queyu (Choyul) is an under-documented Sino-Tibetan language spoken primarily along the Yalong River in Yajiang County, Garzê Tibetan Autonomous Prefecture, Sichuan Province. It has approximately [NUMBER] speakers, all of whom are ethnically classified as Tibetan. The language is currently considered threatened or even moribund \cite{Roche, Zheng}. While recent studies have noted uvularization in the language \cite{Suzuki, Soman, Wangmo}, earlier literature instead characterized this feature as velarization \cite{Nishida}. The data sources for this study are as follows:
The primary linguistic materials for Rongba Queyu were collected firsthand in Litang County between [YEARS].
6. 结论与展望
An Overview of the Phonological System of Rongba Queyu: Monosyllabic words follow a $(C)(C)(C)V$ structure. There are five pairs of vowels distinguished by uvularization, resulting in a specific set of vowel phonemes. These five groups of vowels are organized into distinct uvularized and non-uvularized pairs.
How can we identify the pairings between plain and uvularized vowels? One method is through the auditory similarity of the sounds, such as in the case of $[ź̩̩]$ and $[ź̩̩ʶ]$. Although I have employed different symbols here, they can be considered equivalent in this context.
Regarding verbal person inflection, for the verb $ptʰé$, all other forms remain unchanged. This parallelism in vowel alternation assists us in identifying vowel pairings even when the phonetic quality of the vowels appears dissimilar.
Minimal pairs of plain and uvularized vowels in Rongba Queyu:
6. 结论与展望
Born in Renda Village, Rongba Township, and currently relocated to Litang County. Materials and recordings were collected focusing on minimal pairs. Each word was read in the most natural manner possible. The recording process was completed at the recording studio of the Litang County Converged Media Center. Acoustic parameters were extracted using Praat (Boersma & Weenink).
Minimal pairs used for recording: (C)CV CORONAL sickle.
Acoustic data extraction was performed using Praat for vowel segmentation and annotation. The first three formants ($F_1, F_2, F_3$) were extracted from the stable portion of each vowel. We averaged the data points for each formant. For the first three formants of each vowel phoneme, we obtained statistical data and generated corresponding plots.
Regarding the acoustic correlates of uvularization in Qiangic languages, Evans conducted paired-sample t-tests on the first three formant values for each pair of vowels, finding them to be statistically significant.
44.66 t (14)=8.007 p <0.001
60.44 t (14)=6.240 p <0.001
156.21 t (14)=15.273 p <0.001
(14)=
-
446.20 t (14)= - 14.344 p <0.001
-
904.87 t (14)= - 28.305 p <0.001
(14)=4.340
447.00 t (14)=5.310 p <0.001
980.66 t (14)=12.650 p <0.001
(14)=2.568
142.85 t (14)=7.836 p <0.001
-
479.63 t (14)= - 25.770 p <0.001
-
579.75 t (14)= - 22.577 p <0.001
582.17 t (14)=12.547 p <0.001
677.83 t (14)=8.278 p <0.001
The Acoustic Space of Rongbaque Tibetan
1. Introduction
Rongbaque Tibetan is a unique Tibetan variety spoken in the Rongba region. As a linguistic variety situated within the complex dialectal landscape of the Tibetan plateau, its phonetic and phonological systems provide crucial data for understanding the historical evolution and geographical variation of the Tibetic languages. This study focuses on characterizing the acoustic space of Rongbaque Tibetan, primarily through the analysis of its vowel system and formant structures. By mapping these acoustic properties, we aim to provide a quantitative foundation for further comparative linguistic research.
2. Methodology
To define the acoustic space of Rongbaque Tibetan, we conducted field recordings of native speakers in controlled environments. The primary data consists of monophthongs produced in various consonant contexts.
2.1 Data Collection and Processing
The speech signals were recorded at a sampling rate of 44.1 kHz. We utilized Praat software for the extraction of acoustic parameters, specifically focusing on the first two formants ($F_1$ and $F_2$), which are the primary determinants of vowel quality (height and backness, respectively). To minimize the influence of individual physiological differences, the raw formant values were normalized using the Lobanov method or similar z-score transformations where appropriate.
2.2 Statistical Analysis
The extracted data were analyzed to determine the mean values and standard deviations for each vowel category. The acoustic space is represented by plotting $F_1$ against $F_2$ on a bark scale or Hertz scale (inverted), allowing for a visual representation of the vowel distribution.
[FIGURE:1]
3. Results and Discussion
The acoustic analysis reveals a distinct vowel distribution pattern for Rongbaque Tibetan. The results indicate a system characterized by a specific number of peripheral vowels and potential internal shifts compared to Standard Lhasa Tibetan or neighboring Amdo and Khams dialects.
3.1 Vowel Inventory and Distribution
Based on the formant analysis, the vowel space of Rongbaque Tibetan can be visualized as follows:
[TABLE:1]
The $F_1$ values, which correlate with vowel height (aperture), show a clear distinction between high, mid, and low vowels. For instance, the high vowels /i/ and /u/ exhibit low $F_1$ values, while the low vowel /a/ shows the highest $F_1$ frequency. The $F_2$ values,
6. 结论与展望
Why should this be analyzed as a vowel feature? Phonotactic distribution shows that both sets of vowels can occur with labial and coronal consonants, thereby creating contrast. This differs from consonantal uvularization or velarization, such as that found in Jordanian Arabic \cite{TamimiHeselwood} or the velarized lateral in English \cite{LadefogedMaddieson}, where the feature is tied to specific consonants.
Why should this be analyzed as a vowel feature? Regarding phonotactic distribution, when paired with dorsal consonants, a complementary distribution emerges. Plain vowels only occur with velar consonants, while uvularized vowels only occur with consonants at the uvular place of articulation. In this context, uvular consonants can be viewed as variants of dorsal consonants when paired with uvularized vowels.
Why should this be analyzed as a vowel feature? Uvularization can trigger leftward vowel harmony, transmitting the uvularization of the root to the prefix. Vowels are thus divided into two categories: uvularized vowels, which can trigger harmony, and plain vowels, which undergo it.
The directionality and monovalency of uvularization: Uvularization harmony cannot extend to the right. Furthermore, plain vowels do not trigger "de-uvularization." This suggests that uvularization is a monovalent (unary) feature.
The phonological domain of uvularization: Uvularization can simultaneously affect both the vowel and the consonant of a syllable (e.g., [póʶ]). Since uvularization harmony can cross morpheme boundaries, it raises the question of whether uvularization is a feature of the entire syllable or a suprasegmental feature.
The phonological domain of uvularization.
2. 后依附词
Uvularization on enclitics cannot cross the clitic boundary to affect the host. For example, in the forms lóʶxpe, lóʶxp, and lóʶxp, uvularization may affect multiple segments within a single syllable or phonological word. The scope of this process is determined by various factors and is analyzed as a suprasegmental feature.
6. 结论与展望
From a regional perspective, many dialects of the Queyu language possess uvularized vowels. According to Zheng’s field notes (as cited in Suzuki and Sonam Wangmo), the phonological behavior of velarization in these dialects is remarkably similar to that observed in Pubarong. While the actual phonetic values of the same phoneme vary across dialects—approximating a central vowel in Rongba but sounding closer to a back vowel in Pubarong—the underlying system remains consistent. Beyond Queyu, the phenomena described in this paper bear a strong resemblance to the linguistic patterns found in the Mawo and Yunlinsi dialects of Qiang \cite{Evans}.
From a regional perspective, similar secondary articulatory gestures are also present in other neighboring languages. These include velarization in Puxi Horpa \cite{Evans} and Zhibu Jiarong, as well as pharyngealization in Northern Horpa \cite{Huang}. Other related phenomena include the retracted tongue root in Pengbuxi Muya, and the distinction between tense and lax vowels in Shade Muya \cite{Huang, Huang}. It remains an open question whether these various labels truly represent distinct articulatory mechanisms or are different descriptions of similar phonetic processes.
6. 结论与展望
The vowel system of the Rongbaque dialect features uvularized vowels paired with their plain counterparts. The primary acoustic correlate of these uvularized vowels is a significantly higher $F_1$ value. As a unary (monovalent) feature, uvularization within a sonorant syllable can extend leftward across morpheme boundaries to prefixes; however, it cannot extend to the right. The scope of uvularization is governed by numerous factors. From a regional perspective, its appearance in surrounding areas warrants close attention from field linguists.
To clarify these various phenomena, it is necessary to employ additional experimental tools and methods to investigate the physiological aspects of their articulation. Variations within a speech community—including differences among native speakers, across multiple dialect points, and the ongoing changes observed in the younger generation—merit further study. While acoustic analysis provides an objective basis for understanding the vowel system, the system as a whole must be defined through evidence from phonotactics and relevant phonological processes. Such an approach will facilitate the tracing of the diachronic origins and evolution of uvularized vowels and the vowel system in its entirety.
Thank you for your attention!