Active Syllable Average Limit 1,000 (音涯一千)

This paper mainly studies phonemic cognitive ability through the databases of living spoken languages in the Sino-Tibetan languages including 20 Chinese dialects, 6 Tibetan dialects, 5 Miao dialects, Mian, Zhuang, Thai, Li, Dai, Yi, Burmese, Zaiwa, and, Achang. The methods of statistics and information entropy and the concepts of the actual syllabic space, the syllabic theoretical space and redundancy rate are used and proposed in this paper. The results show that: (1) statistical methods can be used in the study of phonemic cognition; (2) the actual syllabic space in spoken Sino-Tibetan languages reflects the man’s phonemic cognitive ability; (3) the theoretical syllabic space composed of initial, final, and tone in the Sino-Tibetan languages reflects the dynamic process of a phoneme system in language contact and evolution; (4) a redundancy rate of 60% is the bottom limit in oral communication in the Sino-Tibetan languages. Therefore, the conclusion of this study is that Active Syllable Average Limit 1,000 not only reflects man’s phonemic cognitive ability, but also reflects the interdependence of phonemic cognition and semantic cognition, and reveals an important link in the process of a language chain from semantic to phonemic transformation, which has important theoretical significance in the study of language cognition.

End-to-End Speech Synthesis for Tibetan Multidialect

The research on Tibetan speech synthesis technology has been mainly focusing on single dialect, and thus there is a lack of research on Tibetan multidialect speech synthesis technology. This paper presents an end-to-end Tibetan multidialect speech synthesis model to realize a speech synthesis system which can be used to synthesize different Tibetan dialects. Firstly, Wylie transliteration scheme is used to convert the Tibetan text into the corresponding Latin letters, which effectively reduces the size of training corpus and the workload of front-end text processing. Secondly, a shared feature prediction network with a cyclic sequence-to-sequence structure is built, which maps the Latin transliteration vector of Tibetan character to Mel spectrograms and learns the relevant features of multidialect speech data. Thirdly, two dialect-specific WaveNet vocoders are combined with the feature prediction network, which synthesizes the Mel spectrum of Lhasa-Ü-Tsang and Amdo pastoral dialect into time-domain waveform, respectively. The model avoids using a large number of Tibetan dialect expertise for processing some time-consuming tasks, such as phonetic analysis and phonological annotation. Additionally, it can directly synthesize Lhasa-Ü-Tsang and Amdo pastoral speech on the existing text annotation. The experimental results show that the synthesized speech of Lhasa-Ü-Tsang and Amdo pastoral dialect based on our proposed method has better clarity and naturalness than the Tibetan monolingual model.

Multitask Learning with Local Attention for Tibetan Speech Recognition

In this paper, we propose to incorporate the local attention in WaveNet-CTC to improve the performance of Tibetan speech recognition in multitask learning. With an increase in task number, such as simultaneous Tibetan speech content recognition, dialect identification, and speaker recognition, the accuracy rate of a single WaveNet-CTC decreases on speech recognition. Inspired by the attention mechanism, we introduce the local attention to automatically tune the weights of feature frames in a window and pay different attention on context information for multitask learning. The experimental results show that our method improves the accuracies of speech recognition for all Tibetan dialects in three-task learning, compared with the baseline model. Furthermore, our method significantly improves the accuracy for low-resource dialect by 5.11% against the specific-dialect model.

A phonological history of Amdo Tibetan rhymes

In this study, a reconstruction is offered for the phonetic evolution of rhymes from Old Tibetan to modern-day Amdo Tibetan dialects. The relevant sound changes are proposed, along with their relative chronological precedence and the dating of some specific changes. Most interestingly, although Amdo Tibetan, identically to its ancestor Old Tibetan, does not have phonemic length, this study shows that Amdo Tibetan derives from an intermediate stage which, like many other Tibetan dialects, does make the distinction.

Electropalatographic Parameter Analysis of Single Consonant of Xiahe Tibetan Dialects

Targeting at Xiahe Tibetan dialects and with 16 single consonants followed by vowel /a/ as the recording texts, the article analyzes the electropalatographic parameters such as AC, PC, VC, Ant, Pos, CA, CP, CC. The findings are as follows: 1. Plosives show more tongue-palate contact in front part of palate, with strong anteriority and centered tendency: 2. Fricatives show less contact than affricates of the same speech organ; 3. Lateral alveolar consonants have more tongue-palate contact, with strong anteriority and centrality, and weak posterity. 4. Nasals have bigger values in terms of all parameters. 5. The trill has more tongue-palate contact, with stronger posterity and centrality.

Volition and control in Wǎdū Pǔmǐ

In this paper I will argue that verb inflection in the Wǎdū variety of Pǔmǐ is not based on actor-agreement or person-number agreement as has been attested for several other Pǔmǐ varieties, but is based rather on pragmatic notions of volition and control that tie in with evidentiality and egophoricity, similar to that reported for Tibetan dialects.