Abstract:

Despite the rich cultural heritage of the Kazakh language, there exists a significant gap in linguistic resources for it, particularly in the domain of spoken corpora. This presentation introduces the spoken corpus of the Kazakh language. Although a number of corpora projects exist declaring spoken corpora as their components, MultiCorSKL is the first-ever corpus focusing on genuine naturally occurring spoken discourse, addressing a critical need for such data and analysis in computational linguistics and language diversity preservation.

The project created a comprehensive spoken corpus of modern Kazakh language using a team of trained linguists as well as innovative crowd-sourcing techniques to work with the organic interactional language data. Apart from aiming to collect representative spoken language data, we sought to implement reliable methods of annotation of conversational data, facilitating its use in training automated speech recognition (ASR) models, which has not been done with great success up to this point.

Through the crowd-sourcing approach, we engaged speakers of the Kazakh language across diverse regions and social strata to contribute to the corpus. The collected data underwent annotation, at various levels, from simple orthography to deep annotation using Intonation Unit segmentation, orthographic transcripts, Discourse Functional Transcription, IPA notation, morphemic glossing, integrated within ELAN, enabling detailed and accurate linguistic analysis.

In the most recent audit, the corpus contains 180 hours of recorded data, with 80 hours of data annotated, and the data volume is expected to grow significantly through crowd-sourced recording and automation of annotation. Conversational ASR, an important workflow element, used the original project data, segmented into intonation units, resulting in approximately 30000 files, to train and test the models.

The development of the MultiCorSKL represents a landmark achievement in the studies of low-resource languages. It not only enables advanced research in computational and corpus linguistics but also holds profound implications for cultural preservation and technological advancements in Kazakhstan. Various disciplinary domains such as socio- and cognitive linguistics, conversational analysis, comparative morphosyntax, among others will benefit from the access to spoken corpus data, which provides reliable empirical insights into the spontaneous, day-to-day language usage.

Future work will focus on expanding the corpus and refining ASR models to enhance linguistic accessibility and digital inclusivity. We invite researchers and technologists to explore the corpus and contribute to its expansion. Collaboration is essential to harness the full potential of this unique resource for linguistic research in various areas and technological innovation.

Abstract:

People often perceive the speech of individuals from the Western regions of the country negatively. Reports from other regions' speakers claim that understanding Western Kazakh (WK) is difficult because the spoken language is faster than other varieties of Kazakh (Bizhanova, 2022). Possibly for such a reason, WK is perceived to be rude.

A pilot investigation revealed that, while WK speakers' speech rate is not the fastest, WK speakers employ phonological reduction more often. Specifically, Westerners phonologically reduce verbs 50% more than North-Eastern Kazakh speakers. This finding suggests phonological reduction may explain why WK speakers are perceived to be fast speakers.

It now remains to understand how the perception of this phonological feature is extended by listeners to a moral assessment about WK speakers. Iconization is a process through which language users infuse structural linguistic elements with symbolic significance, often reflecting cultural values or moral judgments (Irvine&Gal, 2000). In this study, I will show the steps through which phonological reduction is iconized into rudeness evaluations in the case of WK.

To address this question, I conducted a matched-guise experiment. Participants from different regions of Kazakhstan listened to an audio sample, featuring a WK speaker, without knowing the dialect and had to guess it solely based on the audio. The findings indicate that the half of participants successfully identified the dialect as WK, while the rest identified it as Southern. This suggests that the dialect is not readily distinguishable to native speakers of Kazakh unless they are informed that the speaker is from the Western region, and consequently, speaks WK. The outcome of the performance prompts an inquiry into the nature of the comments evoked thereby.

Subsequently, I analyze the responses of these participants to identify whether they perceive WK negatively and if so, the reasons behind such perceptions. People indicate that the negative perceptions may result from cultural beliefs associating the Kishi Juz, inhabitants of the Western region, with traits like cruelty and aggression due to being historically known for their warrior heritage. Moreover, every participant agrees that the frequent reduction of verbs in the speech affects their comprehension more, which makes them feel irritated. These results indicate that the negative attitudes towards the WK are primarily ideological constructs, which may not be negatively perceived in the absence of relevant information.

Concluding, results suggest that the frequent phonological reduction of WK speakers may affect the negative attitudes about WK being shaped.

Abstract:

Due to political, social, and economic changes, there has been close contact between Chinese and minority languages spoken in China today. This study explores how Chinese content words are accommodated morphosyntactically in Kazakh in the context of Kazakh-Chinese language contact in Xinjiang. Although borrowing strategies have been studied for many languages of the world, there is a gap in the literature regarding spoken Kazakh in Xinjiang and Chinese lexical borrowing in Kazakh. The data used in this study represents the recordings I collected in Xinjiang as a research assistant in the Multimedia Corpus of Modern Spoken Kazakh Language project. After analyzing 6.8 hours of conversational data, a total of 133 cases of insertion of verbs, 21 adjectives, and 7 adverbs were found. Moreover, 250 nouns were found in 3.5 hours of data, which could be extrapolated to over 450 per 7 hours of conversational data. Firstly, the strategies for the accommodation of Chinese verbs were identified according to the typology of verbal borrowings of Jan Wohlgemuth (2009). The results show four different strategies used to accommodate Chinese verbs: light verb strategy, direct insertion, indirect insertion, and paradigm insertion. The light verb strategy is the most attested in Kazakh, which agrees with Wohlgemuth’s (2009) findings that “language with the dependent-head (OV) orientation strongly prefers the Light Verb Strategy” (p. 203). The light verbs used in Kazakh are qyl- ‘to do’ and bol- ‘to be’, where bol- ‘to be’ is used when the subject of the verb does not have an agentive role and is affected by the action, and qyl- ‘to do’ is used when the subject is agentive, and the action is voluntary. In direct insertion, the Chinese verbs are directly inserted into the Kazakh sentence without any morphological accommodation, but in indirect insertion, Kazakh verbalizing affixes are added to monosyllabic Chinese verbs, as in pī-le ‘to approve’ and zū-la ‘to rent.’ The use of paradigm insertion allows the speakers to use the Chinese perfective aspect marker le together with the Chinese verb. Secondly, since adjectives in Mandarin are regarded as stative verbs (Li & Thompson, 1989), adjectives are accommodated in two different ways – light verb strategy and direct insertion into the position of the Kazakh adjectives. Finally, nouns and adverbs are integrated into Kazakh just like native words without any accommodation strategies. Overall, this study provides an overview of the accommodation of Chinese content words in Kazakh language.