4TB: A new tool to study Tocharian A – Old Uyghur parallels
The paper introduces 4TB, a web-based corpus developed to support research on Tocharian A and Old Uyghur. Its aim is to collect and align Tocharian A fragments with their Old Uyghur (and later Sanskrit) parallels, particularly the Tocharian A Maitreyasamiti-Nāṭaka and its Old Uyghur translation Maitrisimit nom bitig. Despite extensive scholarship, no complete editions of these texts are currently published. A Tocharian A edition is in preparation and will serve as the basis for the corpus, while 4TB focuses on assembling Old Uyghur parallel fragments as a foundation for a future full edition of the Maitrisimit nom bitig. Compared to existing digital resources, Tocharian is relatively well represented, whereas Old Turkic corpora remain limited, making even a partial parallel corpus a meaningful contribution. The paper discusses key problems of corpus design and development. Transliteration is omitted for both languages, while transcription is handled differently. For Tocharian A, it follows established conventions; for Old Uyghur, all data are normalized into a unified transcription system, with limited correction of outdated forms. Additional challenges arise from inconsistent transcription practices in the literature. Translations are provided as in the source publications (Russian for Tocharian A, mainly German for Old Uyghur), with plans to add automatic English translations. The corpus is structured around tokens as the smallest unit, but alignment operates on higher-level units due to the lack of reliable word-to-word correspondence. Because sentence segmentation is problematic, a flexible unit called a “passage” is introduced. Passages are grouped into “passage groups” to account for manuscript variation and are then aligned across languages, allowing for discrepancies such as omissions and additions. This approach preserves textual coherence and differs from standard KWIC-based corpus models. The corpus offers several core features. Dictionaries cover all lemmatized tokens and link forms and spellings with grammatical annotation. A concordance displays passages together with their parallels. Search tools support queries by spelling, form, lemma, and grammatical features, including cross-linguistic searches. Editing tools allow modification of texts and lexical data, with semi-automatic lemmatization and potential for further expansion. Overall, 4TB provides a scalable framework that can be used in future corpus projects. It facilitates cross-linguistic research on Tocharian and Turkic languages and improves access to the material for non-linguistic specialists such as Buddhologists.
Сведения
- Журнал
- Turkic Studies Journal
- Опубликовано
- 2026 · С. 77-92
- Язык
- Английский
- ISSN
- 2664-5157 · 2708-7360
- Издатель
- L. N. Gumilyov Eurasian National University