Bashkir Lexicon (Machine Fund) is a machine-readable lexical database of the Bashkir language, compiled from the open digital resource Machine Fund of the Bashkir Language (mfbl2.ru). It contains 34,008 unique lexical entries covering dialectal word forms with their part of speech, dialect, subdialect, literary norm, and Russian translation.

The dataset preserves dialectal diversity (all three major Bashkir dialect groups: Southern, Eastern, North-Western), synonyms (multiple Russian translations for the same Bashkir word), and literary norms, while removing full duplicates within a single dialect.

It is designed to support NLP research and applications for Bashkir, a low-resource Turkic language.

Curated by: Arabov Mullosharaf Kurbonovich, Khaybullina Svetlana Sergeevna (BashkirNLPWorld)
Language(s) (NLP): Bashkir (ba), Russian (ru), Cyrillic script
License: other – see Licensing below.

What BashkirNLPWorld contributed

The BashkirNLPWorld team built this dataset end-to-end on top of the open source:

Parsing of the source HTML pages with BeautifulSoup.
Field extraction into six structured columns.
Normalization of field names to English (bashkir_word, part_of_speech, dialect, subdialect, literary_norm, russian_word).
Deduplication by the key (bashkir_word, russian_word, dialect) — preserving synonyms and dialectal variants, removing only full duplicates within a single dialect.
Machine-readable packaging as UTF-8 JSONL, ready for datasets.load_dataset.
Documentation of fields, dialects, parts of speech, and dataset limitations.

Dataset Sources

Repository: https://huggingface.co/datasets/BashkirNLPWorld/bashkir-lexicon
Original source: http://mfbl2.ru/mfbl/bashdial — Machine Fund of the Bashkir Language
Related datasets:

Licensing

Original lexical content (dictionary entries and their annotations) belongs to the Machine Fund of the Bashkir Language (mfbl2.ru) and its contributors. BashkirNLPWorld does not claim ownership of this content.
The structured compilation, metadata, English field names, deduplication, and documentation created by BashkirNLPWorld are released under CC BY-SA 4.0.
Users are responsible for complying with applicable law when reusing the original content.
Rights holders can request removal of specific entries by contacting the maintainers — we respond within 14 business days and remove disputed content in the next release.

Uses

Direct Use

Linguistic research on Bashkir dialectology, lexicography, and morphology
Building lexical databases, thesauri, and dictionaries for Bashkir
Training and fine-tuning language models (LLM, BERT) for Bashkir
Machine translation (Bashkir ↔ Russian) using dialect-to-literary-norm and dialect-to-Russian pairs
Instruction tuning of generative LLMs for Bashkir
Text classification (dialect, subdialect, part of speech)

Out-of-Scope Use

Re-licensing the original lexical content under a different license
Treating the dataset as a gold-standard linguistic reference without verification against the original source
Any use that violates the rights of the original content creators

Dataset Structure

Data Fields

Field	Type	Description	Missing
`bashkir_word`	string	Bashkir dialectal word form	0%
`part_of_speech`	string	Part of speech (abbreviated, as in the source)	<0.1%
`dialect`	string	Dialect (`көньяҡ`, `көнсығыш`, `төньяҡ-көнбайыш`)	<0.1%
`subdialect`	string	Subdialect (говор)	0%
`literary_norm`	string	Literary norm of the word	0.7%
`russian_word`	string	Russian translation	0.3%

Dialect Distribution

Dialect	Records	%
көньяҡ (Southern)	15,704	46.2%
көнсығыш (Eastern)	12,480	36.7%
төньяҡ-көнбайыш (North-Western)	5,823	17.1%

Part-of-Speech Distribution

Abbreviation	Records	%	Meaning
`и.`	20,725	60.9%	noun
`ҡ.`	6,935	20.4%	verb
`с.`	3,729	11.0%	adjective
`р.`	1,748	5.1%	adverb
`а.`	212	0.6%	pronoun
`ы.`	187	0.5%	interjection
`б.`	158	0.5%	conjunction
`к.`	103	0.3%	particle
other	211	0.6%	mixed / rare

Data Splits

The dataset contains a single split (train) with all 34,008 examples. Users are encouraged to create their own train/validation/test splits.

Example Record

{
  "bashkir_word": "А-А",
  "part_of_speech": "ы.",
  "dialect": "төньяҡ-көнбайыш",
  "subdialect": "гәйнә",
  "literary_norm": "аптырау ымлығы",
  "russian_word": "междометие недоумения"
}

Dataset Creation

Curation Rationale

Bashkir is a low-resource Turkic language with limited digital lexical resources. The goal was to turn the open dialectal lexicon published on mfbl2.ru into a clean, machine-readable, deduplicated dataset suitable for linguistic research and NLP.

Data Collection and Processing

Downloaded the HTML pages of the source (mfbl2.ru).
Parsed each file with BeautifulSoup to extract lexical entries.
Extracted: Bashkir word, part of speech, dialect, subdialect, literary norm, Russian translation.
Renamed columns to English field names.
Deduplicated by (bashkir_word, russian_word, dialect) — preserving synonyms and dialectal variants, removing full duplicates within a single dialect.
Saved as UTF-8 JSONL.

Tools: Python 3, BeautifulSoup4.

Source Data Producers

The original lexical content was produced by the authors, compilers, and contributors of the Machine Fund of the Bashkir Language (mfbl2.ru). BashkirNLPWorld is responsible for parsing, normalization, deduplication, metadata, and packaging.

Annotations

No manual annotations were added. All labels (dialect, subdialect, part of speech, literary norm) come from the original source.

Personal and Sensitive Information

The dataset contains publicly available lexical data and does not intentionally include personally identifiable information.

Bias, Risks, and Limitations

Dialect imbalance: Southern dominates (46.2%), North-Western is less represented (17.1%).
Subdialect imbalance: Top-3 subdialects (урта, ҡыҙыл, дим) account for 36.6% of records.
Part-of-speech imbalance: Nouns dominate (60.9%).
Missing values: 0.7% lack a literary norm; 0.3% lack a Russian translation.
No transcription or stress marks.
Possible overlap with Tatar: Due to close language contact, some entries may coincide with Tatar forms.
Source snapshot: Reflects the state of mfbl2.ru at parsing time.

Recommendations

Account for dialect and subdialect imbalance when training models (class weights, oversampling).
For balanced tasks, filter by dialect or subdialect.
For commercial use, verify the copyright status of specific entries.
Report issues via the Hugging Face repository.

Citation

BibTeX:

@dataset{arabov2026bashkirlexicon,
  author       = {Arabov, Mullosharaf Kurbonovich and Khaybullina, Svetlana Sergeevna},
  title        = {Bashkir Lexicon (Machine Fund)},
  year         = {2026},
  publisher    = {Hugging Face},
  url          = {https://huggingface.co/datasets/BashkirNLPWorld/bashkir-lexicon}
}

APA: Arabov, M. K., & Khaybullina, S. S. (2026). Bashkir Lexicon (Machine Fund) [Data set]. Hugging Face. https://huggingface.co/datasets/BashkirNLPWorld/bashkir-lexicon

Glossary

Dialect – a regional variety of a language (in Bashkir: Southern, Eastern, North-Western).
Subdialect (говор) – a smaller territorial variety within a dialect.
Literary norm – the standard, codified form of a word in the literary language.
Low-resource language – a language with limited digital resources and tools.
Machine Fund – a digital lexicographic resource for the Bashkir language.

Dataset Card Authors

Arabov Mullosharaf Kurbonovich
Khaybullina Svetlana Sergeevna
BashkirNLPWorld

Dataset Card Contact

For questions, takedown requests, or collaboration, please open an issue: https://huggingface.co/datasets/BashkirNLPWorld/bashkir-lexicon/discussions 📧 Direct email: cool.araby@gmail.com

Downloads last month: 18

Total file size:

7.11 MB