Skip to main content
preprint arXiv preprint 2026

BLAD: A Historically Contextualized, Multilingual Dataset of Bangladeshi Legal Acts (1799 to 2025)

Adib Sakhawat

1,484 acts spanning 226 years, decomposed into 35,633 sections and 14,523 footnotes.

BibTeX

@misc{sakhawat2026blad,
  title         = {BLAD: A Historically Contextualized, Multilingual Dataset of Bangladeshi Legal Acts (1799 to 2025)},
  author        = {Sakhawat, Adib},
  year          = {2026},
  eprint        = {2607.17111},
  archivePrefix = {arXiv}
}

Description

The question

Legal NLP has rich resources for common-law, English-language jurisdictions, and almost none for low-resource, civil-law jurisdictions in South Asia. Bangladesh has more than two centuries of statute law, written in English, Bengali and both, with no structured corpus.

What BLAD is

The Bangladesh Legal Acts Dataset: a curated collection of every legislative act enacted between 1799 and 2025, scraped from the official government portal and enriched with historical context.
Acts 1,484
Time span 1799–2025 (226 years)
Sections 35,633
Footnotes 14,523
Languages English, Bengali and mixed-language documents

Each act comes with:

  • its full text, split into structured sections and footnotes;
  • its repeal status;
  • metadata linking it to the governing regime, the head of state and the prevailing legal framework at the time it was enacted.

Design choices

  • Original language preserved. Bilingual and mixed documents are kept as written rather than translated, because in legal text the wording is the data.
  • Built for time, not just search. Linking every act to the legal system in force makes the corpus usable for temporal legal reasoning and historical analysis, not just retrieval.

The paper describes the acquisition and enrichment pipeline, reports descriptive statistics over two centuries of legislation, and outlines the research directions the corpus opens.

Resources