Home ➤ Projects ➤

Sinhala Diachronic Corpus

This project aims to develop a comprehensive diachronic corpus for the Sinhala language, addressing the scarcity of historical linguistic resources for a language spoken by approximately 16 million people in Sri Lanka. Despite its rich literary heritage dating back to the 3rd–2nd centuries BCE, Sinhala remains underrepresented in computational and corpus-based linguistic research. The project seeks to compile a diverse collection of texts spanning multiple historical periods and genres, including books, newspapers, scholarly writings, and other textual materials. Collaboration with Sri Lankan institutions and language scholars enables texts to be annotated according to their original writing dates, allowing for more accurate representation of linguistic evolution over time.
The corpus is constructed through a systematic process that includes the digitisation of printed texts, optical character recognition, and extensive post-processing to correct formatting issues, modernise orthography, and address challenges such as code-mixing and malformed tokens. The resulting collection is carefully filtered to ensure authorship attribution, copyright compliance, and data quality. Texts are organised through layered genre categorisation, distinguishing between fiction and non-fiction while further grouping works into areas such as religious writing, history, poetry, language studies, and medical texts. By creating an openly accessible historical text resource, this project supports research into lexical change, historical syntax, and corpus-based lexicography while contributing to the preservation and study of Sinhala’s linguistic and literary heritage.

Objectives:

• Compile a large-scale historical Sinhala text collection spanning multiple time periods and genres by sourcing materials from books, newspapers, scholarly writings, and other archival and digital sources.
• Digitise and preprocess historical Sinhala texts using optical character recognition and systematic post-processing techniques to correct formatting issues, normalise orthography, and ensure high-quality machine-readable data.
• Annotate texts with reliable temporal and genre metadata, including original writing dates and layered genre classifications, to enable accurate diachronic linguistic analysis.
• Support linguistic and computational research on Sinhala language evolution by providing an open-access resource that facilitates studies on lexical change, historical syntax, corpus-based lexicography, and related historical language research.

Keywords: Sinhala | Natural Language Processing | Sinhala Diachronic Corpus | Corpus | Language Change |

Publications

Team