Projects

Why is more work on the BSL Corpus needed?

A modern linguistic corpus (e.g., the British National Corpus of English) is large, representative, accessible, and importantly, machine-readable (i.e., can be searched by a computer). The dataset that we refer to as the “BSL Corpus” is not yet a true corpus because it is not yet machine-readable. Annotations and translations are needed to make it machine-readable.

Therefore current and future projects that use the BSL Corpus video data include annotation and translation as key elements in the project methodology so that the BSL Corpus can become a true sign language corpus in the future. A more searchable, accessible BSL Corpus will enable more research on the structure and use of BSL. This will in turn bring about improvements to the training of BSL teachers, sign language interpreters and educators of deaf children. Furthermore, annotations and translation enable work comparing BSL with related and unrelated sign and spoken languages elsewhere in the world, and will lead to an improved understanding of human language in general.

The major studies from the original BSL Corpus Project (2008-2011) were on phonological variation (1-handshape study), lexical variation (signs for colours, countries, numbers and UK place names), and lexical frequency (based on 25,000 sign tokens from conversation). See the Publications page for papers on each of these.

BSL SignBank

Under the BSL Corpus Project, a lexical database was created as part of a study on lexical frequency. This lexical database, which documents 50,000 signs from four regions from the BSL Corpus data (Bristol, Birmingham, London and Manchester) was transformed into an online dictionary, BSL SignBank, as part of work on language documentation and language change by the Deafness, Cognition and Language Research Centre between 2011 and 2015. BSL SignBank has been available online since September 2014 but continues to grow and evolve as we research more about BSL and annotate more of the BSL Corpus. More…

Researchers involved: Kearsy Cormier, Jordan Fenlon, Sannah Gulamani, Sandra SmithKatherine Rowley, Robert Adam, Alan Wendt, Bencie Woll

Directional Verbs Project (2012-2014)

The aim of this project is to use the sign language data collected under the BSL Corpus Project, and to conduct an investigation into variation and change in the use of directional verbs in BSL. More…

Digging into Signs Project (2014-2015)

The aim of this project is to develop cross-corpus annotation standards for sign language data, using the BSL Corpus in the UK and Corpus NGT  in the Netherlands, and to improve current software tools in working with sign language corpora. More…

Researchers involved: Kearsy CormierJordan FenlonSannah Gulamani, Sandra Smith

BSL Syntax Project (2016-2021)

The aim of this project is to document and describe word order and non-manual features in different types of BSL sentences. More…

Language Attitudes Project (2017-2019)

The aim of this project is to study language attitudes and language awareness in the British Deaf community. More…

ExTOL Project (2018-2022)

This project (ExTOL: End to End Translation of British Sign Language) aims to take annotated sign language data from the BSL Corpus and other sources and to use this to build a system that is capable of watching a human signing and turning this into written English. In the process, this project will also create computer vision tools to assist with sign language analysis. More…

Enactment Project (2019-2021)

The aim of the project is to investigate how deaf signers of BSL mimetically reproduce the actions, utterances, thoughts and feelings of themselves, other people, animals and things using existing conversation and personal narrative data in the BSL Corpus. More…

EASIER Project (2021-2023)

This project (EASIER: Intelligent Automatic Sign Language Translation) was a Horizon 2020 project aimed at developing a multilingual machine translation system to enable seamless communication between deaf and hearing individuals. The project focused on creating a platform that also supported sign language content creation. By integrating sign language linguistics expertise with advanced technological resources, EASIER developed a robust sign language recognition engine. It also incorporated a signing avatar capable of reflecting sign language grammar and prosody features for more natural synthetic signing. The system used state-of-the-art machine translation technology, processing both annotated and unannotated data, to handle various communication scenarios effectively.