A corpus for signed language interpreting research

Abstract Because of the visual nature of signed language, the compilation of a signed language interpreting corpus along the lines of spoken-language interpreting corpora has been viewed as extremely challenging, if not impossible. This study offers a unique contribution in the construction of a lemmatized, annotated text-based corpus of signed language media interpretations, which allows analysis of interesting features using readily-available concordance software. In this article, characteristics of original (not interpreted) signed language corpora are explored in terms of metadata conventions, transcription and annotation, in order to provide a framework for an interpreting corpus. Within this framework, the decisions and steps taken in the construction of the interpreting corpus are discussed and explained.