Opbygning af et tegnsprogskorpus - problemer og udfordringer: Korpus og ordbog over dansk tegnsprog

Corpora of spoken and written languages often consist of a large amount of digitised text, e.g. books, magazines, or newspapers. Sign languages have no standard written representation, at least none that is easily written and read by laymen, and hence – like it is the case for spoken language corpora – sign language corpora consist of (video) recordings supplied with transcriptions. But whereas the transcription of spoken language to a considerable extent can be performed through rendering uttered words through words of a written language, transcription of signed languages is further complicated by the lack of a standard written language. Furthermore, specialised video tools are needed for performing the transcription and utilising the corpus. Mainly for these reasons, building a sign language corpus of an adequate size and accuracy is a cumbersome process, and larger corpora have emerged only in the last decade, e.g. for the sign languages of Australia (Johnston, 2008), the Netherlands (Corpus NGT), Sweden (Swedish Sign Language Corpus Project), Great Britain (The British Sign Language Corpus), Poland (PJM (PSL) corpus), and Germany (DGS-Corpus).