Interference from Flemish-Belgian Sign Language into the Dutch of Deaf People: Reference tracking

In the written Dutch of many deaf children (and adults) there seem to be traces of Flemish Belgian Sign Language i.e. it would seem that some of the mistakes in their written productions can be attributed to interference from Flemish Belgian Sign Language. Until now this is something that has only been acknowledged intuitively, especially by educators in deaf education, but has not been verified empirically. The main reason is that it is a real challenge to compare a spatial grammar, so typical of signed languages, to a verbal grammar, typical of spoken languages. Nevertheless we believe that we have found an angle from which this endeavour can be undertaken. This study is a first attempt at verifying the interference on an empirical basis. As part of a previously conducted large scale study dealing with the acquisition of Dutch by deaf pupils 119 subjects were asked to watch an animated cartoon and to recount the story in written Dutch. They were all pupils of a school for deaf or partially deaf children, which is a necessary prerequisite for sign language to be acquired. For the current paper we showed the same cartoon to eight signers and asked them to retell the story in Flemish Belgian Sign Language. The subjects were three children, one teenager, and four adults who told their story to a familiar person who had not seen the cartoon. Their narratives were videotaped and afterwards transcribed and analysed with the help of a near-native Flemish Belgian Sign Language user. Although we intend to study a number of different aspects this paper focuses on reference tracking i.e. the ways in which a narrator identifies a certain character/referent as the one being discussed. Since the cartoon contained four main characters who were all male it proved to be an excellent elicitor for this purpose. We especially concentrated on reference by means of full noun phrases, anaphoric reference, role taking and shifted reference and null arguments. The results will be presented in the paper.status: publishe