Final Report Summary - MULTISIGN (Multilingual Behaviours In Sign Language Users)
The MULTISIGN project aimed to establish Sign Multilingualism Studies as a new sub-field of research in linguistics. Research proceeded in three strands: “cross-signing” examines communication between deaf signers who have no shared language in common; “sign-switching” is about the mixing of languages by bilingual signers who are familiar with the same set of sign languages; and in “sign-speaking”, sign language and speech are produced simultaneously but their structures do not match. The international project team conducted research with people co-using Turkish/Turkish Sign Language, Turkish Sign Language/German Sign Language, Hindi/English/Indian Sign Language, Burundi Sign Language/Indian Sign Language, and Indian Sign Language/American Sign Language.
For the cross-signing study, two groups of deaf signers from the UK, Indonesia, Japan, and Jordan, and from India, Nepal, Indonesia, and Jordan, were videotaped repeatedly over a period of 4-6 weeks. This afforded a unique first-time opportunity to track the development of a contact variety (“jargon”) from the very beginning. The pairs of signers co-developed multimodal and multilingual “shared spaces”, using a structured process of introducing a sign, then accommodating its use, and then persisting with using it (IAP-model). In the elicitation task, in which target pictures had to be matched to descriptions signed by the interlocutor, error rates were low (ca. 5%) and did not change over time, but signers became up to 50% faster with the task after spending four weeks together. The signers were also found to adapt their lexicon to each other, exemplified by numerals, colour terms, and signs for animate beings.
The sign-switching participants were fluent in two sign languages and code-switched between them. When groups of signers were together for longer periods of time (e.g. Burundi-Indian signers) they formed a small bilingual “Community of Practice”. The study investigated the shared norms of their bilingual variety, including the proportion of signs from each language, and the use of question words and negation. By contrast, the bilinguals in the German and Turkish sign languages who came together only for a week, did not develop shared linguistic features. The research also revealed that a substantial 40% of “shared” signs could not be clearly assigned to either of the contributing languages. Hence determining points of switching from one language to the other, which is a common issue in spoken language bilingualism, is not possible in sign-switching.
Sign-speaking happens when hearing signers address a mixed group of hearing and deaf people and, as there are no sign language interpreters, the sociolinguistic norms of the setting compel them to produce the same message simultaneously in both modalities. We conducted ethnographic fieldwork in India on situations with participants speaking Hindi (with some interspersed English words) and using Indian Sign Language (ISL). When identifying the relationship between these languages in text samples, semantic and/or syntactic mismatches were found in a very high proportion (48%) of utterances. Sometimes sign-speakers even express two different propositions at the same time, and existing theoretical models of language production cannot account for such data.
A synthesis of findings across the whole project resulted in a model of factors that influence bilingual outputs. These include the typological profile of the languages involved, sociolinguistic norms, individual linguistic backgrounds, the surrounding linguistic environment, and the use of metalinguistic skills. These forces have provided a unifying framework as they apply across different bi-/multilingual situations studied in the project.
The project’s approach to ethics focused on the issue of agency in research, i.e. ensuring that sign language research is carried out with and by deaf people. Therefore, the research team included deaf members from the UK, the US, India, South Korea, and Turkey. Moreover, we aimed to highlight the impressive multilingual and meta-linguistic capabilities of signers, and this positive view of sign languages in terms of their affordances due to the language modality and the linguistic/meta-linguistic skills displayed by their users has been a recurrent theme.
For the cross-signing study, two groups of deaf signers from the UK, Indonesia, Japan, and Jordan, and from India, Nepal, Indonesia, and Jordan, were videotaped repeatedly over a period of 4-6 weeks. This afforded a unique first-time opportunity to track the development of a contact variety (“jargon”) from the very beginning. The pairs of signers co-developed multimodal and multilingual “shared spaces”, using a structured process of introducing a sign, then accommodating its use, and then persisting with using it (IAP-model). In the elicitation task, in which target pictures had to be matched to descriptions signed by the interlocutor, error rates were low (ca. 5%) and did not change over time, but signers became up to 50% faster with the task after spending four weeks together. The signers were also found to adapt their lexicon to each other, exemplified by numerals, colour terms, and signs for animate beings.
The sign-switching participants were fluent in two sign languages and code-switched between them. When groups of signers were together for longer periods of time (e.g. Burundi-Indian signers) they formed a small bilingual “Community of Practice”. The study investigated the shared norms of their bilingual variety, including the proportion of signs from each language, and the use of question words and negation. By contrast, the bilinguals in the German and Turkish sign languages who came together only for a week, did not develop shared linguistic features. The research also revealed that a substantial 40% of “shared” signs could not be clearly assigned to either of the contributing languages. Hence determining points of switching from one language to the other, which is a common issue in spoken language bilingualism, is not possible in sign-switching.
Sign-speaking happens when hearing signers address a mixed group of hearing and deaf people and, as there are no sign language interpreters, the sociolinguistic norms of the setting compel them to produce the same message simultaneously in both modalities. We conducted ethnographic fieldwork in India on situations with participants speaking Hindi (with some interspersed English words) and using Indian Sign Language (ISL). When identifying the relationship between these languages in text samples, semantic and/or syntactic mismatches were found in a very high proportion (48%) of utterances. Sometimes sign-speakers even express two different propositions at the same time, and existing theoretical models of language production cannot account for such data.
A synthesis of findings across the whole project resulted in a model of factors that influence bilingual outputs. These include the typological profile of the languages involved, sociolinguistic norms, individual linguistic backgrounds, the surrounding linguistic environment, and the use of metalinguistic skills. These forces have provided a unifying framework as they apply across different bi-/multilingual situations studied in the project.
The project’s approach to ethics focused on the issue of agency in research, i.e. ensuring that sign language research is carried out with and by deaf people. Therefore, the research team included deaf members from the UK, the US, India, South Korea, and Turkey. Moreover, we aimed to highlight the impressive multilingual and meta-linguistic capabilities of signers, and this positive view of sign languages in terms of their affordances due to the language modality and the linguistic/meta-linguistic skills displayed by their users has been a recurrent theme.