Multilingual behaviours in sign language users (Multisign)

This project examines a range of complex multilingual behaviours in sign language users. The individual studies in this project are called “cross-signing”, “sign-speaking”, and “sign-switching”. Through these three interrelated studies, the project is intended to break new ground with respect to a field of research that can be called “Sign Multilingualism Studies”. This new area of research is only just starting to develop, and this is the first time researchers have ever investigated a complex range of multilingual behaviours in sign language users.

In terms of theoretical grounding, this project seeks to extend known bi- and multilingual phenomena to the domain of sign languages. A better understanding of sign multilingualism can lead to better training for sign language interpreters working in international contexts, as well as improved services for deaf people who move abroad. This project can also be seen as the first step in opening up the new field of Sign Multilingualism Studies for future more extensive and detailed research.

Sign multilingualism studies includes a whole range of inter-related topics, some of which cover known phenomena that also exist in spoken languages, such as code-switching and interpreting between two languages (cf. the “sign-switching” study). However, other areas are particular to sign languages and do not occur in spoken languages, such as the simultaneous use of signing and speaking (the “sign-speaking” study), or the rapid emergence of improvised inter-languages in situations of language contact (the “cross-signing” study).

This project received funding from the European Union’s Seventh Framework Programme for research, technological development and demonstration under grant agreement no 263647. In 2017, the European Research Council provided a follow-on grant for this work, funding our proof-of-concept project called SIGNSPACE.  



Project Lead:

Project Staff

Clients or funders

European Research Council, €1.16 million


Deaf assistants for the cross-signing part of the project visited the iSLanDS Institute in summer 2012 for six weeks of research and training, and a follow-up project is taking place in Jordan involving one of the Jordanian deaf visitors.

Collaborators and Partners

  • Prof Karen Emmorey, San Diego State University, USA
  • Dr Susanne Michaelis, Max Planck Institute for Evolutionary Anthropology, Leipzig, Germany
  • Prof Iclal Ergenc, Ankara University, Turkey
  • Hasan Dikyuva, Ankara University, Turkey
  • Bahtiyar Makaroglu, Ankara University, Turkey
  • Dr Selcuk Issever, Ankara University, Turkey
  • Dr Connie de Vos, Max Planck Institute for Psycholinguistics, the Netherlands
  • Kang-Suk Byun, Max Planck Institute for Psycholinguistics, the Netherlands
  • Dr Samuel Mathew, National Institute of Speech and Hearing, Kerala, India

Read more detailed information on the three Multisign studies below


This is a study of language contact between pairs of signers from different linguistic backgrounds who do not have a shared language between them, communicating, as it were, across language boundaries (hence the term "cross-signing").

Research questions

How do communicative strategies develop between users of two different sign languages?

What factors determine the relative success of communication?

How does this kind of communication compare to what we know about spoken language pidgins, and about International Sign pidgin?


This study draws on existing video data recorded in 2003-2005 at the Max Planck Institute for Psycholinguistics in Nijmegen, the Netherlands. Participants were deaf members of the Sign Language Typology Research Group, led by Ulrike Zeshan, from India, China, Bali, Uzbekistan, the Netherlands, Hong Kong, South Korea, and Turkey.

In addition, in the summer of 2012 additional data were collected when iSLanDS received visitors from Indonesia, Japan, and Jordan, who participated in the research along with a British signer.
Participants were filmed in communication with each other at regular intervals, that is, the very first time they met, and subsequently after one week and after one month, in order to track the development of their communication. Data include both free conversations and elicitation games (director-matcher game with picture stimuli).



Picture for receiver

Picture for signer

Picture stimulus pair for the cross-signing experiment


Cross-signing participants use a wide range of linguistic and communicative resources, including their own and invented signs, fingerspelling, pointing, mouthing, mime, and various representations of writing. They construct shared multilingual-multimodal spaces for the purpose of these conversations, and using Conversation Analysis, this process can be reconstructed. The experimental data from the director-matcher games reveal that over the course of a month, the communication has become 30% faster overall, but with large individual differences between signers.

Analysis of the use of numerals by signers was presented at the eleventh conference on Theoretical Issues in Sign Language Research (TISLR 11) in London in the summer of 2013. A poster presenting the findings can be viewed by clicking the link below.

Multilingual and multimodal aspects of “cross-signing” – A study of emerging communication in the domain of numerals (.pdf, 1MB)


“Sign-speaking” is the simultaneous production of natural utterances in sign and speech, where the structure of each language is largely felicitous rather than adjusted to the other language. Sign-speaking is distinct from notions such as "Signed English" which is an artificial code that produces signs using English syntax. Sign-speaking is also different from the bilingual utterances in which the sign language structure remain intact, but the English has become ungrammatical.

Research questions

  1. What structural characteristics do these bimodal productions have?
  2. What model of language production can account for the simultaneous production of bimodal utterances in sign-speaking?
  3. What can we infer for models of language processing from the difficulties and errors in this type of bimodal communication?


The first phase of data collection for the sign-speaking study was carried out in India in September 2013. These data comprise natural conversations by bilingual participants in Indore, who work in the bilingual environment of a residential deaf school without formal sign language interpreting provision. They all used Indian Sign Language (ISL) with spoken Hindi, but additional English/ISL bilinguals have been identified for further filming. Ankara University has been collecting conversational video data from Turkish/Turkish Sign Language bilinguals. Experimental data will be collected in 2014.


When sign-speakers communicate with hearing non-signers and deaf people at the same time, they aim to convey the same information to both hearing and deaf participants despite the syntactic and semantic differences between ISL and Hindi. This phenomenon is different from other situations where the signing is modified to parallel spoken language structures. Sign-speaking, in contrast, involves various semantic and syntactic mismatches between the two co-produced languages.

The ISL-Hindi sign-speaking data show a range of differences between the simultaneously produced autonomous utterances of ISL and Hindi. These differences can either be syntactic only, semantic only, or both semantic and syntactic as in this example:

The extent to which semantic content and sequential organisation of sign and speech can diverge in sign-speaking are of interest to issues in language processing. Data analysis is on-going.


“Sign-switching” involves the use of two or more sign languages by bi-/multilingual signers within one and the same conversation, and is functionally equivalent to code-switching/code-mixing in spoken languages.

Research questions

  1. What linguistic patterns can be found in sign-switching between two or more sign languages?
  2. How is sign-switching similar to or different from code-switching / code-mixing between spoken languages?
  3. What is the relationship between sign-switching and cross-signing?


Seven bilingual students from Burundi were selected as participants for the sign switching study. These individuals are all studying in India and are bilingual users of Burundi Sign Language (BuSL) and Indian Sign Language (ISL). The first phase of data collection was carried out in October 2011 and resulted in four hours of natural conversation between pairs and groups of sign-switchers including some guiding questions, such as ‘What was your first impression of Indian Sign Language when you arrived in India?’ Linguistic background questionnaires were also completed for each participant. Elicitation materials (a map game) were used to generate focused data on Wh-questions, negation, fingerspelling and numbers.

The Multisign team at Ankara University is filming signers who are bilingual in Turkish Sign Language and German Sign Language, so that the different groups of bilinguals can be compared.

Two of the BuSL/ISL bilingual signers participating in the ‘map game’ elicitation activity.


Signers combine both lexical and grammatical elements of ISL and BuSL in various complex ways, including the two participants who began acquiring ISL less than two years prior to filming. The data reveal very few if any communication difficulties between the signers, even where some utterances are purely in ISL. Receptive competence in ISL is higher than productive competence, especially for the signers who have not been in India very long. There are large individual differences between the signers in terms of language preference.

In addition to ‘insertion’ and ‘alternation’ which are known from spoken language research, we also find more complex patterns with repeated mid-sentence switching between the two languages, as in this example (BuSL in bold, ISL in italics, normal type font for shared signs that occur in both languages).

‘There is technical training. Or you get into a school for English or anything else. Or they have other kinds of high-quality training.’

Signs with equivalent meanings from both BuSL and ISL may appear in the same clause. This happens both for lexical signs (e.g. ENGLISH) and, more interestingly, grammatical signs (e.g. negators). It is often difficult or impossible to determine which language is “in charge” of the grammar in a given utterance.

The project is interested in modality-specific aspects of code-switching/-mixing in sign languages, as will the relationship between general constraints and individual differences, on the basis of a larger data base with more participants. Initial results have been presented at international conferences including the 9th International Symposium on Bilingualism (Singapore, June 2013).

Further Information

For further information about this project, please contact Prof Ulrike Zeshan:

International Institute for Sign Languages & Deaf Studies
Harrington Building HA212
+44 (0)1772 893104