Calle Börstell, Tashi Bradford & Onno Crasborn

Cross-signing and creating comprehensible communication

For spoken languages, it is well known that mutual intelligibility can exist between related languages (Schuppert, Hilton, & Gooskens, 2015). This phenomenon, communicating across what are seen as distinct languages, has remained an under-studied topic for sign languages. This is possibly a consequence of the lack of a linguistic definition of what separates a language from a dialect (the age-old question in linguistics), although a few studies have tried to establish relatedness metrics between sign languages in areal or historical contact using lexicostatistics (Woodward, 1993; Johnston, 2003; Bickford, 2005; Al-Fityani & Padden, 2010; Yu, Geraci, & Abner, 2018) or by measuring mutual intelligibility (Safar et al., 2015). However, sign languages are unique in the sense that it is pos-´ sible to communicate across unrelated languages in so-called cross-signing. This has been shown to be an effective method that makes use of various deictic and iconic strategies and context-based face-to-face accommodations (Byun, de Vos, Bradford, Zeshan, & Levinson, 2018). Thus, the situation is different from direct mutual intelligibility, in which the interlocutors may use their own languages more or less unchanged and still reach a point of (partial) understanding.

In our current project, we expand on previous work on cross-signing by collecting a larger and more controlled set of data from signers of four unrelated sign languages: Sign Language of the Netherlands (NGT; Nederlandse Gebarentaal; n=15), Flemish Sign Language (VGT; Vlaamse Gebarentaal; used in Belgian Flanders; n=2), French-Belgian Sign Language (LSFB; Langue des Signes de Belgique Francophone; used in Belgian Wallonia; n=5), and Chinese Sign Language (CSL; Shanghai variety; n=5), all signers have higher education but minimal exposure to other sign languages. With these data, we investigate how cross-signing success is achieved, and to what extent factors such as shared written language knowledge (Dutch for NGT and VGT) or culture (NGT, VGT, and LSFB) facilitate communication compared to languages that are culturally and geographically distant (NGT and CSL). We use two different methods involving a) cross-signing success and b) cross-linguistic intelligibility. By combining these approaches we can investigate the extent to which a signer can understand signs from a foreign sign language and whether this predicts mutual intelligibility and communicative success in face-to-face interaction.

The first method involves pairing signers from different sign languages and video recording their conversing around general topics (e.g. personal stories) and negotiating specific communicative tasks (e.g. picture-matching tasks).

The second method involves cross-linguistic comprehension tasks in which signers individually match forms and meanings from their own or another sign language. Here, we use an NGT lexical comprehension task with 80 NGT signs presented in a randomized order. Each sign was presented with four illustrations (random 2x2 configuration; Figure 1a) with one match and three distractors (one selected based on form-matching to another concept; two based on shared iconicity with target). Our preliminary findings from this study point to comprehension being correlated with geographical and cultural proximity (Figure 1b).

With these combined methods, we are able to not only analyze the adaptive, ad hoc accommodation strategies used in a face-to-face cross-linguistic conversation, but also pinpoint how shared form or iconic mappings as well as cultural and geographical factors play a role in basic lexical comprehension and mutual intelligibility across sign languages.

Figure 1: Stimulus and results from lexical comprehension task.