Article

Can sign-naïve adults learn about the phonological regularities of an unfamiliar sign language from minimal exposure?

---

Abstract

Research on spoken languages shows that adults can extract phonological regularities from just several minutes of exposure to naturalistic input of an unknown language (Gullberg et al., 2010). In this study, we examined whether such implicit statistical learning mechanisms operate in the visual modality. Specifically, we investigated whether sign-naïve adults can learn about phonological regularities from minimal exposure to a sign language. Participants (N=121) were L1-speakers of English with no prior knowledge of a sign language. The input materials consisted of a continuous sign stream in the form of a four-minute weather forecast video in Swedish Sign Language (Svenskt teckenspråk; STS). Participants were randomly assigned to three groups: two experimental groups who watched the forecast once (N=43) or twice (N=38), and a control group who did not watch it (N=40). To take account of individual differences in task performance, we assessed participants’ English vocabulary knowledge, non-verbal reasoning ability, and demographic and language-learning background. After having watched the forecast, participants completed a ‘surprise’ lexical decision task designed to tap into their awareness of the phonological properties of the core STS lexicon. They viewed individual signs and were asked to indicate whether or not each sign could be a real STS sign. The task comprised four sets: (1) STS signs that were presented in the forecast, (2) STS signs that were not presented in the forecast, (3) Signs that are not STS signs and contain handshapes outside the STS handshape inventory, (4) Signs that are not STS signs and contain sets of phonological features that are dispreferred across sign languages. Response accuracies were compared as a function of sign set, quantity of input, and individual differences, using generalised linear mixed effects models. In short, we found no evidence that participants could distinguish between the signs that are and are not real signs of STS, and therefore those who had viewed the forecast did not demonstrate any learning of STS phonological regularities. ‘Yes’ responses accounted for approximately 58% of responses across all groups and for all sign sets. Considered in conjunction with two companion studies which revealed that participants were able to demonstrate learning of sign forms and their meanings after viewing these same materials, we argue that our findings demonstrate the limits of what can be learnt after just a few minutes of implicit and naturalistic exposure to a new language in an unfamiliar modality: information about specific lexical items is learnable, but information that requires generalisation across items may require greater amounts of, or a different quality of, input.  

Keywords: sign language, implicit learning, first exposure, phonological regularities, iconicity