Compared to the longstanding histories of spoken languages, all known signed languages are considered to be young languages (Meir, Sandler, Padden, & Aronoff, Reference Meir, Sandler, Padden, Aronoff, Marschark and Spencer2010). For this reason, the study of sign languages and the social mechanisms through which they evolve provides a unique opportunity to shed light on the following questions: Which aspects of our communicative abilities are present from the very earliest stages of language emergence; and, by extension, which aspects of our cognition have been selected for as language evolved?

From the 1970s until recently, sign language linguistics focused almost entirely on sign languages that have arisen as deaf people have congregated in the context of government institutions for the deaf, primarily deaf schools (McBurney, Reference McBurney, Pfau, Steinbach and Woll2012). Oftentimes such sign languages have been around for several centuries, such as Old French Sign Language and its descendent American Sign Language, but in a few cases sign linguists have been able to track the emergence of new sign language from the very start (Senghas, Kita, & Ozyurek, Reference Senghas, Kita and Ozyurek2004). From 2005 onward, the field has started to investigate the many sign languages to have emerged in rural areas with a high incidence of deafness (Zeshan & de Vos, Reference Zeshan and de Vos2012). In a handful of cases, such complex gene-culture coevolution has led to longstanding rural signing communities, but in most cases the unique circumstances that lead to emergent signing varieties do not allow them to persist across multiple generations (Mudd, de Vos, & De Boer, Reference Mudd, de Vos and De Boer2020).

Emergent signing varieties are often thought to originate in homesign systems (Senghas et al., Reference Senghas, Kita and Ozyurek2004); that is to say, one-off communication systems that begin and end with just one deaf individual who co-creates a visual–gestural form of communication with their hearing relatives and friends in the absence of a signing community (Goldin-Meadow & Brentari, Reference Goldin-Meadow and Brentari2017). The homesign literature thus far has focused mostly on the genesis of linguistic structures and the cognitive consequences of long-term language deprivation (see Motamedi, Schouwstra, Smith, Culbertson, & Kirby, Reference Motamedi, Schouwstra, Smith, Culbertson and Kirby2019, for a recent overview). Most notably, Gagne and Coppola (Reference Gagne and Coppola2017) found that the four Nicaraguan homesigners who participated in their study were unable to pass standard false belief tasks that require the ability to predict other's beliefs and behaviors. When taken at face value, these findings are problematic for any perspective on language evolution that views our pragmatic abilities as foundational to human language (cf. the target article; Levinson, Reference Levinson2019). In the remaining paragraphs of this commentary, I provide an alternative view based on data from Bali: that, in everyday conversation, homesigners may demonstrate ample evidence of mentalizing abilities.

Crucially, most work on homesign has been based on small-scale case studies elicited from a small number of deaf individuals in Nicaragua and the United States. The data discussed here stem from the newly created Balinese Homesign Corpus, which includes, among other things, conversational data from 14 homesigners and their hearing interlocutors across the province of Buleleng, Bali (see Fig. 1). This area of Bali is of particular value for understanding how homesign may jumpstart the grammars of emergent signing varieties, because it is also home to Kata Kolok, a rural sign language that has been acquired naturally by at least six subsequent generations (de Vos, Reference de Vos2012; Lutzenberger, Reference Lutzenberger2022).

Figure 1. Social interaction patterns of all 14 homesigners recorded for the Balinese Homesign Corpus, including data from the villages of Bebetin, Bulian, Suwug, and Tajun as well as their relation to the Kata Kolok signing community in Bengkala.

While there is plentiful evidence for grammatical complexity in homesign systems, the homesigners and their interlocutors in the Balinese Homesign Corpus communicate using the full range of semiotic strategies available to them (see also Safar, Reference Safar2019). An important communicative strategy to ground the meaning of their utterances builds on the local gesture culture (Nyst, Reference Nyst2019). This includes highly conventionalized forms of quotable gestures, also known as emblems, being co-opted for more diverse meanings. For example, the Balinese conventional gesture to insult men by comparing them with female genitalia has actually attained a neutral, normalized meaning of “woman” or “wife” in Kata Kolok (Marsaja, Reference Marsaja2008). As has been well-documented for spoken languages, such a semantic shift is engendered by interlocutors using this form in conversational settings in which the new intended meaning can be derived from the situational context.

Homesigners also ground their utterances in what they understand to be shared knowledge between themselves and their hearing interlocutors by inventing novel signs based on cultural practices. For example, the way a particular fruit is normally peeled or cut becomes a way of referring to that fruit. Similarly, homesigners will point to locations that they know their interlocutors will make similar associations with, such as pointing to someone's homebase to refer to that individual. Lastly, in conversational settings, deaf homesigners will monitor their interlocutor's understanding of what is being expressed by responding to situations in which positive feedback by way of nodding is discontinued, or in which non-understanding is signaled through a puzzled facial expression by their interlocutor. Subsequent utterances are tailored to deal with the source of communicative trouble, supplementing their initial expression with additional forms to make themselves understood (Safar & de Vos, Reference Safar and de Vosforthcoming). All in all, our observations of homesigners in their everyday social interactions indicate an ability to capitalize on semiotic resources to help build rapport with their interlocutors and re-establish mutual understanding when needed. This means that prior indications of the limited mentalizing abilities of homesigners may be indicative of a task effect, or alternatively a result of the extent to which their hearing interlocutors in Bali engage with homesigners using visual communication from the get-go.

The remaining questions then are to understand the extent to which the quality of social interaction and subsequent cultural learning boost or impede pragmatic competence, and by extension, the extent to which such abilities have indeed been shaped by human evolution. The Balinese Homesign Corpus represents the world's first extensive collection of homesign conversations, which enable us to pursue fundamental questions on the nature and origins of cognitive pragmatic competence.