Short description: Basic word order in silent gesture: initially conditioned semantically, but becomes regular through interaction and iteration

Abstract:

All languages have ways to describe who did what to whom, and many have a fixed order for Subject, Object and Verb. Silent gesture, an experimental paradigm in which adult hearing participants are asked to describe events using only their hands, has proven to be a valuable tool to investigate the origins of word order. It was found that regardless the dominant order of their native language, people prefer to use SOV order for describing extensional transitive events (e.g., boy- ball-throw) (Goldin-Meadow et al., 2008). They deviate from this order, however, when an event has certain semantic characteristics. When events are reversible (the roles of Agent and Patient can be switched; e.g. boy-lift-girl) this leads to various word orders, one of which SVO (Gibson et al., 2013, Hall et al., 2013). When intensional (the Object is non-existent or non-specific; e.g., boy-search- ball), this leads to a preference for SVO order (Schouwstra & de Swart, 2014).

The two orders mentioned above, SOV for extensional and SVO for intensional events, arise independently of the dominant order of the participants’ native language, and we will claim that they represent naturalness: they are cognitively the most intuitive way to impose linear structure on information, reflecting a preference to put Agents first (see Jackendoff (2002)) and more abstract or relational information last. Existing languages do not generally reflect this natural- ness, as they tend not to have word order conditioned on event type. Given the improvisation-situation that favours semantically conditioned word order, and the fully conventionalised situation that favours regularity in word order, we investigated what happens to silent gesture over time when it is used for communication. Will it become more regular, like conventional language?

In experiment 1, 24 adult native speakers of English with no knowledge of any sign language were assigned into dyads, and each dyad was asked to communicate about intensional and extensional events. The set of stimuli consisted of 64 line drawings: 32 intensional and 32 extensional events. Participants alternated between the role of actor and interpreter and engaged in six rounds of 32 trials each (switching roles each trial). As actor they described an image (presented on an iPad) using only their hands, and as interpreter they chose (from an array of 8 images on the iPad) the image they thought was intended by the actor. Each actor described equal numbers of intensional and extensional events. They received immediate feedback after each trial, and were encouraged to increase their speed over rounds and be as quick and accurate as possible overall.

Speed and accuracy increased over the course of the experiment. Moreover, the word orders showed signs of conventionalisation: over the rounds, word order became less conditioned on meaning, and 7 of 12 dyads even converged on a single word order. However, all 7 pairs converged on SVO, the natural order for intensional events but also same order as their native language.

To see if the frequency of event types could influence the word order of the emerging sign system, we conducted a second experiment, in which extensional events were more frequent than intensional events. Experiment 2 was set up the same as experiment 1, except for the proportions of the two kinds of events: each actor described 24 intensional events (25%) and 72 extensional events (75%). The larger proportion of extensional events had an influence on the proportion of SOV orders used throughout the experiment, compared to experiment 1, and on the way in which conventional word order was introduced: although 3 dyads converged on SVO word order, 3 other dyads converged on SOV.

Our experiments show that in silent gesture communication, semantically conditioned word order tends to disappear in favour of more regular word order. The frequency of extensional and intensional events influences the way in which regularisation progresses. This suggests that where pressures for naturalness and regularity are in conflict, languages may start natural, but that naturalness will give way to regularity as signalling becomes conventionalised through repeated usage.

Download Paper (pdf)

Supplementary Materials (zip)

Citation:

Schouwstra M., Smith K. and Kirby S. (2016). From Natural Order To Convention In Silent Gesture. In S.G. Roberts, C. Cuskley, L. McCrohon, L. Barceló-Coblijn, O. Fehér & T. Verhoef (eds.) The Evolution of Language: Proceedings of the 11th International Conference (EVOLANG11). Available online: http://evolang.org/neworleans/papers/67.html