Who did what to whom? Marking event participants in Emerging Languages
A central question in the study of any language is how it marks the role that each argument is playing in an event. A central question in linguistic typology is why some patterns of marking are more common than others. And central questions in language development are how these patterns are acquired, what they tell us about pre-linguistic conceptual structures and the degree to which the acquisition process shapes typology.
Historically, two strong hypotheses have bridged typology and acquisition. The word-order hypothesis proposes that order is a cognitively salient cue, which is available to children before linguistic cues (like case or verbal agreement) and thus appears early in the emergence of a new language, though it may decline in importance over historical evolution as new devices are created. The agent-first hypothesis proposes that agents have a privileged role in event representations–either because they move first or because they are conceptually salient. This leads children to expect them to appear first in the sentence, resulting in a strong bias for agent-first word order in emerging languages, which persists in most languages with a preferred word order.
Both hypotheses make strong predictions for emerging languages. While spoken language creoles conform to these predictions, this could simply reflect the word-order regularities inherited from the contact languages. Work on emerging sign languages offers a more mixed picture. ABSL and KQSL are typically described as SOV languages but have more variable word order when the patient is human. Studies of NSL have reached varying conclusions, in part because the stimuli that were used only rarely resulted in the production of both the agent and patient.
In this talk, I’ll present three lines of work bearing on these hypotheses: a published study on gestural language creation, some new data on transitive constructions in several dialects of Chinese Sign Language (lead by Hao Lin) and ongoing work on Nicaraguan Sign Language (lead by Annemarie Kocab).
The results of all three programs undermine the word-order and agent-first hypotheses (at least in their simplest form). The findings however, strongly suggest that concepts like agent and patient guide the emergence of language. Our ongoing work explores this possibility more systematically.