1. University of Chicago, Department of Psychology, 5730 South Woodlawn Avenue, Chicago, Illinois 60637, USA

Correspondence to: Susan Goldin-Meadow¹ Correspondence and requests for materials to S.G.-M. (e-mail: Email: sgsg@ccp.uchicago.edu).

Abstract

Deaf children whose access to usable conventional linguistic input, signed or spoken, is severely limited nevertheless use gesture to communicate^{1, 2, 3}. These gestures resemble natural language in that they are structured at the level both of sentence⁴ and of word⁵. Although the inclination to use gesture may be traceable to the fact that the deaf children's hearing parents, like all speakers, gesture as they talk⁶, the children themselves are responsible for introducing language-like structure into their gestures⁷. We have explored the robustness of this phenomenon by observing deaf children of hearing parents in two cultures, an American and a Chinese culture, that differ in their child-rearing practices^{8, 9, 10, 11, 12} and in the way gesture is used in relation to speech¹³. The spontaneous sign systems developed in these cultures shared a number of structural similarities: patterned production and deletion of semantic elements in the surface structure of a sentence; patterned ordering of those elements within the sentence; and concatenation of propositions within a sentence. These striking similarities offer critical empirical input towards resolving the ongoing debate about the 'innateness' of language in human infants^{14, 15, 16}.