Published online by Cambridge University Press: 26 September 2008
A test was designed to throw light on how three- to five-year-old Tamil children comprehend two complex sentence types, one of them ambiguous for adult speakers and involving alternating Ns and Vs in surface form (NVNV), the other unambiguous and with clustering Ns and Vs (NNVV). The results indicate that neither of the age groups tested showed adult comprehension of either construction, but apparently relied on certain strategies of perception which are sensitive to features of surface syntactic form rather than to assumed real-world probabilities. The NVNV type yielded the largest proportion of ‘correct’ responses (in terms of the adult model), and also of favoured responses (in terms of syntactic strategies that may be postulated to account for the data); but another relevant variable turned out to be the order in which the constituent structures were probed; probing the internal predicate first enhanced the probability of favoured responses, for both the constructions tested.
This paper reports on one of a number of tests carried out in south India in July–August 1972, forming part of a follow-up study to work done by Dr R. J. Wales, P. D. Griffiths, Dr B. Clayre and myself, in a project – Cross-Linguistic Studies of Language Acquisition – supported by a grant from the SSRC to Dr R. J. Wales at the Cognition Project, Department of Psychology, University of Edinburgh, 1970–1. I thank the authorities of the Nirmala School, Chidambaram, where the testing was done, and the Centre of Advanced Study in Linguistics, Annamalai University, for help and hospitality. My special thanks go to Sri R. Kothandaraman (who carried out all the 1972 tests); and to Dr R. J. Wales, P. D. Griffiths, Dr P. Fletcher and Dr P. Trudgill for their efforts to improve the paper.