Back to list

Detail of contribution

Auteur: John WINWARD

Patterns in article production by learners with non-DP L1

Abstract/Résumé: For many years, three questions have been at the centre of generative approaches to second language acquisition: The argument from the poverty of the stimulus; parameter resetting and post-critical-period access to UG. L2 acquisition of the English article system by learners whose L1s lack articles (hereafter L2AA) is of central importance for several reasons. First, the DP / Np split may be one of the most fundamental in comparative linguistics. Bošković (August 2012) identifies 20+ language phenomena that, he argues, follow from the NP / DP split (to a first approximation, the difference between languages with and without article systems). Secondly, L2AA directly challenges the APS. It is widely recognised that article acquisition is one of the most difficult aspects of L2 learning. Even advanced speakers do not use articles in precisely the same way that native speakers do, and intermediate learners perform at levels around chance, yet input is abundant. English articles are the highest frequency items in the language: 'the' is the single most frequently occurring item in the British National Corpus, accounting for over 6% of all items' 'a/an' is in fifth place at about 2.5%. Thirdly, the errors that L2 speakers make are not random, but map very predictably onto combinations of semantic / pragmatic factors. This has been demonstrated empirically across a wide range of language groups: Japanese, Korean, Serbo-Croat, Russian, Mandarin and Thai. Fourthly, the patterns of L2 errors allow quite precise distinctions to be made between the pragmatic and semantic factors that are influencing acquisition, though these have not been explicitly explored in the empirical literature. Finally, L2AA provides an ideal heuristic for investigating Chomsky's 'three factors' in language acquisition. This paper will present data from a series of empirical acquisition studies involving Thai L2 learners, exploring these themes in detail. The empirical studies will include a cross-sectional analysis of Thai learners at different levels of overall proficiency, a longitudinal study showing that exposure to large volumes of tailored input, unaccompanied by explicit teaching, leads to a radical change in the pattern of L2 errors and an investigation of the clustering of accuracy patterns across a wide range of language structures predicted by typological comparisons between article systems in a variety of different languages.