Editor for this issue: Karolina Owczarzak <karolina
linguistlist.org>
New Dissertation Abstract Institution: Loughborough University Program: Department of Social Sciences Dissertation Status: Completed Degree Date: 2002 Author: Anita Claire Dissertation Title: Language Acquisition Linguistic Field: Language Acquisition Subject Language: Panjabi, Eastern (code: 5011) English (code: 1738) Dissertation Director 1: Chris Hinde Dissertation Director 2: John Connolly Dissertation Abstract: This project investigates acquisition of a new language by example. Syntax induction has been studied widely and the more complex syntax associated with Natural Language is difficult to induce without restrictions. Chomsky conjectured that natural languages are restricted by a Universal Grammar. English could be used as a Universal Grammar and Punjabi derived from it in a similar way as the acquisition of a first language. However, if English has already been acquired then Punjabi would be induced from English as a second language. In both cases the source language restricts the target language. However, if indirect access to the Universal Grammar is provided the second approach could emulate access to the Universal Grammar and provide the restrictions required to adequately induce a Natural Language grammar. Studies in the thesis show that these effects can be observed in humans. The development of the second language takes place by transforming and extending the syntax of the English statements and basing the emergent Punjabi syntax on those syntactic elements. As similar possible structures are discovered the belief in those structures increases as new evidence is presented. Several examples resulting in the same structure will result in belief in that structure being increased and so syntax will build up from positive evidence. The transformations that are used to transform the source language to the target language are omission, addition and rearrangement. Omission omits a selected syntactic element from a structure in the target language. Addition complements omission and proposes an additional syntactic element. Re-arrangement takes the English syntactic elements and re-arranges them to form Punjabi. These transformations take care of all observed syntactic differences. Results have shown that there has been an increase in the belief in words and structures that have re-occurred on several occasions. The system starts with no knowledge of Punjabi, but after being presented with some Punjabi and the corresponding English it learns some elements of Punjabi syntax that differ from English, while inheriting those that are similar.Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue