The course surveys competence and performance issues relating to modelling grammars of free word order languages.
Regarding competence issues, we focus on two fundamental issues adjacency (unbounded dependencies, crossed and nested dependencies), and the power required of the formal apparatus used to explain free word order. In this context, we survey various approaches to modelling free word order in natural language grammars (notably, categorial grammar, G/HPSG, TAG), and argue, following the Prague school of linguistics, that variation in word order is a structural means to realize information structure - just like intonation in English, or the use of specific morphological particles in Japanese.
Regarding performance modelling, we first present a summary of existing research (both theoretical and empirical) on sentence processing in Dutch, German, Japanese, and Korean that focuses on dependency issues (like cross-serial center-embedding, and scrambling). We also present new research on Hindi sentence processing that sheds new light on the cross-linguistic issues currently under investigation in the literature.
We conclude by presenting a coherent computational model of grammars for these languages that takes into account the competence and performance issues and facts discussed earlier in the course.
Universität des Saarlandes
Computational Linguistics, Lehrstuhl Uszkoreit
Postfach 15 11 50
PHONE: +49 (681) 302 4502
FAX: +49 (681) 302 4700
Ohio State University 222
Oxley Hall 1712 Neil Avenue
Columbus, OH 43210-1298, USA
PHONE: +1-614-688-0973 (Eastern Standard Time)