Project A3:
Suboptimal syntactic structures
 |
||
|
||
|
||
|
||
|
Project leader
Prof. Dr. Wolfgang Sternefeld
Project members
- Sam Featherston
- Universität Tübingen
- SFB 441
- Nauklerstr. 35
- D - 72074 Tübingen
- e-mail:
- Tel.: 07071 - 2977154
- SFB 441
- Tanja Kiziak
- Universität Tübingen
- SFB 441
- Nauklerstr. 35
- D - 72074 Tübingen
- e-mail:
- Tel.: 07071 - 2977152
- SFB 441
- Anca Weimer
- e-mail:
- Tel.: 07071 - 2977152
- Tel.: 07071 - 2977152
- Tim Friedrich
- e-mail:
- Tel.: 07071 - 2977152
Advisory partner
- Ilona Steiner (Project B3)
Ex-members
- John Vanderelst
- Ilona Steiner (Project B3)
- e-mail:
- Aria Adli (New York University)
- Monika Toth
Project Outline
The insight that grammaticality judgments are not binary but rather gradient is not new, but this fact has received relatively little attention in syntactic theory. A few approaches have been developed in computational linguistics (e.g. Abney 1997, Riezler 1997), and a few works have discussed this issue in theoretical syntax (e.g. Uszkoreit 1987, Hayes 1998, Fischer 1999, Keller 2000). Project A3 investigates the problem of suboptimality systematically; methodological aspects as well as experimental findings constitute the main issues. The former focus on the data type of grammaticality judgments, the latter concern the survey and the analysis of graded data with respect to grammatical theory. The collection of judgements using strictly controlled experimental techniques guarantees a methodologically sound data base for the implications that we draw for theory. Our theoretical approach is strongly data-driven, but maintains a generative perspective on the issues to be addressed by syntax. Our findings suggest that a generative grammar enriched by a parameter of constraint violation cost strength is an empirically adequate model. Most of the syntactic case studies that we address are classic issues from German, but we address parallel phenomena cross-linguistically when it appears fruitful to do so. A further part of the work of the project is our database of suboptimal structures. The primary aim of this is to allow generative syntacticians easy access to a collection of controversial but theoretically relevant data. This database is also interesting as the systematic instantiation of a generative grammar (Sternefeld 2004) in a tree bank.
Current Projects
Outlines of selected research areas
Our current experiment
Database of Grammaticality Judgements (test version)
from left to right: Anca Weimer, Sam Featherston, Simone Hartung, Wolfgang Sternefeld, Tanja Kiziak, John Vanderelst
Original design by Laura Kallmeyer and Christoph Singer. Maintained by Sam Featherston, last updated 18.05.06
