Robert Levine and Detmar Meurers:
Locality of Grammatical Relations
795K/820, Spring 2003


Course Description

The idea of constituent structure plays a central role in most current syntactic frameworks, and with it the idea that it is possible and useful to define constituents as modular local domains in which to express grammatical relations and constraints. From category selection and semantic role assignment, via case and other government phenomena, to agreement phenomena, there is a general consensus that most grammatical relations can and should be expressed in terms of local trees or the domain of a single head-projection. Apart from the linguistic issue as such, it is also relevant from a computational and psycho-linguistic perspective to determine which grammatical relations need to be ensured in what domains. Language phenomena which cannot be captured locally, such as unbounded dependency constructions (topicalization, wh-questions, ...) or middle distance dependencies (raising, restructuring phenomena like clitic-climbing, ...) have thus received much attention in the literature.

In this seminar, we investigate the nature of the domains that are required to establish different kinds of grammatical relations. Some investigators have recently pointed out a number of phenomena in which traditionally local properties of embedded constituents apparently have to be visible outside of the local head domain these constituents occur in. For example, work on case assignment in German by Meurers and in Polish by Przepiorkowski, on English tag questions by Flickinger and Bender and `tough' complement structure by Levine, all point to the persistence of information about clause-internal constituents at higher levels of phrase structure configuration.

On the basis of a clarification of the different ways in which locality considerations are effective in the various frameworks, the crucial questions we envisage this course addressing include

We believe this topic to be particularly appropriate for a general discussion since the questions of how the locality of grammatical relations is reflected in the architecture, and how the apparent exceptions can be integrated into this picture, clearly involve theoretical and empirical aspects which are relevant independent of the particular formalization.

The course will start out with a short introduction of the traditional HPSG paradigm (Pollard & Sag 1994) and how the issue of locality of grammatical relations has been addressed in it. The main part of the course then is dedicated to the various empirical phenomena which have been argued in the literature to violate the mentioned locality.

The seminar is intended for graduate students from the department of linguistics and the language departments. Prerequisite is an introduction to syntax; some background in HPSG is useful.


Last modified: Tue Mar 25 18:51:56 EST 2003. For questions or comments regarding this page, please contact: Detmar Meurers.