This paper presents an algorithm for identifying the noun phrase antecedents of third person pronouns and lexical anaphors (reflexives and reciprocals). The algorithm applies to the syntactic representations generated by McCord's Slot Grammar parser, and relies on salience measures derived from syntactic structure and a simple dynamic model of attentional state. Like the parser, the algorithm is implemented in Prolog. The authors have tested it extensively on computer manual texts, and conducted a blind test on manual text containing 360 pronoun occurrences. The algorithm successfully identifies the antecedent of the pronoun for 86 % of these pronoun occurrences. The relative contributions of the algorithm's components to its overall success rate in this blind test are examined. Experiments were conducted with an enhancement of the algorithm which contributes statistically modelled information concerning semantic and real world relations to the algorithm's decision procedure. Interestingly, this enhancement only marginally improves the algorithm's performance (by 2%). The algorithm is compared with other approaches to anaphora resolution which have been proposed in the literature. In particular, the search procedure of Hobbs ' algorithm was implemented in the Slot Grammar framework and applied to the sentences in the blind test set. The authors ' algorithm achieves a higher rate of success (4%) than Hobbs ' algorithm. The relation of the algorithm to the centering approach is discussed, as well as to models of anaphora resolution which invoke a variety of informational factors in ranking antecedent candidates. 1.

Abstract not found

Traditionally, vector-based semantic space models use word co-occurrence counts from large corpora to represent lexical meaning. In this article we present a novel framework for constructing semantic spaces that take syntactic relations into account. We introduce a formalization for this class of models which allows linguistic knowledge to guide the construction process. We evaluate our framework on a range of tasks relevant for cognitive science and natural language processing: semantic priming, synonymy detection and word sense disambiguation. In all cases, our framework obtains results that are comparable or superior to the state of the art. 1.

Typology represents an approach to the study of linguistic structure that differs in certain important respects from the generative and the functionalist approaches (Wasow, this volume; Van Valin, this volume), although it is closer in spirit to the latter. The most important difference between typology and these other approaches to linguistic structure

this paper. Towards that end, comments are welcome. 1 (1) S

This dissertation examines the "free" word order or scrambling phenomena in German and Korean from the perspective of constraint interaction in Optimality Theory. To overcome the problems raised in single-component analyses in explaining word order variation, I propose an `interface' approach in which the constraints from several different components of grammar participate, compete, and interact with one another. That is, various word orders are considered to be motivated and constrained by interactions among syntactic, semantic, and discourse principles of these languages. As the constraints from different modules of grammar are highly conflicting, I utilize Optimality Theory to demonstrate how the constraints interact and resolve conflicts among one another. In this approach, each scrambled variant, i.e. a sentence with a particular word order, is conceived of as the "optimal" output, which instantiates the syntactic, semantic, and discourse-contextual information given in the input....

This paper is drawn from my 1994 Stanford dissertation of the same name (copies of which are available from http://kinks.phil.cmu.edu/manning/papers/, or by contacting the author), which should be consulted for further information, acknowledgements and references.

Radical Construction Grammar is a model of morphosyntactic representation that emerged from bringing together the results of typological research and the issues addressed by contemporary theories of syntax. The product of this marriage of typology and syntactic theory is basically a variety of construction grammar, but one that is quite

this paper is how can we go about explaining the observed constraints on variation across languages --- in other words, language universals.

Tradiiaonally, the document summansahon task has been tackled other as a natural language processmg problem, with an mstanhated meanrag tnplate being rndeocl mto cohent prose, or as a passage xiractlon problem, where cetam fragments 0ypcally sentences) of the source document am deemed to be highly represeniative'of its content, and thus dehverod as rneanmgful "apprommahons "of It Balanong the coflctmg requu ments d depth and accurecy of a sunwnary, on h one hand, and document and domam mo dependence on the other, has proven a very hard problem Tius pez descnbes a novel approach to content charactensahon of text documents It domam- and genre-mdependent, by vu-tue of not requmng an m-depth armlyss of the full raean- mg At the same tune, l! remams closer to the core meamu blt chonsm a ddferent granularity of its repmsentahons (phrasal expressions rather than senience or paragraphs), by explotmg a notion of &scourse conhgmty and coherence for the purposes oi'undorm coverage and context mamtenance, and by uhhsmg a strong lagrustic notion of sahence, as a more appropriate and mpmsenta- bye meastue of a documents "aboutne" 1 Capsule overviews The matonty of techmques for "summansaton", as apphed to average-length documents, fall wlthm two, broad categones those that rely on template mstantmton and those that rely on passage extraction Work ut the former framework tracs its mots to some ploneenng research by DeJong [Y]. and Tait [29], more recentl th DARPA-Sponsored TIPSTER programme ([2])---and, m particular, the meage understandrag conferencas {Ivc e g [6] and [1])-have provided fettle ground for such work, by placmg the emphass of document analyss to the identification and extraction of cer- tam core entihes and facts m a document, which are "pac...