Abstract. In this paper the Centering model of anaphora resolution and discourse coherence (Grosz, Joshi and Weinstein, 1983, 1995) is reformulated in terms of Optimality Theory (ot) (Prince and Smolensky 1993). One version of the reformulated model is proven to be descriptively equivalent to an earlier algorithmic statement of Centering due to Brennan, Friedman and Pollard (1987). However, the new model is stated declaratively, and makes clearer the status of the various constraints used in the theory. In the second part of the paper, the model is extended, demonstrating the advantages of the ot reformulation, and capturing formally ideas originally described by Grosz, Joshi and Weinstein. Three new applications of the extended ot Centering model are described: generation of linguistic forms from meanings, the evaluation and optimization of extended texts, and the interpretation of accented pronouns.

Existing software systems for automated essay scoring can provide NLP researchers with opportunities to test certain theoretical hypotheses, including some derived from Centering Theory. In this study we employ ETS's e-rater essay scoring system to examine whether local discourse coherence, as defined by a measure of Rough-Shift transitions, might be a significant contributor to the evaluation of essays. Our positive results indicate that Rough-Shifts do indeed capture a source of incoherence, one that has not been closely examined in the Centering literature. These results not only justify Rough-Shifts as a valid transition type, but they also support the original formulation of Centering as a measure of discourse continuity even in pronominal-free text.

Abstract not found

this paper we follow the version that was rst implemented in the algorithm by Brennan et al. (1987) and is more clearly stated in Walker et al. (1998a). Under this interpretation the Cb(Un) can be unde ned

WITH SPECIAL REFERENCE TO HINDI Rashmi Prasad Supervisor: Ellen F. Prince This dissertation makes a progress towards the generation of referring expressions in Hindi. We first make a proposal to exploit a combination of Gricean implicatures (Grice, 1975) and Centering theory constraints (Grosz et al., 1995) to formulate a generation algorithm for referring expressions whose domain of application is defined in terms of the Centering Transitions. The formulated algorithm is an abstraction over the cross-linguistic variability observed across languages. To set the language-specific parameters of the algorithm, in particular the parameter that decides the relative salience of the discourse entities in an utterance, we propose a corpus-based methodology to identify the ways in which discourse salience is realized linguistically in any language. We apply this method to a Hindi corpus to investigate three possible linguistic reflexes of discourse salience: grammatical role, word order, and information status, and show that Hindi does not display exhibit any correlation between discourse salience and either word order or information status, and that grammatical function emerges as the primary determinant of salience. Using the results of the proposed methodology for Hindi, we provide an analysis of Hindi zero pronouns. We argue that the constraints on the use of zeros in Hindi are neither syntactic (Kameyama, 1985) nor explicable purely in terms of the singular notion of the topic (Butt & King, 1997). Our analysis, provided in terms of Centering transition preferences, shows that pronouns can be dropped in Hindi only when they occur in an utterance following a CONTINUE or a SMOOTH-SHIFT transition, thus demonstrating the importance of the Preferred Center for zero pronoun re...

This paper presents a referent-tracking and pronoun resolution system for Finnish, a free-word-order, articleless language that poses a challenge for algorithms designed for languages like English that have definite/indefinite articles. To track referents and interpret pronouns in Finnish, this algorithm uses the pragmatically-motivated word order tendencies of Finnish to create an ordered register of pegs (where each peg is associated with an entity in the discourse), ranked according to salience. Pronouns are interpreted as referring to the topmost peg in the register. The algorithm aims to extend and adapt notions from Dynamic Semantics (Groenendijk et al. 1996) and Centering Theory (Grosz et al. 1995) to a typologically di#erent language.