Abstract not found

This paper considers the problem of automatic assessment of local coherence. We present a novel entity-based representation of discourse which is inspired by Centering Theory and can be computed automatically from raw text. We view coherence assessment as a ranking learning problem and show that the proposed discourse representation supports the effective learning of a ranking function. Our experiments demonstrate that the induced model achieves significantly higher accuracy than a state-of-the-art coherence model. 1

this paper as well as the competence and interests of its author. Rather, I shall look at just one feature: how subjects differ from nonsubjects in the text. More specifically, I shall investigate the differences between subjects and nonsubjects with respect to one formal phenomenon, definiteness, and one discourse phenomenon, the information-status of the entities that the subjects and nonsubjects represent.

this paper, more specifically a subset of the inferences that correlate with the syntactic form of a sentence uttered. Beginning with the early functional syntax studies by Kuno, Bolinger, and others more than a decade ago, a good deal of research has been carried out that shows that particular syntactic forms correlate with particular non-truth-conditional inferences. Understandably, the work has been mostly descriptive in nature, more concerned with the necessary first steps of establishing what sorts of inferences may correlate with linguistic form than with finding general principles of such correlations. Thus the picture that one currently has is a very fragmented one, within a single language as well as cross-linguistically or universally. That is, it seems as though any syntactic form can in principle correlate with any understanding. This may of course be the case; however, one would like to investigate the possibility that there do exist general principles underlying such correlations, perhaps of a universal nature. In what follows, I shall very tentatively propose one possible universal generalization concerning syntactic form and non-truth-conditional understanding. And my tentativeness is real, not simply a nervously modest hedge: what I shall propose is rough hewn and requires deep crosslinguistic research, but, as I shall try to show, it is plausible on the basis of a sampling of the data and it provides an agenda for further research. 2. 'Information-packaging.'

We present an implemented procedure to select an appropriate connective to link two propositions, which is part of a large text generation system. Each connec- tive is defined as a set of constraints between features of fire propositions it connects. Our focus has been to identify pragmatic features that can be produced by a deep generator to provide a simple representation of connectives. Using these features, we can account for a variety of connective usages, and we can distinguish between similar connectives. We describe how a surface generator can produce complex sentences when given these features in input. The selection procedure is implemented as part of a large functional unification grammar.

In order to provide a principled foundation for the study of discourse, in this paper we propose answers to three basic questions: What are the atomic units of discourse? What kind of structures can be built from the elementary units? How do we interpret the resulting structures semantically? Inferences and the correct interpretation of deixis and anaphors in discourse depend upon both structural and semantic accessibility relations. Structurally, we argue, discourse is context free and accessibility is determined by the coordination and subordination relations specified by the model of discourse presented here. Semantically, accessibility is controlled by relations among a number of modal contexts (interaction, speech event, genre unit, modality, polarity, and point of view) which determine the discourse world relative to which each primitive discourse unit is interpreted. To demonstrate the validity of our approach, the linguistic discourse model developed here is applied to a problem concerning the distribution of a discourse particle in Mocho and to various problems of discourse interpretation.

this article, the facts suggest to us that these sentences lack any lexical predicate, verbal or otherwise. Perhaps this fact, which is unique to equative sentences, is the source of the general blocking of extraction. It does not seem useful to us to give a partial account, as the inversion analysis does, for what is a general failure of extraction.

The processing of NP anaphors in discourse is argued to be grounded in cognitive representation and communicative principles. Derived from the Griceian maxim of quantity, the proposed theory views NP anaphor processing as an optimization process based on the principle that processing cost, defined in terms of conceptual representation, should serve some discourse function -- identifying the antecedent and/or adding new information. In a series of 5 self-pacedreading experiments, anaphors' functionality was manipulated by changing the discourse focus, and their cost was manipulated by changing the conceptual relation between the anaphors and their antecedents. The results show that reading times of NP anaphors reflect their functional justification -- anaphors were read faster when their cost had a better functional justification. These results are incompatible with any theory that treats NP anaphors as one homogeneous class regardless of discourse function and cost.

There is increasing awareness of the large degree of crosslinguistic diversity involved in the structural realisation of information packaging (or information structure). Whereas English and many Germanic languages primarily exploit intonation for informational purposes, in other languages, like Catalan, syntax plays the primary role in the realisation of information packaging and intonation is reduced to a secondary role. In yet another group of languages the primary structural correlate is morphology. This paper provides a contrastive analysis of the structural properties of information packaging in a number of languages. It also contains a discussion of some basic issues concerning information packaging and identifies a set of information-packaging primitives that are applied to the crosslinguistic facts. 1 Introduction Alternative sentential structures, differing in string order, in intonational structure, or in both, may be used to express the same propositional content. Neverthe...

this paper, I shall attempt to illustrate this complexity by examining the functioning in discourse of what is apparently one syntactic form, Left-Dislocation. I can think of few syntactic forms whose discourse function has been mentioned as often in the literature as has Left-Dislocation. However, as I shall try to show, statements of its discourse functioning have often been vague and always grossly incomplete. A careful account of how it actually works will show, I believe, that such form-function correlations must lie squarely within the domain of linguistic competence, attributable neither to common sense reasoning nor to 'iconicity'. 1. Left-Dislocation: form and previously claimed functions.