Abstract not found

Abstract not found

This paper investigates whether the model of local rhetorical coherence suggested in Knott et al. (2001) can boost the performance of the Centering-based metrics of entity coherence employed by Karamanis et al. (2004) for the task of information ordering. Rhetorical coherence is integrated into the way Centering’s basic data structures are derived from the annotated features of the GNOME corpus. The results indicate that (a) the simplest metric continues to perform better than its competitors even when local rhetorical coherence is taken into account, and (b) this extra coherence constraint decreases its performance. Keywords: Information Ordering, Centering Theory, Rhetorical Coherence. 1.

In line with recent studies we propose a model of human sentence processing that is based on Optimality Theory (OT). Rather than explaining parsing preferences through extralinguistically motivated parsing strategies or frequencies in the hearer’s linguistic environment, our model explains these preferences as the intermediate results of the incremental application of our OT grammar. In contrast to most other current OT-approaches to language processing, we use constraints from OT semantics rather than from OT syntax to explain on-line comprehension. We illustrate the workings of our model by investigating the comprehension of coordination, a phenomenon which is ill-understood from a competence perspective and sparsely investigated from a processing perspective. The psycholinguistic evidence that is currently available strongly suggests that the on-line comprehension of coordinate structures is influenced by constraints from many different information sources: pragmatics, discourse semantics, lexical semantics, and syntax. The competence / performance model we propose is able to formalize this cross-modular constraint interaction, and to yield concrete predictions with respect to both intermediate parsing preferences and ultimate interpretations.

Abstract not found

generalised quantifiers and hypothetical contexts

Indefinite noun phrases in certain contexts are unable to support anaphoric coreference to an individual entity, and therefore should be ignored when searching for coreferent antecedents of anaphoric pronouns. However, many algorithms for anaphora resolution utilize noun phrase chunking or shallow parsing, and therefore do not make the needed distinctions to avoid this type of spurious antecedent. This study used simple syntactic criteria to remove indefinite phrases from consideration as antecedents to evaluate the effect of their removal on pronoun resolution. Pronoun resolution performance improved only marginally, revealing some interesting properties of current pronoun resolution algorithms. 1.

Young children often give the impression that they speak or listen from their own perspective, with little or no recognition of the linguistic information that a discourse partner might be using. In relation to discourse reference, children often demonstrate insensitivity to rules that determine the use of discourse pronouns (Karmiloff-Smith 1981). Frequently, discourse development has been explained by such cognitive notions as egocentricity. However, recent eye-tracking studies have shown that very young children are not exclusively egocentrically oriented (Nadig & Sedivy 2002) and that even adults sometimes begin with an egocentric perspective and only later switch to their partner’s perspective (Keysar, Barr, Balin & Paek 1998). The question in the present study is: if children and adults are capable of cognitively taking either perspective, do both children and adults linguistically take into account their partner’s perspective in discourse? To explore this question, children and adults participated in a production experiment and a comprehension experiment aimed at investigating their sensitivity to discourse topic and topic shift. In Section 1, we discuss the theoretical framework of bidirectional Optimality Theory, which is shown to make linguistic predictions as to expected differences between adults and children when they deal with topic shift during discourse. Adult speakers and hearers are predicted to optimize bidirectionally and, therefore, to use specific referring expressions to maintain or shift topics during discourse. If children optimize unidirectionally, they will demonstrate a different use or interpretation of referring expressions. Section 2 introduces a production experiment and a comprehension experiment, designed to test for possible discourse topic differences between adults and children. For production, the participants are asked to describe six-page picture books and for comprehension they listen to structured stories about which a question must be answered. The results of these two experiments are presented in Section 3. Section 4 discusses the differences between adults ’ and children’s use and interpretation of pronouns and noun phrases as signaling topic and topic shift in ongoing discourse. Conclusions are summarized in Section 5.

A recent paper (Beaver, 2004) recasts the Centering algorithm for pronoun resolution (Brennan et al., 1987) in terms of optimality theory (OT). Although the limitations of centering for pronoun resolution are well known, the algorithm's restatement in OT highlights the strengths of centering and produces an algorithm whose behavior is more transparent to the developer. The authors' motivation for developing this algorithm was due to the software engineering characteristics of OT as a specification mechanism, rather than improved pronoun resolution performance. There has been little previous work marrying computational pronoun resolution work with OT. Therefore, this study reports our experience implementing the OT formulation of centering in order to evaluate its desirability as a pronoun resolution specification formalism. Although the expressive capability of OT has some limitations, the benefits of the OT formalism are worth some attention. This paper seeks to highlight those benefits, and indicate how they can be employed within pronoun resolution algorithms, whether or not an OT formalization is employed. 1.

The article aims to give an overview about the application of Optimality Theory (OT) to the domain of pragmatics. In the introductory part we discuss different ways to view the division of labor between semantics and pragmatics. Rejecting the doctrine of literal meaning we conform to (i) semantic underdetermination and (ii) contextualism (the idea that the mechanism of pragmatic interpretation is crucial both for determining what the speaker says and what he means). Taking the assumptions (i) and (ii) as essential requisites for a natural theory of pragmatic interpretation, section 2 introduces the two main views conforming to these assumptions: Relevance theory, Levinson’s theory of presumptive meanings, and the Neo-Gricean approach. In section 3 we explain the general paradigm of OT and the idea of bidirectional optimization. We show how the idea of optimal interpretation can be used to restructure the core ideas of these three different approaches. Further, we argue that bidirectional OT has the potential to account both for the synchronic and the diachronic perspective on pragmatic interpretation. Section 4 lists relevant examples of using the framework of bidirectional optimization in the domain of pragmatics. Section 5 provides some general conclusions. Modeling both for the synchronic and the diachronic perspective on pragmatics opens the way for a deeper understanding of the idea of naturalization and (cultural) embodiment in the context of natural language interpretation. 1