[Paper Review] Integrating Gricean and Attentional Constraints
This paper proposes an integrated model combining Gricean principles of informativeness and centering theory to predict and explain choices between definite pronouns and phrasal noun phrases in discourse anaphora. Analyzing 1,233 anaphoric expressions in narrative monologues, it shows that both Gricean constraints and attentional focus (via centering) jointly determine pronoun use, and extends Dale and Reiter's generation model to include understanding, offering a unified framework for discourse processing.
This paper concerns how to generate and understand discourse anaphoric noun phrases. I present the results of an analysis of all discourse anaphoric noun phrases (N=1,233) in a corpus of ten narrative monologues, where the choice between a definite pronoun or phrasal NP conforms largely to Gricean constraints on informativeness. I discuss Dale and Reiter's [To appear] recent model and show how it can be augmented for understanding as well as generating the range of data presented here. I argue that integrating centering [Grosz et al., 1983] [Kameyama, 1985] with this model can be applied uniformly to discourse anaphoric pronouns and phrasal NPs. I conclude with a hypothesis for addressing the interaction between local and global discourse processing.
Motivation & Objective
- To understand the factors governing the choice between definite pronouns and phrasal noun phrases in discourse anaphora.
- To investigate how Gricean constraints on informativeness interact with attentional focus in discourse processing.
- To extend Dale and Reiter's generation model to support understanding of anaphoric expressions.
- To integrate centering theory with Gricean principles for a unified account of discourse anaphora.
- To propose a hypothesis for reconciling local and global discourse processing mechanisms.
Proposed method
- Analyzes a corpus of ten narrative monologues containing 1,233 discourse anaphoric noun phrases.
- Applies Gricean principles to assess informativeness constraints on pronoun use.
- Uses centering theory to model attentional focus and its influence on anaphoric reference.
- Augments Dale and Reiter's generation model to support understanding by incorporating centering and Gricean constraints.
- Empirically validates the model's predictions against observed data in the corpus.
- Proposes a unified framework where both Gricean and attentional constraints jointly determine anaphoric form.
Experimental results
Research questions
- RQ1What factors determine whether a discourse anaphor is realized as a definite pronoun or a phrasal noun phrase?
- RQ2To what extent do Gricean constraints on informativeness govern anaphoric choice in natural discourse?
- RQ3How can centering theory be integrated with Gricean principles to model both generation and understanding of anaphora?
- RQ4What is the role of attentional focus in predicting anaphoric reference form?
- RQ5How can local anaphoric choices be reconciled with global discourse coherence?
Key findings
- The choice between definite pronouns and phrasal noun phrases in discourse anaphora is largely governed by Gricean constraints on informativeness.
- Centering theory provides a complementary mechanism that models attentional focus and predicts anaphoric reference patterns.
- The integration of Gricean and centering constraints enables a unified model that supports both generation and understanding of anaphoric expressions.
- The extended model of Dale and Reiter successfully accounts for the full range of data observed in the corpus.
- Empirical analysis confirms that both local (immediate context) and global (discourse-level) processing factors influence anaphoric reference choice.
- The proposed framework offers a hypothesis for resolving the interaction between local anaphoric decisions and global discourse coherence.
Better researchstarts right now
From reading papers to final review, dramatically reduce your research time.
No credit card · Free plan available
This review was created by AI and reviewed by human editors.