The Role of the Gricean Maxims in the Generation of Referring Expressions

What it is about

Grice's maxims are usually read as directives to a speaker. This short paper argues that, for natural language generation, they are better read as post-hoc descriptions. A goal-driven generator, which adds information only when a goal calls for it and conforms to the conventions of the genre, behaves as the maxims describe without an explicit check. For referring expressions, the authors list goals such as identifying the referent and alerting the hearer to a property. A modifier that serves a second goal is read by the hearer as having a purpose, and so it does not produce a false implicature if the genre uses that attribute for identification or for other purposes. The conclusion is that "do not include elements that do nothing" is what the maxims mean for a generator. The paper has no experiments.

Ideas

• I1. Goal agenda: represent each request as a list of goals (identify the referent; state a fact) and remove a goal once a realisation satisfies it. In the school world, the answer generator would then emit only content that serves a listed goal. Measure answer length on the school questions.
• I2. Genre-preferred attributes: keep, for each lesson genre, the attributes used for identification, and let the realiser use only those. This is a table, not a search.
• I3. Hearer check: an answer with an extra non-identifying attribute is read as having a purpose only when the genre uses that attribute for other purposes. Encode this as a lookup and test with the implicature rules in prag.
• I4. Negative control: a generator that adds random extra attributes must be flagged as an implicature error more often than the goal-driven one.

For Mova

Chapters: 33. Implicature and presupposition, 37. Register and variation, 35. Reference and coreference. Code: world/src/qa.rs (answers), prag/src/infer.rs (implicature rules). Measure: number of content units per answer and the flagged-implicature rate, with the random-attribute control; two runs must match.

Topics: Implicature

Cite

The paper itself, not this card. BibTeX from the library's record of the work. arXiv cmp-lg/9604006

@misc{dale1996role,
  title = {{The Role of the Gricean Maxims in the Generation of Referring Expressions}},
  author = {Dale, Robert and Reiter, Ehud},
  year = {1996},
  eprint = {cmp-lg/9604006},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/cmp-lg/9604006}
}

References

11 references, 1 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Douglas Appelt [1982] Planning Natural Language Utterances to Satisfy Multiple Goals. Technical Note 259, SRI International, Menlo Park, California. find
2. Robert Dale [1989] Cooking Up Referring Expressions. In Proceedings of 27th Annual Meeting of the Association for Computational Linguistics, Vancouver bc. find
3. Robert Dale and Ehud Reiter [1995] Computational Interpretations of the Gricean Maxims in the Generation of Referring Expressions. Cognitive Science, 19(2), pp233–263. find our card
4. H Paul Grice [1975] Logic and conversation. In P Cole and J Morgan, editors, Syntax and Semantics: Vol 3, Speech Acts, pages 43–58. New York: Academic Press. find
5. Anthony Jameson [1987] How to Appear to be Conforming to the Maxims Even if you Prefer to Violate Them. Chapter 2 in G Kempen (ed), Natural Language Generation, Dordrecht: Martinis Nijhoff Publishers. find
6. Anthony Jameson and Wolfgang Wahlster [1982] User Modelling in Anaphora Generation: Ellipsis and Definite Description. In Proceedings of the Fifth European Conference on Artificial Intelligence, Pisa, Italy. find
7. Willem Levelt [1989] Speaking: From Intention to Articulation. Cambridge, ma: mit Press. find
8. David McDonald [1980] Natural Language Generation as a Process of Decision Making Under Constraints. PhD Thesis, mit. find
9. Hans-Joachim Novak [1988] Generating Referring Phrases in a Dynamic Environment. Chapter 5 in M Zock and G Sabah (eds), Advances in Natural Language Generation, Volume 2, pp76–85. Pinter Publishers. find
10. Ehud Reiter [1990] Generating Appropriate Natural Language Object Descriptions. PhD Thesis, Harvard University. find
11. Jacques Robin [1994] Revision-Based Generation of Natural Language Summaries Providing Historical Background. PhD Thesis, Columbia University. find

This page in Markdown → https://md.mova.elamur.ai/free/en/docs/papers/dale-1996-role-gricean-maxims-generation-referring/