Preventing False Temporal Implicatures: Interactive Defaults for Text Generation

What it is about

Readers infer temporal relations between clauses from their order, from causal knowledge and from discourse rules. If a generator always mirrors event order, the text reads as narrative, but a good text can put the main event first and its cause after it, so text order and event order differ. The authors model the hearer's default inferences with defeasible rules (Narration, Explanation, a causal law; the more specific rule wins). They apply the Interactive Defaults strategy: a generator takes a candidate utterance, derives what the hearer will infer, and rejects the candidate if the inference is incoherent, unreliable or ambiguous. If the hearer lacks the causal law, the generator must reorder the clauses, add a clue word such as because, or use the pluperfect. The paper is a theoretical framework with worked examples and no quantitative evaluation.

Ideas

• I1. Filter candidates: after en/src/lin.rs produces orderings of a clause set, run the prag inference on each and keep only the orderings whose derived temporal order matches the intended one. Count the rejected candidates per text.
• I2. Knowledge-gap rule: when the hearer model lacks a causal rule, insert a clue word or the pluperfect instead of reordering. Test on the paper's contrast pairs (opened the door / switched off the light).
• I3. Store the defaults as an ordered rule list in which the more specific antecedent wins, in a BTreeMap-ordered table so the result is deterministic.
• I4. Negative control: remove one causal rule from the knowledge base; the number of unreliable realisations must rise. If it does not, the check is not reading the rule.

For Mova

Chapters: 36. Discourse relations and coherence (temporal and causal relations), 33. Implicature and presupposition, 29. Ellipsis, word order and information structure (clause order). Code: default inference in prag/src/infer.rs, frames in prag/src/frames.rs, candidate orderings from en/src/lin.rs. Measure: share of candidate texts whose inferred temporal order matches the intended one, with and without the causal rule (the negative control); two runs must match.

Topics: Implicature

Cite

The paper itself, not this card. BibTeX from the library's record of the work.

@inproceedings{oberlander1992preventing,
  title = {{Preventing False Temporal Implicatures: Interactive Defaults for Text Generation}},
  author = {Oberlander, Jon and Lascarides, Alex},
  booktitle = {COLING 1992 Volume 2: The 14th International Conference on Computational Linguistics},
  year = {1992},
  url = {https://aclanthology.org/C92-2108/}
}

References

17 references, 0 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Allen, J. F. & Perrault, C. R. [1980] Analyzing intention in Dialogues. Artificial Intelligence, 115, 143 178. find
2. Asher, N. & Morreau, M. [1991] Comnton Sense Entailment: A Modal Theory of Nomnouotonic Reasoning. In Proceedings o] the 12th International Joint Con]cr-ence on Artificial Intelligence, Sydney, Australia, August 1991. find
3. Bach, E. [1986] The algebra of events. Linguistics anti Philosophy, 9, 5-16. find
4. Cohen, P. It. & Perrault, C. R. [1979] Elements of a Plan-Based Theory of Speech Acts. Cognitive Science, ACRES DE COLING-92, NANTES, 23-28 AO(n' 1992 3, 177-212. find
5. Cohen, P. R., Morgan, J. & Pollack, M. E. [1990] Intentions in Communication. Cambridge, MA: hilT press. find
6. Gtice, H. P. [1975] Logic and Conversation. In Cole, P. and Morgan, J. l,. reds.) Synlaz and Semantics, Vol-rune 3: Speech Acts, pp41-58. New York: Academic Press. ilinrichs, E. [1086] Temporal Anaphora in Discourses of English. Linguistics and Philosophy, 9, 63--82. llobbs, J. it. [1985] On the Coherence and Structure of 1)iscourse. Report OSL1-85 ,37, Center for the Study of Language and htformation, Stanford, Ca., October, 1985. find
7. Hobbs, 3., Stickel, M., Martin, P. & Edwards, D. [1988] Interpretation as Abduction. In Proceedings of the 261h Annual Meetin 9 of the Association ]or Compu° rational Linguistics, suNY, lluffMo, N.Y., June, 1988, pp95 103. flobbs, J., Sticket, M., Appelt, D. & Martin, P. [1990] Interpretation as Abduction. Technical Note No. 499, sl~l lnternationM, Menlo Park, Ca., l)ecember 1990. llovy, E. [1990] Pragmatics and Natural Language Generation. Artificial Intelligence, 43, 153-197. find
8. Joshi, A., Webber, B. & Weisehedel, R. M. [19849] Preventing I"Mse Inferences. In Proceedings of the lOth h~ternatlonal Con.[vrence on Computational Linguistics and the $2nd Annual Meeting o] the Association .[or Computational Linguistics, Stanford University, Stanford, Ca., 2-6 July, 1984, pp134~138. find
9. Joshi, A., Webber, B. & Weischedel, It. [1984b] Default reasoning in interaction. In Proceedings o.[ the Non-Monotonic tleasoning Workshop, AAAI, N.Y., Oc tober, 1984, pp144 150. find
10. Joshi, A., Weblrer, B. & Weischedel, R. [1986] Some Aspects of Default Re~oning in Interactive Discourse, lteport MS-C1S-86-27, University of Pennsyl-vasia. find
11. Konolige, K. [1991] Abduction vs Closure in Causal Theories. Forthcoming Research Note in A rtificial In. telligeacc. Page references to ms. find
12. I,ascarides, A. & Asher, N. [1991] Discourse Relations and Couuoon Sense Entailment. Submitted to Journal o] Logic, Language and In.formation. DYANA deliverable 2.5b, Centre for Cognitive Science, University of Edinburgh. find
13. Lasearides, A. & Oberlander, J. [19929] Temporal Coherence anti Defeasible Knowledge. Theoretical Linguistics, 18. find
14. Lascarides, A. & Oberlander, J. [1992b] Abducing Temporal Discourse. In Dale, R. tlovy, E. RSsner, D. and f<t,ck, O. reds.) Aspeets o] Automated Natural Language Generation. Berlin: Springer-Verlag. find
15. Scott, l). R. & Souza, C. S. [19911] Getting the Message Across in nsT-based Text Generation. In It. I)Me, C. Mellish and M. Zock reds.) Current Research in Nat-oral Langua~le Generation. London: Academic Press. find
16. Sibun, P. [1992] Generating Text without Trees. To appear in Computational Intelligence: Speciol Issue on Natural Language Generation, 8. find
17. I)ROC. OF COLING-92, N^NTES, AUo. 23-28, 1992 find

This page in Markdown → https://md.mova.elamur.ai/free/en/docs/papers/oberlander-1992-preventing-false-temporal-implicatures-interactive/