Elamur Research⌕
☰
home / papers / 2019 / An Incremental Iterated Response Model of Pragmatics

An Incremental Iterated Response Model of Pragmatics

Summary

Rational Speech Acts (RSA) and other Iterated Response (IR) models treat pragmatics as speakers and listeners recursively reasoning about one another. They usually reason over whole utterances, yet psycholinguistic evidence shows that people draw pragmatic inferences while a sentence is still unfolding. The authors build an incremental variant of RSA in which reasoning happens word by word, with a STOP token marking the end of an utterance. They derive an incremental semantics from an ordinary truth-conditional one by counting which full continuations of a partial string are true of each referent. A word-level speaker is then combined through the chain rule into an utterance-level speaker (S1 UTT-IP), which they contrast with the global speaker (S1 UTT-GP). Simulations show the two models make different predictions. The incremental model accounts for two findings the global one cannot: cross-linguistic word-order asymmetries in over-informative reference and anticipatory implicatures triggered by contrastive modifiers. On the TUNA corpus, it more often favors two-word descriptions under a length penalty. The approach also suggests a scalable route for RSA in NLP.

Key points

  • The model sits within the RSA paradigm. The incremental literal listener L0 WORD is derived from an utterance-level semantics, and greedy unrolling of S1 WORD lets the speaker generate utterances efficiently.
  • In a toy reference game (red dress, blue dress, red hat), the incremental listener who hears "red" already leans toward the red hat (probability 0.64).
  • Two empirical phenomena are captured only by the incremental model: the adjective–noun vs. noun–adjective asymmetry in over-informativity (Rubio-Fernández, 2016) and anticipatory contrastive inferences (Sedivy, 2007).
  • The TUNA experiment used the furniture and people single-referent domains and was limited to utterances of up to two words, which make up 32% of that data. Each word cost 1.0 and STOP cost 0.
  • In the TUNA results, S1 UTT-IP marked many more two-word utterances as optimal than S1 UTT-GP:
    • People domain (114 trials): 287 vs. 120.
    • Furniture domain (83 trials): 149 vs. 88.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

Fields: Linguistics & languages

Cited by

12 papers of our library cite it (of the 140673 whose bibliographies we read); 2 of them have a page here, 10 not yet.

Cite

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.7275/cprc-8x17

@inproceedings{cohngordon2019incremental,
  title = {{An Incremental Iterated Response Model of Pragmatics}},
  author = {Cohn-Gordon, Reuben and Goodman, Noah and Potts, Christopher},
  booktitle = {Proceedings of the Society for Computation in Linguistics (SCiL) 2019},
  year = {2019},
  doi = {10.7275/cprc-8x17},
  url = {https://doi.org/10.7275/cprc-8x17}
}

References

26 references, 5 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Noah D. Goodman and Andreas Stuhlmüller. 2013. Knowledge and implicature: Modeling language understanding as social cognition. Topics in Cognitive Science, 5(1):173–184. ↗
  2. Jacob Andreas and Dan Klein. 2016. Reasoning about pragmatics with neural listeners and speakers. In Proceedings of the 2016 Conference on Empirical Methods in Natural Language Processing, pages 1173–1182. Association for Computational Linguistics. find our card
  3. H. Paul Grice. 1975. Logic and conversation. In Peter Cole and Jerry Morgan, editors, Syntax and Semantics, volume 3: Speech Acts, pages 43–58. Academic Press, New York. ↗
  4. Leon Bergen, Roger Levy, and Noah D. Goodman. 2016. Pragmatic reasoning through semantic inference. Semantics and Pragmatics, 9(20). ↗
  5. Theo Herrmann and Werner Deutsch. 1976. Psychologie der objektbenennung. Huber. find
  6. Gerhard Jäger. 2007. Game dynamics connects semantics and pragmatics. In Ahti-Veikko Pietarinen, editor, Game Theory and Linguistic Meaning, pages 89–102. Elsevier, Amsterdam. find
  7. Reuben Cohn-Gordon, Noah Goodman, and Christopher Potts. 2018. Pragmatically informative image captioning with character-level inference. In Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 2 (Short Papers), pages 439–443. Association for Computational Linguistics. find our card
  8. Gerhard Jäger. 2012. Game theory in semantics and pragmatics. In Claudia Maienborn, Klaus von Heusinger, and Paul Portner, editors, Semantics: An International Handbook of Natural Language Meaning, volume 3, pages 2487–2425. Mouton de Gruyter, Berlin. find
  9. Kees van Deemter, Ielka van der Sluis, and Albert Gatt. 2006. Building a semantically transparent corpus for the generation of referring expressions. In Proceedings of the Fourth International Natural Language Generation Conference, pages 130–132. Association for Computational Linguistics. find
  10. Justine T. Kao, Jean Y. Wu, Leon Bergen, and Noah D. Goodman. 2014. Nonliteral understanding of number words. Proceedings of the National Academy of Sciences, 111(33):12002–12007. ↗ our card
  11. Paul E Engelhardt, Karl GD Bailey, and Fernanda Ferreira. 2006. Do speakers and listeners observe the gricean maxim of quantity? Journal of Memory and Language, 54(4):554–573. ↗
  12. Will Monroe, Robert X. D. Hawkins, Noah D. Goodman, and Christopher Potts. 2017. Colors in context: A pragmatic neural model for grounded language understanding. Transactions of the Association for Computational Linguistics, 5:325–338. ↗ our card
  13. Michael C. Frank, Andrés Gómez Emilsson, Benjamin Peloquin, Noah D. Goodman, and Christopher Potts. 2016. Rational speech act models of pragmatic reasoning in reference games. Ms., Stanford University. ↗
  14. Will Monroe and Christopher Potts. 2015. Learning in the Rational Speech Acts model. In Proceedings of 20th Amsterdam Colloquium, Amsterdam. ILLC. ↗
  15. Christopher Potts, Daniel Lassiter, Roger Levy, and Michael C. Frank. 2016. Embedded implicatures as pragmatic inferences under compositional lexical uncertainty. Journal of Semantics, 33(4):755–802. ↗
  16. Michael C. Frank and Noah D. Goodman. 2012. Predicting pragmatic reasoning in language games. Science, 336(6084):998. ↗ our card
  17. Paula Rubio-Fernández. 2016. How redundant are redundant color adjectives? an efficiency-based analysis of color overspecification. Frontiers in psychology, 7:153. ↗
  18. Michael Franke. 2009. Signal to Act: Game Theory in Pragmatics. ILLC Dissertation Series. Institute for Logic, Language and Computation, University of Amsterdam. find
  19. Julie C Sedivy. 2007. Implicature during real time conversation: A view from language processing research. Philosophy compass, 2(3):475–496. ↗
  20. Michael Franke and Gerhard Jäger. 2014. Pragmatic back-and-forth reasoning. In Salvatore Pis-toia Reda, editor, Pragmatics, Semantics and the Case of Scalar Implicatures, pages 170–200. Palgrave Macmillan UK, London. ↗
  21. Ramakrishna Vedantam, Samy Bengio, Kevin Murphy, Devi Parikh, and Gal Chechik. 2017. Context-aware captions from context-agnostic supervision. In Computer Vision and Pattern Recognition (CVPR), volume 3. find
  22. Albert Gatt, Anja Belz, and Eric Kow. 2009. find
  23. The TUNA-REG Challenge 2009: Overview and evaluation results. In Proceedings of the 12th European Workshop on Natural Language Generation (ENLG 2009), pages 174–182, Athens, Greece. Association for Computational Linguistics. find
  24. Adam Vogel, Max Bodoia, Christopher Potts, and Dan Jurafsky. 2013. Emergence of Gricean maxims from multi-agent decision theory. In Human Language Technologies: The 2013 Annual Conference of the North American Chapter of the Association for Computational Linguistics, pages 1072– 1081, Stroudsburg, PA. Association for Computational Linguistics. find
  25. Noah D Goodman and Michael C Frank. 2016. Pragmatic language interpretation as probabilistic inference. Trends in Cognitive Sciences, 20(11):818– 829. ↗
  26. Erica J. Yoon, Michael Henry Tessler, Noah D. Goodman, and Michael C. Frank. 2016. Talking with tact: Polite language as a balance between kindness and informativity. In Proceedings of the 38th Annual Conference of the Cognitive Science Society, pages 2771–2776. Cognitive Science Society. find

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/gordon-2019-incremental-iterated-response-model-pragmatics/