Elamur Research⌕
☰
home / papers / 2022 / Emoji semantics/pragmatics: investigating commitment and lying

Emoji semantics/pragmatics: investigating commitment and lying

Paper · 2022 · Benjamin Weissman

Summary

The paper asks how directly emoji carry speaker meaning. It uses two pragmatic diagnostics: whether a sender counts as committed to what a single emoji conveys, and whether a false, deceptive emoji counts as a lie. Free-response pretests measured how much people agree on each emoji's meaning. From these, the author picked two sets of 20. One set held concrete objects, foods and animals, all with high agreement. The other held facial-expression and gesture emoji with widely varying agreement. In Experiment 1, Prolific participants rated text-message exchanges where the reply was one emoji, judging commitment and deniability. Experiment 2 added context that made the reply true or false, with a motive to deceive, and compared emoji replies against word replies. Concrete emoji produced strong commitment and could count as lies. For expressive emoji, commitment and lie ratings rose with meaning agreement. Emoji overall were judged less of a lie than equivalent words, and the gap was larger for expressive ones. The conclusion is that an emoji's contribution to meaning depends on how conventionalized or lexicalized it is, not on emoji as a single category.

Key points

  • Pretests:
    • Non-expressive set: 80 emoji rated by 49 participants. The 20 chosen all had agreement above 0.75, with a mean of 0.88.
    • Expressive set: 75 emoji rated by 28 participants. The 20 chosen averaged 0.62 and ranged from 0.25 to 0.96.
  • Experiment 1 (100 participants):
    • On a 1–7 scale, commitment averaged 6.09 for non-expressive emoji and 5.17 for expressive ones (linear mixed effects model, F(1,99) = 114.2).
    • Within the expressive set, pretest agreement significantly predicted commitment.
  • Experiment 2 (202 recruited, 195 kept after filler-based exclusion):
    • Emoji replies were rated as lies less strongly than word replies.
    • The word–emoji gap was larger for expressive items, a reported type × presentation interaction.
  • Across experiments, expressive emoji with higher Experiment 1 commitment received higher Experiment 2 lie ratings (F(1,432) = 15.13).
  • Design details:
    • Each participant saw 4 counterbalanced lists of 20 items, half expressive and half non-expressive.
    • Experiment 1 mixed matching and mismatching target meanings.
    • Experiment 2 crossed true/false context with word/emoji presentation.

Interesting ideas

  • I1. Commitment can be treated as a graded quantity that tracks how conventional a form is, rather than a yes/no property of a whole category (emoji versus words).
  • I2. A false statement counts as a lie only when the speaker is committed to its content; otherwise it is merely misleading.
  • I3. Meaning agreement can be measured from free-text glosses: synonyms are binned together, near-neighbours like "happy" and "smiling" are kept apart, and the top bin's share becomes the score.
  • I4. Each experiment includes mismatching-meaning fillers that should score near the floor, plus a matched word-versus-emoji condition. These provide built-in sanity checks and a baseline.
  • I5. Two complementary questions, one direct ("did I commit?") and one inverted ("could I deny it?"), are merged into one score. This makes the judgement more stable.

For Mova

How these ideas could serve Mova, our small open language model — its files, stages and tests:

  • I1 (frames): prag could attach a commitment strength to each speech-act frame, based on the conventionality of the carrying form. Implicatures and ambiguous symbols would then receive weaker speaker commitment than literal, lexicalized wording.
  • I2 (frames): the deception/irony module could separate "lied" from "misled" by checking whether the false proposition was asserted directly or only implicated.
  • I3 (tokens): emoji and other symbol lexicon entries could store such an agreement score. Downstream stages could then tell how much interpretive weight a symbol can bear.
  • I4 (learn): pragmatic evaluations could pair every probe with a deliberately wrong reading and a plain-word control. This would catch a model that rates everything as committed.
  • I5 (answer): commitment-related QA checks could ask the reader both the positive and the negated question and require consistent answers.

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

📚 Shelves 4 Computer Science · 9,086 works Arts and Humanities · 9,296 works Language and Linguistics · 9,206 works Human-Computer Interaction · 61 works

Cite

Cite as Benjamin Weissman. 2022. Emoji semantics/pragmatics: investigating commitment and lying. Proceedings of the Fifth International Workshop on Emoji Understanding and Applications in Social Media.

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/2022.emoji-1.3

@inproceedings{weissman2022emoji,
  title = {{Emoji semantics/pragmatics: investigating commitment and lying}},
  author = {Benjamin Weissman},
  booktitle = {Proceedings of the Fifth International Workshop on Emoji Understanding and Applications in Social Media},
  year = {2022},
  doi = {10.18653/v1/2022.emoji-1.3},
  url = {https://doi.org/10.18653/v1/2022.emoji-1.3}
}

References

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

  1. Ariel, M. (2019). Different prominences for different inferences. Journal of Pragmatics, 154, 103–116. ↗
  2. Beißwenger, M., & Pappert, S. (2019). How to be polite with emoji: A pragmatic analysis of face work strategies in an online learning environment. European Journal of Applied Linguistics, 7(2), 225–253. find
  3. Bonalumi, F., Scott-Phillips, T., Tacha, J., & Heintz, C. (2020). Commitment and communication: Are we 25 5 committed to what we mean, or what we say? Language and Cognition, 12(2), 360–384. to make meaning in digital messages. New Media and Society, (September). find
  4. Boogaart, R., Jansen, H., & van Leeuwen, M. (2021). “Those are your words, not mine!” Defence strategies for denying speaker commitment. Argumentation, 35(2), 209–235. find
  5. Maier, E. (2021). Emoji as pictures. Under review. https://ling.auf.net/lingbuzz/006025. ↗
  6. Mazzarella, D., Reinecke, R., Noveck, I., & Mercier, H. (2018). Saying, presupposing and implicating: How pragmatics modulates commitment. Journal of Pragmatics, 133(June), 15–27. ↗
  7. Boulat, K., & Maillat, D. (2017). She said you said I saw it with my own eyes: a pragmatic account of commitment. In Formal models in the study of language (pp. 261–279). Springer. ↗
  8. McCulloch, G., & Gawne, L. (2018). Emoji grammar as beat gestures. CEUR Workshop Proceedings, 2130, 3–6. find
  9. Cohn, N., Engelen, J., & Schilperoord, J. (2019). The grammar of emoji? Constraints on communicative pictorial sequencing. Cognitive Research: Principles and Implications, 4(1). find
  10. Miller, H., Kluver, D., Thebault-Spieker, J., Terveen, L., & Hecht, B. (2017). Understanding Emoji Ambiguity in Context: The Role of Text in Emoji-Related Miscommunication. Proceedings of the Eleventh International AAAI Conference on Web and Social Media, 152–161. find
  11. Dainas, A. R., & Herring, S. C. (2021). Interpreting emoji pragmatics. Approaches to Internet Pragmatics, 107-144. ↗
  12. Doran, R., Ward, G., Larson, M., McNabb, Y., & Baker, R. E. (2012). A novel experimental paradigm for distinguishing between what is said and what is implicated. Language, 88(1), 124–154. ↗
  13. Moeschler, J. (2013). Is a speaker-based pragmatics possible? Or how can a hearer infer a speaker’s commitment? Journal of Pragmatics, 48(1), 84–97. find
  14. Pfeifer, V. A., Armstrong, E. L., & Lai, V. T. (2022). Do all facial emoji communicate emotion? The impact of facial emoji on perceived sender emotion and text processing. Computers in Human Behavior, 126, 107016. find
  15. Foltz, K. A., & Fray, J. (2020). Emoji goes to court: An analysis of emoji in court proceedings and implications for legal practice. Communication Law Review, 20, 109-125. find
  16. Garcia, C., Țurcan, A., Howman, H., & Filik, R. (2022). Emoji as a tool to aid the comprehension of written sarcasm: Evidence from younger and older adults. Computers in Human Behavior, 126(August 2021), 106971. ↗
  17. Reins, L. M., & Wiegmann, A. (2021). Is lying bound to commitment? Empirically investigating deceptive presuppositions, implicatures, and actions. Cognitive Science, 45(2). find
  18. Robertson, A., Liza, F. F., Nguyen, D., McGillivray, B., & Hale, S. A. (2021). Semantic journeys: quantifying change in emoji meaning from 2012-2018. Emoji2021. http://workshop-proceedings.icwsm.org/pdf/2021_07.pdf ↗
  19. Ge, J., & Herring, S. C. (2018). Communicative functions of emoji sequences on Sina Weibo. First Monday, 23(11). ↗
  20. Grosz, P., Kaiser, E., & Pierini, F. (2021). Discourse anaphoricity and first-person indexicality in emoji resolution. Proceedings of Sinn Und Bedeutung, 25. ↗
  21. Rodrigues, D., Prada, M., Gaspar, R., Garrido, M. V., & Lopes, D. (2018). Lisbon Emoji and Emoticon Database (LEED): Norms for emoji and emoticons in seven evaluative dimensions. Behavior Research Methods, 50(1), 392–405. ↗
  22. Herring, S., & Ge, J. (2020). Do Emoji Sequences Have a Preferred Word Order? 14th International AAAI Conference on Web and Social Media, (June), 1–5. find
  23. Saul, J. M. (2012). Lying, Misleading, and What is Said: An Exploration in Philosophy of Language and in Ethics. Oxford: Oxford University Press. find
  24. Holtgraves, T., & Robinson, C. (2020). Emoji can facilitate recognition of conveyed indirect meaning. PLoS ONE, 15(4), 1–13. ↗
  25. van Tiel, B., van Miltenburg, E., Zevakhina, N., & Geurts, B. (2016). Scalar diversity. Journal of Semantics, 33(1), 137–175. find
  26. Logi, L., & Zappavigna, M. (2021). A social semiotic perspective on emoji: How emoji and language interact Viebahn, E. (2019). Lying with pictures. British Journal of Aesthetics, 59(3), 243–257. find
  27. Vullioud, C., Clément, F., Scott-Phillips, T., & Mercier, H. (2017). Confidence as an expression of commitment: why misplaced expressions of 26 6 confidence backfire. Evolution and Human Behavior, 38(1), 9–17. find
  28. Was, C. A., & Hamrick, P. (2021). What Did They Mean by That? Young Adults’ Interpretations of 105 Common Emoji. Frontiers in Psychology, 12(May), 1– 7. find
  29. Weissman, B. (2019a). The role of linguistic meaning in the concept of lying. PhD Dissertation. University of Illinois at Urbana-Champaign. find
  30. Weissman, B. (2019b). Peaches and eggplants or. . . something else? The role of context in emoji interpretations. Proceedings of the Linguistic Society of America, 4(1), 29. find
  31. Weissman, B., & Tanner, D. (2018). A strong wink between verbal and emoji-based irony: How the brain processes ironic emoji during language comprehension. PLoS ONE, 13(8). find
  32. Weissman, B., & Terkourafi, M. (2019). Are false implicatures lies? An experimental investigation. Mind & Language, 34(2), 221–246. find
  33. Yang, X., & Liu, M. (2021). The pragmatics of text-emoji co-occurrences on Chinese social media. Pragmatics, 31(1), 144–172. Emoji Meaning Agreement 😡 🤪 😎 😮 🥳 🤐 😘 😇 😐 😁 😭 😫 🥱 😨 😰 😶 😬 😑 💁‍♀️ 🥴 angry silly cool shocked celebrate lips are sealed kiss angel neutral happy crying upset tired afraid stressed no words cringe annoyed sassy goofy .964 .964 .893 .857 .786 .786 .750 .750 .679 .643 .643 .571 .536 .500 .429 .429 .358 .286 .286 .250 *Table 1: expressive emoji used in both experiments* ↗

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/weissman-2022-emoji-semantics-pragmatics-investigating-commitment/