Elamur Research⌕
☰
home / papers / 2018 / Linguistic Features of Sarcasm and Metaphor Production Quality

Linguistic Features of Sarcasm and Metaphor Production Quality

Paper · 2018 · Stephen Skalicky and Scott Crossley

Summary

Most computational work on figurative language asks whether a text is figurative. This paper asks a different question: can surface linguistic measures explain how good a metaphor or a sarcastic remark is? The authors built their own corpus. Sixty-one university participants wrote metaphors in response to topic–description prompts and two open scenarios. They also wrote sarcastic lines for blank speech bubbles in twelve cartoons. Two trained raters scored each response on a 1–6 analytic rubric, with a third person adjudicating disagreements. For metaphors the subscales were conceptual distance, novelty and mirth; for sarcasm they were incongruity, novelty and mirth. TAALES supplied lexical sophistication indices and TAACO supplied semantic cohesion indices. Mixed-effects models then related these indices to the ratings. The features explained only a small but significant share of variance. They did better on the creativity subscales than on the theory-specific ones, and better for metaphor than for sarcasm. The authors read this as evidence that sarcasm quality depends heavily on situational context that the words alone do not carry.

Key points

  • Participants: 61 people (46 female, 15 male), aged 17–63.
  • Metaphor prompts: 22 conventional items, each a Topic plus a Description whose words could not be reused, and 2 novel scenarios (a boring class, a disgusting food or drink).
  • Sarcasm prompts: 12 comics in three types, four of each:
    • Rosenzweig Picture Frustration cartoons, taken via Huang et al.
    • Adapted Bizarro! desert-island strips.
    • Original three-panel comics.
  • Corpus and agreement: after removing inattentive answers, 1304 metaphors and 716 sarcastic responses remained. Adjudicated kappa ranged from .777 (sarcasm mirth) to .872 (metaphor conceptual distance).
  • Main findings:
    • Greater metaphor conceptual distance went with more specific, sophisticated vocabulary.
    • Higher sarcasm incongruity went with words that have fewer associations.
    • For sarcasm novelty/mirth, MRC Familiarity and Brysbaert Concreteness interacted with prompt type, separating desert-island from three-panel prompts. These features explained about 6.8% of variance.

Interesting ideas

  • I1. Quality is split into a construct-specific dimension (distance or incongruity) and general creativity dimensions (novelty, mirth), which shows which parts surface features can actually reach.
  • I2. The same lexical feature had opposite usefulness depending on how much context the prompt supplied: sparse scenes favoured concrete, familiar words, while richer comics let context fill gaps.
  • I3. The authors argue that a bare "thank you" can be sarcastic purely through its situation, so word-level measures will miss it.
  • I4. The corpus was built by elicitation, with a defined task and controlled prompts, instead of being scraped. This gave guaranteed intent labels plus extra metadata such as production time.
  • I5. The authors warn that averaging per-word indices over very short answers can bias the measures.

For Mova

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

  • I1 (frames): The irony module could report "is it ironic" separately from "how strong is the clash with the situation," keeping the two signals explainable on their own.
  • I2 (frames): Irony detection could weight lexical cues by how much situational context the world model holds, leaning on world-state contradiction when context is rich.
  • I3 (world): Mova could flag sarcasm candidates by checking a positive speech act against the tracked state of the text world, such as a frustrating event that just happened, instead of relying on wording.
  • I4 (learn): Mova could use controlled prompts to generate irony or metaphor test sets with known intent, then use them as positive cases against literal negative controls.
  • I5 (learn): Evaluation reports should stratify scores by sentence length, so that short-utterance artefacts are not mistaken for real gains.

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

 Institutions

🎓Georgia State UniversityAtlanta · United States3 works · 252 citations

📚 Shelves 6The field of science this work belongs to (OpenAlex fields).Computer Science · 9,086 worksArtificial Intelligence · 8,889 worksArts and Humanities · 9,296 worksLanguage and Linguistics · 9,206 worksPsychology · 798 worksExperimental and Cognitive Psychology · 463 works

🧭 Topics 2Research subjects and languages the work is about.Sentiment & emotion · 608 worksIdioms & multiword expressions · 155 works

Cite

Cite as Stephen Skalicky and Scott Crossley. 2018. Linguistic Features of Sarcasm and Metaphor Production Quality. Proceedings of the Workshop on Figurative Language Processing.

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/w18-0902

@inproceedings{skalicky2018linguistic,
  title = {{Linguistic Features of Sarcasm and Metaphor Production Quality}},
  author = {Stephen Skalicky and Scott Crossley},
  booktitle = {Proceedings of the Workshop on Figurative Language Processing},
  year = {2018},
  doi = {10.18653/v1/w18-0902},
  url = {https://doi.org/10.18653/v1/w18-0902}
}

References

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

  1. Acar, Selcuk and Mark A. Runco. 2014. Assessing associative distance among ideas elicited by tests of divergent thinking. Creativity Research Journal 26(2). 229–238. ↗
  2. Khodak, Mikhail, Nikunj Saunshi, and Kiran Vodrahalli. 2017. A large self-annotated corpus for sarcasm. arXiv preprint arXiv:1704.05579. https://arxiv.org/abs/1704.05579 (15 June, 2017). ↗
  3. Beaty, Roger E. and Paul J. Silvia. 2013. Metaphorically speaking: cognitive abilities and the production of figurative language. Memory & Cognition 41(2). 255–267. ↗
  4. Kintsch, Walter. 2008. How the mind computes the meaning of metaphor. In Raymond W. Gibbs Jr (ed.), The Cambridge handbook of metaphor and thought, 129–142. Cambridge, MA: Cambridge University Press. ↗
  5. Brysbaert, Marc, Amy Beth Warriner, and Victor Kuperman. 2014. Concreteness ratings for 40 thousand generally known English word lemmas. Behavior Research Methods 46(3). 904–911. ↗
  6. Kintsch, Walter and Anita R. Bowles. 2002. Metaphor comprehension: What makes a metaphor difficult to understand? Metaphor and Symbol 17(4). 249– 262. ↗
  7. Chiappe, Dan L. and Penny Chiappe. 2007. The role of working memory in metaphor production and comprehension. Journal of Memory and Language 56(2). 172–188. ↗
  8. Kuperman, Victor, Hans Stadthagen-Gonzalez, and Marc Brysbaert. 2012. Age-of-acquisition ratings for 30,000 English words. Behavior Research Methods 44(4). 978–990. ↗
  9. Colston, Herbert L. 2017. Irony and sarcasm. In Salvatore Attardo (ed.), The Routledge handbook of language and humor, 234–249. New York, NY: Routledge. ↗
  10. Kyle, Kristopher and Scott A. Crossley. 2015. Automatically assessing lexical sophistication: Indices, tools, findings, and application. TESOL Quarterly 49(4). 757–786. ↗
  11. Coltheart, Max. 1981. The MRC psycholinguistic database. The Quarterly Journal of Experimental Psychology 33(4). 497–505. ↗
  12. Kyle, Kristopher, Scott Crossley, and Cynthia Berger. 2017. The tool for the automatic analysis of lexical sophistication (TAALES): version 2.0. Behavior Research Methods. ↗
  13. Crossley, Scott A., Kristopher Kyle, and Danielle S. McNamara. 2016. The tool for the automatic analysis of text cohesion (TAACO): Automatic assessment of local, global, and text cohesion. Behavior Research Methods 48. 1227–1237. ↗
  14. Martin, Rod A. 2007. The psychology of humor: An integrative approach. Burlington, MA: Elsevier Academic Press. find
  15. Davies, Mark. 2008. The corpus of contemporary American English. BYE, Brigham Young University. find
  16. Rosenzweig, Saul. 1945. The picture‐association method and its application in a study of reactions to frustration. Journal of personality 14(1). 3–23. ↗
  17. Dumas, Denis and Kevin N. Dunbar. 2014. Understanding fluency and originality: A latent variable perspective. Thinking Skills and Creativity 14. 56– 67. ↗
  18. Runco, Mark A. and Garrett J. Jaeger. 2012. The standard definition of creativity. Creativity Research Journal 24(1). 92–96. ↗
  19. Gerrig, Richard J. and Raymond W. Gibbs. 1988. Beyond the lexicon: Creativity in language production. Metaphor and Symbol 3(3). 1–19. ↗
  20. Salsbury, Tom, Scott A. Crossley, and Danielle S. McNamara. 2011. Psycholinguistic word information in second language oral discourse. Second Language Research 27(3). 343–360. ↗
  21. Gibbs, Raymond W. 1994. The poetics of mind: Figurative thought, language, and understanding. Cambridge, MA: Cambridge University Press. find
  22. Skalicky, Stephen and Scott Crossley. 2015. A statistical analysis of satirical Amazon.com product reviews. The European Journal of Humour Research 2(3). 66–85. ↗
  23. Glucksberg, Sam. 2001. Understanding figurative language: From metaphor to idioms. Oxford University Press. find
  24. Skalicky, Stephen, Scott A. Crossley, Danielle S. McNamara, and Kasia Muldner. 2017. Identifying creativity during problem solving using linguistic features. Creativity Research Journal 29(4). 343– 353. ↗
  25. Hoffman, Paul, Matthew A. Lambon Ralph, and Timothy T. Rogers. 2013. Semantic diversity: A measure of semantic ambiguity based on variability in the contextual usage of words. Behavior Research Methods 45(3). 718–730. ↗
  26. Huang, Li, Francesca Gino, and Adam D. Galinsky. 2015. The highest form of intelligence: Sarcasm find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/skalicky-2018-linguistic-features-sarcasm-metaphor-production/