Elamur Research⌕
☰
home / papers / 2013 / Character-to-Character Sentiment Analysis in Shakespeare's Plays

Character-to-Character Sentiment Analysis in Shakespeare's Plays

Paper · 2013 · Eric T. Nalisnick and Henry S. Baird

Summary

The paper asks whether a computer can follow how characters in a narrative feel about one another, not just the mood of a whole text or of one character. Nalisnick and Baird use the explicit dialogue structure of plays, working from Jon Bosak's XML editions of Shakespeare. Their method is deliberately shallow and has no domain tuning. Each uninterrupted speech gets a score equal to the sum of the AFINN lexicon valences of its words. That score is then credited to the relationship between the speaker and whoever spoke just before. Adding these scores gives static rankings of each character's allies and enemies. Tracking the running total across line numbers gives dynamic curves for any pair of characters. For Hamlet, Othello, Romeo and Juliet, and The Taming of the Shrew, the curves mostly agree with standard readings of the plays. In one case they drew attention to a turning-point scene. The evaluation is qualitative only. Its value lies in showing that a crude lexicon combined with a structural guess about the addressee already yields interpretable relationship dynamics.

Key points

  • Method: Valence comes from the AFINN lexicon (Nielsen, 2011) and is summed over each continuous speech. The result is credited to the preceding speaker, who is treated as the addressee.
  • Hamlet's final rankings: Claudius scores lowest at −27, while Gertrude scores a surprisingly high +24.
  • Act III Scene IV ("The Queen's Closet", near line 2,250): In this scene, Hamlet's running score toward Gertrude rises from −1 to 22. In the same scene, Gertrude's score toward Hamlet falls from 1 to −19.
  • Other relationships analysed:
    • Othello→Desdemona peaks early and then collapses.
    • Romeo and Juliet are strongly positive toward each other.
    • Petruchio is negative toward Katharina during roughly lines 1200–1700, his "taming" phase.
    • An interactive viewer called ShakespeareExplorer covers every pair of characters in every play.
  • Known failures and future work:
    • Sentiment about characters who are not on stage gets misattributed to whoever spoke last.
    • Iago's flattery hides his hatred of Othello.
    • The authors propose detecting references to offstage characters and handling irony and deceit.
    • For novels, they propose first running Elson and McKeown's speaker attribution.

Interesting ideas

  • I1. Credit a speech's sentiment to a directed speaker→addressee pair rather than to the text as a whole, turning a polarity score into an edge in a relationship graph.
  • I2. Use a cheap structural default for the addressee (the previous speaker) and record where that default fails, such as offstage targets, as a separate repair problem.
  • I3. Follow a cumulative score over the course of the text and treat sudden jumps as markers of pivotal scenes.
  • I4. Name the cases a shallow signal cannot see, such as deceptive flattery, as a known weakness instead of hiding them.
  • I5. Judge the output against well-known reader interpretations of famous works when no gold labels exist.

For Mova

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

  • I1 (frames): Speech-act frames in prag could carry an explicit addressee slot filled from turn order, so attitude becomes a who-feels-what-toward-whom relation.
  • I2 (frames): Coreference could take "previous turn-holder" as a weak prior for second-person reference, overridden when a mention points to someone absent.
  • I3 (world): Timestamped attitude states on entity pairs would let the world model flag events where a relation flips, and those events can serve as explainable hooks for questions about why something changed.
  • I4 (learn): Iago-style deceptive praise makes a good negative control, where surface polarity and true attitude diverge, for testing whether the irony and implicature layers add value.
  • I5 (answer): Fables and tales with canonical moral readings, such as who is the villain, could serve as a sanity-check set for relationship questions answered from the world graph.

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

🎓Lehigh UniversityBethlehem · United States10 works · 1,573 citations

 Fields, topics and concepts

📚 Shelves 5The field of science this work belongs to (OpenAlex fields).Computer Science · 18,614 worksArtificial Intelligence · 17,965 worksArts and Humanities · 17,450 worksLanguage and Linguistics · 17,296 worksLiterature and Literary Theory · 541 works

🧭 Topics 1Research subjects and languages the work is about.Sentiment & emotion · 1,035 works

Cite

Cite as Eric T. Nalisnick and Henry S. Baird. 2013. Character-to-Character Sentiment Analysis in Shakespeare's Plays. Proceedings of the 51st Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers).

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

@inproceedings{nalisnick2013character,
  title = {{Character-to-Character Sentiment Analysis in Shakespeare's Plays}},
  author = {Eric T. Nalisnick and Henry S. Baird},
  booktitle = {Proceedings of the 51st Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers)},
  year = {2013},
  url = {https://aclanthology.org/P13-2085/}
}

References

19 references, 3 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. A. Agarwal, A. Corvalan, J. Jensen, and O. Rambow. 2012. Social network analysis of alice in wonderland. NAACL-HLT 2012, page 88. find
  2. Cecilia Ovesdotter Alm and Richard Sproat. 2005. Emotional sequencing and development in fairy tales. In Affective Computing and Intelligent Interaction, pages 668–674. Springer. find
  3. Alina Andreevskaia and Sabine Bergler. 2007. Clac and clac-nb: knowledge-based and corpus-based approaches to sentiment tagging. In Proceedings of the 4th International Workshop on Semantic Evaluations, SemEval ’07, pages 117–120, Stroudsburg, PA, USA. Association for Computational Linguistics. find
  4. Nathanael Chambers and Dan Jurafsky. 2009. Unsupervised learning of narrative schemas and their participants. In Proceedings of the Joint Conference of the 47th Annual Meeting of the ACL and the 4th International Joint Conference on Natural Language Processing of the AFNLP: Volume 2-Volume 2, pages 602–610. Association for Computational Linguistics. find
  5. Micha Elsner. 2012. Character-based kernels for novelistic plot structure. In Proceedings of the 13th Conference of the European Chapter of the Association for Computational Linguistics, EACL ’12, pages 634–644, Stroudsburg, PA, USA. Association for Computational Linguistics. find
  6. David K Elson and Kathleen R McKeown. 2009. Extending and evaluating a platform for story understanding. In Proceedings of the AAAI 2009 Spring Symposium on Intelligent Narrative Technologies II. ↗
  7. D.K. Elson and K.R. McKeown. 2010. Automatic attribution of quoted speech in literary narrative. In Proceedings of AAAI. find
  8. D.K. Elson, N. Dames, and K.R. McKeown. 2010. Extracting social networks from literary fiction. In Proceedings of the 48th Annual Meeting of the Association for Computational Linguistics, pages 138–147. Association for Computational Linguistics. ↗
  9. Michael Gamon. 2004. Sentiment classification on customer feedback data: noisy data, large feature vectors, and the role of linguistic analysis. In Proceedings of the 20th international conference on Computational Linguistics, COLING ’04, Stroudsburg, PA, USA. Association for Computational Linguistics. find
  10. Soo-Min Kim and Eduard Hovy. 2004. Determining the sentiment of opinions. In Proceedings of the 20th international conference on Computational Linguistics, COLING ’04, Stroudsburg, PA, USA. Association for Computational Linguistics.
  11. Saif M Mohammad and Peter D Turney. 2008. Crowdsourcing the creation of a word–emotion association lexicon. find
  12. S. Mohammad. 2011. From once upon a time to happily ever after: Tracking emotions in novels and fairy tales. In Proceedings of the 5th ACL-HLT Workshop on Language Technology for Cultural Heritage, Social Sciences, and Humanities, pages 105–114. Association for Computational Linguistics. find
  13. P. Mutton. 2004. Inferring and visualizing social networks on internet relay chat. In Information Visualisation, 2004. IV 2004. Proceedings. Eighth International Conference on, pages 35–43. IEEE. ↗
  14. F. Å. Nielsen. 2011. Afinn, March. find
  15. Bo Pang and Lillian Lee. 2008. Opinion mining and sentiment analysis. Foundations and trends in information retrieval, 2(1-2):1–135. ↗
  16. Bo Pang, Lillian Lee, and Shivakumar Vaithyanathan. 2002. Thumbs up?: sentiment classification using machine learning techniques. In Proceedings of the ACL-02 conference on Empirical methods in natural language processing - Volume 10, EMNLP ’02, pages 79–86, Stroudsburg, PA, USA. Association for Computational Linguistics. ↗
  17. Stephen Ramsay. 2011. Reading Machines: Toward an Algorithmic Criticism. University of Illinois Press. find
  18. Antonio Roque. 2012. Towards a computational approach to literary text analysis. NAACL-HLT 2012, page 97. find
  19. C. Strapparava and A. Valitutti. 2004. Wordnet-affect: an affective extension of wordnet. In Proceedings of LREC, volume 4, pages 1083–1086.

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/nalisnick-2013-character-character-sentiment-analysis-shakespeare/