Modelling the interplay of metaphor and emotion through multitask learning

Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing, pages 2218–2229, Hong Kong, China, November 3–7, 2019. c⃝2019 Association for Computational Linguistics 2218 Modelling the interplay of metaphor and emotion through multitask learning Verna Dankers1 Marek Rei2,3 Martha Lewis1 Ekaterina Shutova1 1Institute for Logic, Language and Computation, University of Amsterdam 2The ALTA Institute, Computer Laboratory, University of Cambridge 3Department of Computing, Imperial College London vernadankers@gmail.com, marek.rei@cl.cam.ac.uk, {m.a.f.lewis,e.shutova}@uva.nl Abstract Metaphors allow us to convey emotion by connecting physical experiences and abstract concepts. The results of previous research in linguistics and psychology suggest that metaphorical phrases tend to be more emotion- ally evocative than their literal counterparts. In this paper, we investigate the relationship be- tween metaphor and emotion within a com- putational framework, by proposing the first joint model of these phenomena. We experi- ment with several multitask learning architec- tures for this purpose, involving both hard and soft parameter sharing. Our results demon- strate that metaphor identification and emotion prediction mutually benefit from joint learning and our models advance the state of the art in both of these tasks. 1 Introduction Metaphors allow us to reason about abstract con- cepts by linking them to our physical experiences (Lakoff and Johnson, 1983). Metaphorical lan- guage arises through systematic association be- tween two distinct semantic domains — the source and the target — as illustrated by the sentence “The news leaked out despite the secrecy”, where a term from the source domain of liquids is used to describe information (the target domain). This metaphorical association widely manifests itself in language, e.g. we can similarly talk about “being engulfed by a stream of bad news”. Metaphorical associations allow us to project knowledge from the source domain to the target, inviting new rea- soning frameworks and connotations to emerge. Much previous research on metaphorical lan- guage in fields such as linguistics (Blanchette et al., 2001; Kövecses, 2003), cognitive psychol- ogy (Crawford, 2009; Thibodeau and Boroditsky, 2011) and neuroscience (Aziz-Zadeh and Dama- sio, 2008; Jabbi et al., 2008) points to its prevalent affective content. Linguistic expressions describ- ing one’s emotional state have a relatively high incidence of figurative language and metaphor in particular (Fainsilber and Ortony, 1987; Fussell and Moss, 1998; Gibbs Jr et al., 2002), as illus- trated by the phrase “My mind was seething and boiling”. On the other hand, a stronger emo- tion appears to be conveyed through the associa- tion of source and target domains more generally. Mohammad et al. (2016) found that metaphori- cal phrases are consistently perceived as carrying more emotion than their literal paraphrases and the literal uses of the same source domain words. For instance, “leaking information” conveys an im- plicit judgement, as compared to the more neu- tral paraphrase “disclosing information”. Their re- sults also suggest that the emotional content of the metaphor is not due to the properties of individual source and target domains, but rather arises com- positionally through their interaction. …

Content truncated to 500 words