Harnessing Cross-lingual Features to Improve Cognate Detection for Low-resource Languages

Proceedings of the 28th International Conference on Computational Linguistics, pages 125–138 Barcelona, Spain (Online), December 8-13, 2020 125 Appraisal Theories for Emotion Classification in Text Jan Hofmann1, Enrica Troiano1, Kai Sassenberg2,3, and Roman Klinger1 1Institut f¨ur Maschinelle Sprachverarbeitung, University of Stuttgart, Germany 2Leibniz-Institut f¨ur Wissensmedien, T¨ubingen, Germany 3University of T¨ubingen, Germany {jan.hofmann,enrica.troiano,roman.klinger}@ims.uni-stuttgart.de k.sassenberg@iwm-tuebingen.de Abstract Automatic emotion categorization has been predominantly formulated as text classification in which textual units are assigned to an emotion from a predefined inventory, for instance following the fundamental emotion classes proposed by Paul Ekman (fear, joy, anger, disgust, sadness, surprise) or Robert Plutchik (adding trust, anticipation). This approach ignores existing psycho- logical theories to some degree, which provide explanations regarding the perception of events. For instance, the description that somebody discovers a snake is associated with fear, based on the appraisal as being an unpleasant and non-controllable situation. This emotion reconstruction is even possible without having access to explicit reports of a subjective feeling (for instance expressing this with the words “I am afraid.”). Automatic classification approaches therefore need to learn properties of events as latent variables (for instance that the uncertainty and the mental or physical effort associated with the encounter of a snake leads to fear). With this paper, we propose to make such interpretations of events explicit, following theories of cognitive appraisal of events, and show their potential for emotion classification when being encoded in classification models. Our results show that high quality appraisal dimension assignments in event descriptions lead to an improvement in the classification of discrete emotion categories. We make our corpus of appraisal-annotated emotion-associated event descriptions publicly available. 1 Introduction The task of emotion analysis is commonly formulated as classification or regression in which textual units (documents, paragraphs, sentences, words) are mapped to a predefined reference system, for instance the sets of fundamental emotions fear, anger, joy, surprise, disgust, and sadness proposed by Ekman (1999), or by Plutchik (2001), which includes also trust and anticipation. Machine learning-based models need to figure out which words point to a particular emotion experienced by a reader, by the author of a text, or a character in it. Depending on the resource which has been annotated, the description of an emotion experience can vary. On Twitter, for instance, other than direct reports of an emotion state (“I feel depressed”), hashtags are used as emotion labels to enrich the description of events and stances (“I just got my exam result #sad”). In news articles, emotional events are sometimes explicitly mentioned (“couple infuriate officials” (Bostan et al., 2020)) and other times require world knowledge (“Tom Cruise and Katie Holmes set wedding date”, labeled as surprise (Strapparava and Mihalcea, 2007)). In literature, a sequence of events which forms the narrative leads to an emotion in the reader. In this paper, we focus on those texts which communicate emotions without an explicit emotion word, but rather describe events for which an emotion association is evident. Such textual examples became popular in natural language processing research with the use …

Content truncated to 500 words