Who Feels What and Why? Annotation of a Literature Corpus with Semantic Roles of Emotions

5402

Authors

nan

Abstract

AbstractMost approaches to emotion analysis in fictional texts focus on detecting the emotion expressed in text. We argue that this is a simplification which leads to an overgeneralized interpretation of the results, as it does not take into account who experiences an emotion and why. Emotions play a crucial role in the interaction between characters and the events they are involved in. Until today, no specific corpora that capture such an interaction were available for literature. We aim at filling this gap and present a publicly available corpus based on Project Gutenberg, REMAN (Relational EMotion ANnotation), manually annotated for spans which correspond to emotion trigger phrases and entities/events in the roles of experiencers, targets, and causes of the emotion. We provide baseline results for the automatic prediction of these relational structures and show that emotion lexicons are not able to encompass the high variability of emotion expressions and demonstrate that statistical models benefit from joint modeling of emotions with its roles in all subtasks. The corpus that we provide enables future research on the recognition of emotions and associated entities in text. It supports qualitative literary studies and digital humanities. The corpus is available at http://www.ims.uni-stuttgart.de/data/reman .

Keywords

nan

Full Content

Proceedings of the 27th International Conference on Computational Linguistics, pages 1345–1359 Santa Fe, New Mexico, USA, August 20-26, 2018. 1345 Who Feels What and Why? Annotation of a Literature Corpus with Semantic Roles of Emotions Evgeny Kim and Roman Klinger Institut für Maschinelle Sprachverarbeitung University of Stuttgart, Pfaffenwaldring 5b, 70569 Stuttgart, Germany evgeny.kim@ims.uni-stuttgart.de roman.klinger@ims.uni-stuttgart.de Abstract Most approaches to emotion analysis in fictional texts focus on detecting the emotion expressed in text. We argue that this is a simplification which leads to an overgeneralized interpretation of the results, as it does not take into account who experiences an emotion and why. Emotions play a crucial role in the interaction between characters and the events they are involved in. Until today, no specific corpora that capture such an interaction were available for literature. We aim at filling this gap and present a publicly available corpus based on Project Gutenberg, REMAN (Relational EMotion ANnotation), manually annotated for spans which correspond to emotion trigger phrases and entities/events in the roles of experiencers, targets, and causes of the emotion. We provide baseline results for the automatic prediction of these relational structures and show that emotion lexicons are not able to encompass the high variability of emotion expressions and demonstrate that statistical models benefit from joint modeling of emotions with its roles in all subtasks. The corpus that we provide enables future research on the recognition of emotions and associated entities in text. It supports qualitative literary studies and digital humanities. The corpus is available at http://www.ims.uni-stuttgart.de/data/reman. Title and Abstract in German Wer fühlt was und warum? Annotation eines Literaturkorpus mit Semantischen Rollen von Emotionen Die meisten Ansätze in der Emotionsanalyse in Literatur beschränken sich auf die Erkennung der Emotion. Wir nehmen in dieser Arbeit an, dass dies eine starke Vereinfachung darstellt. Es wird ignoriert, welche Figur die Emotion empfindet und wodurch sie ausgelöst wurde. Dies ist ungünstig, da Emotionen eine entscheidende Rolle bei der Interaktion zwischen Figuren und mit Ereignissen spielen. Allerdings war bisher kein annotiertes Korpus verfügbar, welches all diese Komponenten erfasst. In diesem Aufsatz präsentieren wir das Korpus REMAN (Relational EMotion ANotation), welches diese Lücke füllt. Es basiert auf Ausschnitten von Texten aus dem Projekt Gutenberg, welche auf Phrasenebene mit Emotionen sowie dem Empfindenden, dem Ziel sowie der Ursache der Emotion annotiert sind. Wir präsentieren eine Analyse des Korpus und stellen erste Ergebnisse eines automatischen Vorhersagemodells vor, welches die Grenzen von Wörterbuch-Verfahren aufzeigt. Des Weiteren zeigen wir, dass statistische Modelle von einer gemeinsamen Modellierung der verschiedenen Teilaufgaben profitieren. Unser Korpus unterstützt die Literaturwissenschaften sowie digitalen Geisteswissenschaften und ermöglicht die Erstellung von Modellen zur feingranularen automatischen Vorhersage von Emotionen. Das Korpus ist verfügbar unter http://www.ims.uni-stuttgart.de/data/reman. This work is licenced under a Creative Commons Attribution 4.0 International Licence. Licence details: http:// creativecommons.org/licenses/by/4.0/ 1346 1 Introduction Emotions are one of the crucial aspects of compelling narratives (Oatley, 2002; Ingermanson and Economy, 2009; Hogan, 2015). Not only do emotions help readers in literature comprehension (Barton, 1996; Robinson, 2005) but they also improve readers’ abilities of empathy and …

Content truncated to 500 words

Labeled Text

AbstractMost approaches to emotion analysis in fictional texts focus on detecting the emotion expressed in text. We argue that this is a simplification which leads to an overgeneralized interpretation of the results, as it does not take into account who experiences an emotion and why. Emotions play a crucial role in the interaction between characters and the events they are involved in. Until today, no specific corpora that capture such an interaction were available for literature. We aim at filling this gap and present a publicly available corpus based on Project Gutenberg, REMAN (Relational EMotion ANnotation), manually annotated for spans which correspond to emotion trigger phrases and entities/events in the roles of experiencers, targets, and causes of the emotion. We provide baseline results for the automatic prediction of these relational structures and show that emotion lexicons are not able to encompass the high variability of emotion expressions and demonstrate that statistical models benefit from joint modeling of emotions with its roles in all subtasks. The corpus that we provide enables future research on the recognition of emotions and associated entities in text. It supports qualitative literary studies and digital humanities. The corpus is available at .

Publication Details

Paper ID:
5402
Year:
2018
Publication:
TIPSTER
Publication Type:
conf
DOI:
nan