StoryARG: a corpus of narratives and personal experiences in argumentative texts

Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics Volume 1: Long Papers, pages 2350–2372 July 9-14, 2023 ©2023 Association for Computational Linguistics StoryARG: a corpus of narratives and personal experiences in argumentative texts Neele Falk and Gabriella Lapesa Institute for Natural Language Processing, University of Stuttgart {neele.falk,gabriella.lapesa}@ims.uni-stuttgart.de Abstract Humans are storytellers, even in communica- tion scenarios which are assumed to be more rationality-oriented, such as argumentation. In- deed, supporting arguments with narratives or personal experiences (henceforth, stories) is a very natural thing to do – and yet, this phe- nomenon is largely unexplored in computa- tional argumentation. Which role do stories play in an argument? Do they make the ar- gument more effective? What are their narra- tive properties? To address these questions, we collected and annotated StoryARG, a dataset sampled from well-established corpora in com- putational argumentation (ChangeMyView and RegulationRoom), and the Social Sciences (Eu- ropolis), as well as comments to New York Times articles. StoryARG contains 2451 tex- tual spans annotated at two levels. At the ar- gumentative level, we annotate the function of the story (e.g., clarification, disclosure of harm, search for a solution, establishing speaker’s au- thority), as well as its impact on the effective- ness of the argument and its emotional load. At the level of narrative properties, we annotate whether the story has a plot-like development, is factual or hypothetical, and who the protago- nist is. What makes a story effective in an argument? Our analysis of the annotations in StoryARG uncover a positive impact on effectiveness for stories which illustrate a solution to a problem, and in general, annotator-specific preferences that we investigate with regression analysis. 1 Introduction Narratives and argumentation are deeply related: this is a well established observation in psychol- ogy and social science. Although stories per se express something individual and concrete, they allow people to draw conclusions about matters of general interest, for example, social problems and injustices - something general is expressed through something concrete and can thus often be better understood (Fisher, 1985). In addition, stories have a unique effect on the recipient(s) (e.g., the other participants in a discussion): they offer room for interpretation, therefore encourage reflection, and precisely because they are individual, the recipi- ent is required to take on the perspective of the other (Polletta and Lee, 2006; Hoeken and Fikkers, 2014), a quality that is becoming increasingly im- portant in times of growing political polarization. On the side of computational argumentation re- search, however, the role of narratives and personal experiences has barely been investigated, since in argumentative contexts they are often regarded as rather second-class (not logical, not verifiable). With our paper, the resource it presents, and the analysis we carry out, we aim at building a fine- grained empirical picture of this phenomenon, cru- cial both in terms of its persuasiveness within an argument and its contribution to interpersonal com- munication. While there are existing datasets that make it possible to develop classification methods to de- tect …

Content truncated to 500 words