Fallacies play a prominent role in argumentation since antiquity due to their contribution to argumentation in critical thinking education. Their role is even more crucial nowadays as contemporary argumentation technologies face challenging tasks as misleading and manipulative information detection in news articles and political discourse, and counter-narrative generation. Despite some work in this direction, the issue of classifying arguments as being fallacious largely remains a challenging and an unsolved task. Our contribution is twofold: first, we present a novel annotated resource of 31 political debates from the U.S. Presidential Campaigns, where we annotated six main categories of fallacious arguments (i.e., ad hominem, appeal to authority, appeal to emotion, false cause, slogan, slippery slope) leading to 1628 annotated fallacious arguments; second, we tackle this novel task of fallacious argument classification and we define a neural architecture based on transformers outperforming state-of-the-art results and standard baselines. Our results show the important role played by argument components and relations in this task.

Political debates are one of the most salient moments of an election campaign, where candidates are challenged to discuss the main contemporary and historical issues in a country. These debates represent a natural ground for argumentative analysis, which has always been employed to investigate political discourse structure and strategy in philosophy and linguistics. In this paper, we present DISPUTool 2.0, an automated tool which relies on Argument Mining methods to analyse the political debates from the US presidential campaigns to extract argument components (i.e., premise and claim) and relations (i.e., support and attack), and highlight fallacious arguments. DISPUTool 2.0 allows also for the automatic analysis of a piece of a debate proposed by the user to identify and classify the arguments contained in the text. A REST API is provided to exploit the tool's functionalities.