Abstract
Political debates offer citizens a unique opportunity to appreciate the position of political representatives on the most controversial issues of the day. In view of the active expression of the various players in political life, these debates constitute a source of information that needs to be capitalized on in order to gain a better understanding of societal dynamics. Given their innate argumentative quality, these exchanges constitute a suitable application scenario for the implementation of computational methods of argument extraction. Argument mining is a line of research studied in the field of natural language processing, the aim of which is to extract and automatically identify argumentative structures from a natural language text using computer programs. The analysis of argument structures is a complex task, involving the study of argument components and patterns, relations between arguments and counter-argumentation strategies. In this lecture, I will detail the steps required to automate the analysis of political discourse using argument mining methods. First , I will present the approaches dedicated to the identification of argumentative structures and their relationships. Next, I will describe the strategies deployed in the automatic identification of fallacious arguments, notably through the analysis of different forms of argumentation and the detection of strategic maneuvers in argumentative discourse.
