Computational Approaches to Fallacy Detection in Natural Language Arguments

One way to improve the quality of an argument and the reasoning employed is to look at the ways in which it has fallen short or erred. Fallacy theory provides us with a set of guidelines to appraise an argument and discover the most common circumstances in which these errors occur.
The goal need not be, primarily, to criticise an argument and call it a day, but rather, to develop a starting point from which a sound argument can be developed. Additionally, an understanding of fallacies provides the audience of an argument with a set of tools to prevent them from being misled and to critically appraise the arguments presented.
This paper presents the goal of our research, namely, to employ fallacy theory in a computational context. More specifically, the work aims to: develop artificial intelligence models to automatically identify fallacies; and develop a computationally focused fallacy compendium to promote further work and collaboration in this field.