Analysis of the Persuasiveness of Argumentation in Popular Science Texts

Abstract

The paper discusses the methods of modeling and assessing the quality of the argumentation used in popular science texts, as well as the software supporting them. First, the authors study the aspects of argumentation persuasiveness, i.e. the validity of conclusions presented in articles. Argumentation modeling is performed using the argumentation ontology based on the AIF format (Argument Interchange Format), which was adopted by the international community as a standard notation for describing arguments and argumentation schemes. The authors have supplemented this ontology with the facilities necessary for modeling and analyzing the quality of argumentation in popular science discourse. In particular, we have introduced into the ontology facilities allowing us to assign the estimates of persuasiveness (degree of the truth) to the arguments and statements and to model the target audience. Thanks to these facilities, it has become possible to analyze the persuasiveness of the argumentation regarding different target audiences. To solve this problem, the authors propose a model and an algorithm for calculating the persuasiveness of arguments allowing taking into account conflicts between the arguments. The paper also provides an example of constructing a network of arguments and calculating the degree of their persuasiveness using the software system developed.