Argumentative Annotation of the Scientific Internet-Communication Corpus: Genre Analysis and Study of Typical Reasoning Models based on the ArgNetBank Studio Platform

The article presents a corpus of texts with argumentative annotation pertaining to different types of scientific Internet communication. The corpus is annotated using the ArgNetBank Studio platform. The genres and functional characteristics of scientific Internet communication texts are discussed. The purpose of creating the corpus and the principles of selecting the data to be included are formulated, the types of selected texts are defined. The peculiarities of argumentative annotation of scientific and popular science articles with commentaries, which, unlike other types of texts included in the corpus, are internet-dialogues, are considered. This determines the specificity of their argumentative structure. We also discuss the issue of identifying argumentative schemes that arises within the process of annotating the texts of different types. A trial experiment has shown that for a number of pairs of schemes the choice is difficult. We propose taking these difficulties into account while developing a methodology of annotation and defining distinguishing semantic features of argumentative schemes; this approach is illustrated by analyzing the schemes from one of the considered pairs. In the course of annotating the analytical texts, the necessity of adding new reasoning schemes and supplementing some of the existing ones has arisen. The paper provides formalized descriptions of such schemes and illustrative examples from the analyzed corpus texts.