Smorgunov L.V. Network Characteristics of Topological Argumentation on Digital Platforms. Russian Foundation for Basic Research Journal. Humanities and social sciences. 2026. No. 1 (124). Pp. 97-109.

Abstract

The process of searching for justifications – reasoning – is not limited exclusively to a logical approach but also includes topological argumentation, which is associated with a specific segment of a network or corresponds to the structural balance of its environment. The article argues that, within topological argumentation, the process of justifying judgments is intimately connected to the concept of network communication. Networks are not just channels for transmitting information; they also generate a structure that enables the selection of arguments based on particular features of communication organization, such as how vertices and edges are arranged relative to one another. The structural approach was examined by means of network metrics, encompassing both individual influence indicators (eigenvector centrality, average path length, closeness centrality) and the extent of network cohesion (density, eigenvector centrality, assortativity, modularity). Three types of communicative subgraphs are further considered: dialogical, discourse‑oriented, and commentator‑driven. Network analysis was conducted by means of Gephi tools and the Python programming language. The DeepSeek neural network was used to identify communication subgroups. The structure of argumentation in network communication is studied using the example of the Russian social network Mash on the Moika (on the VKontakte platform). The process of searching for justifications – reasoning – is not limited exclusively to a logical approach but also includes topological argumentation, which is associated with a specific segment of a network or corresponds to the structural balance of its environment. The article argues that, within topological argumentation, the process of justifying judgments is intimately connected to the concept of network communication. Networks are not just channels for transmitting information; they also generate a structure that enables the selection of arguments based on particular features of communication organization, such as how vertices and edges arearranged relative to one another. The structural approach was examined by means of network metrics, encompassing both individual influence indicators (eigenvector centrality, average path length, closeness centrality) and the extent of network cohesion (density, eigenvector centrality,assortativity, modularity). Three types of communicative subgraphs are further considered: dialogical, discourse‑oriented, and commentator‑driven. Network analysis was conducted by means of Gephi tools and the Python programming language. The DeepSeek neural network was used to identify communication subgroups. The structure of argumentation in network communicationis studied using the example of the Russian social network Mash on the Moika (on theVKontakte platform).