Three-Valued Modal Logic for Reconstructing the Semantic Network Structure of a Corpus of Coded Texts

Abstract

Researchers doing qualitative content analysis often face situations, where the preliminary results of the coding-process are inconsistent, too complex, and at a relatively low level of abstraction. This article presents a research method, by which these problems may successfully be tackled. It assumes, that the analyzed collection of texts is a manifestation of the cognitive map of the author of the texts, which in turn represents a relatively abstract, consistent, and parsimonious description of the author’s environment. Hence, the article proposes to reconstruct this cognitive map by systematically transforming the initial coding of the texts. The article describes this process and presents appropriate heuristics for this purpose. At the end of the mentioned transformation of coded texts, there is a semantic network, which can be visualized by graphical means. Its nodes are the concepts of the described cognitive map and the links are implications from three-valued propositional logic. Hence, for describing cognitive maps, a third logical value “possibly true” is being used in this paper. This allows to define new types of implications between concepts, which are unknown in binary logic. Moreover, the third truth-value also helps to solve the mentioned problems with inconsistent or contradictory coding. The practical usefulness of three-valued logic for representing semantic networks of cognitive maps is illustrated by an example from a cooking book with Swiss specialties. The article analyzes the ingredients of typical Swiss soups and synthesizes the recipes in a semantic network structure, which is assumed to correspond to the cognitive map of the authors of the book.