?
Text content variables as a function of comprehension: Propositional discourse analysis
The impact of the complexity of a foreign-language text on the amount of information recalled and the choice of metadiscursive strategies remains under-researched, largely because addressing these issues requires the integration of multidisciplinary data. The purpose of this study is to determine how the lexical and informational complexity of a text—as well as the discursive differences between Russian and English—influence the volume and specific characteristics of the content recalled by Russian-speaking readers when engaging with an English-language text. The study involved 94 participants who use English as a foreign language. The source text (which the participants read) and their subsequent retellings were compared at both the informational and linguistic levels. The informational complexity of the source text—a journalistic article corresponding to the C2 level of the Common European Framework of Reference for Languages (CEFR)—was assessed using propositional analysis. Linguistic complexity was measured through descriptive text parameters (word length, sentence length, and the proportion of long words), a readability index, lexical complexity, and the range of metadiscursive markers employed. A comparison between the source text and its retellings revealed a slight decrease in the metrics for all linguistic parameters. The study confirmed that the complexity of the source text serves as a predictor for the number of propositions reproduced: when condensing information in their retellings, readers with a high level of linguistic proficiency reproduce approximately 60% of the main propositions while omitting more than 70% of the secondary propositions from the original text. When retelling texts, native speakers of Russian tend to alter the metadiscursive model: they omit hedges, boosters, and markers of evidentiality, while introducing coordinating conjunctions that were not present in the original source text. This study deepens our understanding of cross-linguistic differences in the use of metadiscursive markers. The findings are of particular practical significance for the automation of propositional analysis and natural language processing.