Cognitive Mechanisms of Contrast Recognition and Interpretation and Their Functional Analogues in Artificial Intelligence Systems

Abstract

The interpretation of explicit and implicit contrast is a complex psycholinguistic process that requires the integration of linguistic, cognitive, and contextual information. This study examines the mechanisms underlying contrast recognition and interpretation in contemporary Ukrainian and English-language poetry and compares them with the computational operations of Large Language Models (ChatGPT, Gemini, and Claude). The research combines a controlled psycholinguistic experiment with comparative AI analysis within a unified interdisciplinary framework. The experimental corpus comprised thirty authentic twenty-first-century poetic excerpts, including twenty Ukrainian and ten English-language texts representing both explicit and implicit contrastive structures. Human participants and the three LLMs analysed the same material under identical experimental conditions, while representative cases were selected for detailed qualitative analysis. The study proposes an original seven-stage psycholinguistic model of contrast interpretation and demonstrates that the relationship between human cognition and transformer-based language models is best understood through the concept of functional analogues rather than cognitive equivalence. The findings show that both humans and LLMs accurately recognise explicit contrast, whereas implicit contrast exposes fundamental differences in interpretative strategies. Human readers rely on semantic memory, cultural knowledge, contextual integration, and inferential reasoning, while LLMs primarily exploit contextual representations and probabilistic prediction. The proposed framework extends psycholinguistic research on literary discourse and provides a methodological basis for evaluating the interpretation of implicit meaning by contemporary artificial intelligence systems.