The article actualizes the enduring institutional meanings of the ancient cultural heritage, traditionally considered by the classical “school of humanity”. The empirical material of this interdisciplinary study, based on selected methods, is textual pastoral typifications characteristic of the discursive practices of Hellenism. The communicative aspect, which focuses on the pragmatics of the text, is primary, but it is not the most important aspect in relation to the more significant axiological aspect, as well as the equally important sociocultural aspect, which summarizes the observations. By combining textual, philosophical, and cultural studies, and by applying different perspectives to the problem, the goal is to identify and describe the specific thematic "repertoires" of bucolic idylls in the historical context of the unique culture of everyday life in the Hellenistic era, as well as to understand its specific mental and other structures. The genre aspect limits the study to the field of historical and literary studies, and while it does not provide much insight into the conceptual realm, it can still serve as a subject for specialized philological research. The acceptable scope of work does not imply exhaustive conclusions as the process moves from the particular to the general, but it does allow for verified conclusions about the nature of poetic textual typifications.
The article addresses the problem of objectively evaluating the pragmatic competence of Large Language Models regarding their ability to interpret conversational implicatures. The relevance of this study is driven by active integration of AI into social practices and the need to verify its capabilities to understand implicit intentions. The theoretical framework of research is based on Cooperative Principle and Gricean maxims. The pilot experiment tested three Russian language models (YandexGPT-5.1-pro, GigaChat-2-Max и T-pro-it-2.0). They were tested in 28 communication situations that required binary choice of interpretations. The key objective was to determine the speaker's intention when the Cooperative Principle is observed or violated. Results showed high formal accuracy, but qualitative analysis revealed the phenomenon of simulated understanding. Even when models successfully detect implication, they often construct interpretation based on incorrect logical assumptions. Therefore, study supports the need for paradigm shift in assessment: transition from classification to generative tasks with expert validation.
