The relationship between humans and machines in education is analyzing the characteristic features of the LLM artificial language as a tool for teaching the creation of scientific written text through a critical analysis of errors made in texts created by man and machine. This departure from traditional ideas about learning to write in a foreign language reflects progressive processes in teaching, since identifying parallels between human learning and model training based on mistakes refers to a high-tech approach focused on the student rather than the teacher. LLM, as a tool capable of correcting and creating generated text, acts as an incentive for the development of critical thinking when analyzing inaccuracies or evaluating the text “appropriatness”. LLM’s ability to generate different results can be used as a error-based methodology for teaching academic writing. In this regard, academic writing training is a critical analysis and editing of text by visualizing errors, creating a sense of competition between a machine and a human and stimulating mental activity. This seems to be a clear example of how inconsistency and ambiguity can be used for development and creation.
The situation created by the rapid development and widespread availability of artificial intelligence in general, and large language models in particular, which can generate high-quality and coherent true or truthful texts in different languages, is a pressing problem. By studying interdisciplinary phenomena at the junction such as “semantic entropy”, “simulation of meaning” and “perplexity”, based on the theoretical bases of quantitative linguistics, philosophy and information technology, the authors hypothesize that the consequence of such changes is a radical transformation of the informational picture of the world. This situation, on the one hand, leads to a global semantic devaluation and skepticism towards newly created written speech works, and, on the other hand, increases the value of authentic human-made texts (supported by their authority and factual basis). The article also discusses the features of the latest technical and semantic mechanisms for detecting the use of neural networks in text creation, and attempts to describe further trends in this area. The authors consider the phenomenon of generative artificial intelligence (LLM) through the prism of the concept of Jean Baudrillard’s simulation. The technical nature of the neural networks is analyzed as a probabilistic distribution of tokens, creating an illusion of semantic coherence. The examples show a gap between syntactic correctness and the lack of intentionality (meaning-making), which allows AI products to be classified as “pure form” that lacks a referent in reality.
The relevance of research in edusemiotics and psycholinguistics of educational texts is determined by the multicultural nature of contemporary higher education. Entering their profession, students from different ethnocultural backgrounds become involved in a unified professional discourse that encompasses a variety of multimodal educational texts. However, the ethnically specific features of students' semantic perception of educational texts are often not taken into account, which may result in superficial comprehension of the learning material. The aim of this study is to test the hypothesis that there exist specific invariant and variant speech structures whose actualization through the medium of the Kazakh language affects the semantic perception of the educational text. Theoretical analysis of the problem of educational text perception allowed the authors to identify the methodological basis for the study of the obligatory relationships between the linguistic units of an educational text and their psychological correlates. The described problematic task formed the basis for the development of a research strategy for semiotic, psychological and psycholinguistic patterns of perception of the educational text. The article presents some results of a pilot experiment to study the psychological and psycholinguistic features of semantic perception of educational texts within Kazakh linguoculture. During the experiment, the hypothesis of the existence of special invariant and variable speech structures was tested, the actualization of which by means of the Kazakh language affects the semantic perception of the educational text. The hypothesis was tested using appropriate methodological tools: associative and psychosemantic experiments. The experiment involved first-year humanities students at Al-Farabi Kazakh National University. The results of the experiment made it possible to identify general cultural (invariant) and ethnospecific (variable) semantic universals that determine the semantic perception of the educational text. This is followed by a psychological interpretation of the experimental findings based on key characteristics of the psychology of the Kazakh ethnic group and signs have been identified indicating the presence of proposed universals.
The relevance of the study of a critical stage of text processing in the era of the dominance of large language models is due to the systemic divergence between statistically identified token boundaries and linguistically justified morpheme boundaries in tokenizers, which is particularly evident when working with rich morphology languages. The scientific novelty of the study lies in considering the token not as technical part, but as a linguistic artifact - an algorithmic ‘shadow’ of a morpheme that has lost its semantic motivation but preserved its external resemblance. The aim of the research is to identify linguistic losses that occur during statistical segmentation of words and to assess the degree of divergence between token and morpheme boundaries. Research methods: comparative linguistic analysis, corpus analysis, critical discourse analysis and typological modeling. During the research, a typology of morphological distortions in BPE tokenization has been developed and empirically substantiated, describing three types of distortions and divergence between token and morpheme boundaries: fragmentation, contamination, and aberration. Using the lexemes ‘podorozhnik’ and ‘vysokotekhnologichny’ as material, the three-stage degradation of morphological transparency in the transition from monolingual to bilingual and multilingual BPE models has been experimentally proven. It has been established that multilingual tokenizers completely lose morphological properties in relation to a particular language multiplying statistical artifacts that are not identified by a classical set of morphological units and native speakers of this language.
The aim of the study is to clarify the conceptual boundaries of the media text as a variety of polycode text and media discourse under conditions of digital convergence, taking into account the transformation of the contemporary media environment, media consumption, and media production, as well as to substantiate the vertical video text as a new form of media text functioning within a platform-based environment. The article examines the conceptual boundaries of the media text as a type of polycode text and media discourse in the context of digital convergence. The methods employed include functional-stylistic analysis, discourse analysis, and content analysis, as well as a semiotic analysis of the interaction between verbal, iconic, and melodic codes; the empirical corpus consists of 200 media texts from Russian mass media, brand media, and civic journalism, presented in the formats of vertical videos, cards, and video messages (“circles”) in the Telegram messenger. The results demonstrate that the content of the category “media text” is specified as a polycode formation shaped by screen culture, clip thinking, stream consciousness, the decentralization of mass media, and the “explosion of authorship”; it is shown that the media text loses its institutional anchoring and functions within a unified flow of journalistic, user-generated, advertising, and corporate utterances; the format is substantiated as a supragenre unit of media production that determines the visual-communicative structure of the media text and the mode of its perception. The conclusions suggest that the contemporary media text should be regarded as a type of polycode text functioning within digital media discourse; the vertical video text is defined as one of the most indicative forms of its semiotic and pragmatic transformation under the platform logic of content distribution.
