Papers reviewed and determined not to be word norm studies. Use the flag icon to report errors or suggest re-inclusion.
16504 papers
Background and objective Navigating interprofessional team dynamics is essential for high-quality patient care in pediatric settings. This study involved medical students on a pediatric clerkship who explored the characteristics of high- and low-performing clinical teams by considering drivers and barriers to effective team performance. By analyzing these reflections, the study aimed to identify key facilitators and barriers to effective team-based care. Methods Survey evaluations and narrative reflections were completed by third-year students (M3s) at a single US allopathic medical school during their pediatric clerkship after receiving training in TeamSTEPPS® and Institute for Healthcare Improvement (IHI) Open School, two programs that support quality improvement (QI) in healthcare. Descriptive statistical and inductive thematic analyses were conducted on the resulting 183 narratives. A valence rating system was employed to quantify narrative responses as positive/attractive or negative/aversive, with a Cronbach alpha of 0.958 between two independent reviewers. Results Inductive thematic analysis generated 40 themes that we grouped under the five TeamSTEPPS® skill domains (situation monitoring, communication, leadership, team structure, mutual support) into thematic conceptual models. High-performing teams demonstrated open communication, role clarity, shared understanding, and organized task delegation. Low-performing teams displayed a lack of information exchange, uncertain team roles, unhealthy power dynamics, and disorganized task delegation. Conclusions After instruction in QI methods, pediatric clerkship students identified consistent drivers of and barriers to effective team performance. The themes within the narrative reflections can provide insights into improving patient care delivery, specifically around situation monitoring, communication, and team structure.
Perceptual responses are related to long-term exercise adherence. This within-subjects study compared blood lactate concentration (BLa) and perceptual responses to reduced exertion high intensity interval (REHIT) between rowing and cycle ergometry. Twenty healthy, active adults (age = 27 ± 6 yr) underwent a VO 2 max test followed by completion of REHIT on the rower or cycle ergometer. Exercise consisted of three “all-out” 20 s sprints separated by active recovery. BLa, affective valence, rating of perceived exertion (RPE 6 – 20), and enjoyment were acquired during exercise. Results showed no difference in peak HR (98 ± 6 vs. 95 ± 4 %HRmax, p =.06) or maximal workload (191 ± 34 vs. 204 ± 29 %Wmax, p =.25) between rowing and cycling REHIT. Rowing REHIT exhibited significantly lower BLa ( p <.001) at 2, 30, and 60 min post-exercise (7.0 ± 2.2 vs. 11.6 ± 2.8 mM, p <.001, d = 1.9; 3.1 ± 1.5 vs. 6.2 ± 2.8 mM, p <.001, d = 1.4; and 2.0 ± 0.7 vs. 3.1 ± 1.3 mM, p =.02, d = 1.2) versus cycling. Results also showed significantly greater enjoyment (101 ± 12 vs. 89 ± 17, p <.001, d = 0.84) and lower change in RPE (7.8 ± 2.0 vs. 9.5 ± 2.1, p =.001, d = 0.85) and affective valence (−1.2 ± 1.5 vs. −2.4 ± 2.7, p =.036, d = 0.54) with rowing versus cycling REHIT. Overall, rowing REHIT elicits a more positive perceptual response versus cycling.
The exponential growth of Portuguese-language legal documents has renewed interest in Question Answering (QA) systems capable of returning concise, legally sound answers to natural-language queries. This study presents a systematic literature review, conducted according to PRISMA 2020 guidelines, that synthesises current evidence on QA techniques applied to Lusophone legal texts. Searches, without temporal restrictions, were executed in nine databases (ACM, El Compendex, ISI Web of Science, Periódico Capes, Scielo, Science@Direct, Scopus, Sol SBC and Springer Link) using a string that combine jurisprudential, linguistic and methodological terms. After duplicate removal, independent screening and quality appraisal, ten primary studies met the inclusion criteria (peer-reviewed publications developing or evaluating QA pipelines over Brazilian or Portuguese legislation). Publication activity is recent: more than 70% of the papers appeared between 2023 and 2025 and focus on Brazilian statutes and court decisions. Most pipelines adopt hybrid retrieval—BM25 or symbolic regex filters coupled with BERT-family dense encoders fine-tuned on legal corpora, while Retrieval-Augmented Generation with GPT-class models emerges in the latest research. Reported exact-match scores range from 0.60 to 0.83 and F1 from 0.75 to 0.87; however, only a quarter of the studies release code or data, hindering reproducibility. Common gaps include limited handling of the temporal validity of norms, scarce evaluation by legal specialists, and the absence of benchmark datasets for Portuguese. Overall, QA research for Lusophone law is accelerating yet remains fragmented; future work should prioritize shared resources, temporally aware models, and metrics that capture legal soundness beyond lexical overlap.
The article comprehensively analyzes the modern Ukrainian scientific language in its oral and written forms with an emphasis on the internal organization, functional-stylistic, lexical, grammatical, communicative-pragmatic features of texts of various genres, as well as in plane of academic ethics and the implementation of speech strategies and tactics, which enabled a multidimensional interpretation of the object under study. The focus is made on typical models of professional communication in the scientific field report, scientific message, dispute, monograph, article, theses, etc., their linguistic (lexical, morphological, and syntactic units), compositional and logical structure. Normative and non-normative words and their compounds, attested in scientific works of various genres and in oral monological and dialogical professional speech, are highlighted. Deviations from the stylistic norms of the Ukrainian language are identified and described, with an emphasis on stylistic figures and tropes, the sphere of expression of which is mainly oral professional speech. The specificity of lexical units that serve as a means of linguistic manipulation in disputes, as well as those that give emotionality, expressiveness, unorthodoxy to public speeches, and attract the attention of listeners, is emphasized. It is traced that the oral and written forms of scientific language, despite the presence of common features, in particular, objectivity, accuracy, argumentation, etc., have a number of different parameters. The oral form of scientific language is characterized by extensive syntactic variation, spontaneity, and the presence of some elements that give speech emotionality. In contrast, written works of various genres are characterized by a higher level of completeness, normativity, and a clear, logically and structurally motivated construction of sentences. It was found that mastery of the norms of the Ukrainian scientific language, a high level of professional communication culture, and skillful use of vocabulary serve as important factors in forming the image of a modern highly qualified researcher who is able to analyze and objectively evaluate the achievements of specialists in a certain field, effectively argue own position, and present own research results in an orderly, accurate, and understandable manner.
Artificial Intelligence (AI) refers to systems that can perform tasks that typically require human intelligence, such as learning from data, understanding natural language, recognizing patterns and making decisions (Russell and Norvig, 2009). Unlike traditional computer programs that follow explicit instructions, AI systems are featured in digesting big data, adapting to new inputs and improving performance over time. At its core, AI is powered by algorithms that are sets of mathematical and logical instructions specifying how computers analyze information and make decisions. Modern AI relies on deep learning techniques, particularly massive neural networks, to process complex and high-dimensional inputs such as images, speech and natural language and extract patterns from relevant big data (Mienye et al., 2024; Razavi, 2021). The rapid-evolving generative AI represents a signature move of contemporary AI paradigms (Sengar et al., 2024). Tremendous industrial and academic resources have been investing to advance AI alignment, interpretability and efficiency and seek to balance innovation with safety, transparency and responsible deployment across global industries.In the context of the sport business, AI is not just a technological add-on but a transformative force that provides the tools to analyze complex data, automate operations and create new value for fans and stakeholders. This deep connection extends to sport business research, where AI has been reshaping how sport business knowledge is created, validated and applied. The impact of AI on sport business research can be understood from forging novel research agendas, advancing methodological paradigms and improving research efficiency.The most profound impact of AI is its role as a catalyst for novel research agendas. As AI technologies become deeply embedded in the everyday lives of sport consumers and rapidly reshape the operations of sport organizations, they create emergent phenomena that existing theories may not adequately explain. This presents fertile ground for sport business researchers to explore, potentially leading to conceptual refinements and the development of entirely new theoretical frameworks. Below we outline several research agendas for sport business scholars engaging with AI.How Fans Respond to Sport Products Integrating AI. Sport products are increasingly AI-supported, where AI serves as a supplemental component to enhance existing functions or even AI-powered, where AI acts as the core engine of the product itself (Naeem et al., 2024). This AI transformation ranges from the integration of AI for enhancing traditional game-day experience to the full adoption of AI in delivering personalized content recommendation, powering emerging sport-betting platforms and revolutionizing chatbot-based interactions. A new field of research is emerging to explore the psychological and sociological dimensions of how sport consumers interact with AI. This line of inquiry could further refine the conceptual models in the field or even lead to new theoretical frameworks explaining human–AI dynamics.How sport businesses utilize AI. The deployment of AI is reshaping sport organizations by transforming both day-to-day operations and high-level strategic roles, including but not limited to public relations, marketing planning, resource allocation, talent management and organizational design. As AI assumes a more central role in shaping strategy and governance, sport organizations may need to adapt their culture, capabilities and decision structures to remain competitive. A growing body of research aims to assess the effectiveness of adopting AI technologies in organizational operations and strategic management and explore organizational structures needed to address the challenges of AI transformation.How governing bodies and society adapt. The prevalence of AI raises significant questions for governance and ethics. Critical frontiers for research include investigating algorithmic bias in fan profiling, ensuring data privacy and establishing fair regulations for AI in sports betting. This line of inquiry is essential for developing new governance models and ethical guidelines for the sport industry.Sport business research has progressed from employing AI primarily as an analytical tool or methodological enhancement to developing AI-centered research agendas that examine how key stakeholders respond to various applications of AI in the sport industry. At present, consumer responses have received greater scholarly attention, whereas the perspectives of business entities and the implications for governance and policy remain comparatively underexplored.Beyond creating new topics, AI is reshaping the methodology of sport business research. It provides powerful new ways to advance existing research agendas by advancing data variety, data volume, data collection, analytics and experiment simulation, which benefits both correlational and experimental studies.Augmented data features. With the assistance of natural language processing (NLP) and computer vision, researchers are no longer limited to structured numerical data such as surveys and official statistics. We now can tap into vast and varied unstructured data sources like natural language from social media, images from fans and video feeds from games. The data with high breadth, granularity and contextual richness enable extraction of sentiment, emotion, contextual meaning and relationships that traditional datasets could not capture (Mao, 2025). The volume and velocity of available data have expanded exponentially, offering a richer, more holistic view of the sport ecosystem (Mamo et al., 2022).New quantitative solutions. Machine learning and language processing models provide powerful additions to the researcher's arsenal, which largely enhance our capacity for analyzing complex unstructured data, modeling non-linear relationships, uncovering latent structures and elevating prediction power that are challenging to achieve with the traditional analytics paradigm. AI can also assist in cleaning and pre-processing large datasets, detecting anomalies and suggesting data transformations. The analytical advancement empowers researchers to develop conceptual models or test established theories with a level of rigor and predictive accuracy that was previously impossible (Chen and Chen, 2024).AI-enhanced qualitative approaches. AI and NLP have also transformed traditional qualitative research methodologies, significantly expanding researchers' ability to process large-scale textual data and automate time-intensive coding processes (Hitch, 2024). They help researchers identify thematic patterns across extensive textual corpora, facilitate rapid comparison across multiple data sources and enhance reproducibility of interpretive analysis in ways that manual approaches often fall short of (Nelson, 2020). This paradigm shift allows researchers to engage with both the breadth and depth of consumer experience simultaneously (Mao et al., 2024).Innovative computational simulation. AI facilitates the creation of sophisticated simulations that explore and examine the behavior of key stakeholders (e.g. fans, athletes, sport organizations and general businesses) within sport business ecosystems. Notably, generative AI gives researchers unprecedented power to design and tailor realistic experimental stimuli such as synthetic commentary, virtual sport environments or tailored promotional messages, enhancing the rigor and ecological validity of experimental designs.AI also benefits sport business research at a broad level by reshaping and streamlining fundamental early-stage tasks, thereby improving research workflow efficiency and enabling scholars to devote more time to higher-level analysis and interpretation. For example, AI-powered platforms such as Semantic Scholar and Sourcely have significantly improved the efficiency of literature review and information search, sorting and synthesis. These tools interpret the context of queries rather than relying solely on keywords, automatically identify related papers, summarize key findings and generate conceptual maps of research areas, enabling scholars to more quickly evaluate existing literature and identify gaps. By automating these foundational tasks, AI not only improves efficiency but also enhances reproducibility in the iterative research cycle where new knowledge builds upon prior work.Five studies in this special issue explore consumer response to AI transformation in various sport consumption settings, ranging from the deeply personal (AI-supported wearable devices) and the interactive (AI chatbots) to the persuasive (generative AI in ads) and the high-stakes (AI-driven sports betting). This collection of work provides crucial managerial implications for navigating the AI transition and enriches theoretical frameworks by illuminating the complex factors, such as emotion, technology anxiety, perceived trust and subjective norms that ultimately determine consumer adoption.Lee et al. (2025) investigated how consumer evaluations of AI-generated sports ads are affected by AI awareness timing, advertisement model type and source-message incongruence. The results show that AI awareness generally have a positive impact, particularly when consumers are aware of the AI's role after viewing the ads. Virtual Human models are rated the lowest compared to Digital Twin and Human models and source-message incongruence negatively influenced evaluations. The study offers insights for practitioners on optimizing AI ads by strategically timing disclosures and selecting appropriate models and provides references for effective AI integration in sport advertising practices.Gerke et al. (2025) empirically examined consumer responses to the AI-supported wearable devices based on the Artificially Intelligent Device Use Acceptance Model (AIDUA). Their findings highlight a significant intention–behavior gap, as emotions were found to predict the intention to use but not actual consumption. The study also nullified a common assumption that consumers' appraisal of AI anthropomorphism influences their performance or effort expectancies. These insights are critical for sports managers and marketers aiming to improve the adoption of AI-supported sports services.Grounded in the parasocial interaction and the social exchange theories, Choi and Lee (2025) investigated how anthropomorphized AI chatbots in sports enhance social presence to boost consumer loyalty and reduce technology anxiety. Key findings indicate that anthropomorphism successfully increases social presence, which in turn positively influences loyalty while negatively affecting technology anxiety. This study also identifies technology anxiety as a partial mediator, showing that a heightened social presence can mitigate anxiety's negative impact on user loyalty. This research effort extends theory by demonstrating that social presence is a key mechanism for reducing user anxiety, offering practical insights for sports marketers using AI to enhance consumer engagement.Buechner et al. (2025) scrutinized whether the source of a sports betting recommendation (AI or human) affects consumer perceptions of expertise and their likelihood to follow the advice. Through three lab experiments, this research consistently found that compared to human resources, AI recommendations significantly decreased consumers' perceptions of expertise. This lower perceived expertise in turn reduced participants' likelihood of following the betting recommendation. As one of the pioneer studies in this area, the findings suggest that despite technological advances, consumers currently exhibit lower trust in AI for sports betting advice, perceiving human sources as more credible.Dinç et al. (2025) assessed ChatGPT adoption's impact on soccer bettors' behavioral intention and word of mouth. Survey results show that attitude and subjective norms are strong predictors of behavioral intention. Specifically, perceived ease of use and usefulness positively shaped attitude. The effect of usefulness on intention was indirect, mediated entirely by attitude. Social influence significantly drove word of mouth via subjective norms and behavioral intention. This research broadens the applicability of existing theoretical frameworks by examining AI adoption among soccer bettors, while simultaneously offering AI developers actionable strategies to improve user acceptance within this evolving market.Perspectives on how sport organizations utilize AI show organizations' capacity to digest AI techniques to compete, profit and grow in today's market. This research body is essential for understanding the strategic, operational and economic implications of AI in the sport ecosystem. The current special issue highlights two studies by five scholars.Du et al. (2025) examined whether AI can assist in training sports salespeople by evaluating their interactions with prospective ticket buyers. Using topic modeling and sentiment analysis on transcribed National Basketball Association (NBA) sales calls, the research identified several key predictors of success. Findings show that agents with greater lexical diversity and a moderate speaking pace generated more positive customer sentiment and achieved higher sales success. Additionally, a positive association was found between asking more open-ended questions and effective information gathering. Guided by the Technology-Task-Fit theory, this study provides evidence that AI can effectively analyze sales conversations to deliver applicable, data-driven feedback, supporting its integration into modern salesperson training programs.Fortunato and Kosterich (2025) examined how Amazon Web Services (AWS) uses its functionally congruent sponsorship with the National Football League (NFL) to demonstrate its performance capabilities. AWS provides both on-field (e.g. player health and safety) and off-field (e.g. game scheduling) services to the NFL and promotes this deep integration via major marketing communications (e.g. AWS websites, in-game elements and TV commercials) to showcase its brand reliability, which is a key factor in business-to-business marketing. The core message of sponsorship implies that if AWS can handle complex tasks for the NFL, it can certainly do the same for a potential client's business. This study provides a timely, practical example of how AI brands leverage sports sponsorship to communicate and position their advanced technical services.As previously discussed, AI is largely reshaping the sport data frontier, altering data attributes (variety, volume and velocity), collection methods, analytical routines and experimental simulation. The ripple effects of this data revolution on sport business research are profound and pervasive. This impact is comprehensively exemplified by four articles featured in the current special issue.Anagnostopoulos et al. (2025) utilized natural language processing (GPT-4) to analyze how companies on the Qatar Stock Exchange reported their “corporate social responsibility (CSR) through sport” initiatives within annual reports from 2006 to 2022. By automating information retrieval from all 46 listed companies, the analysis identified 672 distinct CSRs through sport initiatives, revealing a significant upward trend over the period. The primary contribution lies in its human–AI framework, which provides a novel and efficient method for systematically analyzing how publicly listed companies communicate their CSR activities in the sport sector. This offers a new perspective on corporate philanthropy in the Middle East.Ryu et al. (2025) used AI to analyze how player performance in professional women's volleyball affects fan emotions, measured via sentiment analysis of Instagram comments. It also examined the moderating roles of superstar status and facial attractiveness. The results confirmed fan emotions are tied to game outcomes but revealed a complex beauty bias. Specifically, attractive players received a beauty premium (less negativity) for errors, yet faced a “beauty penalty” for scoring points, a dynamic not observed with non-all-star players. This research advances the use of AI to demonstrate how attractiveness moderates fan reactions to on-court performance, expanding prior work that focused primarily on off-court factors like salary.Bian and Cork (2024) developed a machine learning model capable of predicting the of and identified the key these machine learning for fan offers a data-driven to traditional The model identify sponsorship as the most by and fan The research provides a novel for understanding how fan are offering an actionable for practitioners to enhance sponsorship et al. (2025) the of sport consumption by consumers' perceived and This offers a data-driven method for understanding consumer in a by AI and deep learning approaches to analyze The study highlights that is by a of and social It provides foundational evidence and practical insights for brand managers navigating the sport and fan have become data sources et al., et al., of it which in turn AI data collection is more and as interaction by social and is to that could be used for marketing et al., 2021). This is further by a of through wearable technologies that extensive and behavioral data et al., The questions of data and the of and significant ethical and are to assess the evolving of fan and data as as of technological and frameworks be needed to of personal information in sport consumption. these represents an emerging yet essential of research, one with broad implications for the social between sport entities and their stakeholders in today's data-driven of algorithmic which refers to the systematically and outcomes by has been increasingly across and It can in of data resources, advertising and and impact et al., 2024). the sport AI particularly on datasets, may to and of the contemporary sport leading to or even and 2021). For an for results if it has not been adequately on the of the new and ecosystem. AI models may not the such as the rapid of sport betting or of fan or management tools to be based on an understanding of the market. most these algorithms can and even existing by an AI model for talent or fan marketing is on data that it bias its potentially leading to outcomes and the of from key sport agendas. when effects of AI it may on these the of bias in the AI of intelligence, and business presents a critical governance for the modern sport which is currently by a significant AI innovation is advancing more rapidly than the development of or 2024). This is in several high-stakes the of performance and fan the of algorithms that can in talent and the deployment of that can the in like sport betting and This an on organizational business to and novel that or For are the ethical and of when creating or the of sport is the of when an AI model promotes betting offers to fans A and is the sport a and governance to responsible AI and the of the AI is not a for it is a fundamental in the and of the sport industry. AI on broad datasets, may be with context knowledge to effectively in of the sport ecosystem. This fundamental is At the the challenges include data privacy and for It also when evaluating complex sport phenomena (e.g. talent and of sport ethical like to and operational of that require navigating the from full interpretability to AI systems are and based on a in sport such as datasets and to perform relevant data, algorithms and models could AI systems to achieve higher greater efficiency and be to theoretical frameworks and practical guidelines for developing relevant AI critical research lies in the creation and impact of new sport products by AI The most one be the and AI capable of complex and tasks on of et al., Using fan consumption as an example, an AI can as a personal fan that is to the based on personal and the at the appropriate This AI shift from analytical tools to agents a potentially research economic and operational are particularly on potential to established sport business models in and also be needed to assess the significant operational of high-stakes and decisions in sport business to the development of new frameworks for in the of this research is embedded with ethical and governance It a of the ethical for fan influenced by sports betting or is the new strategic for sport industry. The articles in this special issue have advanced this by and empirically examining key from sport AI The of AI and the relevant critical research in this as a for to further explore, and in sport business, the for transformation across the sport ecosystem. we like to our profound to all of the their time and expertise to this special Their and the scholarly of this special
The fourth technological revolution in language, caused by the invention and active introduction of artificial intelligence into social life, leads to the reformatting of the modern media communication language space. The emergence of the technological dimension in media communication along with the humanitarian one leads to the fact that both the humanitarian communicative model and the technological one are formed and function in the linguistic space of media communication. In relation to the linguistic norm, two types of humanitarian communicative model can be distinguished, one of which is normative and the other creative. The authors of the normative model are professional journalists, the creative model is produced by the Internet users in the comment sections, and the users themselves become the ‘collective Pushkin of social networks’. The normative model in media communication is oriented towards linguistic, stylistic and communicative norms, which contributes to the preservation of the literary language in the media as an absolute communicative value. The creative model of the ‘collective Pushkin’ generates a creative usus that contributes to the renewal of language through the creative energy of its speakers. The norm in the creative model exists implicitly, and its conscious violation demonstrates the expressive possibilities of the national, not only literary language, contributes to filling lacunas, expanding the repertoire of stylistic means, creating a multimedia code in media communication, providing an opportunity to express emotional nuances. The technological model, authored by artificial intelligence, is based on the norms of literary language, as technological authorship in media is not widely advertised. The technological model is normative; its task is to mimic the humanitarian model of professional journalists. The coexistence of the author-journalist and the technological author (AI) in the linguistic space of modern media not only reformats this space in terms of norms and creativity, but also changes its humanitarian component, as it ceases to be unconditionally human.
Effective Aanaphora Resolution (AR) is essential for computational linguistics, as it underpins coherent text analysis, information processing pipeline, and the development of advanced language technologies. Pronominal Anaphora Resolution plays a crucial role in analyzing and understanding large text collections, enabling discourse understanding systems and enhancing the performance of related applications like text summarization, sentiment analysis, and machine translation. This paper proposes and evaluates a novel hybrid method for Hindi AR. The proposed method uses the rule-based method to resolve reflexive and locative pronouns, whereas it uses supervised classifiers to resolve demonstrative and relative pronouns. We investigate two machine learning and one deep learning classifiers- the Distributed Random Forest classifier, Stacked ensemble classifier, and Multi-Layer Perceptron. The performance evaluation is done on two standard datasets: the Hindi tourism dataset and the Hindi Dependency Treebank Data (HDTB). The stacked ensemble model outperforms all other models investigated in this paper on the Hindi Tourism dataset with an accuracy of 76.33%. The deep learning model performs better than stacked ensemble and random forest on the HDTB dataset with an overall accuracy of 75.96%. The proposed hybrid method outperforms most of the earlier reported work on Hindi AR. This research demonstrates the potential of DL and ML classifiers in developing an automatic entity linking system in Hindi text, which is necessary for the correct semantic interpretation of text..
The article presents a comprehensive analysis of the impact of Internet memes and social media discourse on the evolution of contemporary English, focusing particularly on the processes of lexical innovation, semantic shift, and pragmatic adaptation of linguistic units.In the digital age, communication increasingly takes place through multimodal platforms such as TikTok, Twitter/X, Instagram, and Reddit, which serve as powerful instruments of sociolinguistic dynamics.These platforms not only facilitate the rapid dissemination of slang and emerging expressions but also create new contexts in which users engage in collective linguistic creativity.Internet memes are interpreted as cultural and semiotic units that combine text, imagery, audiovisual effects, and situational context, ensuring the effective circulation and consolidation of new linguistic forms within the global communication space.The article emphasizes that memes function as mechanisms of social identification and linguistic play, reflecting the values, moods, and trends of youth subcultures.The study demonstrates that memes are not merely reflections of linguistic change but active catalysts of language evolution, fostering the emergence of new lexical items, neologisms, and grammatical patterns.The growing influence of Internet culture is shown to shape new norms of speech behavior, where the boundaries between spoken and written, formal and informal language are increasingly blurred.The linguistic transformations driven by digital environments reveal a profound reconfiguration of English lexis, syntax, and pragmatics, as well as a reconsideration of the role of context in communication processes.The conclusion highlights the necessity of an interdisciplinary approach to this phenomenon,
The article provides a comprehensive analysis of Viktor Gavrilov's idiosyncrasy as a modern Ugra writer with a unique, hybrid style close to the Leningrad underground. Special attention is paid to the «language game» technique, as one of the dominant ones in some works from the collection “Times. Coda”. Based on the interpretations of the approaches of several scientific schools, based on the material of the poem “(M)art”, the characteristic features of Gavrilov's poetics, synthesizing elements of playful postmodernism and existential lyrics, are comprehensively investigated. The methodological basis of the research combines elements of structural analysis (identification of lexical and syntactic features) and an intertextual approach (identification of allusions and reminiscences). Special attention is paid to parody strategies (the use of reduced vocabulary, neologisms, allusions to classical texts), existential motives (freedom, time, creative act), a specific synthesis of colloquial intonation and philosophical reflection. The results of the study demonstrate that Gavrilov's idiosyncrasy is characterized by a dialectic of playful and serious principles, intertextual saturation, semantic polyphony, and a special type of lyrical subject. These artistic techniques allow the writer to show his linguistic freedom, to manipulate the norms of language to create an aesthetic and expressive effect. The analysis contributes to the study of the poetics of modern writers, the transformations of postmodern text in Russian literature, and the specifics of modern idiosyncrasy. The scientific novelty of the work consists in introducing the texts of a modern Ugra author into scientific circulation, identifying the features of the regional idiosyncrasy and influences, and determining the place of Viktor Gavrilov's work in the modern literary process. The article reflects a holistic approach to the study of regional text, which corresponds to modern trends in the study of individual works in the general Russian literary process.
The study aims to determine the specific features of the representation of conceptual metaphors of death in Russian translations of Dylan Thomas’s poetry, in comparison with the poetic and philosophical content of the original texts. The research focuses on the verbalization methods of metaphorical images related to the theme of death and the analysis of poetic devices used by translators to preserve the semantic depth and emotional impact of the original. The scientific novelty of the research lies in establishing the characteristic strategies for conveying the conceptual metaphors of death in poetic translations of D. Thomas’s works into Russian and identifying the specifics of lexical-semantic and rhythmic-intonational transformations. The Russian versions of the poems “And death shall have no dominion” and “Do not go gentle into that good night”, translated by V. Betaki and M. Koreneva, were selected as research material. The analysis revealed that key metaphors and images related to the theme of death undergo semantic and stylistic transformations in the translations, due to both the individual interpretation of the translators and the peculiarities of the target language culture. The translations that successfully combine the preservation of the original conceptual metaphor with its adaptation to the poetic norms of the Russian language demonstrate the greatest semantic accuracy and poetic expressiveness.
This article presents a lexicothematic study of nouns and adjectives that characterize a person, based on the Primer of Tatar and Arabic script, published in 1802 by Bukhara native Niyaz Baki Atnometyev. These lexical units represent one of the main thematic groups in any language. The introduction of lexemes that functioned in the speech of Tatars in the Tobolsk Province at the turn of the 18 th and 19 th centuries provides insights into the formation of norms of the Tatar language within a specific region. This article examines four lexicothematic groups: 1) a person as a living being; 2) a person as a sensing and desiring being; 3) a person as a thinking and speaking being; and 4) a person as a social being. We have established that the majority of the lexemes are part of the vocabulary of modern literary Tatar. Besides lexical dialectisms, constituting 23.8% of the total number of units, phonetic dialectisms are recorded (e.g., devoicing: зур – сур, таз – тас; b → m: мукру – бөкре ). The Primer of Tatar and Arabic script represents a unique linguistic source. It preserves a valuable layer of vocabulary that refers to ancient Turkic roots (e.g., әпиәм – abït, исең – ïsrïm, юрчу – jurč ). The scientific perspective lies in the study of nouns, adjectives and verbs in a comparative-historical aspect.
The article considers lexical and grammatical features of the translation of English-language tourist texts into Ukrainian. A tourist text is a form of advertising discourse aimed at getting potential consumers interested in tourist services and encouraging them to order a tour. This requires the creation of an emotionally positive image of the destination. A translator of tourist texts should take into account the main characteristics of tourist language, namely: extensive use of imperatives and adjectives, commonly used phrases to meet the personal and cultural expectations of potential customers focusing on the service and its benefits, repetition of words, adherence to a special rhythm, selection of vocabulary with an exclusively positive connotation. The analysis of tourist texts from the English-language website World Travel Guide allowed us to identify lexical and grammatical features of their translation into Ukrainian. In terms of vocabulary, tourist texts include commonly used words, tourist terms, proper and geographical names, names of cultural realia and stylistic expressive means, which cause the greatest difficulties in translation. The reproduction of such vocabulary in Ukrainian is complicated by the difference in language norms and the lack of direct counterparts for figurative vocabulary, different stylistic traditions of English and Ukrainian, the need to preserve the emotional effect when adapting to the Ukrainian language context. Equivalent translation and the use of lexical translation transformations (modulation, adaptation, use of analogues, compensation, descriptive and synonymous translation, alliteration) are the main ways to ensure the stylistic expressiveness when rendering English language texts into Ukrainian. The grammatical correspondence of the Ukrainian text to its English-language original is ensured by grammatical translation transformations: substitution, addition, omission, integration and partitioning of sentences.
This study aims to analyze the interference of Indonesian in Arabic translation among students of the Arabic Language Education program, focusing on morphological, syntactic, and lexical errors. The research employed quantitative, qualitative, and descriptive approaches, with data collected through a document study of students’ theses translated from Indonesian into Arabic. The analysis was conducted to identify the types of errors, their frequency, and the underlying factors affecting translation quality. The findings indicate that Indonesian interference occurs at multiple linguistic levels, affecting the coherence, cohesion, and stylistic appropriateness (Uslub) of the translated texts. Morphological errors included word-for-word translation, incorrect verb conjugation, and gender disagreement; syntactic errors involved word order, misuse of conjunctions, and improper clause combination; while lexical errors consisted of inappropriate word choice, literal translation of idiomatic expressions, and inaccurate use of technical terms. These results underscore the need for targeted training in linguistic rules, stylistic norms, and discourse practice, alongside the development of cultural and pragmatic awareness. The study concludes that Indonesian interference significantly influences Arabic translation, manifesting at morphological, syntactic, and lexical levels.
Taboo concept founded as basiс notion in linguistic worldview originated from an ancient beliefs. Taboo is a forbidden word and action. The article contains views on taboo meaning based on principles, norms of beliefs, traditions, customs of Turkic people, determined by sociocultural, eсological environment. This study is devoted to prohibited lexics, its place in worldview of various cultural representatives, implementation of linguocognitive analysis.The value of the study identify national, cultural features of prohibitions in Turkic languages. Facing taboo specifics in real use and materials, based on conclusions related life, culture, civilization, describes theories of moral, religious norms supporting by examples. The main goal is focusing on taboo in cognition, to consider prohibition concepts of Turkic and Kazakh people. Determine taboo’s linguocognitive, linguocultural, educational value based on the continuity of cognition and traditions. During the study methods of selecting, collecting, classifying, analyzing materials were used. The theoretical part based on scientists works studying lexicology and ethnolinguistics; showed the specificity, originality of prohibitions by prism of Turkic world languages. The practical significance is the use of materials from Turkic languages allows to create taboo dictionary.Тhe result of the study presented groupings of taboo types in Kazakh and Turkic people which preserved to present days. Analyzed linguistic manifestation of taboos in the Turkic languages, identified characteristics and common features. Taboos and linguistic taboos in Turkic languages can serve for analysis of linguistic units, create dictionaries, as a source of linguocognitive works.
This study examined linguistic deviances (LDs) as stylistic resources in the Nigerian music industry, through a semiotic analysis of Adeleke’s “Funds” and Apata’s “Hustle”. Linguistic deviance, a hallmark of creative language use in artistic expression, is explored here as a deliberate semiotic act that encodes cultural, ideological, and socio-economic meanings within the contemporary Nigerian popular music. A qualitative research design was adopted in the analysis of the selected songs. The study draws from Barthes’ semiotic theory of denotation, connotation, and myth, as well as Leech and Short’s stylistics theory as frameworks for unpacking how deviation from linguistic norms constructs stylistic identity and social commentary. Findings from the study showed that LDs in both songs transcend mere artistic play; they index resistance to linguistic hegemony, assert sociolectal authenticity, and project the artists’ personae as voices of economic struggle and self-affirmation. The results further showed that LDs are deliberate strategies that enhance rhythm, meaning, and cultural identity. The use of code-switching by the artists fosters hybridity, slang, and neologisms that reflect youth culture, while NPE ensures inclusiveness. Repetition and phonological stylization are also found to strengthen emphasis and musicality. The study, while concluding that LDs are powerful stylistic and semiotic devices that enrich Nigerian music, negotiate cultural identity, and index the lived realities of the youth, recommends that further research be conducted across other artists and genres, as well as the documentation of emerging linguistic innovations in other African music. This study has implications for theoretical studies and contributes to the growing body of scholarship at the intersection of stylistics, semiotics, and sociolinguistics, highlighting how popular music mediates between local linguistic practices and global stylistic trends.
Foregrounding is a linguistic and stylistic phenomenon that intentionally deviates from conventional language norms to create emphasis, aesthetic appeal, or emotional impact. This paper conducts a comprehensive contrastive analysis of foregrounding techniques in English and Uzbek, examining grammatical structures, lexical innovations, and stylistic devices in literary and media texts. The study reveals that English foregrounding frequently relies on syntactic rearrangements, phonetic patterns, and lexical creativity, whereas Uzbek employs morphological flexibility, proverbial parallelism, and culturally embedded metaphors. By comparing these strategies, the research highlights how linguistic typology and cultural context shape rhetorical expression. The findings contribute to cross-linguistic stylistics, offering insights into how different languages manipulate form and meaning for artistic and communicative effects.
We present the morphosyntactic annotation of Nheengatu as spoken in the 19th century in the Lower Amazon region. The annotated data expand the UD Nheengatu-CompLin treebank, the first for Nheengatu in the Universal Dependencies project, by incorporating forms and syntactic patterns characteristic of that region and time. We describe the corpus source, the orthographic normalization process, and the main annotation strategies used. So far, 345 sentences have been annotated, with 310 already integrated into the current version of the treebank. This historical data annotation enhances the lexical and morphosyntactic coverage, supporting the documentation and computational modeling of Nheengatu.
Mental health issues such as depression, stress, anxiety, and personality disorders are increasingly prevalent, particularly within online communities. This study proposes a lightweight and efficient multi-class classification framework to identify five mental health conditions using Reddit user-generated posts. While previous studies predominantly rely on conventional CNNs or standard machine learning techniques for binary classification, our work introduces a novel Bidirectional Long Short-Term Memory (BiLSTM) model integrated with an attention mechanism. The architecture is further enhanced by synonym-based data augmentation using the WordNet lexical database, which improves semantic diversity and enhances model robustness, particularly for underrepresented classes. Unlike prior works that focus narrowly on binary classification or employ transformer-based models with high computational demands, our model offers a lightweight, high-performance architecture optimized for multi-class detection and real-world deployment. Experimental results demonstrate that the proposed model achieves a peak validation accuracy of 95.02%, along with precision 95.08%, recall 95.02%, and F1-scores of 95.03%. These findings support the advancement of efficient AI-driven diagnostic systems in mental health analytics and lay the groundwork for future integration into mobile or resource-constrained platforms.
Vocabulary acquisition is essential to second language learning, as it underpins all core language skills. Accurate vocabulary assessment is particularly important in standardized exams, where test items evaluate learners' comprehension and contextual use of words. Previous research has explored methods for generating distractors to aid in the design of English vocabulary tests. However, current approaches often rely on lexical databases or predefined rules, and frequently produce distractors that risk invalidating the question by introducing multiple correct options. In this study, we focus on English vocabulary questions from Taiwan's university entrance exams. We analyze student response distributions to gain insights into the characteristics of these test items and provide a reference for future research. Additionally, we identify key limitations in how large language models (LLMs) support teachers in generating distractors for vocabulary test design. To address these challenges, we propose the iterative selection with self-review (ISSR) framework, which makes use of a novel LLM-based self-review mechanism to ensure that the distractors remain valid while offering diverse options. Experimental results show that ISSR achieves promising performance in generating plausible distractors, and the self-review mechanism effectively filters out distractors that could invalidate the question.
This is a corpus based sociolinguistics study of language variation (between American and British English) and attitudinal alignment of The Dawn. The main aim of this study is to understand which language variety is more frequent in The Dawn and what this frequency reveals about The Dawn attitudinal orientation toward modernity or tradition. For data collection a specialized corpus was compiled of about 817,946 words by taking news from different sections of The Dawn newspaper online platform. Beside compiled corpus, a reference list, containing word pairs of American and British English vocabulary and spellings, was also prepared for finding frequencies in compiled corpus. To derive frequencies of American and British lexical items (vocabulary and spellings) AntConc 3.5.9 was used. In order to interpret attitudes, two sociolinguistic frameworks were used; Giles’s (2007) communication accommodation theory and Bourdieu’s (1991) concept of linguistic capital. Findings of the study suggest that although both varieties are present in The Dawn newspaper articles the frequency of British English is relatively higher. Moreover, the results also shows that The Dawn as one of the oldest and most read online newspaper of Pakistan, not only aligns with traditional and established norms but also shows a slight alignment with new innovative American linguistic trends making it more relevant and appealing for international audience.
Beyond their incentive value, visual sexual stimuli are thought to have intrinsically rewarding properties that may contribute to the rising prevalence of problematic pornography use. However, whether excessive consumption of visual sexual stimuli fits classic models of addiction and involves reinforcement-based learning remains controversial. To address this question, the present study focused on the interplay of individual differences in trait sexual desire (specifically, the drive to engage in solitary sexuality) with stimulus modality in appetitive Pavlovian conditioning. 62 heterosexual participants (final sample, 36 women) underwent two sessions of differential conditioning, spaced one week apart. During one learning session, neutral cues were reinforced (50 %) by presentation of visual sexual stimuli, while auditory sexual stimuli served as unconditioned stimuli during the other session. Indexing both sexual arousal and appetitive learning, pupil dilation (as well as startle modulation) was used to track the acquisition of conditioned responses. Results revealed that solitary sexuality was associated with blunted differential pupillary responses to cues predicting visual (yet not auditory) sexual stimuli and less sensitization across trials, presumably reflecting reduced anticipatory arousal (consistent with self-report findings) and/or altered processing of uncertainty. At the same time, both enhanced startle habituation and valence ratings suggest that the preference for erotica was unaffected in individuals high in solitary sexuality. Fitted computational models provide additional evidence for a link to divergent learning trajectories. Taken together, our findings underscore the special nature of visual sexual stimuli (compared to auditory sexual stimuli) and support the view that excessive consumption may reflect a dispositional reward deficiency that drives individuals to seek out more intense stimulation.
The growing capabilities of Large Language Models are accompanied by high memory & computational needs, posing significant challenges for deployment in resource-constrained environments such as edge devices. Traditional Post-Training Quantization, Quantization-Aware Training & Low-Rank Adaptation (LoRA) methodologies exhibit limitations in accuracy, memory & inference efficiency respectively. Recently proposed Low-Rank Quantisation-Aware Training (LR-QAT), serves as a lightweight, memory & inference efficient general extended pretraining QAT method. This paper presents an elemental adaptation of LR-QAT, making it feasible to train large models like GPT-2 Medium & BERT-Base-Uncased on low resource consumer grade GPUs. Our method freezes pretrained weights, injects simulated low-bit noise in accordance to LR-QAT, trains LoRA adapters and quantization parameters within this grid, fuses the adapters into a single full-precision checkpoint, and finally emits BF16, INT8 & NF4 models from one training run. Experimental evaluations on benchmark Natural Language Processing tasks such as Stanford Sentiment Treebank (SST-2) and Question-Answering Natural Language Inference (QNLI) from the General Language Understanding Evaluation (GLUE) dataset demonstrate that the proposed method edges past both traditional Post-Training Quantization and customized Quantized LoRA (QLoRA) in performance metrics, while maintaining comparable memory usages at reduced bit-widths.
This Profile looks at two technologies that were developed to make source texts in the original Greek, Latin and, indeed, any language directly accessible to audiences who have not yet studied – and may never study – the language itself: (1) translations aligned at the word and phrase level with the original text and (2) rich linguistic annotations explaining the part of speech, regularised dictionary form and syntactic function of each word in a corpus (typically called treebanks, because the syntactic structure is commonly visualised as an inverted tree).
Large language models (LLMs) are widely deployed in settings where both reliability and efficiency matter. We present a calibrated, seed‑robust empirical comparison of an encoder fine‑tuned model (bidirectional encoder representations from transformers (BERT)‑base) and a decoder in‑context model (generative pre-trained transformer (GPT)‑2 small) across Stanford question answering dataset v2.0 (SQuAD v2.0) and general language understanding evaluation (GLUE)-multi-genre natural language inference (MNLI), Stanford sentiment treebank 2 (SST‑2). Beyond accuracy, we assess reliability (expected calibration error with reliability diagrams and confidence–coverage analysis) and efficiency (latency, memory, throughput) under matched conditions and three fixed seeds. BERT‑base yields higher accuracy and lower calibration error, while GPT‑2 narrows gaps under few‑shot prompting but remains more sensitive to prompt design and context length. Efficiency benchmarks show that decoder‑only prompting incurs near‑linear latency/memory growth with k‑shot exemplars, whereas fine‑tuned encoders maintain stable per‑example cost. These findings offer practical guidance on when to prefer fine‑tuning versus prompting and demonstrate that reliability must be evaluated alongside accuracy for risk‑aware deployment.
Abstract The “sleep to forget and sleep to remember” hypothesis states that sleep attenuates the emotional tone of a memory while strengthening its factual content. However, previous experimental research has yielded inconsistent results, associating sleep with the reduction, enhancement, or maintenance of the emotional tone of memories. Although the hypothesized process may necessitate multiple nights of sleep, most studies have relied on single-night protocols. To address this, we further investigated whether immediate sleep diminishes emotional reactivity triggered by memory reactivation after one week. In a karaoke paradigm, we recorded participants’ singing of two songs and played back one of their recordings (rec1) to induce an embarrassing episode either in the early afternoon (delayed sleep group; N = 25) or the evening (immediate sleep group; N = 25). One week later, we assessed participants’ emotional reactions to the re-exposed recording (rec1) and a newly introduced recording (rec2). Emotional reactivity was assessed using facial blushing as a primary physiological measure and subjective ratings of embarrassment, valence, and blushing. Sleep was monitored using diaries. While the embarrassing episode was successfully induced, Bayesian mixed-effects models revealed reduced facial blushing and more negative valence ratings from initial exposure to re-exposure (rec1) after both a shorter and longer interval to sleep. These changes were nonspecific to the reactivated recording (rec1) and were also observed for the new recording (rec2). Other subjective measures remained unchanged. This study demonstrates that neither the time interval to sleep following encoding nor memory reactivation influenced long-term emotional reactivity, leaving sleep’s role in emotional memory processing elusive.
Apple Inc. is a globally leading company in the electronics and technology industry, whose remarkable success is attributed to the high quality of its products. However, the popularity of any product cannot be achieved without its advertising. Linguistic deviation is a common phenomenon and language strategy in English advertisement. English advertisement, as a means to attract consumers, uses a lot of language deviation, which makes the advertising language novel and unique, and stimulates people’s strong desire to make a purchase. This paper adopts a qualitative analysis method and studies six types of linguistic deviations in Apple’s advertisement----phonological, lexical, graphological, grammatical, semantic deviation and deviation of register under the guidance of Leech’s language deviation model and further explores the social factors behind these linguistic deviations in Apple’s advertisement. The study finds that phonologically, Apple’s advertisements frequently employ rhetorical devices such as alliteration, repetition, and consonance to create a harmonious and catchy rhythm. Lexically, the advertisers coin novel terms to communicate the innovative nature of the products. Graphologically, they offer readers a fresh experience by altering the visual form of words, even if it involves intentional misspellings. Grammatically, Apple’s advertisements tend to favor simple sentences and imperatives, often eschewing traditional grammatical norms. Semantically, the advertisers excel at utilizing personification and metaphor to bridge the psychological gap between the audience and the products. Finally, at the register level, they introduce unconventional expressions, despite their apparent mismatch with the electronics context, to highlight the uniqueness of Apple’s uniqueness.
Tourism terminology is an important communication tool for both users of tourism services and specialists working in the field of tourism.This is explained by the fact that tourism terminology is based on a generally accepted literary norm, supplemented by special terms.The terminology of any field of activity is conventionally divided into three groups: general scientific, interdisciplinary and highly specialized terms.Tourism terminology has a specific structure, which consists of two levels: conceptual and linguistic.At the conceptual level, a term is a reflection of a certain concept, and at the linguistic level -it is a word or phrase that denotes this concept.Thus, the term acts not only as an element of the vocabulary of a certain language, but also as a separate link in the system of scientific and conceptual apparatus.Analysis of English dictionaries shows that most often neologisms are formed using two methods of derivation: lexical and semantic.Lexical derivation is the process of forming new words by adding prefixes, suffixes, or the base of another word to existing words.Semantic derivation is the process of forming new words by changing the meaning of an existing word.In addition to lexical and semantic derivation, other productive ways of forming new terms include telescoping, abbreviation, and borrowing.Telescoping is a way of forming new terms by combining two or more bases of several words.Abbreviation is a way of forming new terms by shortening words or word combinations.Borrowing is the process of forming new terms by borrowing them from other languages.These ways of forming terms are the most productive, as they allow you quickly and effectively create new terms, which meet the needs of scientific and professional activity.The international component of vocabulary is based on the use of the same words with the same meanings in a wide range of languages.The internationalization of vocabulary is associated with the internationalization of social life.Internationalisms have spread over large geographical areas -as a result of the linguistic embodiment of common concepts of modern science, culture, technology, policy.
This paper presents a critical examination of how linguistic diversity, social identity, and equity intersect to shape inclusion within modern, multicultural, and technologically mediated societies. The primary aim is to uncover the complex dynamics through which language serves both as a bridge to empowerment and a barrier to participation. Adopting a conceptual and interdisciplinary methodology, the study synthesises scholarship from linguistics, education, digital technology, and governance to develop a comprehensive framework that situates multilingualism as central to advancing fairness and social justice. The analysis reveals that language functions as a powerful marker of identity and belonging, mediating access to opportunities, resources, and representation. Through an intersectional lens, the study demonstrates that linguistic hierarchies frequently interact with social structures of power, such as class, gender, and ethnicity, to reproduce or challenge inequality. Multilingualism, when embraced within institutional, educational, and digital contexts, emerges as a transformative tool that promotes inclusion, empathy, and cultural understanding. However, the persistence of dominant linguistic norms continues to marginalise minority voices, reinforcing asymmetries of knowledge and power within global communication systems. The findings affirm that the equitable integration of multilingual and intersectional frameworks is essential for realising sustainable social progress. The study recommends that policymakers and educators adopt inclusive language strategies that reflect the realities of linguistic diversity while ensuring ethical and culturally responsive communication in both physical and digital spaces. Collaboration among linguists, technologists, and institutional leaders is further advocated to construct systems that prioritise linguistic justice as a foundation for democratic participation and equity. By reframing language as a medium of empowerment and transformation, the research contributes meaningfully to the global discourse on identity, inclusion, and social justice in the twenty-first century.
Word-level psycholinguistic norms lend empirical support to theories of language processing. However, obtaining such human-based measures is not always feasible or straightforward. One promising approach is to augment human norming datasets by using Large Language Models (LLMs) to predict these characteristics directly, a practice that is rapidly gaining popularity in psycholinguistics and cognitive science. However, the novelty of this approach (and the relative inscrutability of LLMs) necessitates the adoption of rigorous methodologies that guide researchers through this process, present the range of possible approaches, and clarify limitations that are not immediately apparent, but may, in some cases, render the use of LLMs impractical. In this work, we present a comprehensive methodology for estimating word characteristics with LLMs, enriched with practical advice and lessons learned from our own experience. Our approach covers both the direct use of base LLMs and the fine-tuning of models, an alternative that can yield substantial performance gains in certain scenarios. A major emphasis in the guide is the validation of LLM-generated data with human "gold standard" norms. We also present a software framework that implements our methodology and supports both commercial and open-weight models. We illustrate the proposed approach with a case study on estimating word familiarity in English. Using base models, we achieved a Spearman correlation of 0.8 with human ratings, which increased to 0.9 when employing fine-tuned models. This methodology, framework, and set of best practices aim to serve as a reference for future research on leveraging LLMs for psycholinguistic and lexical studies.
Machine learning relies heavily on language modeling to understand detailed information in modern applications of natural language processing. In addition, Deep learning methodologies have gained traction in various fields. Recurrent neural networks have emerged as powerful tools in this context, excelling at sequence modeling. Notably, Long-Short-Term Memory (LSTM) layers have become fundamental to language modeling. Additionally, Temporal Convolutional Networks (TCNs) are a recent and promising addition to Deep Learning. Meanwhile, meta-heuristic optimization has been adopted in several research areas due to its effectiveness. This study proposes a TCN-based language model, optimized by the Arithmetic Optimization Algorithm (AOA). AOA is used in this study to optimize parameters for LSTM and TCN models. The proposed language model is compared with the Exponential Linear Unit LSTM and the Binary Input Gate Recurrent Unit, two important models in modern language processing. The performance of these models is examined and compared using the Penn Treebank (PTB) dataset. The obtained results highlight the ability of AOA to accelerate the convergence of parameters during training and highlight the better performance of the proposed model. By improving Sparse Categorical Cross Entropy loss and perplexity, the proposed model outperforms other models in the PTB language modeling task.
BACKGROUND: Social anxiety disorder (SAD) in youth is associated with significant psychosocial impairments; however, the cognitive and neural mechanisms that maintain it, particularly during childhood and adolescence, remain underexplored. Cognitive models emphasize the role of altered face processing, and neutral facial expressions may be perceived as threatening. Due to their ambiguous nature, contextual cues may play a particularly important role in interpretation. METHODS: We presented neutral child faces paired with social context information varying in valence (negative, neutral, positive) while continuous EEG was recorded. Subjective valence ratings and neural responses (P100, N170, and LPP) were assessed in children and adolescents aged 10-15 years with SAD (n = 53), clinical controls with specific phobias (SP; n = 41), and healthy controls (HC; n = 61). RESULTS: Overall, context information affected both the subjective and neural responses to neutral faces in all children and adolescents, for example, more negative ratings for negatively contextualized faces. Further, participants with SAD generally rated all faces as more negative compared to HCs. Neurally, they showed lower N170 amplitudes compared to both control groups in response to all neutral faces, independent of the context valence. However, only younger children (aged 10-12 years) with SAD showed higher LPP amplitudes than younger HCs. CONCLUSIONS: Processing biases seem to be already present in children and adolescents with SAD, both at the subjective and neural level. Social context information influences neutral face processing but is independent of psychopathology. Future studies examining age effects are needed to investigate whether childhood reflects a particularly sensitive period for the development of processing biases.
ABSTRACT Colonial monolingual norms are a present oppressive force within schooling spaces, with a direct assimilative target on the linguistic practices of historically marginalized peoples, histories, and knowledge systems. For racially minoritized multilingual refugee learners, the space of in‐school science learning can be experienced as an involuntary detachment from linguistic wholeness, intergenerational ways of knowing, and community practices. Our work offers a disruption of the deficit‐based narratives that leave the brilliances associated with multilingualism invisibilized. Based on facilitator reflections and video‐based interaction analyses, our findings shed light on the pedagogy of whole languaging hearts that actively resist colonial and racialized linguistic norms and center social interactions and relations transcending divides across languages, geopolitical locations, and species. The pedagogy of whole languaging hearts was enacted to embrace how children and teachers located themselves within land‐based networks of interrelated communities. We offer to look at this network intersectionally, lovingly, and relationally to imagine otherwise, refusing the predefined disciplinary possibilities for eco‐ and socially just science futures together. Extending further the literature on translanguaging, our article focuses on the transformative possibilities of languaging emerging from our work with diasporic refugee children hailing from the “Middle East,” which transcends the borders of the named languages of Arabic, Kurmanji, and English. With the pedagogical enactment of languaging with whole hearts, we demonstrate how teachers can open space for science teaching grounded in the lives and worlds of multilingual refugee children and in embodied interactions on the land. We argue that ways of coming to know sciences are inseparable from our sense of community, stories, and relations.
This study assesses community-based tourism as an innovation mechanism for protected areas by examining Agdal commons, traditional seasonal resource-closure systems, in Oukaimden (Toubkal National Park) and Tiout (Argan Grove Biosphere Reserve), Morocco. Through exploratory fieldwork comprising eleven semi-structured interviews with forest rights holders, cooperative members, and local governance councils (Jmaa), complemented by correspondence factor analysis, hierarchical classification, and lexical visualization techniques, the research identifies three organizing dimensions of stakeholder discourse: ecological stewardship anchored in Agdal practices, territorial relationships linking place-people-resources, and institutional coordination through cooperatives and associations. Findings reveal that while Agdal terminology pervades local narratives as shorthand for intergenerational ecological knowledge and adaptive regulation, tourism-driven livelihood diversification risks fragmenting external tourism expertise from indigenous governance norms unless deliberately integrated through certified agro-sylvo-pastoral value chains such as cooperative-marketed argan oil. Oukaimden demonstrates underutilized ecotourism capacity beyond its winter-sports focus, whereas Tiout exhibits more mature community tourism adoption. The study proposes actionable policy frameworks that couple community-based tourism development with product certification, cooperative strengthening, and environmental safeguards to prevent resource degradation and social displacement, thereby advancing equitable transitions toward locally controlled, low-impact tourism models within Morocco's biosphere reserves. Although tourism and natural resource management are often perceived as distinct areas, they are interconnected and have significant implications for local populations. Tourism in Oukaimden and Tiout can provide development opportunities but can also put pressure on natural resources. The paper highlights the community approach through a significant observation and interviews carried in both communes. Developing a system of measures and policies to promote the adoption of community tourism, along with implementation mechanisms, positioning it as a key strategy for ensuring a successful transition.
The paper focuses on the translational approach to terminologies and special lexicon in the optional course of translation of official documentation. The aim of the paper is to propose didactic techniques in teaching specialized translation, in particular, translation of official documents. We suggest to precise the term “content-based instruction” and propose the term “content-based meta-linguistic instruction/teaching” in the context of teaching of translation where the object of studies comprises translation peculiarities of linguistic units. We emphasize the importance of knowledge of special vocabulary for future translators. The translation difficulties related to polysemic, homonymic and synonymic terms are solved by the context and discourse analyses, the frequency and the tradition of the use of lexical units. The search of translation equivalents is facilitated by the study of functional or pragmatic peculiarities of special lexicon and the structure of official documents in Spanish and Ukrainian. We recommend paying special attention to international terms which look synonymic, though they are translated differently depending on the context, such as “certificate”, “diploma”, “license”, which, in many cases, are not translated literally depending on certain document and on the Spanish-speaking country. Translators must consider the clichés used in Spanish official communication which seems less brief and dry than the Ukrainian one. The use of Subjunctive Mood and Future Simple in Spanish laws is transmitted in Ukrainian with the Present Simple of the Indicative Mood according to the established norms of use.
Mental time travel involves mental imagery to recollect past experiences and envision future events, eliciting anticipatory emotional responses that motivate goal-directed behaviour. However, the temporal dynamics of neural, physiological, and affective processing of mental time travel remain elusive. This study examined late positive potential (LPP), skin conductance responses (SCR), and behavioural affect ratings in response to mental time travel. Forty-eight participants (52% female) viewed 16 neutral, positive, and negative stimuli from the International Affective Picture System ("encoding task"). Participants then vividly imagined the stimuli ("recall task") and imagined a scenario involving the presented stimuli as if it might occur after leaving the lab ("prospection task"). Results showed enhanced LPP amplitudes when recalling negative and prospecting positive experiences, alongside elevated self-reported affect and arousal during these emotional recall and prospection tasks. These findings suggest that mental time travel through emotionally salient events is associated with increased LPP amplitudes akin to the processing of immediate experiences. This might reflect a neural mechanism of anticipatory affective responses to mental representations.
R) subunit regulation in altering sensitivity to neuroactive steroids, a theory difficult to test in humans. We examined peripheral mRNA expression of GABRD, GABRG2, and GABRA5 in a transdiagnostic sample of 112 participants (75 psychiatric outpatients with suicidality; 37 controls) to capture a dimensional range of MCAC severity. Participants provided daily affect ratings and five LH-timed blood samples using PAXgene tubes for RT-PCR analysis. Replicating preclinical findings, GABRG2 expression decreased during the midluteal phase. Critically, individual differences in both person-mean expression and perimenstrual change (AUCi) were correlated with cyclical symptom trajectories. Greater perimenstrual increases or higher average expression were associated with classic luteal-phase worsening of anxiety (GABRG2, GABRA5), irritability (GABRG2), and suicidal ideation (GABRG2, GABRA5, GABRD). However, relative to those with moderate or stable expression, perimenstrual decreases or lower average expression were also linked to cyclical worsening of anxiety and suicidal ideation that emerged around menses and persisted into the follicular phase. Overall, more stable subunit expression was associated with less cyclical symptom change. These findings could be consistent with a loss of homeostatic stability in GABAergic plasticity, rather than a simple unidirectional differences in levels or changes, as a potential correlate of MCAC. Peripheral GABAAR gene expression warrants further investigation, particularly in experimental studies, to determine whether there are causal associations between subunit gene expression and MCAC.
This article presents a comparative analysis of figurative language in Italian and Azerbaijani, arguing that idioms operate as semio-cognitive devices linking embodied experience to cultural-historical norms. Integrating conceptual metaphor theory, mental spaces and blending, frame semantics, and semiotics, the study models cross-linguistic mappings across FIRE / HEAT, CONTAINER, JOURNEY, VERTICALITY, and ENERGY. The aim of the study is to identify the semio-cognitive models of figurative language and to determine their similarities and differences within a cultural-historical context, to establish how these models are shaped by the typological structures of both languages, and, ultimately, to develop new theoretical generalizations. The study is based on the synthesis of cognitive-semantic, semiotic, and comparative-historical methods. The corpus concentrates on canonical texts (Dante, Petrarch, Ariosto, Manzoni; Nasimi, Fuzuli, Vagif, Zakir) and on established Azerbaijani phraseology. Methodologically, idioms are treated as stabilized sign- packages with layered denotative, connotative, and symbolic values; each item is mapped from source to target domains and situated in discourse frames that mediate social roles and ethical normativity. Findings reveal robust universals alongside salient local filters. Universally, emotional and evalua- tive meanings recruit shared schemas—ANGER IS FIRE / HEAT, MIND / SELF IS A CONTAINER, LIFE / LOVE IS A JOURNEY, SOCIAL POWER IS UP / SUBMISSION IS DOWN, LOVE IS ENERGY / FORCE—realized through language-specific resources: graded thermal lexicons (Italian caldo→bollente→rovente; Azerbaijani isti→qızmar→ yandırıcı→büryan), container states and overflow, and vertical-motion patterns. Locally, Italian discourse foregrounds visual-theatrical staging and intratextual legitimization of folk wisdom as pro- verbio (e.g., Ariosto’s “cader de la padella ne le brage”; Manzoni’s “Ambasciator non porta pena”), encod- ing institutional and juridical norms. Azerbaijani classics privilege mystical-soteriological semantics around can / od / könül and exploit agglutinative morphology to fine-grade intensity (e.g., sinə büryan olmaq, içi yanmaq, can vermək, ayağına düşmək – En.literal “to have one’s breast burned,” “to burn inside,” “to give one’s soul / life,” “to fall at someone’s feet”). Parallel readings align Petrarch’s “picciol foco” with Fuzu- li’s “eşq atəşi,” (the fire of love) and Manzoni’s fuori di sé (to be outside oneself) with Azerbaijani özündən çıxmaq (to go out of oneself). Analytically, idioms function on two interconnected planes. At the psychosemantic level, they crys- tallize embodied experience into portable mappings; at the socio-discursive level, they normalize roles and values, turning proverbs and fixed epithets into interpretive institutions. The proposed three-axis frame- work — (i) source→target mapping, (ii) semiotic stratification, (iii) discursive-cultural embedding—yields a reproducible procedure for comparison. Practical implications span didactics (model-based teaching of idioms), translation (equivalence by mapping rather than lexical parity), lexicography (dictionary fields for conceptual model, discourse function, scalar degree), and NLP (annotation layers such as THERMAL_ SCALE, CONTAINER_STATE, VERTICAL_MOVE, INSTITUTION_ROLE). Limitations include a classical-text fo- cus and the absence of psycholinguistic testing; future work will expand to modern media, broaden Ro- mance/Turkic coverage, and evaluate model-guided annotation in multilingual transformer pipelines. In sum, figurative language in both traditions emerges as a semio-cognitive nexus: universals provide the skeleton (FIRE / HEAT, CONTAINER, JOURNEY, VERTICALITY, ENERGY), while local codes dress it in culturally spe- cific forms. Idioms and metaphors thus act not as ornament but as operational mechanisms that bridge cogni- tion and cultural order, offering a unified account of how poetic form encodes social and ethical meaning. In addition, the article operationalizes its claims with a replicable annotation scheme that links idi- om tokens to explicit source–target mappings, scalar degrees, and discourse functions, enabling quantita- tive corpus work and explainable NLP. By aligning parallel examples (e.g., Petrarch—Fuzuli; Ariosto—Zakir; Manzoni—Vaqif), the study demonstrates how identical conceptual skeletons yield distinct stylistic realiza- tions under divergent cultural constraints. This dual lens—universal templates filtered by local codes—of- fers practical guidance for curriculum design, translator training, and culturally aware language technolo- gies. As such, the framework advances an integrated, testable agenda for future research at the intersec- tion of cognitive linguistics, semiotics, and comparative philology.
This study examines the representation of feminism in Linda Howard’s novel Cry No More using Sara Mills’ Feminist Stylistics model. Situated within the romance thriller genre, the novel offers a productive site to explore how female agency is negotiated between empowerment and vulnerability. The data consist of selected narrative units featuring the protagonist, Milla Edge, which are analyzed through lexical and syntactic categorization focusing on Subject–Object positioning and reader positioning. The analysis reveals three dominant patterns: Milla is constructed as an active Subject through high-transitivity verbs and agentic nominalizations in her role as leader of the “Finders” organization; she is simultaneously rendered an Object via passive structures and a lexicon of suffering; and key resolutions of the plot often depend on male protectors. These findings demonstrate that Milla’s agency functions as “negotiated empowerment,” thereby contributing to feminist Critical Discourse Analysis of popular fiction by showing how patriarchal norms are both challenged and reproduced at the level of linguistic choice.
The arti cle analyzes internet discourse as an innovative phenomenon of modern communication that significantly influences linguistic practices, cultural processes, and social dynamics.The study examines the key characteristics of internet discourse, such as multimodality, interactivity, hypertextuality, and anonymity, which fundamentally transform traditional forms of communication.Particular attention is paid to sociolinguistic aspects, specifically the impact of internet communication on the formation of new linguistic norms, sociocultural identities, and lexical innovations, which serve as important indicators of globalization.The research findings indicate that internet discourse facilitates the spread of global linguistic trends while simultaneously affecting local languages.The analysis shows that phenomena such as borrowings, neologisms, and abbreviations are not only means of communication but also markers of cultural transformations.The study highlights examples of the adaptation of the Ukrainian language to digital realities, particularly through the integration of terms such as dystantsiika ('remote learning'), zumytysia ('to join a Zoom meeting'), and others, which reflect current societal changes.These lexical units result from the influence of global processes and illustrate how local languages integrate into the global communicative space.It is specifically emphasized that internet discourse creates new opportunities for intercultural dialogue, where language interaction occurs through the flexibility and innovativeness of communication formats.In particular, multimodal forms of communication, including text, graphics, audio, and video, allow for maximum adaptation of communication to the needs of the audience.Moreover, the use of memes and emojis demonstrates how visual elements complement the textual component, ensuring broader reach and mutual understanding among speakers of different languages.The study also reveals that internet discourse is a means of transforming social norms, particularly through its influence on public opinion, community mobilization, and the formation of new rules of communicative behavior.Examples of active discussions on linguistic and social issues in social networks, forums, and blogs confirm that internet discourse is becoming an important space for reflection on cultural and linguistic processes.The research findings demonstrate that internet communication blurs the boundaries between oral and written forms, creating new hybrid interaction models that shape the foundation for further changes in linguistic and cultural practices.
Cognitive reappraisal and attentional distraction constitute two core strategies for regulating emotions. Prior studies have largely focused on young adults regulating simple laboratory stimuli, with few direct comparisons of brain regions that differentiate or mutually implement these strategies. Here, we expanded the typical age range of participants, compared reappraisal and distraction within participants, and used ecologically valid autobiographical memories as regulatory targets. Sixty-two healthy adults aged 35-75 years generated cue words for negative and neutral autobiographical memories and were trained to either reappraise, distract, or let their emotions flow naturally in response to cued memories. Strategy-specific contrasts were derived from whole-brain fMRI data using univariate analyses. For reappraisal, relative to flow, we observed activity in bilateral occipital cortex, right cerebellum, and cingulate cortex and primarily left-sided frontal, temporal, and parietal cortices. Distraction, relative to flow, engaged bilateral lateral prefrontal, medial parietal, cingulate, occipital, and retrosplenial regions and left cerebellum. Common areas of activation included midline occipital and posterior cingulate cortices. Direct comparisons yielded strategy differences across multiple cortical areas: distraction engaged paralimbic areas (insula and left parahippocampal gyrus), dorsolateral and ventrolateral PFC, and right inferior frontoparietal cortex, whereas reappraisal engaged dorsomedial PFC, left ventrolateral PFC, anterior temporal cortex, and left posterolateral PFC. In-scanner valence ratings verified the efficacy of the experimental manipulation and revealed a negative impact of age on reappraisal success, which was correlated with greater visual cortical processing. These findings extend knowledge regarding the neural mechanisms of emotion regulation across the adult lifespan for autobiographical events.
The rapid development of Neural Machine Translation (NMT) and Large Language Models (LLMs) marks what can be described as an Algorithmic Turn in translation studies.This shift fundamentally reconfigures translation from a primarily human-centered act of linguistic mediation to a process increasingly shaped by computational systems and algorithmic logic.This paper critically examines the growing influence of Artificial Intelligence on intercultural narratives and on the process of semantic transfer in an interconnected global context.While AI-driven translation technologies offer unprecedented speed, scalability, and accessibility, their widespread adoption also raises significant concerns related to cultural authenticity, linguistic diversity, and the preservation of meaning.Drawing on a critical-theoretical framework, the study argues that AI systems function not merely as neutral tools but as active co-authors in the construction of crosscultural meaning.The analysis focuses on two central issues: first, the ways in which intercultural narratives are shaped by AI models trained on culturally imbalanced datasets, often privileging dominant linguistic norms and contributing to cultural homogenization; and second, the inherent limitations of AI in handling semantic nuance, particularly in relation to irony, implicit power relations, and culturally embedded meanings.The paper ultimately contends that the Algorithmic Turn produces a paradoxical outcome-a translation that is linguistically fluent yet culturally hollow.Addressing this paradox requires moving beyond accuracy metrics toward a critical evaluation of AI's ethical, cultural, and epistemological implications, reaffirming the essential role of the human translator as an intercultural mediator whose expertise complements, rather than competes with, technological innovation.
As part of a Critical Discourse Analysis (CDA) project, which focuses on a case study, this study examines how lexical choices influence gendered ideologies through a study of film reviews of Barbie (2003). The study looks at the language employed in critiques and how gender roles, stereotypes, and feminist themes are written and spoken in the discourse of the film. It looks at how Barbie and Ken get empowered and how Ken is subordinated through certain key choices of lexical meaning, as well as the way Barbie and Ken subvert the traditional norms of gender. It demonstrates that the study revolves around the same lines as those of empowerment vs. subordination, stereotypes, and resistance to gender norms. Barbie is always shown as an empowered and independent figure who challenges traditional femininity, whereas Ken always represents a passive and subordinate identity that mirrors traditional masculinity. While both characters seem to cast their roles at first, Barbie struggles with societal beauty standards, and Ken struggles with a crisis of masculinity. Further, the analysis also discusses the pressure between feminism and commercialization, pointing out how the film criticizes patriarchal structures even when the film takes place in a consumer-obsessive culture. Overall, the study indicates the extent to which language can bring out gendered meaning and contributes to our understanding of how media discourse influences social perception of gender.
The article is devoted to the linguotextual study of the translated Life of Saint Athanasius the Athonite and the identification of the basic principles of lexical editing found in the copies of the δ text group of the First Edition of the source. The Greek original and 15 Slavic copies of the 14th-16th centuries of the First Edition of the Life helped to characterise the lexical substitutions attested in the copies of the δ text group. By using the material of the Second Edition of the Life in the “Sobornik” by Nilus of Sora, it is shown how significantly the editing principles differ in the texts of the δ group of the First Edition and in Nilus’ edition. Firstly, in the texts of the δ group, lexical substitutions are sporadic and are not conditioned by the belonging of words to the linguistic tradition of any book school, while Nilus gives preference to ancient Ohridisms and eliminates pre-Slavicisms. Secondly, the creator of the prototype of the δ text group, when correcting archaisms for frequently used lexemes, selects partial synonyms, unlike Nilus, who uses exact equivalents. A common feature of the lexical editing of the δ group texts of the First Edition and Nilus’ edition is the desire to adapt the original text for the perception of a scribe of the late 15th - early 16th centuries by reflecting the progressive changes in the lexical norm of the literary Slavonic language in the language of the Life.
Modern social networks are a significant factor in transforming communication processes, contributing to the formation of a unique network language. The combination of oral and written communication elements ensures rapid information transmission, inevitably influencing linguistic norms. The given article explores the role of social networks in language transformation, particularly their impact on vocabulary and the functioning of network language.The paper analyzes the concept of “social network” and identifies key categories of vocabulary innovations actively used in online communication: abbreviations, acronyms, and neologisms. Examples of their spread in popular social networks (Twitter, Instagram, Facebook) are provided. For instance, Twitter has enriched the English language with lexemes such as attwicted, twitamin, egotwistical, and retweet. Among the most common abbreviations used on this network are Ab/abt, BFN, HAND, etc. Thanks to Instagram, acronyms such as OOTD (Outfit of the Day) and WCW (Woman Crush Wednesday) have become widespread. The name of the social network Facebook itself is a commonly used abbreviation (FB).Additionally, Facebook has contributed to the establishment of abbreviations such as IMO (in my opinion) and TBT (Throwback Thursday) in the English language.The article examines their impact on traditional linguistic norms, particularly changes in the lexical composition of English and the tendency toward simplified spelling rules. The research methodology includes a descriptive method, comparative analysis, linguistic and stylistic analysis, contextual analysis, and statistical methods. The study analyzes 300 lexical units selected through the continuous sampling method. It has been found that the majority of network neologisms originate from English due to globalization and the dominance of English in digital communication. The study results demonstrate that network language is a dynamic phenomenon that continuously expands the vocabulary.Popular social networks generate unique neologisms and abbreviations, which gradually integrate into users’ speech and become embedded in their communicative practices.
Introduction. The article is devoted to the problems related to the gender reform of the German language. Started on the wave of feminist movement of the 70s of the XX century, the transition to gender-neutral language in recent decades has become one of the most discussed topics in both socio-political and scientific circles in Germany, dividing politicians, lawyers, linguists and ordinary citizens into supporters and opponents of gender-neutral language. Methodology and sources. The article examines legal documents regulating the use of a gender-neutral language, highlights the opinions of participants in the discussion about gender correctness, based on the “myth of the invisible woman”, analyzes gender-oriented transformations used in German, and identifies problems related to gender-oriented language correction. Results and discussion. The starting point of linguistic distortions in the field of gender politics was the confusion of the concepts of grammatical gender (Genus), biological sex (Geschlecht) and gender (Gender/ soziales Geschlecht). The refusal of gender reform proponents to use the forms of generic masculine gender (generisches Maskulinum), which includes a wide range of meanings, and the introduction of gender-oriented transformations into the language provoked problems in the field of linguistic word usage, associated with both distortion of meaning and violation of grammatical norms of the German language. Conclusion. Gender reform has had a significant impact on various spheres of public life in Germany. The gender reform of the German language, dictated by the political agenda, has generated many linguistic and extra linguistic problems. The proposed artificial language changes aimed at achieving gender neutrality actually complicate communication and lead to a violation of the linguistic norms of the German language.
This article explores lexicalization as a fundamental, dynamic process in Romanian vocabulary enrichment, emphasizing that it is not a separate mechanism of word formation but an outcome of various internal procedures such as derivation, compounding, conversion, semantic specialization, and contamination. Lexicalization is defined as the diachronic transition from analyzable syntactic or grammatical constructions to stable, autonomous lexical units with consistent form and meaning. The article distinguishes between formal and semantic lexicalization and analyzes how gradual institutionalization and usage lead to complete or partial integration of new words and idiomatic expressions into the lexicon. Ultimately, lexicalization serves to diversify the vocabulary, resolve ambiguities, and reflect evolving communicative norms, acting as a collaborative effect of multiple linguistic processes rather than a distinct method.
Word order difference between source and target languages is a major obstacle to cross-lingual transfer, especially in the dependency parsing task. Current works are mostly based on order-agnostic models or word reordering to mitigate this problem. However, such methods either do not leverage grammatical information naturally contained in word order or are computationally expensive as the permutation space grows exponentially with the sentence length. Moreover, the reordered source sentence with an unnatural word order may be a form of noising that harms the model learning. To this end, we propose an Implicit Word Reordering framework with Knowledge Distillation (IWR-KD). This framework is inspired by that deep networks are good at learning feature linearization corresponding to meaningful data transformation, e.g. word reordering. To realize this idea, we introduce a knowledge distillation framework composed of a word-reordering teacher model and a dependency parsing student model. We verify our proposed method on Universal Dependency Treebanks across 31 different languages and show it outperforms a series of competitors, together with experimental analysis to illustrate how our method works towards training a robust parser.
The article emphasizes that speech is an important means of communication, development of thinking, self-expression of a child and his/her successful socialization.At the same time, as practice shows, preschool children have speech disorders.Thus, in middle preschool age, various speech disorders are observed that require correction.Pedagogical correction of speech is a complex process that includes special methods and techniques for the development of correct speech in children.The article defines pedagogical correction as a purposeful influence of the teacher on the correction of children's speech errors caused by violation of phonetic-orthoepic, lexical and grammatical norms of the native language through specially organized pedagogical activities.It is emphasized that the leading activity of preschool children is play, which is integrated with various activities, including communicative ones.Communicative and game activity is a two-component formation that involves solving educational speech tasks, forming communicative qualities (correctness, logic, appropriateness, expressiveness, imagery, accuracy) by means of various types of games.The article highlights the causes of children's speech disorders at the preschool stage (biological, psychological, social), principles (motivational support, complex selection of methods and techniques, visualization, situational and communicative orientation, speech activity, interconnection of the artistic word and speech), individualization and differentiation, emotional saturation of correctional and game tasks) and the content of the implementation of support for the pedagogical correction of speech of middle-aged children in communicative and game activities (interactive learning platforms, various games, exercises, situations; recommendations for parents).
The article examines the issue of translating idioms in I.S. Turgenev's novel Fathers and Sons in the context of cultural and historical influences on translation decisions. A comparative analysis of three English translations of the novel, produced in the 19th, 20th, and 21st centuries, is conducted to identify changes in the translation of idiomatic expressions. Special attention is given to the concept of «cultural time» as a factor affecting translation strategies. The study demonstrates that each generation of translators adapts the text according to the prevailing cultural and linguistic norms of its time. The analysis highlights how the perception of Russian idioms evolves and how cultural shifts influence the choice of equivalents in the target language. The findings of this research may be valuable for specialists in literary translation and intercultural communication.
Introduction Emotional and stress-related disorders pose a growing threat to global mental health, emphasizing the critical need for accurate, robust, and interpretable emotion recognition systems. Despite advances in affective computing, existing models often lack generalizability across diverse physiological and behavioral datasets, limiting their practical deployment. Methods This study presents a dual deep learning-based framework for mental health monitoring and activity monitoring. The first approach introduces a framework for stress classification based on a 1D-CNN trained on the WESAD dataset. This model is then fine-tuned using the ScientISST-MOVE dataset to detect daily life activities based on motion signals, and it is used as transfer learning for a downstream task. An explainable AI technique is used to interpret the model’s predictions, while class imbalance is addressed using focal loss and class weighting. The second approach employs a temporal conformer architecture combining CNN and transformer components to model temporal dependencies in continuous affective ratings of emotional states based on valence, arousal, and dominance (VAD) using the DREAMER dataset. This method incorporates feature engineering techniques and models temporal dependencies in ECG signals. Results The deep learning classifier trained on WESAD biosignal data achieved 98% accuracy across three classes, demonstrating highly reliable stress classification. The transfer learning model, evaluated on the ScientISST-MOVE dataset, achieved an overall accuracy of 82% across four activity states, with good precision and recall for high-support classes. However, the explanations produced by Grad-CAM appear uninformative and do not clearly indicate which parts of the signals influence the prediction. The conformer model achieved an R 2 score of 0.78 and a rounded accuracy of 87.59% across all three dimensions, highlighting its robustness in multi-dimensional emotion prediction. Discussion The framework demonstrates strong performance, interpretability, and real-time applicability in personalized affective computing.