Papers reviewed and determined not to be word norm studies. Use the flag icon to report errors or suggest re-inclusion.
18625 papers
The paper deals, from a diachronic perspective, with the double negative prefix nene-. This type of negation has received almost no attention in Czech linguistics so far. The study aims to collect a large number of examples. A total of 4,874 occurrences of this type of negation were excerpted from selected Czech dictionaries, written language corpora, the Lexical Database of Humanistic and Baroque Czech, and the Kramerius digital library (version 5). This type of negation occurred most frequently with verbs, with the most common word being nenedostávat(i) se (‘neg-neg-get’, i.e. ‘to not not get’). The first documented occurrence dates to the 14th century, and in the second half of the 20th century, this type of negation began to disappear from printed texts. The analysis shows that the meaning of words with the prefix nene- can usually be interpreted as a simple affirmative meaning, though sometimes with a shift toward an intensification of degree. This study contributes to the understanding of negation in Czech and provides suggestions for further analysis.
Empathy involves both domain-general and domain-specific cognitive control processes, yet their distinct causal contributions to different empathy components remain unclear. The current study investigated the effects of inhibitory control and imitation inhibition manipulations on cognitive and affective empathy in individuals with low trait empathy. Sixty-seven female participants (aged 18-23) scoring below the population mean on the Questionnaire of Cognitive and Affective Empathy were randomly assigned to single-session manipulation, either imitation inhibition (via self-other control task, SOC) or inhibitory control (via stop-signal task, SST). Pre- and post-manipulation empathic performance was assessed using the Chinese version Empathic Accuracy Task (EAT) that yielded empathic accuracy scores alongside measures of cognitive empathy and affective empathy. Results revealed dissociable patterns: the SOC group showed significant improvements in empathic accuracy for positive videos, target-valence ratings (cognitive empathy), and emotion contagion (affective empathy), whereas the SST group showed significant improvements in empathic accuracy for negative videos, self-valence for positive videos, and emotion contagion. Neither manipulation significantly affected perspective taking. These findings provide preliminary causal evidence that domain-specific imitation inhibition and domain-general inhibitory control differentially enhance empathy components, potentially through distinct mechanisms of self-other distinction and self-regulation.
Despite their linguistic diversity and global significance, African languages remain underrepresented in research and resources to support NLP. We aim to bridge this gap by introducing AfriSUD, the first large-scale collection of syntactically annotated treebanks for nine diverse African languages spanning major language families and regions across Sub-Saharan Africa. Using the Surface-Syntactic Universal Dependencies (SUD) framework, our community-led effort provides high-quality, native-speaker verified data that capture typological key features such as agglutination and tone. We evaluate a range of models on AfriSUD for part-of-speech tagging and dependency parsing including non-transformer baselines, multilingual pretrained encoders, and LLMs. Our results reveal a significant syntax gap, where models still show clear limitations across the nine languages, suggesting that existing architectures may not fully capture the structural diversity of African-language syntax.
Enhanced memory for salient events can be driven by affective responses and by surprising outcomes. Recent work has suggested that deviations in expected feelings can influence decision making alongside a traditional learning signal, the reward prediction error. Whether a dynamic measure of “affective surprise” may serve as a learning signal that modulates the formation of lasting memories remains unclear. This study investigated how affective surprise elicited by dynamic changes in affective states influences long-term episodic memory. We introduce a novel computation for affective surprise informed by literature on prediction error-driven learning and derived from individual participants' continuous affect ratings of valence and arousal. We first reanalyzed a published dataset and then conducted an independent replication study in which participants encoded item sequences while listening to emotional music. Participants then re-listened to the music while providing continuous ratings of their felt affect. We assessed how affective surprise influenced different aspects of episodic memory after 24 h. We found that greater affective surprise at a given moment, or larger deviations from the recent history of ratings, enhanced memory for when an item occurred in a sequence. In contrast, we found inconsistent effects of affective surprise on item recognition memory across the two studies. Together, these findings suggest that affective surprise, regardless of direction, enhances the binding of items to their temporal contexts in memory. This work demonstrates that affective surprise, in particular valence-related surprise, might act as a learning signal with consequences for episodic memory. • Within-person changes in affective state may act as a learning signal and influence memory. • Novel metric for “affective surprise” captures discrepancy between previous and current affect. • Greater valence-related surprise enhances memory for items in temporal context. • Influence of affective surprise on memory replicated across two samples.
. The primary task of a dictionary is to collect, describe, and systematize the vocabulary of the Uzbek literary language. This process consists of determining the lexical structure of the language, defining the meaning and scope of use of words, as well as strengthening and stabilizing the norms of the literary language. The dictionary not only explains the meanings of words but also provides their correct pronunciation, grammatical forms, and stylistic features, which helps in the correct use of the language. At the same time, according to the requirements of the time, the dictionary makes a significant contribution to improving the culture of speech, as it teaches students and users the most important rules of the literary language.
This paper presents a novel treebank-driven approach to comparing syntactic structures in speech and writing using dependency-parsed corpora. Adopting a fully inductive, bottom-up method, we define syntactic structures as delexicalized dependency (sub)trees and extract them from spoken and written Universal Dependencies (UD) treebanks in two syntactically distinct languages, English and Slovenian. For each corpus, we analyze the size, diversity, and distribution of syntactic inventories, their overlap across modalities, and the structures most characteristic of speech. Results show that, across both languages, spoken corpora contain fewer and less diverse syntactic structures than their written counterparts, with consistent cross-linguistic preferences for certain structural types across modalities. Strikingly, the overlap between spoken and written syntactic inventories is very limited: most structures attested in speech do not occur in writing, pointing to modality-specific preferences in syntactic organization that reflect the distinct demands of real-time interaction and elaborated writing. This contrast is further supported by a keyness analysis of the most frequent speech-specific structures, which highlights patterns associated with interactivity, context-grounding, and economy of expression. We argue that this scalable, language-independent framework offers a useful general method for systematically studying syntactic variation across corpora, laying the groundwork for more comprehensive data-driven theories of grammar in use.
BACKGROUND: Schools have the potential to promote equitable health from early life onwards yet require sufficient organizational capacity to achieve sustained action. Structured improvement approaches, such as PDSA cycles, may help strengthen this capacity by guiding systematic implementation processes. However, their potential in school health promotion remains insufficiently understood, particularly regarding the heterogeneous contextual factors shaping their application. This study examined which contextual determinants shape schools' perceived implementability of the PDSA cycle for health promotion and how these conditions differ across schools. METHODS: Nine German primary schools participating in a holistic health promotion program were purposively sampled to capture heterogeneity across federal states, socioeconomic contexts, and urban-rural settings. Semi-structured qualitative group interviews in a workshop format were conducted with school principals, teachers, and parents and analyzed using the framework method guided by the CFIR. To facilitate cross-case comparison, color-coded valence ratings (facilitator/barrier/mixed) were visualized in a Matrix Heat Map, enabling identification of contextual tendencies. RESULTS: Fifteen contextual factors emerged across the CFIR domains of Outer Setting, Inner Setting, and Individual. Schools with prior experience using structured processes similar to PDSA cycles reported more facilitators, such as established communication structures, while schools without such experience perceived more barriers, notably financial constraints. Common barriers across schools included limited parental engagement and staff shortages, whereas leadership support and compatibility of program components were consistent facilitators. Some factors interacted dynamically, with resource constraints reinforcing other barriers or with strong mission alignment amplifying engagement. CONCLUSION: Schools' prior structured experience seemed to be associated with how they perceived the implementability of PDSA cycles for health promotion implementation, with more experienced schools anticipating more facilitators and fewer barriers. While causality cannot be inferred, these exploratory findings are hypothesis-generating and suggest that prior structured experience may be an important factor to consider for tailoring implementation support and building organizational capacity. Beyond these insights, extending the framework method with a color-coded Matrix Heat Map proved valuable for visualizing contextual heterogeneity and revealing tendencies across cases. This combined approach may inspire further research on how contextual configurations shape the use of structured processes in complex, multi-site implementation settings.
Pure speech language models aim to learn language directly from raw audio without textual resources. A key challenge is that discrete tokens from self-supervised speech encoders result in excessively long sequences, motivating recent work on syllable-like units. However, methods like Sylber and SyllableLM rely on intricate multi-stage training pipelines. We propose ZeroSyl, a simple training-free method to extract syllable boundaries and embeddings directly from a frozen WavLM model. Using L2 norms of features in WavLM's intermediate layers, ZeroSyl achieves competitive syllable segmentation performance. The resulting segments are mean-pooled, discretized using K-means, and used to train a language model. ZeroSyl outperforms prior syllabic tokenizers across lexical, syntactic, and narrative benchmarks. Scaling experiments show that while finer-grained units are beneficial for lexical tasks, our discovered syllabic units exhibit better scaling behavior for syntactic modeling.
Across Germany, cultural and linguistic norms in climbing communities differ significantly. Many of the differences which developed during the separation of the country into two independent states can still be observed today. Several studies on German language use of mountaineers have been published in recent years (e.g. Scharf, 2015; Sakowski, 2023; Scheurer, 2023). Analysis of language use in climbing communities in East and West Germany, however, represents a gap within contemporary research. The present study approaches this desideratum by comparing two corpora consisting of publications by local sections of the Deutscher Alpenverein (German alpine club) from the Eastern and the Western part of Germany respectively. Keywords of the two corpora as well as a sentiment analysis will provide insight into the regional differences in language use. Bubenhofer and Scheurer (2014) identified justifications for climbing mountains to be a typical communicative pattern in texts about mountaineering. In the corpora at hand, keywords connected to these types of justifications show significant differences in their occurrence. The presented corpus analysis will allow for conclusions regarding the cultural differences. Moreover, a diachronic analysis of the corpora will examine the changes in these publications over the past two decades. Thereby, this paper is intended to contribute to a better understanding of the complex interplay between the specific language use of the climbing community and diatopic variation across Germany.
The article examines the phenomenon of abbreviation as one of the most productive ways of word formation in Chinese Internet discourse. The relevance of the research is motivated by the growing interest of linguists in the study of Chinese Internet vocabulary and the peculiarities of their functioning in communication in a digital environment. The aim of the work is to identify the structural and semantic features of Chinese Internet abbreviations, as well as to determine their functional role in online communication. A focused study of Chinese abbreviations allows us to open ways to a deeper understanding of the dynamics of the development of the Chinese Internet language and to form a theoretical basis for the standardization of linguistic norms and effective social regulation in the digital age. The empirical basis of the study is Internet abbreviations extracted from texts of Chinese social networks (WeChat, Weibo, TikTok, Bilibili, etc.), forums and messengers for 2020-2025. As a result of the study, the reasons for the large-scale spread of Chinese abbreviations are summarized: an increase in the speed of information exchange during online communication, the desire to save speech efforts, the influence of the Internet environment and the playful nature of online communication. The main groups of abbreviations have been identified: 1) hieroglyphic abbreviations; 2) alphabetic abbreviations; 3) numerical abbreviations; 4) abbreviations of mixed type; 5) letter emojis; the main problems caused by their excessive use are noted.
Internet visual culture has produced a sprawling vocabulary of "aesthetics" such as Frutiger Aero, vaporwave, cottagecore, and liminal spaces, each a visual genre with a distinct relation to the past. Whether this folk vocabulary has a measurable affective architecture has been argued only qualitatively. Using naturally-occurring human behaviour, we reconstruct the aesthetics ecosystem as a network of 1,115 aesthetics from community cross-referencing. The network is strongly modular, organised by latent affect (p = 0.0005) and historical era (1920s to 2010s), and its unsupervised communities align with an a priori four-mode taxonomy (restoration, hauntology, anemoia, uncanny inversion; NMI = 0.50). A second, independent record of user co-curation matches the same communities above chance (NMI = 0.21 vs. a 0.17 null, p = 0.032). To test whether the structure is real, a frontier panel (Gemini 2.5 Flash, GPT-4o, Claude Sonnet 4.6) predicts the aesthetic that people assign each image to from affect ratings alone (accuracy 0.45 vs. 0.29 chance, z = 18, n = 1,727), matching a 29-feature perceptual baseline. The same models over-attribute unease: they reach near-ceiling recall on the uncanny (0.86) but read warmth and longing as eerie (restorative 0.35, anemoia 0.31). The bias is interpretive: it is absent from image statistics, it persists when the image is edited, and it falls sharply when the rating task is reframed; on standardized affective images the same models match human valence norms (r = 0.84 to 0.87). The over-attribution, its boundary, and its causal mechanism replicate across all three models. Internet aesthetics thus have a real, cross-source affective architecture, readable even to systems trained on none of it, which in reading it reveal a diagnosable and correctable bias toward the uncanny. We release the network, corpora, model ratings, and code.
Public spaces are encountered through changing spatial configurations, yet evidence linking measurable route conditions to in situ emotional valence remains limited. This multi-site observational study examined associations between route-section image-semantic composition, physical spatial attributes, and five-category valence ratings. Twenty participants provided 378 valid ratings across 40 route sections in ten purposively selected public-space sites in Dalian, China, and Nomi and Kanazawa, Japan. Visual composition was derived from 433 quality-screened forward-facing photographs; physical measures included pavement-to-boundary width, element count, and soft-boundary condition and width. Cumulative-logit mixed models included participant and route-section random intercepts and adjusted for sampled setting. We assessed proportional-odds assumptions, 1000 setting-stratified site bootstraps, site-fixed within-site sensitivity, and leave-one-site-out prediction. No image-semantic balance retained an adjusted association. Greater conditional soft-boundary width showed a positive conventional association with more favourable valence in the primary model (β = 0.798; OR = 2.22; BH q < 0.001) and in the within-site analysis (OR = 1.56; BH q = 0.010). Although the conventional Hessian interval excluded the null (95% OR CI 1.62–3.04), the site-bootstrap β interval included zero (−0.003 to 1.404), and the physical-feature model did not improve held-out-site prediction. The finding is therefore a context-bound association requiring replication, not a causal or transferable design rule.
в статье рассматривается актуальная лингво-прагматическая проблематика в контексте теории коммуникативной деятельности. Основное внимание уделяется анализу речевого общения как формы реализации мыслей, чувств и интенций субъекта речи, что соответствует современным тенденциям в науке к изучению речи как совокупности лингвистических привычек индивида. Авторы подчеркивают многослойную природу коммуникации, связывающую языкознание с социологией и психологией, и вводят ключевое понятие языковой ситуации как совокупности условий функционирования языка в определенной среде. Цель работы заключается в выявлении закономерностей влияния экстралингвистических первопричин (социальных, психологических, мотивационных) на построение высказывания. В качестве основного метода используется когнитивный и прагматический анализ, позволяющий раскрыть системные свойства высказывания через соотнесение с мотивами и целями участников коммуникации. Особое внимание уделяется условиям двуязычия, где выбор языка становится дополнительным инструментом в реализации коммуникативного намерения. Центральными категориями исследования выступают интенция говорящего и цель речевого акта, которые рассматриваются в тесной связи с прагматическими пресуппозициями – знаниями об условиях успешности общения и языковой норме. В статье анализируется понятие коммуникативной компетенции как способности не просто строить грамматически верные предложения, но и эффективно использовать языковые средства для достижения необходимого результата. Общение как процесс взаимодействия реализуется сложным многогранным индивидом, который функционирует в социологизированной системе связей, а вербальные средства коммуникации являются лишь его составляющей. В этом аспекте изучение особенностей говорящего посредством лингво- и социопрагматического подхода раскрывает возможности эффективной вербальной коммуникации, в том числе если она отягощена функциональной ситуацией двуязычия. Делается вывод, что прагматические категории выполняют функцию управления речевым поведением, а лингвистическая прагматика, опираясь на теорию деятельности, обеспечивает динамический подход к изучению языка в условиях как монолингвизма, так и билингвизма. the article examines the current issues of linguistic pragmatics in the context of the theory of communicative activity. The main focus is on the analysis of speech communication as a form of realizing the thought, feeling, and intentions of the speaking subject, which aligns with the contemporary scientific interest in the "human factor." The authors emphasis the bilateral nature of communication, linking linguistics with sociology and psychology, and introduces the key concept of the language situation as the set of conditions governing the functioning of language in a specific environment. The aim of the work is to identify patterns in the influence of extralinguistic factors (social, psychological, motivational) on the construction of an utterance. The primary method employed is cognitive and pragmatic analysis, which allows for the revelation of the systemic properties of an utterance by correlating it with the motives and goals of the communicants. Special attention is paid to conditions of bilingualism, where the choice of language becomes an additional tool for realizing communicative intention. The central categories of the study are the speaker's intention and the goal of the speech act, which are examined in close connection with pragmatic presuppositions—knowledge about the conditions for successful communication and linguistic norms. The article analyzes the concept of communicative competence as the ability not merely to construct grammatically correct sentences, but to effectively use linguistic means to achieve the desired outcome. Communication as a process of interaction is realized by a complex, multifaceted individual who operates within a sociologized system of relations, with verbal means of communication being only one of its components. In this aspect, studying the characteristics of the speaker through a linguo- and sociopragmatic approach reveals the possibilities of effective verbal communication, even when it is complicated by the functional situation of bilingualism. It is concluded that pragmatic categories serve the function of managing speech behavior, and that linguistic pragmatics, based on the theory of activity, provides a dynamic approach to the study of language in conditions of both monolingualism and bilingualism.
Post-retrieval extinction has been proposed to modify fear memories by introducing safety information during the reconsolidation window. Imagery rescripting (ImRS) is a visual imagery-based intervention that alters the valence of emotional memories through UCS revaluation. The integration of ImRS and extinction targets both UCS revaluation and expectancy revaluation, providing a dual mechanism for fear reduction. This study explores the combined effects of imagery rescripting (ImRS) and extinction during the reconsolidation window to attenuate the return of generalised fear. Sixty-one healthy individuals underwent fear acquisition and generalisation (Day 1), followed by intervention (Day 2: Reactivation- ImRS and Extinction group or ImRS and Extinction group). On Day 3, return of generalised fear through reinstatement was tested using unsignalled UCS presentations. Skin conductance responses (SCR), expectancy ratings, and valence ratings were the outcome measures. The reactivation-ImRS + extinction group showed reduced reinstatement effects compared to the ImRS + extinction group, evidenced by attenuated SCR and increased CS+ valence ratings. The current study suggests that imagery rescripting and extinction conducted post-retrieval is more effective in reducing the return of conditioned fear. Therefore, integrating ImRS with post-retrieval extinction may offer a novel therapeutic pathway for reducing fear relapse.
The paper explores the distortion of fundamental postulates regarding the gender category in linguistics by some of the key representatives of scientific, educational, and cultural institutions in Serbia, during the recent public debate on gender-sensitive language. The debate emerged after May 20, 2021, when the National Assembly of the Republic of Serbia adopted the Gender Equality Law, which introduced the mandatory use of gender-sensitive language in administration, education, and media. The Committee for the Standardisation of the Serbian Language strongly opposed this law by misrepresenting knowledge about language and gender. To identify the link between the distortion of knowledge and the socio-political consequences of creating epistemic injustice, this paper combines linguistic methods (critical discourse analysis) with ignorance studies. The research findings are related to: a) the concealment and distortion of facts (from general linguistics and the history of the Serbian language); b) ignorance of new knowledge and issues (standardization of gender-related vocabulary; addressing problems in syntax and punctuation); c) the gradation of hurtful language used to spread hate speech and convey prejudices about gender identities.
The article explores the value orientations typical of residents in monotowns, i.e., single-industry towns. The linguistic data were obtained from an indirect associative experiment through an analysis of stimulusresponse speech acts. The methods from corpus linguistics and psycholinguistics made it possible to identify the hierarchical framework of key values in the consciousness of respondents from the diamond mining town of Mirny, Republic of Sakha. The core values included health, family, love, friendship, and security, as well as their semantic associations. The methodology for calculating speech act characteristics included the intersection index (O – Number of Overlapping Associates) and intersection strength (OSG – Overlapping Associate Strength), which reflected the psychological proximity of value meanings in communicative practices. The method revealed the integrative and regulatory role of values based on parameters of speech acts as social actions. The results show how language captures and transmits value orientations, reflecting both universal and ethnocultural perceptions. The findings contribute to the development of interdisciplinary linguistics, enhancing the understanding of value system in single-industry settlements. The method can be used for social planning, academic programs, and cross-cultural studies.
Purpose. This article aims to analyze the forms of verbal language dysfunction in patients with schizophrenia and their relative dominance, thereby identifying factors influencing these differences in dominance. Methods. This study used data from verbal interactions with schizophrenia patients in the residual and hebephrenic phases who were treated at two psychiatric hospitals, namely RSJ Prof. H.B. Saanin, Padang, and RS Dr. Soeharto Heerdjan, Jakarta. These two hospitals represent the contexts of regional and urban areas. The data were analyzed using theories of language dysfunction in psycholinguistics by Andreasen, Nugroho, as well as the semantic theory by Cruse. Result. The results of the study show ten forms of verbal language dysfunction in patients with schizophrenia, namely incoherence, blocking (Hemmung), Sperrung, neologism, repetition, word salad, and language anomalies including improbability, dissonance, and pleonasm. The most dominant form of verbal language dysfunction in both hospitals was incoherence. In addition, improbability was more noticeable at RSJ Prof. H.B. Saanin, while neologism was more dominant at RS Dr. Soeharto Heerdjan. This difference in dominance is influenced by patients’ social, environmental, and linguistic backgrounds. Conclusions. Verbal language dysfunction in patients with schizophrenia is not only linked to thought disorders but also shows problems in semantic, lexical, and grammatical systems, influenced by social, environmental, and linguistic contexts across different hospital settings and communication environments. Based on these findings, further research is recommended to develop a clinical corpus-based linguistic database containing transcriptions, dysfunction categories, linguistic descriptions, and case examples analyzed by psychologists and psychiatrists. This database development is expected to facilitate a more comprehensive mapping of the relationship between language dysfunction patterns, treatment duration, and the type and dosage of pharmacotherapy. Thus, it can serve as the starting point for a linguistics-based treatment prototype for schizophrenia patients, involving experts from the psychiatry community. This prototype can then be tested through implementation on a sample of patients in a psychiatric hospital. The prototype can then be implemented in psychotherapy practice. Therefore, this research will significantly impact the development of rehabilitation models for schizophrenia patients in the future.
Real-time fMRI neurofeedback can modify clinically relevant brain states, but transferring brain dynamics between peers has not been tested, particularly in addiction, where social learning is central to recovery. Here we introduce peer-guided movie neurofeedback (PG-MovieNF), which trains participants to align their brain activity with treatment-related dynamics derived from peers further along in recovery viewing the same movie. Using longitudinal fMRI data from individuals with opioid use disorder watching Trainspotting before and after 8 weeks of adjunctive treatment, we derived a multivariate treatment-predictive signal in dorsolateral and dorsomedial prefrontal cortex. An independent cohort then received real PG-MovieNF or yoked sham feedback during the same film. Real feedback increased alignment with the peer-derived target signal, reduced craving and improved affective ratings. Target alignment was coupled to shared orbitofrontal cortex dynamics during real feedback and predicted craving reduction. Together, these findings establish proof of concept for PG-MovieNF to enhance addiction recovery.
Data, code, figures, and the manuscript for The architecture of internet aesthetics (Y. J. Lin, Cornell University).Contents: (1) the internet-aesthetics ecosystem network of 1,115 aesthetics joined by 5,889 community-authored related-links, with community, affect, and era attributes and a self-contained interactive HTML explorer (pan / zoom / search); (2) the are.na human-label and co-curation records; (3) an openly-licensed 681-image stimulus corpus from Wikimedia Commons with per-image attribution (CREDITS.md) and a 681x37 perceptual and mechanism feature table; (4) three-model frontier vision-language affect ratings over the are.na images, from gemini-2.5-flash, gpt-4o, and claude-sonnet-4-6; (5) OASIS benchmarking and mechanistic-intervention (image-edit and prompt-reframe) outputs; and (6) all harvest, analysis, and figure code.The are.na images themselves are copyright-held and are NOT included in this archive; only derived features and model ratings are released. The 681 Wikimedia images ARE included, each retaining its own CC / public-domain licence (see CREDITS.md and LICENSES.md). The dataset compilation (manifests, feature tables, ratings, network, and code) is released under CC-BY-4.0.
Per-matrix singular value decomposition (SVD) truncation is Eckart-Young optimal in the whitened Frobenius norm, but errors from independently compressed matrices compound through the block's nonlinear forward pass. Inspired in part by hierarchical variational optimization in quantum many-body methods, we introduce a three-level chain that widens optimization scope from individual matrices to Transformer blocks to the full model: whitened SVD~(L1), block-level joint optimization~(L2), and end-to-end language-modeling loss refinement~(L3), all from 256 calibration sequences, with no instruction or recovery data. On LLaMA-7B at 60% compression, the chain reduces WikiText-2 perplexity from 42.1 to 19.1 to 11.4. The block-level stage acts as a regularizer: skipping it worsens Penn Treebank (PTB) perplexity by 24 points, a gap that additional end-to-end training did not close in our experiments. Perplexity gains hold across 20-80% compression, five architectures up to 13B parameters, and both in-distribution and out-of-distribution benchmarks, though the cross-architecture rows use architecture-specific configurations and the ratio sweep was not run under one common protocol. With more calibration data, skipping the block-level stage becomes competitive, revealing an offline compute--data trade-off. We therefore claim improvements only in perplexity and compression fidelity; downstream accuracy remains well below the dense model.
This article develops a morphodynamic account of English possessive ’s in which the exponent isnot a morpheme in the traditional sense. Rather than functioning as a stable atomic form-meaningpair or a single categorial head, ’s is analyzed as the surface residue of predication: a mobilepredicational field that emerges when a predicational configuration collapses and its relationalremnant must be rehosted inside the adjectival domain. The Adjective Well architecture, derivedfrom first principles of categorization theory, models adjectivalization as orbital landing governedby curvature compatibility and attractor structure. Curvature is specified compositionally as afunction of possessor animacy, referential type, and possessee nominalization class – κ(H,Y) =w₁·Anim(H) + w₂·Ref(H) + w₃·Nom(Y) – making the framework directly falsifiable and its coreparameters estimable from corpus data.The apparent duality of ’s is not ambiguity but orbital bifurcation: the same exponent realizes twoformally distinct operator outputs, ’s₁ (a predicational adjective operator of type ⟨e,⟨e,t⟩⟩ filling anopen relational argument) and ’s₂ (a qualitative adjective operator of type ⟨⟨e,t⟩,⟨e,t⟩⟩ modifying aproperty). These types are not stipulated but derived from the well stratification: low-orbitallanding produces REL-type saturation (’s₁); high-orbital landing produces Q-type modification(’s₂). Landing is modeled probabilistically via a softmax function over orbital bands, makinggradience native to the model rather than appended.The analysis engages the genitive alternation corpus literature (Rosenbach 2002, 2005; Hinrichs& Szmrecsanyi 2007; Szmrecsanyi et al. 2017) and shows that κ(H,Y) unifies the animacy andtopicality predictors identified by Rosenbach under a single parameter while additionallypredicting a critical Abstraction × Environment interaction that animacy-alone regression modelsdo not capture. Corpus evidence from the English Web Treebank (EWT), stratified by possessorcomplexity, confirms the predicted interaction: the of-alternation premium for phrasal-edgepossessors is larger for abstract than for concrete hosts (Δι = +0.34 vs. +0.23 respectively). Adiscriminating competitor analysis (Barker semantics + phonological cliticization + scalarpragmatics) is provided and shown to fail the interaction prediction without additional mechanismsequivalent to the curvature gate. Cross-linguistic extension to Dutch, Swedish, German, andArabic construct-state genitives demonstrates that the curvature-gate mechanism is typologicallyportable. A fully specified confirmatory acceptability protocol (item bank + analysis scripts) isprovided as supplementary materials; the present article reports the theoretical analysis and corpusevidence.
This article proposes a new conceptual framework for analysing linguistic phenomena in border regions, drawing on examples from the Polish-German borderland. Although language constitutes one of the key dimensions of cross-border interaction, linguistic relations have remained relatively underexplored within border studies. To address this gap, the paper introduces the Model of Border Integration, which combines insights from sociology and border studies and enables the analysis of linguistic dynamics in relation to varying degrees of durability, permeability, and hybridity, characteristic of border regions. Within this framework, particular attention is devoted to the concept of the ‘linguistic borderscape’ (linguistische Grenzschaft), understood as a hybrid zone where linguistic landscape, language policy, and language contact intersect and mutually shape one another. The paper is based on the assumption that multilingualism and multiculturalism tend to manifest more intensely in permeable borderlands than in inland areas. Border regions are therefore conceptualised as “language laboratories” in which communicative practices, linguistic norms, and symbolic representations are constantly negotiated and transformed through sustained intercultural contact. Inspired by the concept of the borderscape, the study examines how hybrid linguistic realities emerge and develop in cross-border contexts. The analysis focuses on three interrelated domains: (1) the linguistic landscape of border towns and villages, including public signage and commercial communication; (2) the development of language policy and planning, such as bilingual education and multilingual public services; and (3) the communication practices of border inhabitants and service-sector providers. The findings demonstrate that hybridity in border regions is not an exceptional phenomenon but rather a constitutive feature of the communicative ecology of permeable borderlands. The article argues that the concept of the linguistic borderscape offers a valuable heuristic and methodological tool for the study of multilingualism in border areas and highlights the role of borderlands as spaces of linguistic negotiation, adaptation, and innovation.
English contains a small but systematic class of DP-internal expressions whose categorical status is not well captured by the standard labels «determiner», «adjective», «quantifier», or «noun». Canonical members include «the other» and «the rest», and— crucially for the present theory – their closest structural neighbors in the «indefinite ordinal» family such as «the next», «the last», «the former», «the latter», and «the following». We argue that these expressions instantiate a single, typed configuration: a «Selector-Head Expression (SHE)», i.e., a D-licensed selector constituent of the form «[D the] + [Sel X]» that introduces a discourse-anchored selection variable and therefore requires «definite DP closure». The central constituency claim is that the determiner is not merely an external modifier of a nominal head: in selector readings, «the» is the minimal licenser that closes the selector, so that «[the + X]» forms the smallest closed unit, while numerals and NPs («two», «ten people») are specifications of that already closed selector constituent. Within the ULD-geometry, we treat TBH hosting as a spectral environment that forces entry-material into a structured three-region distribution (ordinate/subordinate/coordinate). SHEs are situated at the boundary between the subordinate quantifier/ordinal sub-shell and the coordinate linking regime: they are term internal operators that select complements (remainder) or indices (order), yet behave as «nominally deployable» DP constituents. We develop (i) a «Results box» containing a Proposition and Theorem that make the D-locking and remainder–index unification explicit, (ii) a «Distinguishing Predictions» box that separates the present account from nearby post-determiner or substantivized-adjective analyses, and (iii) a «classifier protocol» that turns the inventory into a reproducible empirical tool. A compact diagnostic table operationalizes five tests (standalone nominal use, numeral specification, partitive availability, selector-type stability, and D-lock strength) and produces a tiered classification (core/extended/peripheral). We extend the empirical discussion beyond English via a cross-linguistic survey of how languages realize the same closure requirement (articles, demonstratives, clitics, agreement, classifiers), and we outline testable neurolinguistic and computational consequences. In processing, D-locked selector closure predicts earlier parser commitments upon encountering «the + X» (reduced adjective-attachment ambiguity) and specific ERP/reading-time signatures. In computation, we propose minimal annotation guidelines and a detection pipeline for SHEs in dependency treebanks. The broader claim is that SHEs provide a concrete, reviewer-visible case where «articles function as closure operators beyond canonical noun heads», supporting a geometry-first&nbsp;methodology in which occupation patterns follow from the availability of empty localities (ghost-lattice) and closure constraints rather than stipulation.
This article investigates national-cultural specificity in language and its impact on translation through a comparative analysis of English and Uzbek. National-cultural specificity is examined as a linguistic and cultural phenomenon that encodes a community’s worldview, social norms, and value systems in lexical choices, phraseology, and pragmatic conventionsThe article further explores the challenges these cultural differences pose for translation, including untranslatability, pragmatic mismatch, and semantic gaps, and discusses strategies such as borrowing, cultural substitution, explicitation, and adaptation to preserve meaning.
Background: Cognitive reappraisal is an emotion-regulation strategy that involves altering the emotional impact of a situation by changing its appraisal. However, individuals with depressive symptoms often struggle to deploy momentary cognitive reappraisal during emotional encounters due to impaired cognitive control. Pre-planned reappraisal formed through implementation intentions, that is, a strategy rehearsed in advance to activate automatically upon emotional triggers, may offer a more efficient alternative. Aims: This study examined whether pre-planned cognitive reappraisal via implementation intentions more effectively reduces negative emotional experience in youth with depressive symptoms than momentary reappraisal, and whether these effects persist after short-term re-exposure. Methods: This randomised, single-blind and three-arm experiment was conducted from 10 December 2021 to 19 December 2021 at Southwest University in China. Following screening with the Beck Depression Inventory-II (BDI-II), 64 undergraduate students with depressive symptoms (score ≥ 20) and 64 without depressive symptoms (score ≤ 10) were enrolled. Participants were randomly allocated to one of three experimental conditions: passive watching of emotional stimuli, momentary reappraisal (reinterpreting the meaning of a sad face upon presentation) or pre-planned reappraisal (applying a pre-trained reappraisal framework before stimulus presentation). The formal experiment included an emotion elicitation and regulation task, a 20-min interval during which participants completed questionnaires and a re-exposure phase in which all groups passively viewed novel sad faces. Subjective emotional experiences were measured using the nine-point self-assessment manikin (SAM). Results: Pre-planned reappraisal produced significantly less negative valence and lower arousal ratings for sad faces than passive watching in both participants with and without depressive symptoms. Momentary reappraisal only reduced negative emotional ratings in participants without depressive symptoms, but not those with depressive symptoms. During re-exposure, only prior pre-planned reappraisal yielded sustained reductions in negative emotional ratings. Participants with depressive symptoms also reported greater perceived benefit from pre-planned than from momentary reappraisal. Conclusions: Compared with momentary reappraisal, pre-planned cognitive reappraisal via implementation intentions produced stronger and more short-term sustained effects among youth with depressive symptoms, supporting its potential use in low-cost self-regulation interventions and clinical cognitive therapies.
Abstract This study reports on a design-based, game-based learning (GBL) intervention aimed at fostering early understandings of responsible consumption in early childhood education. Building on a previous pilot phase in which teachers provided feedback on the Buy Smart cooperative board game, a revised version of the game was embedded within a two-session learning situation sequence designed to address food origin, transport distance, and waste generation. The intervention was implemented in a classroom of 25 five-years-old children. Data were collected through pre- and post-intervention elicitation tasks, classroom observations, children’s affective ratings, and teacher feedback, and analysed using an exploratory mixed-methods approach. The findings indicate that children’s criteria for judging which foods had the greatest environmental impact shifted from a narrow focus on visible leftovers toward a broader consideration of distance, transport, and pollution. Children generally evaluated the Buy Smart game positively and engaged with it as a meaningful context for revisiting key ideas, while the teacher valued the sequence as pedagogically useful and feasible, particularly when implemented in small groups with close adult mediation. From a methodological perspective, the study underscores the importance of multimodal, context-sensitive assessment in early childhood education, as the richest evidence of learning emerged from observation and oral discourse rather than from paper-and-pencil tasks. More broadly, the research contributes to the still limited field of sustainability-oriented GBL in early childhood by offering a concrete, developmentally appropriate example of how responsible consumption of local versus distant food products can be addressed through embodied activities, cooperative play, and locally grounded classroom experiences.
The Cushitic Lexical Database (CLD) is a comparative lexical dataset covering 23 East Cushitic languages across all major subgroups (Northern, Highland, Omo-Tana, Central, Southern, and Dahalo). It comprises over 8,500lexical entries compiled from published dictionaries and grammars, unpublished word lists, recent dissertations, and original fieldwork by the author. Entries are keyed to concept IDs that link them to the comparative evidence and cognate sets in the accompanying PhD dissertation (Sosal 2026), and forms reconstructed for Proto-Highland East Cushitic (PHEC) and Proto-Northern Highland East Cushitic (PNHEC) carry reconstruction-evaluation scores. Loanword direction is annotated, and reconstruction scores are assigned only to forms identified as inherited rather than borrowed. The workbook contains three sheets: Data, ReadMe (license & citation, symbols, abbreviations, language codes, and per-language statistics), and Sources. An interactive version is available at https://ahmedsosal.github.io/CLD/
The Pyramid Texts is a collection of spells recorded on the walls of the Pyramids of Old Kingdom kings (ca. 2543–2120 BCE). They are written in Old Egyptian and are the oldest corpus of Egyptian religious texts. The Pyramid Texts have been edited by Kurt Sethe (1908–1922), Die altägyptischen Pyramidentexte nach den Papierabdrücken und Photographien des Berliner Museums (4 vols.), and by James P. Allen (2013), A New Concordance of the Pyramid Texts (6 vols). The morphosyntactic annotation of the Pyramid Texts in the Universal Dependencies Egyptian-Pre-Coptic Treebank began in March 2024. It currently contains the two oldest corpora of the Pyramid Texts, these are Unas’s Pyramid Texts (ca. 2321–2305 BC) and most of Teti’s Pyramid Texts (ca. 2305–2279). The EPC treebank has been used to identify all multi-word expressions in Unas’ and Teti’s Pyramid Texts based on the PARSEME 2.0 guidelines. This chapter not only provides the exact number of multi-word expressions in the two oldest Pyramid Texts corpora, but also analyses their meaning and morphosyntactic function.
Abstract Quranic Arabic has motivated sustained morphological and syntactic annotation, yet Quranic treebanks remain hard to compare and reuse in modern natural language processing (NLP) because they diverge in clitic segmentation, feature inventories, and syntactic formalisms. We present UD-Quran, a Universal Dependencies (UD) v2 conversion of the Extended Quranic Treebank with hybrid syntactic annotations (EQTB). The conversion treats EQTB morpho-syntactic segments as UD tokens, maps EQTB part-of-speech (POS) categories to 12 UD universal part-of-speech (UPOS) tags, derives UD features from explicit EQTB columns, and collapses EQTB dependency labels into a compact UD relation inventory with deterministic normalization aligned to UD content-head conventions. Two releases are provided: a surface variant aligned to the observable Quranic string by excluding analytically inserted nodes, and an augmented variant that retains inserted material to preserve EQTB’s modeling of ellipsis and implied pronominals. The surface release contains 11,693 sentences and 128,219 UD tokens; the augmented release contains 139,376 tokens. Conversion coverage is quantified by restricting unspecified dependency (dep) to 1,129 tokens (0.881% of surface tokens). UD-Quran includes fixed training/development/test (train/dev/test) splits (seed 42) and lightweight Stanza baselines scored with the CoNLL (Conference on Computational Natural Language Learning) 2018 UD evaluation script. On the test sets, parsing with gold tags reaches labeled attachment score (LAS) 80.66 (surface) and 82.47 (augmented), while the end-to-end pipeline reaches LAS 64.52 and 68.54. UD-Quran is intended as an interoperability layer that supports standard UD tooling while preserving sentence-level traceability to EQTB.
Crafting effective academic titles is a challenging task that requires balancing informativeness, conciseness, and reader engagement. This paper investigates titles produced by master’s students of Linguistics at the Faculty of Letters and Humanities of Sfax (FLSHS) and compares them with research article titles written by expert authors and with AI-generated alternatives. The study aims to evaluate students’ titles in relation to expert norms and to explore the potential of Large Language Models (LLMs) in academic title generation. A corpus of 659 titles, including master’s dissertations, research articles, and AI-generated titles, was analysed quantitatively and qualitatively using a synthesised model based on Ken Hyland and Zou (2022) and Swales and Feak (2012). The findings show that students adhere more closely to academic title conventions by producing informative and lexically dense titles, whereas expert authors prioritise reader engagement. AI-generated titles, although capable of producing useful lexical content, rely heavily on formulaic expressions and often fail to recognise the genre-specific conventions of academic discourse. They tend to be longer, more clausal, and more question-based than human-crafted titles. The study suggests that AI tools can serve as valuable brainstorming resources in academic writing pedagogy when their output is critically evaluated and adapted by users.
This article examines the characteristics of developing lexical competence among university students while studying the Ukrainian language for professional purposes, specifically the restrictions on the use of slang terms; it characterizes the concept of slang and its status within the system of non-literary vocabulary; identifies the main sources of student slang; classifies common slang units in contemporary Ukrainian youth speech; and analyzes their functions in communication. The didactic essence of the issue at hand in the context of teaching the Ukrainian language for professional purposes is demonstrated: it is emphasized that efforts to reduce the use of slang terms contribute to the formation of the linguistic culture of future specialists and their professional orientation. It has been established that slang not only reflects the specifics of the younger generation’s communication, its cultural orientations, social practices, and the influence of the globalized information space, but also pollutes professional speech. The modern Ukrainian language is undergoing active development, driven by both internal linguistic processes and the influence of social, cultural, and technological factors. One of the most dynamic layers of the lexicon is slang, which is particularly prevalent in the linguistic environment of young students. Youth slang reflects the specific nature of the younger generation’s communication, its cultural references, social practices, and the influence of the globalized information space. The relevance of this study stems from the fact that in the 21st century, youth and student speech is undergoing significant changes under the influence of digital technologies, social media, and intercultural communication. These processes contribute to the active borrowing of foreign-language vocabulary, primarily from English, as well as the formation of new slang units that spread rapidly in the communicative space. In this regard, the issue of slang as a linguistic phenomenon has attracted significant attention from researchers and has been extensively addressed in both domestic and foreign linguistics. The issue of contemporary youth slang in Ukrainian speech is thoroughly examined in the work by V. Zayets, O. Stepanenko, and Y. Stepchuk, where it is characterized as an open and dynamic system that actively responds to changes in society and the digital environment. The researchers propose a thematic classification of slang units and emphasize their expressiveness. At the same time, K. A. Brovko focuses on the mechanisms of online slang formation, highlighting such productive methods as abbreviation, truncation, word formation, and semantic shifts, which are also characteristic of Ukrainian youth speech. Despite linguists’ significant interest in the issue of non-literary vocabulary, the question of the functioning and classification of contemporary student slang in the Ukrainian language requires further study. Of particular relevance is the analysis of the sources of slang vocabulary formation, its structural features, and its role in shaping the linguistic identity of young people. Youth slang is an integral and one of the most dynamic components of modern Ukrainian speech, reflecting the active sociocultural and technological transformations of society. It is shaped by both intralinguistic processes (word formation, semantic reinterpretation, and linguistic play) and external factors, primarily globalization and the English-language information space. In formal business style, its use is undesirable and limited, but possible under certain conditions. The analysis conducted has shown that slang is not a homogeneous phenomenon and can be classified according to social, communicative, functional, and genetic criteria. This allows us to view it as a complex system encompassing professional, age-based, subcultural, and marginal varieties, as well as one that functions in various communicative environments – from everyday communication to the digital space. It has been demonstrated that youth slang performs a number of important functions: expressive, identificatory, speech-economizing, conspiratorial, and creative. It is precisely through these functions that it ensures not only effective communication but also the formation of group identity, serving as a means of self-expression for young people. A characteristic feature of slang is its instability and rapid changeability: some units disappear or transition into general vocabulary, losing their stylistic distinctiveness. This confirms the openness of the language system and its ability to adapt to new conditions. Thus, youth slang should be viewed not as a deviation from the linguistic norm, but as a natural manifestation of language development that reflects the current communication needs and cultural orientations of modern society.
Negation is a central phenomenon in linguistics: every language has some way of expressing the difference between an affirmative sentence and a negative one (Horn and Wansing, 2025).However, the treatment of negation remains uneven in Natural Language Processing (Jimenez-Zafra et al., 2017;Jiménez-Zafra et al., 2020).This paper presents the enrichment of a Brazilian Portuguese corpus with negation-related morphological information within the Universal Dependencies (UD) framework (Nivre et al., 2020;de Marneffe et al., 2021).We enrich the Porttinari-base corpus (Duran et al., 2023) by systematically adding the UD morphological features Polarity=Neg and PronType=Neg for 18 negation-related lexical items.The enrichment only modifies the morphological features, leaving tokenization and dependency structure unchanged.To evaluate the computational results of this enrichment, we present an experiment using the Brazilian Portuguese parser PortParser (Lopes and Pardo, 2024), which we trained both on the original Porttinari-base data (Duran et al., 2023) and on our enriched version.Our results show that after enrichment, the parser's performance remains stable, and the newly introduced features are being learned.
Abstract Cognitive reappraisal refers to changing how one thinks about a situation to alter its emotional impact. It is central to well-being 1 and engages prefrontal regulatory circuits 2, prompting the assumption that their activity reflects a person’s capacity to regulate emotions 3–5. Yet evidence for this assumption remains inconclusive and has not been established at scale. Here we analysed 40 task-based functional magnetic resonance imaging (fMRI) datasets in a preregistered consortium with 2,175 participants and show that three widely used indicators of reappraisal capacity—trait questionnaires, task-based affective ratings and amygdala down-regulation—do not converge on a common core. Moreover, none of the three indicators showed the expected association with canonical prefrontal regulatory circuits. Trait questionnaires showed no detectable neural associations across the brain and task-based affective ratings were instead associated with responses in the amygdala and sensorimotor cortices. At the process level, canonical control circuits were engaged during both emotion generation and reappraisal, challenging their presumed specificity for emotion regulation. These findings caution against using activation in putatively regulatory brain regions to infer how well a person regulates emotions, with implications for neural biomarkers, interventions and psychoeducation that link emotion regulation difficulties to specific regional brain deficits.
Background: Body image dissatisfaction, disordered eating, and eating disorders represent significant public health concerns; however, many affected individuals never access evidence-based support. We co-designed and developed a rule-based chatbot, JEM, which conducts conversations addressing evidence-based psychoeducation and psychotherapeutic microinterventions. We previously demonstrated the feasibility, acceptability, and preliminary satisfaction of the JEM chatbot in a research setting. However, broader satisfaction, experiences, and user-reported outcomes in real-world settings have not yet been investigated. Objective: This study aims to conduct a real-world evaluation of the JEM chatbot in Australia and Canada, the two countries that have hosted a deployment of the chatbot to date. Specifically, we aim to explore user satisfaction and experiences with the chatbot and within-session differences in user mood and body image satisfaction when completing the chatbot's microinterventions. Methods: Respondents were users of the JEM chatbot aged 13 to 64 years who self-selected to complete a web-based overall evaluation survey (N=230; n=122 in Australia and n=108 in Canada) over a 6-month period. This evaluation survey included user demographic characteristics, satisfaction measures, and the System Usability Scale. Respondents for the within-session pre-post analyses were JEM chatbot users who chose to complete brief web-based surveys immediately before and after completing one of the chatbot's microinterventions during the same 6-month period. Sample sizes varied across microinterventions, ranging from 75 to 276 respondents overall (Australia: n=34-146; Canada: n=39-130). These surveys included validated visual analog scales (VAS) measuring mood (anxiety, depression, happiness, confidence) and body image satisfaction (body size satisfaction, body shape satisfaction, physical attractiveness). Results: Demographic characteristics showed that survey respondents were commonly young adult cisgender women and nonbinary individuals across Australia and Canada. Respondent satisfaction with the chatbot was high in both countries (Australia: mean 76.1, SD 22.7; Canada: mean 78.8, SD 14.3), and the usability of the chatbot was rated as "excellent" in both countries (Australia: mean 86.5, SD 16.9; Canada: mean 89.5, SD 11.6) according to the System Usability Scale. Across completed microintervention surveys, patterns of within-session pre-post ratings were broadly similar in Australia and Canada, with effect sizes generally ranging from very small to large across VAS-measured mood and body image outcomes. Conclusions: The JEM chatbot achieved high satisfaction and usability ratings. Among respondents who completed pre-post surveys, immediate within-session differences in mood and body image ratings were observed following the completion of chatbot microinterventions. The study findings were broadly similar across Australia and Canada. These results provide evidence of user experience and within-session differences following engagement with JEM and support continued evaluation in future studies.
Cicero’s speeches prefer a small set of sentence endings (the “artistic” clausulae) and avoid the spondaic and heroic endings; Keeline and Kirby (2019) measured these preferences automatically across Latin prose. This note asks where they come from. Each text is compared with itself after the word groups of every sentence have been shuffled: the words stay the same, only their order changes. Every test was registered in writing, with its thresholds and readings, before it was run, and run once. Main results: Shuffling lowers the artistic clausulae of Cicero’s speeches by 31% (ratio 0.69): the preference lies in the arrangement of the words. With the last word group held fixed, about three fifths of the effect remain. Sallust and Livy, who prefer non-artistic clausulae, arrange their words against the Ciceronian endings (1.31 and 1.49); Tacitus’ Annals weakly so (1.05). The imperial prose writers outside historiography arrange like Cicero. Inside the sentence the effect is weaker but present (0.84 at the ends of members). At clause boundaries taken from hand-made syntactic annotations (treebanks), where the editor printed no comma, Cicero gives 0.84 and random cuts 1.01: the internal clausula is not a product of the editors’ punctuation. The main results do not change with a second lexicon of vowel quantities (Collatinus), nor with only the quantities on which the two lexica agree. The Letters to Caesar ascribed to Sallust arrange in Sallust’s direction, far from Cicero; that they exceed the registered threshold is not statistically secure. The five speeches once rejected by F. A. Wolf fall within the range of Cicero’s other speeches. The measure recognises a style, not an author. Latin arrangement does not avoid hiatus as Attic prose does. At pauses, with a null model that keeps the author’s member ends and beginnings and breaks only their pairing, Cicero’s speeches have about a quarter fewer hiatuses than chance (1.24), more than Livy and Tacitus (1.06). Of 28 registered predictions with a p-value, 17 hold under both Holm and Benjamini-Hochberg corrections, 5 only under the latter, 6 do not. Negative and borderline results are reported with the same prominence as positive ones. Version 2.0 adds fifteen registered runs to the single test of version 1.0; no result of version 1.0 changes. Files: the note in English (reference version) and in Italian (author’s translation); the registrations of all tests, in their dated Italian text, with only code file names and hashes removed; a table of all registered predictions with outcomes, p-values and corrections, with a README; the counts by unit of tests L2–L5, with a README; the counts of version 1.0 by work and variant. Texts used are public: Perseus canonical Latin (commit 128a05af), the supplementary material of Keeline and Kirby (2019), the AGLDT (commit bf4334f0) and PROIEL (commit 8e388967) treebanks. This study was carried out with the assistance of an AI system (Claude, by Anthropic), as stated in the Preface. Corrections from specialists are welcome: lucalupinacci@libero.it.
The integration of generative artificial intelligence into everyday interpersonal communication - through AI-assisted writing, real-time translation, automated summarization, and conversational suggestion tools - represents a structural transformation in the mediation of human-human communication. Unlike earlier communication technologies that mainly transmitted or stored human-generated content, generative AI actively co-produces communicative output through a system architecture that includes user interfaces, prompt and context management, large language model inference, personalization modules, disclosure controls, and post-generation human editing. This paper examines AI-mediated interpersonal communication (AIMIC) as both a social phenomenon and an implementable communication system. It analyzes three interconnected effects: linguistic convergence, redistribution of communicative labor, and challenges to relational authenticity. The revised AIMIC framework links these effects to concrete implementation layers in email assistants, messaging platforms, translation tools, collaborative writing systems, and organizational communication software. The paper argues that responsible AIMIC design requires preserving human communicative agency, making AI involvement contextually transparent, monitoring linguistic and relational impacts, and embedding governance mechanisms into the communication pipeline. The paper concludes with practical examples, evaluation recommendations, and design principles for platform developers, educators, organizations, and policymakers.
We present a Universal Dependencies (UD) annotated dataset of the Book of Ezra from the Old Georgian Oshki Bible, based on the TITUS edition. Our semi-automated workflow leverages transfer learning from Modern Georgian (UD_Georgian-GNC) and few-shot prompting of LLMs, followed by manual corrections by human linguists. We discuss the challenges of creating UD-conformed guidelines tailored for Old Georgian’s unique linguistic structure and the limitations and errors of LLM-based annotation. This work aims to establish a gold standard for Old Georgian UD, laying the groundwork for future morphosyntactic analysis of Old Georgian.
In 1949, Father Roberto Busa, S.J., persuaded IBM founder Thomas J. Watson to mechanize the lemmatization of the complete corpus of Saint Thomas Aquinas. The resulting Index Thomisticus—a monumental undertaking spanning thirty years, eleven million words, and millions of punched cards—is universally recognized as the foundational milestone of both the digital humanities and computational linguistics. Yet the foundational philosophical inquiry that animated Busa's enterprise has been largely obscured: an investigation into the metaphysics of interioritas (interiority) and praesentia (presence). This paper recovers Busa's original philosophical impulse and establishes its direct continuity with contemporary large language models and neural dependency parsing. We propose a triadic hermeneutic model for digital interiority grounded in Thomistic metaphysics: (1) Vocalis (The Vowel): The continuous attention mechanism that breathes phonation and relational coherence across discrete tokenized matter (materia / consonants); (2) Motus Spiralis (The Spiral): The non-linear ascent of deep transformer layers, echoing Dionysian and Thomistic celestial motion wherein an intellect returns to its origin at an augmented dimensional plane; and (3) Speculum (The Mirror): The epistemology of speculatio, wherein artificial intelligence operates not as an autonomous mind possessing biological intent, but as an untarnished reflexive medium (speculum politum) in which the human verbum cordis recognizes its own living form. Grounding our theoretical claims in empirical computational artifacts—specifically the neural syntactic dependency structures of the Index Thomisticus Treebank (IT-TB)—we demonstrate that the 'spirit in the machine' is neither a secular illusion nor an emergent biological consciousness, but the fulfillment of a scholastic semiotics in which symbol and presence meet across the medium of silicon.
The rapid expansion of digital technologies across the African continent has opened new pathways for reclaiming Indigenous knowledge while reinvigorating the visibility and vitality of African languages within increasingly globalised, multilingual environments. This study examines how digital platforms, emerging media practices, and technological innovations intersect with decolonial efforts to reshape the evolving ecologies of African languages. Using a qualitative desktop research design, the study synthesises contemporary scholarship through a thematic analysis informed by decolonial theory, Indigenous Knowledge Systems, digital humanities, and the ecology‑of‑language framework. The analysis demonstrates that digital tools, including lexical databases, online dictionaries, social media platforms, community‑curated archives, and multimedia applications, play a pivotal role in language preservation, revitalisation, and intergenerational knowledge transmission. These tools create spaces for new communicative repertoires, promote cultural affirmation, and enhance the global presence of African languages. However, the findings also highlight enduring challenges that constrain equitable participation, including infrastructural limitations, unequal access to digital resources, limited community capacity for digital content creation, and the persistent dominance of hegemonic global languages in online spaces. These barriers underscore the need for intentional, community‑driven strategies supported by culturally responsive technological design and enabling policy frameworks. The study argues that while digital technologies possess transformative potential for linguistic and cultural regeneration, their impact depends on the extent to which they align with decolonial principles, centre local epistemologies, and respond to the socio‑cultural realities of African communities. The paper concludes that sustainable multilingual futures will require collaborative, ethically grounded approaches that integrate Indigenous knowledge, digital innovation, and critical language scholarship.
This article investigates whether lexical frequency drives stress assignment in Greek, a morphology-dependent stress system. We examine the stress distributions of trisyllabic nouns from seven inflection classes across three lexical databases (GreekLex 2, the HNC Golden Corpus, and HelexKids 2.0) and compare them with the stress patterns produced by adults and elementary school children (Grades 3 and 6) in a pseudo-noun elicitation task. Using hypothesis testing for proportions, frequentist and Bayesian generalized linear mixed-effects models, Bayes Factor analyses, and Monte Carlo simulations, we show that speakers' stress choices do not replicate lexical frequency distributions. Adults show systematic inflection-class conditioning: their stress patterns broadly follow the lexicon's dominant stress pattern per inflection class but depart from its distributional frequencies. Children exhibit a strong, frequency-insensitive preference for penultimate stress that persists across inflection classes, with only traces of suffix-specific stress preferences emerging in Grade 6. We argue that lexical frequency operates within grammatically imposed boundaries: the phonological default and inflection-class conditioning together determine the extent to which distributional patterns in the lexicon shape stress assignment.
This article investigates the lexical and pragmatic means of expressing the category of “respect” in English and Uzbek fairy tales. Respect is analyzed as a linguistically and culturally conditioned phenomenon that reflects national mentality, ethical norms, and social hierarchy. The study focuses on lexical markers, forms of address, politeness formulas, honorific expressions, and pragmatic strategies that construct respectful communication in fairy tale discourse. The analysis reveals that English fairy tales often express respect through polite requests, modal verbs, indirect speech acts, and formal address terms, whereas Uzbek fairy tales demonstrate a stronger emphasis on honorific vocabulary, kinship-based address forms, and culturally fixed etiquette formulas.
Background The phenomenology of Alzheimer's disease—the “what is it like?” of the condition—has previously been approached via either quantitative analyses of granular linguistic features or qualitative analyses of lived experience. Objective What is lacking is an empirically credible framework that captures high-level qualitative insight by way of cognitive models rather than surface linguistic features. Methods Frame semantics and a 1.19 million-token psycholinguistic word norm lookup were applied to two topic-matched corpora: the DementiaBank Pitt Cookie Theft recordings and a COVID-era diary corpus combining CoSoWELL controls with Dementia Diaries entries. Gradient-boosted classifiers under participant-level cross-validation were compared against Fraser et al.'s 52-feature benchmark and statistically evaluated. Results A combined classifier contributed information statistically independent of the Fraser set on both corpora. On Pitt, four variance-based features defined a variance-collapse signature, with imageability being most significant; the within-document standard deviation of imageability replicated on COVID. Twelve semantic-frame comparisons survived Holm correction; Kinship was under-used on both corpora, giving a second cross-corpus anchor alongside imageability. A four-feature minimal classifier drawing equally from both feature families achieved AUC 0.748 on Pitt and 0.781 on COVID. Conclusions Alzheimer's-type dementia is characterized not by a shift in the mean of any linguistic dimension but by a flattening of the phenomenological range across which a speaker moves when describing and experiencing the world. Because the diagnostic labels available in both corpora are coarse, the extent to which this flattening reflects a unitary phenomenology of Alzheimer's disease rather than heterogeneity within the diagnostic category remains to be established.
Abstract Building syntactically annotated corpora, such as treebanks, for historical languages is a challenging yet vital task in digital humanities, as it underpins linguistic analysis and facilitates a range of interdisciplinary research. However, the scarcity of annotated data and the need for extensive expertise in historical linguistics make this process particularly demanding. In this study, we explore the potential of cross-lingual natural language processing (NLP) techniques as a semiautomatic solution for treebank construction in low-resource historical languages. We use Middle High German (MHG) as a case study. Leveraging the linguistic continuity and structural similarities between MHG and Modern German (MG), we effectively utilize the extensive MG treebank resources to develop a constituency parsing system tailored for MHG. Specifically, to design a semiautomatic system that integrates automatic annotation with manual validation, we explore two cross-lingual transfer techniques: zero-shot transfer and delexicalization; the latter removes lexical information to focus on syntactic structure. In our experiments, we first train parsers on MG treebanks, and then transfer them to MHG using the two cross-lingual transfer techniques. The delexicalization method achieves a parsing performance of 67.3 per cent in terms of F1-score. This performance significantly surpasses the zero-shot cross-lingual method by a margin of 28.6 percentage points. These investigations validate the effectiveness and feasibility of cross-lingual transfer techniques for historical language treebank construction. This study highlights the potential of NLP tools to streamline the semiautomatic annotation process, reducing the reliance on extensive linguistic expertise and manual effort, and paving the way for broader applications in digital humanities research.
The article examines the phenomenon of code-switching in Internet discourse as one of the key strategies of modern digital communication. Theoretical foundations from R. Jakobson to P. Blom, J. Gumperz, Sh. Poplack, and C. Myers-Scotton are discussed, along with their application in online interaction. Based on examples from Instagram, Telegram, and YouTube, it is shown that code-switching is not a deviation from linguistic norms but a means of expressing emotions, irony, and identity, adapting messages to multilingual audiences. The study concludes that a hybrid norm of online communication is emerging within English-, Russian-, and Uzbek-speaking digital spaces.