We present RoBaseLM-S (125 M) and RoBaseLM-M (260 M), two compact Romanian decoder-only language models trained from scratch on a 4.3 B-token curated corpus. Architecturally, they follow a modern LLaMA-style recipe with pre-norm RMSNorm, rotary position embeddings, SwiGLU feed-forward blocks, grouped-query attention, and 4 k-token context windows. We release both full-precision (FP16) and post-training 5-bit (Q5_K_M) checkpoints in GGUF format for lightweight local inference. The 5-bit variants fit under 500 MB and generate text in real time on a Jetson Nano 4 GB, enabling fully offline Romanian text generation on consumer-grade edge hardware. We evaluate the models intrinsically (multi-domain perplexity across news, literary prose, poetry, and heterogeneous web text) and extrinsically (LaRoSeDa sentiment classification and RO-STS sentence similarity). Relative to Romanian GPT-2–style baselines at similar parameter scales, RoBaseLM-S and RoBaseLM-M reduce perplexity substantially, e.g., from 30.7 to 15.9 on our held-out news split. The 5-bit post-training quantized checkpoints remain within FP16 performance across all reported tasks. To our knowledge, these are the first Romanian small language models explicitly optimized for long-context inference, post-training quantization, and low-power on-device deployment.

This study examines how international students learning Romanian interpret and apply the Cooperative Principle in everyday and academic interaction. The research is grounded in the observation that pragmatic competence often develops unevenly in second-language learning, particularly in multilingual environments where learners rely on norms carried over from their first language. To investigate these dynamics, a small spoken and written corpus was compiled from classroom activities, recorded peer interactions, and informal conversations with students enrolled in Romanian language courses. The data were annotated for instances of maxim observance, weakening, and flouting, as well as for implicatures that required contextual inference. The analysis shows recurring patterns of pragmatic transfer, especially in the interpretation of relevance and quantity, and highlights areas where learners systematically misinterpret or underproduce implicatures. Several examples also reveal successful adaptation to Romanian communicative expectations, suggesting that exposure to diverse interactional settings supports the refinement of pragmatic cues. The findings contribute to a clearer understanding of how the Cooperative Principle operates in cross-cultural learning contexts and point to practical implications for teaching Romanian as a foreign language.

Globalisation and rapid technological expansion continue to reshape linguistic structures, identities, and communicative functions across Romance and Germanic languages. These processes intensify multilingual interaction, alter digital visibility, and transform the relationship between demographic prominence and technological representation. The present study aims to examine how linguistic identity, digital inclusion, and structural resilience evolve under these global dynamics, with particular emphasis on the changing position of European languages in the digital environment. This research is designed as a systematic review following PRISMA principles, supplemented by secondary statistical analysis. While global indicators are used to position Romance and Germanic languages within the broader digital hierarchy, the analytical focus centres primarily on the European linguistic context. Therefore, the distinction between global and Europe-specific datasets is explicitly maintained throughout the study to ensure scope consistency and interpretive clarity. The review covers publications from 2018 to 2025 and draws on databases including Scopus, Web of Science, LLBA, and ERIC. It incorporates transparent inclusion and exclusion criteria, multi-stage screening, and a coding protocol grounded in linguistic, sociocultural, and technological indicators. Statistical data were retrieved from UNESCO UIS, W3Techs, Eurostat, and Ethnologue, with validation conducted in January–February 2025. The results show that while English maintains disproportionate dominance in digital spaces, Romance and Germanic languages demonstrate strong adaptive capacity through structural stability, hybridisation practices, and expanding digital resources. The synthesis highlights enduring grammatical resilience, uneven digital representation, and the importance of multilingual educational ecosystems. These findings indicate that sustainable linguistic development depends on inclusive digital infrastructures and equitable technological support that preserve linguistic diversity while enabling global participation.

ABSTRACT The study explores how translation can be used as part of pedagogical translanguaging to facilitate multilingual learners’ understanding of the way tempo-aspecutal meanings are conveyed. Two sets of data are analysed, focusing on tense-aspect marking: (1) a novel excerpt originally written in Italian is compared with its professional translations in Romance (French, Spanish) and Germanic (English, German) languages; (2) retranslations from these languages to Italian, provided by L1 speakers of Italian and learners of Italian with French, Spanish and German as their L1 are compared with each other and with the original Italian text. Professional translations provide a window to crosslinguistic variation in the use of tense-aspect marking and to complex linguistic phenomena such as the aoristic drift in Romance languages. Retranslations from these languages back to Italian capture some differences in the way speakers develop a tempo-aspectual representation of situations. Similar analyses, guided by the teacher, can be conducted in multilingual classrooms as part of pedagogical translanguaging. Analysing the same text passage in multiple languages and engaging in its retranslations fosters crosslinguistic awareness and promotes joint exploration and co-construction of linguistic knowledge.

After World War II, Romanian-Polish relations developed under the influence of Moscow, both states becoming part of the socialist bloc. Academic cooperation was formalized through bilateral treaties, a significant example being the 1970 agreement between the „Al. I. Cuza” University of Iaşi and the „Maria Curie-Skłodowska” University of Lublin. This partnership facilitated the sending of lecturers from Iaşi (such as Ioan Sârbu, Ioan Lobiuc or Natalia Cantemir) to the universities of Lublin and Krakow. The role of the professors from Iaşi was a complex one. First, it is the educational and cultural one that involved teaching the Romanian language, history and literature, along with organizing events to promote Romanian culture. Second, it is the political role, because at that time the lecturerships also served as instruments of state propaganda. Although the 1970s were a period of intense academic growth and exchange, the following decade marked a decline due to economic difficulties and bureaucratic barriers. Despite these obstacles, the activity of the lecturers was considered fruitful, managing to train specialists and maintain a constant interest in Romanian culture in Poland.

In this paper we explore the interaction between Maghrebi Darijas and Romance languages from the perspective of both historical and contemporary evidence. As Caubet 2002, following Lahlou 1991 argues, among contemporary educated populations in urban centres across the Maghreb, bilingual/multilingual interaction is the norm not the exception. Historical evidence tells us this was also the case in 11th century al-Andalus, though it is of course impossible down the centuries to reconstruct the actual interaction. We will however argue that certain surviving texts can provide an indication when analysed in terms of the constraints on conversational codeswitching such as is provided by Aabi. We start from the position that Maghrebi Darija is a special case of linguistic permeability due to its politico-geographic location on the frontier and given its thousand year history of close contact with Romance. To investigate this phenomenon in both its historical and contemporary manifestations we draw on the current construct of translanguaging, an alternative perspective on multilingual interaction to code switching as expounded in Baynham & Lee (2018) and the notion of convivencia as elaborated by Bossong in his study of linguistic conviviality and coexistence in mediaeval Andalusian poetry (Bossong 2010). We then go on to analyze this in two time slices: examining evidence of the productive convivencia/coexistence of romance and dialectal Arabic i) in the kharjas of 11th century al-Andalus as discussed by Bossong and others and ii) in the modern Maghreb music and performance scene (cf. Caubet 2002; Baynham & Gintsburg 2022). We do this here through analysis of a song by the Algerian singer Talyani and a performance of the Moroccan comedian Hanane el-Fadhili, using in both time slices translanguaging and Bossong’s notion of convivencia in our analysis. We then conclude by arguing as Heath (2020) does that for effective research into such varieties as Maghrebi Arabic, both currently and historically, it is necessary for cross disciplinary work between researchers in Arabic and its Romance contact languages, in order to fully address its sociolinguistics. We understand this as a form of disciplinary translanguaging to be undertaken in order to establish the dynamics of the convivencia/coexistence of Arabic and Romance elements in this type of data.

Research objectives: This case study aims to describe different modalities of inclusive language, i.e., gender-balanced, gender-sensitive, or gender-neutral language, which are used in the official communication of Comenius University Bratislava. We link these modalities with the results of an online survey carried out in August 2024, which showed how native Slovak speakers perceive inclusive alternatives for generic masculine nouns. Research method: A critical text analysis was performed on online statements of the faculty members of Comenius University Bratislava and the online survey on inclusive language. Process of argumentation: In the past decade, gender-sensitive or inclusive language has permeated the communication of many actors in civil society, be it the media, public institutions, or various associations. It reflects the post-structuralist idea that language is a tool of power and makes women invisible, especially in professional life and various other roles. In Romance and Slavic languages, the “invisibilization” is caused by generic masculine nouns. The implementation of forms other than the generic masculine is based on psycholinguistic research (mainly in the German language). According to these researchers, this leads to women becoming invisible in the language. Comenius University Bratislava adopted a gender equality plan, financed by EU-funded programs. The use of inclusive or gender-sensitive language is a part of this action plan. Research findings and their impact on the development of educational sciences: The critical analysis of online statements of Comenius University Bratislava and its faculty members shows that only in the Faculty of Arts and the Faculty of Pedagogy is gender-sensitive language systematically used (doublets or neuter nouns). The main information channel of the University and the profile of the Faculty of Law almost never apply inclusive language. The online survey which was carried out in August 2024 shows that the majority of Slovak native speakers consider the variants of gender-sensitive language (both lexical and graphical) to be strange and redundant. Conclusions and/or recommendations: According to the survey, the generic masculine form is perceived by the majority of Slovak native speakers as a neutral form which denotes both genders. These findings could be correlated with the use of inclusive language (especially split forms) in statements on the social networks of different faculties of Comenius University Bratislava. The reasons for this situation could be the fact that the administrators who author the content are not informed about the action plan adopted at the university level (about the use of inclusive or gender-balanced language) and continue to use the classical generic masculine forms of nouns because the split form bears the characteristics of non-conventionality. A major shift in the use of gender-sensitive language could be enacted by a top-down approach in academia, but this will not solve the practical questions related to its functionality in the stylistics of the Slovak language.

BackgroundEvidence consistently indicates a disadvantage in bilinguals' speech production compared to monolinguals in healthy individuals, particularly for low‐frequency words. However, research on this phenomenon in clinical populations, such as those with Mild Cognitive Impairment (MCI) and Alzheimer's Disease (AD), remains scarce. This study investigates the impact of bilingualism on naming within the Catalan‐Spanish language pair, two typologically similar Romance languages, while also comparing active and passive bilinguals.MethodA total of 124 patients with MCI, 66 patients with AD, and 58 healthy older adults participated in the study. Participants were classified as either Catalan‐Spanish early and highly proficient bilinguals (active bilinguals) or Spanish‐dominant speakers with low proficiency in Catalan (passive bilinguals). A picture‐naming task in the participants’ dominant language was used, with stimuli selected based on frequency (high vs. low) and phonological overlap (cognates vs. non‐cognates). Naming performance and accuracy were analyzed using linear mixed models and logistic regressions, respectively.ResultFor naming latency, the frequency effect (high vs. low) increased significantly across groups. There was a 521 ms increase (95% CI: [423, 691], z = 10.46, p < .001) from healthy older adults to MCI participants and a 580 ms increase (95% CI: [481, 678], z = 11.06, p < .001) from MCI participants to AD patients. For accuracy, the frequency effect also increased significantly between groups, with a 6.7% decrease in accuracy (95% CI: [5.3, 8.2], z = 9.97, p < .001) from older adults to MCI participants and a further 14.0% (95% CI: [11.9, 16.0], z = 13.58, p < .001) from MCI participants to AD patients. However, the frequency effect was not influenced by the type of bilingualism.ConclusionThe results from Catalan‐Spanish bilinguals with MCI and AD do not align with the weaker link hypothesis or competitive models of bilingual language production. A possible explanation lies in the language context: passive bilinguals, who are regularly exposed to a second language, may differ fundamentally from monolinguals examined in prior studies. This continuous exposure to a second language could influence cognitive and linguistic processing in ways not accounted for in monolingual comparisons.

This paper explores the integration of cultural elements in the instruction of Romanian as a foreign language, emphasizing the interplay between language, culture, and intercultural competence. Drawing on prior empirical research and a comparative analysis of Romanian language textbooks, the study reveals significant deficiencies in the representation of cultural content and the predominance of textual over multimodal cultural forms. Findings indicate that teachers frequently serve as the primary mediators of cultural knowledge, compensating for the lack of comprehensive cultural materials. The paper argues that authentic cultural content, particularly artistic, folkloric, and musical dimensions, must be systematically integrated into Romanian language teaching through innovative, multimodal, and student-centered approaches to foster intercultural awareness and communicative competence.

In the context of globalization and increased mobility, plurilingual and pluricultural competences are essential for communication and integration. Teaching Romanian as a foreign language can leverage these competences through activities that promote intercomprehension, mediation, and linguistic transfer, in line with the CEFR – Companion volume descriptors. The paper outlines five types of activities designed to foster the developments of students’ plurilingual and pluricultural competences. Exercises such as linguistic portraits enable learners to raise awareness of their plurilingual skills, while interactive games and multilingual greetings encourage language comparison and recognition of shared elements. Plurilingual comprehension tasks and parallel texts foster meaning deduction and structural analysis, consolidating vocabulary and grammar while developing intercultural sensitivity. A set of 65 Romanian words similar to Slovene, paired with images, supports lexical acquisition, correct pronunciation, and plurilingual strategies. These activities enhance autonomy, confidence, and intercultural cooperation, demonstrating that the proposed methods facilitate both Romanian language learning and the development of plurilingual and intercultural competences for active international communication.

The present study aims to introduce the book Ducere de mână către cinste și direptate, adecă la copii rumuneştii neuniții cei ce în şcole cele mici să învață spre cetanie rănduită carte [Leading by the Hand towards Honour and Righteousness, that is, a Book Appointed for the Teaching of Reading to Orthodox Romanian Children in the Primary Schools], located at the Matica Srpska Library in Novi Sad, Serbia, in the library's digitized collection, in the Foreign 8section, 18th-19th centuries. Its author is Johann Ignaz von Felbiger, the most important pedagogue and reformer of the education system during the time of Empress Maria Theresa. The first part describes the cultural context in which the education reform took place in the Habsburg Empire and the personality of Felbiger, who managed, through his monastic and Augustinian formation, to create an era, that is, a true movement to improve education, from the establishment of schools, the issuance of laws, the employment of a rigorously selected staff, the writing of textbooks, methodology and pedagogy books, with the full support of Empress Maria Theresa. The second part of the study is a linguistic analysis of the language at morphological level, the book belonging to the period 1640-1780 and reflecting its linguistic particularities. The study contributes to the broadening of investigations on the Romanian language of the second half of the 18th century, offering numerous examples of grammatical categories still in the process of consolidation. The linguistic analysis retains the most important facts regarding the norms of the era and the changes that occurred from a morphological point of view during the researched period. In the last part, the study presents some new aspects of the book (principles of pedagogy, rhetoric), opens some new directions of research (the book contains a true code of good manners from the 18th century and footnotes), which could be interesting for the fields of pedagogy, anthropology or philology. The book Ducere de mână către cinste și direptate is a complex text, not a simple ‘reading book’, as its author describes it, and can be studied from different and interdisciplinary perspectives.

The Italo-Albanian communities of Southern Italy were formed in the 15th century when whole populations migrated from the territory of Southern Albania due to Turkish pressure. Today we find 50 communities in which Albanian variety is still spoken. The history of these communities is lively and engaged on the political-literary front: the Arbëreshë intellectuals contributed with their ideas and their work to the Italian Risorgimento and to the Albanian Rilindja. The Italo-Albanian varieties are of the Tosco type; they faithfully maintain the original morphosyntactic and phonological structure, while presenting loans and forms of contact with the adjacent Romance dialects. The vitality of the Arbëreshë dialects is sensitive to Italian sociolinguistic conditions, subject to the strong pressure of the standard language. Law 482 protects the use and autonomy of minority languages according to the lines of the UN and the European Union.

In this paper the question is addressed whether conversion (referring to word-class change without addition of derivational affixes) is suitable as a comparative concept in cross-linguistic and contrastive research on word-formation. Controversies concerning the scope of conversion in English and in languages with rich inflection are highlighted. Subtypes of conversion, such as partial conversion and syntactic conversion, are identified. Other terms which are used with reference to affixless word-formation operations are mentioned, such as zero-derivation, paradigmatic derivation or transflexion. A brief overview is given of selected recent contrastive (comparative) studies of conversion/derivation in English and in selected Slavonic or Romance languages, focusing on meaning-based and corpus-based research.

Romanian biblical texts are a fundamental source for documenting the old Romanian lexicon and for capturing the semantic evolution of numerous words, playing a decisive role in the compilation of the Thesaurus Dictionary of the Romanian Language (Dicționarul-tezaur al limbii române). These texts provide both specific theological terms and elements of common vocabulary, contributing to the reconstruction of the graphic, phonetic, and morphological forms that were characteristic of the era. Although some lexical references extracted from these texts are not always scientifically rigorous, their theological and cultural importance remains essential. Numerous lexical units identified in the biblical corpus represent the first attestations of words or meanings, which were later integrated and developed in the main lexicographical works dedicated to the Romanian language. At the same time, biblical texts provide very early attestations, often extremely close to the first appearances of a term, thus supporting the diachronic analysis of Romanian vocabulary. The use of biblical texts in the Thesaurus Dictionary has not only documentary value, but also helps by providing a complex framework for understanding the evolution of the Romanian language in all its dimensions (lexical, stylistic, and cultural).

In this paper, I discuss a new construction in Brazilian Portuguese, whereby the negative marker não ‘not’, which is syncretic with other types of negation in the language, appears in the slot [Aux_V] in periphrases, a position that is not common in other Romance languages. Among its properties, this type of negation must receive intonational stress, that is, it is a focus element and, as such, it conveys a specific meaning. A comparison with other types of low negation in the same position in Italian, Catalan and French shows that the new negator in Brazilian Portuguese is not the same as the ones that may occur in those languages. Observing this negation marker in relation to low adverbs and to the lexical verb in progressive and perfective periphrases in Brazilian Portuguese, I propose a syntactic position based on the nanosyntax approach, which can explain its focus nature and semantic/pragmatic properties.

Throughout the Romance languages, a definite clitic serving as a pro-predicate for an adjectival or indefinite nominal predicate is insensitive to the category and φ-features of its antecedent and local subject, resisting concord. This is remarkable in light of the obligatory φ-feature concord with its antecedent when the clitic serves as an argument, and the robustness of φ-feature con­cord between a [+N] full predi­cate and its subject in Romance. In this short paper, Déchaine &amp; Wiltschko’s (2002) analysis of French le will be amended for Romance pro-predicate definite clitics, which will be shown to be best treated as pro-N rather than pro-φ elements. The Romance data are analysed here against the background of the properties of predicate proforms in Hungarian.

In the present study, which constitutes a part of a broader research project, we seek to examine and illustrate–through selected textual excerpts–the ways in which students of Romanian origin enrolled at Sapienza University, who pursue studies in Romanian language and literature as a foreign language, negotiate their sense of identity and construct an image of Romania within a collective imaginary shaped by the distance of migration. This analysis is contrasted with the data emerging from the anonymous questionnaires administered throughout the academic year. The discussion draws upon both the responses to these questionnaires and the excerpts from the “open letters” the students addressed to Romania on December 1st, the National Day of Romania. It should be noted that this issue has already been the subject of previously published research, to which reference will be made throughout the present article in the interest of scholarly accuracy and methodological rigor.

The promotion and study of Romanian as a foreign language has a well-established tradition and has evolved over time in accordance with the language policies implemented by the Romanian state. The case of Romanian language study in Portugal is particularly paradoxical: despite its long-standing tradition, the number of students currently choosing to learn Romanian remains low. This study attempts to provide a more informed and nuanced account of this paradox by applying Hofstede’s 6-D model of national cultures, using data about Romania and Portugal as presented in Hofstede’s Globe. The analysis reveals that, although the two nations share several cultural dimensions, they also exhibit divergent values and cultural tendencies that reflect deeper, often invisible aspects of daily life – factors that may help explain the low interest shown by Portuguese students in studying Romanian.

In this paper, we analyze the nominal group in Dosoftei’s writings, with a focus on the possessor, adopting a diachronic approach. The paper is divided into two parts: the first outlines the theoretical framework, and the second presents the analysis of the corpus un-der investigation. In the theoretical section, we first discuss the con-cepts of nominal group and possessor. After defining the nominal group and describing its constituents, we examine the means of ex-pressing possession in Romanian—primarily the genitive, but also dative clitics, possessive adjectives, relative clauses, and prepositional phrases equivalent to the genitive. We then consider the peculiari-ties of the possessor as a syntactic function introduced in the Basic Grammar of the Romanian Language, focusing on the parts of speech by which it can be expressed. Regarding the corpus, we first address the substitution class of the possessor. We then describe complex nominal groups in which possessors co-occur with other syntactic positions, namely deter-miners, modifiers, and quantifiers. We discuss the internal structure of each nominal group and provide examples to support our analy-sis. Overall, the examination of the corpus reflects the complex possibilities for structuring the nominal group and illustrates the di-versity of syntactic constructions in Old Romanian with respect to the possessor.

This article reconsiders the evidence for an original ternary complementizer system in the Italo-Greek dialects of southern Italy and how such a system was replicated in the neighbouring Romance dialects under contact. Although the evidence is in many cases fragmentary, given the often recessive nature of the ternary system today, it is demonstrated how an original ternary system was often reanalyzed as a binary system in the modern dialects and how such developments gave rise to different formal outcomes across various Greek and Romance dialects.This article reconsiders the evidence for an original ternary complementizer system in the Italo-Greek dialects of southern Italy and how such a system was replicated in the neighbouring Romance dialects under contact. Although the evidence is in many cases fragmentary, given the often recessive nature of the ternary system today, it is demonstrated how an original ternary system was often reanalyzed as a binary system in the modern dialects and how such developments gave rise to different formal outcomes across various Greek and Romance dialects.