Abstract This paper proposes a syntactic analysis of the discourse marker ti : in Tunisian Arabic, traditionally viewed as a truth-value irrelevant verbal disfluency. Contrary to this perception, the study reveals ti : as a communicatively salient element with distinctive properties that are best explained through a syntactic framework that enables the mapping of the discourse effects of ti: within a structural configuration beyond CP. Specifically, the analysis demonstrates that ti: is both grammatically multicategorial and pragmatically polysemous, reflecting complex interface dynamics. These dynamics suggest that its interpretive effects interact internally with various communicative domains and, crucially, that these interactions can be formalized within the syntax-discourse interface. This syntactic approach to ti: uncovers multiple inherent interface relationships and contributes to broader crosslinguistic research on discourse markers hin syntax-discourse interactions.
Abstract The paper presents an intragenetic typological investigation of verbal colexifications. Data from 34 languages of the East Caucasian family were collected and analyzed to detect colexifications of the concept ‘build’. By comparing the results of this intragenetic investigation with those of larger cross-linguistic investigations, I show that small-scale, manually compiled and carefully verified databases can contribute to validating and complementing information available from large-scale typological databases. In particular, investigations of under-documented languages allow detecting patterns that, although seemingly rare cross-linguistically, are relatively common at a microtypological level. The colexification patterns attested in my sample can all be explained by well-known metonymic and metaphorical shifts, but only one of them is reported to be cross-linguistically common. While some of the shared patterns could be attributed to a common origin of the verbs in questions, others are likely to have spread by language contact.
Abstract This study introduces a Bayesian hierarchical model for investigating the discourse functions of tail–head linkage (THL). THL is a cross-linguistically widespread discourse pattern that connects units of text through verbatim repetition or anaphoric reference. Previous research has largely concentrated on the structural properties of THL, while its proposed functions – such as cohesion and coherence, processing facilitation, and discourse structuring – have typically been inferred from qualitative analyses of individual examples. To address this gap, the present study applies a quantitative statistical model to a corpus of ten narrative texts from five speakers of the Papuan language Muyu. The results indicate that THL in Muyu is not used for referential coherence but rather contributes to maintaining coherence across narrative scenes set in different locations. No evidence was found for systematic stylistic variation among speakers. Beyond the case study, the paper demonstrates the methodological potential of Bayesian hierarchical modelling for discourse analysis. The approach allows for flexible inclusion of additional predictors, hierarchical grouping factors, and cross-linguistic data. It thus provides a framework for quantitatively exploring how discourse strategies such as THL contribute to textual organisation and coherence across languages.
Abstract This article revisits a mechanism of the first two primary palatalisations of velar stops before unrounded front vowels and in the clusters *ki̯ and *gi̯ in Common Slavic. I ground the discussion in the analysis of the new acoustic data from Kryvorivnja (Southeast Hutsul) Ukrainian that I collected during my 2019–2022 fieldwork in the Carpathian Mountains. Similar to Common Slavic, the Kryvorivnja dialect lacks a phonological contrast between /k/ and /k j / and highly ranks one and the same phonological constraint, namely, the tendency towards intrasyllabic harmony Constraint, which is responsible for, e.g., velar palatalisation in Common Slavic. I show how this constraint changed from a constraint on surface forms in Common Slavic to a constraint on underlying forms in the Kryvorivnja dialect. I argue that the intermediate stages for the primary velar stop palatalisation were ambiguous [k j ∼t j ] and [g j ∼d j ], consequently targeted by spirantisation, and that, having become allophones of /t/ and /d/, they imposed the later primary palatalisation of the Proto-Ukrainian *ti̯, *di̯ sequences. In turn, the surfacing of /t j / → [k j ∼t j ] 2 and /d j / → [g j ∼d j ] 2 in the Kryvorivnja dialect illustrates the case of hyper-correction based on [k j ∼t j ] 1 attributed to the fourth palatalisation of velars.
This study focuses on word-formation motivation, analysed through the lens of the onomasiological theory of word formation. The approach presented here extends beyond the selection of motivating constituents at the morphematic level, viewing word-formation motivation as a multi-level phenomenon based on creative decisions of a coiner made at the conceptual, onomasiological, and morphematic levels. The theoretical principles are supported by an analysis of the motivation underlying 150 English compounds from the LADEC database and their Slovak equivalents denoting the same objects. The data are analysed in terms of three types of structural relations: onomasiological structure, morphematic structure, and onomasiological type. The study addresses a research question: In which types of motivating structural relations do the languages converge or diverge? The findings reveal a high degree of correspondence in the onomasiological structure and the determining mark, with Onomasiological Type 3 emerging as the most frequent matching type. This points to a shared tendency to omit the Actional constituent (the determined mark) and prioritise economy of expression. Differences appear in the representation of the onomasiological base, largely reflecting structural divergence between English compounds and Slovak derivatives. The analysis highlights the procedural nature of word-formation motivation, with nearly half of the pairs exhibiting a high degree of motivational identity.
Tundra Nenets is an SOV language that appears not to fully prohibit postverbal constituents. Apart from brief remarks in prior literature, the nature of these occurrences has not been examined in detail. This study offers a closer empirical investigation of postverbal phrases in Tundra Nenets, drawing on data from corpora, grammatical descriptions and our fieldwork. Given that clause-internal postposing of constituents in verb-final languages is cross-linguistically constrained by factors such as complexity, information structural status and syntactic function, we examine whether any such asymmetries are attested in Tundra Nenets. We argue that the absence of asymmetries follows if all postverbal phrases are clause-external satellites, functioning either as right dislocation or as afterthought (as conjectured by Nikolaeva 2014. A grammar of Tundra Nenets. Berlin: De Gruyter). This analysis is supported by data indicating that postverbal phrases, with the possible exception of adjuncts, are consistently associated with a preverbal correlate, either overt or covert. The clause-external satellite analysis correctly derives the choice between an overt and a null correlate from independent principles, and it also accounts for the unacceptability of the postverbal placement of non-referential manner and measure adjuncts. We conclude that Tundra Nenets falls in the relatively small group of rigidly verb-final languages.
Abstract This study explores the intersection of phraseology and morphology within the postulates of usage-based Construction Grammar, focusing on the construction [ a N-suffix singular.blow limpio ]. The peculiarity of this syntactic construction is that it embeds productive morphological formations with the suffixes -azo , -ada and -ón , all of which convey the notion of ‘blow’ or ‘impact’. Through a corpus-based analysis of 3,300 tokens from the esTenTen18 corpus (Sketch Engine), we offer a semantic/pragmatic description of the construction and its realizations, examining how lexicalized and non-lexicalized slot fillers interact with the construction’s lexically filled elements. The analysis aims to investigate the degree of morphological creativity and semantic coercion present in the construction, identifying how evaluative suffixes encode metaphorical or compositional meanings depending on the context. By distinguishing between frequent, conventionalized forms and low-frequency forms or hapax legomena, we shed light on the gradient between F-creativity and E-creativity, and on how morphologically derived lexemes can be reanalyzed within a semi-schematic construction. Ultimately, this research contributes to understanding how evaluative and metaphorical meanings emerge at the intersection of morphological schemas and syntactic patterns.
This paper addresses the diachrony of two closely-related Cariban languages, Ikpeng and Arara. I reconstruct the stop consonant inventory, and a set of alternations, for Proto-Kampot (pK), their shared ancestor. Ikpeng innovated by merging part of the set of pK voiced (or lenis) stops with two homorganic approximants. As a result, cognate morphophonological alternations involve stops in Arara, but stops and approximants in Ikpeng. PK *g, which lacked a homorganic counterpart in the approximant series, was retained in Ikpeng, yielding its relatively marked inventory with g alone as a voiced obstruent. Next, I discuss the origins of the pK segments involved in these developments. On the basis of evidence from other Cariban languages, I find support for the idea that contrastive intervocalic fortis stops evolved from previous consonant clusters in Bakairi, the closest relative of the Kampot languages. For Ikpeng and Arara, however, these clusters are retained, and loanwords seem to furnish the bulk of the examples of intervocalic fortis consonants. Finally, I show how Ikpeng broadened the distribution of g to include word-initial contexts by a change that affected the pre-vocalic allomorph of the pK first person prefix *i-. A section of conclusions discusses the significance of these findings.
This paper examines the correspondences between affirmative and negative members of phasal polarity systems, with particular focus on systems that consistently adopt one scope relation between negation and operator to the exclusion of the other. To this end, I draw on a global convenience sample of 150 languages representing 101 phyla. Consistent with earlier, smaller-scale studies, the findings reveal a pronounced asymmetry in the distribution of these correspondences. Inner negation (negation within the scope of the operator) is globally predominant, and systems that consistently employ inner negation far outnumber their wide-scope counterparts. Among inner-negating strategies, the use of 'still not' to express 'not yet' is more common than that of 'already not' for 'no longer'; what is more, the presence of the latter pattern nearly universally predicts that of the former. These results offer further evidence against L & ouml;bner's duality hypothesis as a comprehensive account of phasal polarity systems. In terms of explanation, I argue that the observed asymmetries are best explained by evolutionary biases rooted in semantic proximity, which are also reflected in discourse-pragmatic considerations and lexical source patterns.
This paper provides a reconstruction of the personal pronouns of the Dogon language family of central Mali. We expose, juxtapose and contrast the available synchronic evidence from 19 Dogon languages in order to establish etyma inherited from the presumed proto-language: Proto-Dogon (PD). We present sound correspondences and morphological change between Proto-Dogon and the daughter languages, demonstrating that clear proto-forms can be established for 1sg, 2sg and 3sg pronouns, and to a lesser degree, plural pronouns, which show greater internal variation. Furthermore, internal reconstruction demonstrates that the 'independent', 'possessive' and 'subject-affixal' uses of the pronouns have evolved from a single set of shared inherited morphemes. Evidence from the reconstruction of the Proto-Dogon pronominal paradigm weights favourably into the genetic cohesion of the Dogon languages and provides strong systematic evidence for the internal phylogenetic classification of these languages.
In this study, two corpora were compiled to compare the subject types, reference types and phraseological patterns of reporting clauses in Chinese MA theses in applied linguistics and international applied linguistics journal articles. Differing from previous studies, we modified the classification criterion for reference types, dividing them into identifiable and non-identifiable categories. Under this framework, identifiable references encompass integral and non-integral citations, whereas non-identifiable references denote general references. The results revealed that: 1) Regarding subject types in reporting clauses, authors of Chinese MA theses demonstrated a stronger preference for non-human and it subjects while employing fewer human subjects compared to authors of international journal articles. 2) As for reference types in reporting clauses, authors of Chinese MA theses exhibited a significantly higher frequency of general references but fewer integral citations than authors of international journal articles. No significant differences were observed in the use of non-integral citations between the two author groups. 3) Two dominant reporting patterns in both corpora were found: integral citation + human subject and non-integral citation + non-human subject. The results of the study could provide pedagogical implications for academic writing as well as EAP pedagogy.
This study investigates the principle of Dependency Length Minimization (DLM) in German from a diachronic perspective, looking at the period from 1600 to 1950. It aims to assess whether any changes in dependency length (DL) can be observed over time. Challenging the standard assumption about diachronic DLM, we argue that DLM effects are modulated by the extralinguistic context in which the language is used and thus do not assume that DL as a measure of syntactic complexity decreases over time, but rather put forward the hypothesis that DL is characterized by fluctuations. That is, changes towards a reduction of DL, hence of syntactic complexity, might be observed in environments of increased processing pressure. On the contrary, if the extralinguistic context favors variants that increase syntactic complexity, for example due to normative pressure, DL might even increase. Using a novel diachronic dependency corpus of German newspaper texts, the analysis targets both overall DL and the dependency relation between the lexical verb and the auxiliary specifically, offering a window into changes affecting the size of the middlefield in German.
This article investigates the alternation between indicative and subjunctive moods in complement clauses introduced by verba putandi (e.g. 'think' or 'believe') in contemporary spoken Italian. Drawing on spontaneous conversations from the KIParla corpus, the study tests whether mood selection is governed by degrees of epistemic certainty, by extralinguistic factors, or by linguistic properties internal to the construction. Multivariate statistical models evaluate the contribution of several predictors, including the grammatical person and tense of the subordinate verb, its adjacency to the complementizer, the lemma involved, register, and speakers' educational background. The results show that mood alternation is not primarily driven by epistemic stance or education level, but by usage-based and structural parameters: third-person, present-tense, and adjacent clauses strongly favour the subjunctive, while first- and second-person contexts, past tense, and non-adjacency favour the indicative. Register exerts a measurable influence, whereas educational background does not. Overall, the findings challenge traditional evidential accounts of mood in Italian and suggest that the indicative/subjunctive alternation in spoken language is best understood as a gradient, probabilistic phenomenon shaped by frequency patterns, morphosyntactic dependency, and register variation.
Synthetic verbs, e.g. the Italian verb passeggiare, 'to stroll', or decide, 'to decide', express meaning in a single inflected form, while their analytic counterparts, like fare una passeggiata, 'to take a stroll', distribute predication across multiple lexical elements, typically a light (support) verb plus a noun phrase, such as in light verb constructions. Although analytic verbs show greater structural complexity when compared to synthetic verbs, it has not yet been definitively established whether they require a higher cognitive processing load. A self-paced reading experiment using the moving-window paradigm was conducted to address this issue. Reading times and comprehension accuracy scores for sentences containing both synthetic and analytic verbs were analyzed. Increased reading time was observed in sentence-final regions for analytic verbs, but not for synthetic verb suggesting a possible delayed integration effect. No differences emerged between the two types of verbs in comprehension. These findings contribute to our understanding of how predicate structure affects real-time sentence reading and comprehension as they partially support the view that structural complexity predicts processing difficulty.
Examining Polish-English bilingual speakers, we investigate the degree of phonetic synchronicity between C1 and C2 in stop-sonorant and /s/-initial clusters. In the first experiment, articulatory data - gathered with electromagnetic articulography (EMA) - reveal longer target-to-target lags in Polish than English for stop-initial clusters, but no effect for /s/-initial clusters. A lesser amount of articulatory overlap for L1 Polish stop-initial clusters was also observable in the acoustic duration of C2, which was longer in L1 Polish than in L2 English. The second experiment gathered acoustic C2 duration data from a larger number of speakers, replicating the effects described in Experiment 1 for stop-initial clusters, and also revealing a less robust effect for /s/-initial clusters. A phonological interpretation of the results is presented within the Onset Prominence representational framework, whose phonotactic mechanisms allow for distinct structural representation of the "same" cluster across languages.
This paper seeks to substantiate the utility of AI-driven software in multimodal research. It endeavours to elucidate the pre-eminence of two specific pieces of software over alternatives used in linguistic analyses of this kind. The present study thoroughly scrutinises the eligibility of two sorts of audio and two types of visual data for automatically investigating the correlation of verbal and non-verbal means of expressing emotive modality in the German language. In doing so, it examines how the sentential scope typical of certain lexemes influences the applicability of the data types under investigation. The empirical analysis corroborated the necessity to conduct research of this kind employing entire sentences featuring a lexeme characterised by sentential scope. Intonation patterns and nonmanual features recognised and investigated by the software can occur throughout uttering a sentence which features a lexeme under study. This conclusion is drawn on the basis of the greatness of the quantity of emotions detected by the software vis-& agrave;-vis both the audio and the visual data employed in the analysis.
