Archives

• The evolution of perfect aspect markers in Vietnamese

  Vol. 10 No. 10 (2026)

  Trang Phan, Tuan-Cuong Nguyen and Masaaki Shimizu. This paper examines the diachronic evolution of perfect aspect markers in Vietnamese from the 13th to the 20th century, focusing on the development and interaction of đã, rồi, and the now-extinct đoạn. Utilizing a unique corpus that combines both primary texts in the traditional Nôm script and later texts in the modern Quốc Ngữ script, this study identifies seven distinct stages of development. In the earliest stage (up to the 13th century), đã was the sole perfect marker, primarily used preverbally and occasionally as a verb meaning ‘to finish’, with neither rồi nor đoạn present. By the 14th century, rồi emerged as a postverbal telic marker, and đoạn appeared mainly as a verb meaning ‘to finish’. Over the centuries, these markers coexisted and competed: đã expanded its functions to mark past tense, occasionally function as a conjunction (a usage that is no longer available), and sometimes serve as a sentence-final imperative marker ‘first’. Meanwhile, rồi was initially used as a verb and an immediately postverbal telic/completive marker. Over time, its usage expanded, with an increasing frequency as a perfect marker, while it also took on additional functions as a conjunction, marking temporal or logical transitions between events. On the other hand, đoạn transitioned from a verb into an immediately postverbal telic/completive marker and then into a perfect marker (and also a conjunction) before gradually disappearing by the late 19th century. The early 19th century marked the peak coexistence of all three markers as perfect markers. However, by the late 19th century, đoạn had vanished, leaving đã and rồi as the dominant markers, a pattern that persisted into the 20th century. Importantly, while đã has lost its conjunctional function in modern Vietnamese, rồi retains its conjunctional “and then” use – reflecting a broader Southeast Asian areal trend also attested in Lao lèèw4 and Thai lɛ́ɛw (Enfield 2007: 183; Olsson 2013). This study sheds light on the historical evolution of Vietnamese grammar, highlighting the complex interaction and competition between these markers. The gradual disappearance of đoạn and the increasing specialization of đã and rồi as perfect markers reflect broader trends in grammaticalization, where lexical items undergo reanalysis, semantic narrowing, and functional shifts along the functional spine from lexical to grammatical and discourse functions. The research provides valuable insights into grammaticalization theory, particularly in the areas of aspect and tense marking, and deepens our understanding of the dynamic processes that have shaped Vietnamese grammar over centuries.

• The Innovation of Inalienable Nominalisers in Meto

  Vol. 10 No. 9 (2026)

  Tamisha L Tan. This paper investigates the diachronic origin and synchronic representation of inalienability in the Meto languages of West Timor, focusing on the development of two types of nominal suffixes. Drawing on comparative and language-internal evidence, it proposes that the innovation of an (in)alienable possession system in Meto involved i) the morphosemantic reanalysis of a situational inalienable suffix in *-k > -ʔ, originally developed due to contact with non-Austronesian languages, and ii) the syntactic Spec-Head reanalysis and grammaticalisation of inherited genitive pronoun enclitics into inalienable possessor agreement morphology (e.g. 3SG *=na > -n).The analogical extension of these two types of n head categorisers resulted in their semi-idiosyncratic use with etymologically *V# nouns, and also facilitated the aggressive resegmentation of etymologically *C# nouns, resulting in the illusion of productive subtractive morphology in the language.

• Modelling the acquisition of labile verbs in Middle English

  Vol. 10 No. 8 (2026)

  Jordan Kodner, Achim Stein & Carola Trips. This paper applies current mathematical models of learnability (Yang 2005, 2016, Irani 2019, Li 2024) to historical data to gain a precise understanding of how children in the past learned formal aspects of language. We take the learnability of the causative-anticausative alternation as an example (Irani 2019) and address the question of how children learned the class of labile verbs in Middle English. We investigate the contact hypothesis, which argues that the well-known influx of Old French verbs of change-of-state coming into Middle English (Ingham 2020) had an effect on learning the alternation as a productive rule. Drawing on work on learnability in diachrony (Kodner 2019a, 2020, 2023b, Trips & Rainsford 2022), we present a study which applies the Tolerance and Sufficiency Principle(s) (TSP; Yang 2016) to the lexicons of Present-Day English first-language learners and to the lexicon of Middle English change-of-state and motion verbs. We evaluate the conditions under which learners would posit that all of these verbs are potentially labile using frequency-weighted sampling. Significantly, we show that both morphological change (loss of distinct causative morphology) and contact (copying of French verbs) contribute to the rise in lability. Overall, we present a new methodological pathway for historical linguistics: the combination of historical corpus linguistics with cognitively motivated computational modeling to reveal how the historical child learned grammatical rules.

• Anaphors before intensifiers: nafs-x in Hebrew and Arabic

  Vol. 10 No. 7 (2026)

  Noa Bassel. The history of English x-self has informed both diachronic and synchronic accounts of anaphors by showing that its source is the focused adjective known as intensifier (as in the queen herself came). I present two corpora surveys that show a reversed historical trajectory for the Semitic anaphor nafs-x. The findings suggest a refined explanation to the cross-linguistic homophony between anaphors and intensifiers and to the role of complex anaphors in grammar.

• Double object order in historic German: a unified processing perspective with new data from pronominal constructions

  Vol. 10 No. 6 (2026)

  Sophia Voigtmann & Joanna Wall. Modern Standard German double object constructions with two pronominal objects display an almost completely fixed order, in which the accusative object precedes the dative one. In contrast, in historic German, such double object constructions display variation between an unmarked order in which the accusative object precedes the dative object, and a marked one in which the dative precedes the accusative. In double object constructions in which both objects are instead full noun phrases, there is parallel word order variation between marked and unmarked orders in both historic German and modern Standard German. However, it has been previously proposed that these two cases of word order variation are determined by distinct factors, despite arising in a single double object construction. Going against this well-established view, we propose a uniform, processing-based account of pronominal and nominal double object constructions. This is evidenced by a novel empirical study of the former in the GerManC corpus. First, by assessing information uniformity, we show that use of the marked order is associated with a grammatical processing advantage in the sentences in which it is used compared with using the alternative, unmarked order in those same sentences. Second, we present empirical evidence that this processing advantage relates to mental accessibility through givenness, salience and animacy, and the position of the full verb.

• (Not) Anything goes: The syntax and diachrony of the German 'gehen'-middle

  Vol. 10 No. 5 (2026)

  Benjamin L. Sluckin & Eric Fuss. This paper offers a syntactic and diachronic analysis of the German [‘gehen + zu’-infinitive] construction. Previously misclassified as a periphrastic anticausative or conflated with canonical modal passives, the configuration represents a structurally distinct periphrastic realization of the middle voice. Synchronically, modal ‘gehen’ operates as a functional raising head that strictly selects a voiceless, eventive vP and a resultative configuration — a selectional restriction that directly accounts for the construction’s rigid lexical-semantic preference for Accomplishments. The raising head carries a formal edge feature forcing internal argument raising, which feeds dispositional evaluation at the conceptual-intentional interface. The resulting feature valuation simultaneously bleeds temporal auxiliary insertion and triggers post-syntactic ‘zu’-insertion as a morphotactic repair. Historical corpus data spanning the 15th to 20th centuries show that the ‘gehen’-middle emerged rapidly across the 18th and 19th centuries via a two-stage grammaticalization pathway: the lexical motion verb first underwent a semantic shift toward possibility, extending analogically into clausal control infinitives; ambiguous control structures were then reanalysed as monoclausal raising configurations. This reanalysis stripped the infinitival complement of its clausal architecture, reducing it to a bare vP, and bleached the matrix verb of its thematic subject, producing the constrained functional raising head attested in Present-Day German.

• The Panará clause and the emergence of polysynthesis in the Jê language family

  Vol. 10 No. 4 (2026)

  Bernat Bardagil. This article presents the syntactic reconstruction of the development of a highly synthetic morphosyntax in Panará, a Jê language spoken in Brazilian Amazonia. Languages of the Jê family present a characteristically analytic morphosyntax, with an ergative-accusative alignment split, and a templatic verb-final clause. Panará, while clearly a Jê language in all other respects, presents strikingly non-Jê morphosyntax, being polysynthetic, uniformly ergative, and verb-medial, differences which have often been attributed to external influences from a non-Jê language in an intense contact scenario. I lay out a diachronic reconstruction of an innovation of head movement of the verb to INFL, which in turn triggered generalized cliticisation of phrasal heads on INFL, leading to polysynthogenesis. 

• The Diachronic Development of the Serbian Complementizer 'Kako'

  Vol. 10 No. 3 (2026)

  Alberto Frasson. This paper examines the diachronic development and synchronic distribution of the Serbian form 'kako' (’how’), showing that its interrogative and declarative uses derive from a single inherited composite structure. Drawing on data from Old Serbian texts spanning the twelfth to the eighteenth century, I distinguish two major functions: interrogative 'kako', a manner adverb that participates in regular wh-movement, and declarative 'kako', a complementizer introducing propositional, eventive and irrealis subordinate clauses. I argue that 'kako' originates from a Common Slavic phrasal configuration consisting of a wh-element k- and a manner nominal WAY, whose internal structure became opaque over time. Diachronic evidence shows a gradual increase in the use of 'kako' as a complementizer, accompanied by semantic bleaching, loss of interrogative features, and structural reanalysis in line with syntactic economy principles. A comparison with Romance, particularly the divergent developments of Latin 'quomodo', Friulian 'cemût che' and Salentino 'cu', provides an independent model for understanding how a composite wh-expression can undergo grammaticalization into a C-head. The Serbian data thus offer a well-attested case of a wh-expression grammaticalizing into a multifunctional complementizer, illustrating cross-linguistic pathways between interrogative syntax and complementizer systems.

• The development of modal particles and its implication in the syntax: the case of Basque ahal

  Vol. 10 No. 2 (2026)

  Sergio Monforte. The Basque language marks modality by means of functional elements traditionally termed 'modal particles'. Two of those particles are the epistemic particle ahal and the question particle al; these two particles and the modal verb ahal ('to be able to') share the same phonetic form namely [al] in most varieties. Nevertheless, they are distinguishable since they have separate behaviour regarding some syntactic properties such as: a) cliticisation, b) distribution concerning negation and aspect, c) occurrence in contexts lacking illocutionary force, and d) degree of pragmaticalisation. Based on the aforementioned properties and on their geographical distribution and history, I argue that they are related diachronically as follows: the modal verb ahal gave rise to the modal particle ahal and the latter resulted into the question particle al. Such grammaticalisation patterns are found cross-linguistically (Wegener 2002; Hack 2014). Therefore, the reanalysis of those elements driven by the ambiguity in the interpretation brought a new distribution and a higher syntactic position for the grammaticalised elements and, ultimately, a simpler syntactic structure avoiding move operations.

• Tracing form and function in the history of German intensifier selbst/selber

  Vol. 10 No. 1 (2026)

  Lea Schäfer. This paper investigates the diachronic and synchronic distribution of the German reflexive intensifiers selbst and selber. While traditionally considered free variants, the two forms exhibit subtle differences in stylistic register, syntactic scope, and focus association. Based on historical corpora (including the edition of Weistümer by Jacob Grimm (1840–1869) and texts by Martin Luther, Paul Jacob Marperger, and Jean Paul) and contemporary spoken data (a small podcast corpus featuring Anne Will), the study combines qualitative analysis with quantitative statistical methods. The results show that over time, the formal overabundance of selbst and selber developed into a functional distinction. Three historical stages can be identified: (1) a fully inflected intensifier (selb) in Old and Middle High German, (2) the emergence of invariant forms (selbst, selber) in Early New High German, and (3) a functional split in Modern German, whereby selbst is increasingly restricted to narrow syntactic scope and single-focus contexts. These findings support a syntactic reinterpretation of intensifier variation and con- tribute to broader debates on exaptation and grammatical differentiation.

• The cartography of the DP from Middle High to Modern German: Positions and features of selected quantifiers, possessives, and adjectives

  Vol. 9 No. 21 (2025)

  Elliott Evans, Christopher D. Sapp & Rex A. Sprouse. Based on data from a new parsed corpus, we propose a cartography of the DP in medieval and early-modern German. Our cartography expands on Roehrs’s (2020) structure for the Modern German DP and on claims about the relative orders of different classes of adjectives by Cinque (1994) and Scott (2002). We propose that the DP in earlier stages of German has at least two underlying positions for (different kinds of) quantifiers and a position between D and N for adnominal genitives. Furthermore, we argue that there are two types of DP-internal movement, resulting in some surface deviation from the underlying cartographic order. Diachronically, in earlier German certain determiner-like elements could remain in their underlying positions, but in Modern German they move to a position at the front of the DP. We identify changes in the definiteness features of these determiner-like words as the catalyst for this structural change. 

• An observation on the distribution of correlative relative clauses in Cicero

  Vol. 9 No. 20 (2025)

  Philomen Probert. Part of the Special Collection "Latin and Sabellic relative clauses". This contribution presents an observation on the distribution of sentences with correlative relative clauses (also known as correlative sentences or relative-correlative sentences) in Cicero’s speeches: e.g. quae semper in ciuili uictoria sensimus, ea te uictore non uidimus (‘The things which we have always experienced in a civil war victory, those things we did not see when you were the victor’, Cicero, Deiot. 33). This construction occurs with considerably higher frequency in some speeches than others, and this variation turns out to be structured: in forensic speeches the frequency of the construction is inversely correlated with the length of the speech, while deliberative speeches appear to follow a different pattern. Focussing on forensic speeches in particular, this article argues that the distribution of correlative relative clauses in these speeches can be attributed to the way in which the construction helps to structure information, and the rhetorical purposes for which Cicero exploits this information-structuring device.

• Review of Bloom Ström et al. (eds.) (2024), Morphosyntactic Variation in Bantu

  Vol. 9 No. 19 (2025)

  Elisabeth J. Kerr

• 'Given to understand' as a model for the indirect/recipient passive in English?

  Vol. 9 No. 18 (2025)

  Cynthia L. Allen. The advent in English of passives like 'he was given a book' has long been a source of interest. Such passives, in which what is traditionally analysed as an indirect object is passivized, are not found in Old English. Macleod, Anagnostopoulou, Mertyris & Sevdali (2023) suggest that they may have developed as an extension of sentences like 'he was given to understand', with what was originally an indirect object being reanalysed as a direct object because of the rather opaque nature of 'given to understand'. A corpus-based study rules this construction out as a model for the new passive, which is found two hundred years earlier than 'given to understand'.

• Les relatives dans les vers saturniens latins

  Vol. 9 No. 17 (2025)

  Antoine Viredaz. Part of the Special Collection "Latin and Sabellic relative clauses". The metre in which the first two Latin epic poems and certain Latin verse inscriptions dating mainly from the 3rd-2nd centuries BC were composed is called Saturnian. Both the exact boundaries of the Saturnian corpus and the primary principle of versification of this metre are debated. We can now be certain, however, that a set of eight Latin funerary or dedicatory inscriptions and around 80 epic fragments by Andronicus and Naevius are definitely Saturnian. Studying the linguistic variation within this limited corpus allows us to identify certain indications of metrical constraints. The aim of this paper is to examine whether the variations observed in relative constructions can be exploited for this purpose. I will therefore study the relative clauses attested in Saturnian texts from the epigraphic and literary traditions. It is my aim to highlight the variation in these texts with regard to the position of the relativiser and that of the relative clause. I will attempt to determine what conditions the appearance of different structural types of relative clauses. Two avenues will be explored: that of semantic differences (e.g. restrictive or non-restrictive relative clauses), and that of constraints related to literary genre or metre.

• Latin and the comparative reconstruction of QU-exclamatives

  Vol. 9 No. 16 (2025)

  Michiel de Vaan. Part of the Special Collection "Latin and Sabellic relative clauses". While there are many synchronic studies on exclamatives, diachronic studies into the origin and development of exclamatives are rare. In this paper, I study the Latin evidence in order to arrive at a method for investigating the historical development of a subset of exclamatives, viz. exclamatives headed by interrogatives (QU-). I propose a classification of these exclamatives which takes into account their syntactic behaviour and their semantics. The specific distributional patterns found in Latin support existing etymological insights and allow for a few typological generalizations.

• Basque V2 effects in diachrony

  Vol. 9 No. 15 (2025)

  Maia Duguine & Georg Kaiser. This paper explores the evolution of verb-second (V2) patterns in Basque wh-interrogatives from the Archaic period (14-16th c.) to later periods. While in Modern Basque a systematic “residual” V2 system in wh-interrogatives is observed, in Archaic Basque the patterns appear to be mixed: synthetic verbs display systematic V2 order, whereas analytic verbs allow intervening material between the wh-phrase and the verbal complex. We put forth a multifactorial analysis of Archaic Basque interrogatives, assuming that in analytic constructions independent syntactic and morphophonological properties of auxiliaries interact with the syntax of V2 in such a way that they ‘hide’ its effects on word order. The change from Archaic to Modern Basque involves a simple change in the properties of T, independently of the syntax of V2.

• Moods in relative clauses in Plautus and Terence

  Vol. 9 No. 14 (2025)

  Wolfgang de Melo. Part of the Special Collection "Latin and Sabellic relative clauses". This contribution discusses relative clauses in four comedies by Plautus and one by Terence, as a representative sample of early Latin comedy. The main questions are, firstly, to what extent modal usages are the same as in classical Latin, and, secondly, how any differences can be explained. In defining and non-defining relative clauses alike, subjunctives make up 25% of the total, but they are used for somewhat different reasons in the two types of relative clauses. In defining relative clauses, the subjunctive can be the result of indirect speech, modal attraction, or the desire to express purpose. In classical Latin, the subjunctive is also employed if the head noun is indefinite and non-specific; this type is still spreading in early Latin, being almost obligatory in presentative constructions, but not yet elsewhere. In non-defining relative clauses, subjunctives can be used to express wishes or doubts, or they can be the result of indirect speech or modal attraction. In classical Latin, the subjunctive is also common if the relative clauses have causal or concessive nuances; in early Latin, on the other hand, the subjunctive is more restricted in such contexts. In 'causal' relative clauses, it mostly occurs if the superordinate clause contains a negative evaluation, and I argue that we are dealing with a reanalysis of relative clauses expressing purpose here. The subjunctive has barely begun to spread to concessive relative clauses, and the further spread to cum-clauses only really happens after Plautus and Terence.

• The emergence of the hoi epêlthe ptareîn impersonal construction in Ancient Greek

  Vol. 9 No. 13 (2025)

  Diego Luinetti. This paper focuses on the spread of a particular impersonal construction featuring an argument structure with a dative or accusative coded argument and a noun clause in Ancient Greek. It aims to demonstrate that the construction under analysis expands in its functional domain considerably over time, both in terms of productivity and frequency. Furthermore, it argues that the spread of this impersonal construction was facilitated by a bridging context involving neutral alignment, where some subject properties are lost both on a semantic (agentivity, referentiality) and on a morphosyntactic level (dedicated encoding, verb agreement), such as when a neuter pronoun takes the subject function.

• Relative clauses in Oscan and Latin: the development of new syntactic structures through language contact

  Vol. 9 No. 12 (2025)

  James Clackson. Part of the Special Collection "Latin and Sabellic relative clauses".  This paper takes as its starting point an Oscan inscription which shows several remarkable syntactic features: a relative–correlative order; repetition of the antecedent in the relative and the main clause; and fronting of two constituents out of the relative clause. This inscription been cited by some scholars as an example of archaic Indo-European syntactic patterns but others view these features as the result of language contact with Latin. Through comparison of material both in Oscan and other languages in the Sabellic family, as well as consideration of the evolution and use of relative–correlative syntax in Latin, I show that the language contact explanation is to be preferred. I further argue that inherited Oscan syntax can help explain the repetition of the antecedent.

• Types of relativization and relative heads in the Sabellic languages

  Vol. 9 No. 11 (2025)

  Jasmim Drigo & Yexin Qu. Part of the Special Collection "Latin and Sabellic relative clauses". This paper offers a novel analysis of the formal morphological and syntactic features of the Sabellic languages. We show that Sabellic correlatives align syntactically with other Indo-European branches in terms of headedness and the use of relative pronouns. Our main contribution is that Sabellic correlatives are base generated in the left periphery. Additionally, we compare the semantics of the relative pronouns in the Sabellic languages and Latin. Unlike Latin, where pronouns derived from *kwó- can refer to both animate and inanimate referents, the Sabellic languages restrict their use to animate referents in free relative clauses. Finally, we find that Sabellic languages counterexemplify a proposed universal of relative clauses, which claims that languages with relative pronouns do not have internally-headed relative clauses. This demonstrates a broader diversity in Sabellic relative clause formation than previously assumed.

• Proceedings of the 23rd Diachronic Generative Syntax (DiGS) Conference

  Vol. 9 No. 2-10 (2025)

  Selected papers from the 23rd Diachronic Generative Syntax (DiGS) Conference, which was held at New York University in June 2022. Guest editors Ailís Cournane and Gary Thoms.

• V2 to V1 in Welsh: The role of preverbal particles and fronted adverbials

  Vol. 9 No. 1 (2025)

  Oliver Currie. A peculiarity of V2 in Middle Welsh is the morphosyntactic marking of fronted constituents in positive declarative main clauses (PDMCs) using preverbal particles: 'a' for subjects or direct objects and 'y(d)' for adverbial or prepositional phrases. Unlike other medieval V2 languages which subsequently lost V2 and developed SVO, such as English and French, Welsh became VSO. In a formal analysis of the loss of V2 in Welsh, Willis (1998) argues that a widespread loss of both preverbal particles in the 16th century through phonological erosion was a key factor in the change, though does not provide corroborative quantitative data. Based on a corpus study of Early Modern Welsh (c.1550-c.1750), the present article shows that there is no evidence of a widespread loss of the preverbal particles 'a' and 'y(d)' by the end of the 16th century, only a partial and gradual decline in their use over a longer, more than two century period. There does seem, however, to be a link between the change in use and partial omission of 'y(d)' after fronted adverbial or prepositional phrases and the increase in use in Early Modern Welsh of a specific V1 construction, Absolute V1, where a finite verb comes in absolute-initial position in a PDMC.

• Introduction: Towards a comparative historical dialectology

  Vol. 8 No. 7 (2024)

  Ann-Marie Moser, Lea Schäfer & Sophie Ellsäßer. Part of the Special Collection "Towards a comparative historical dialectology". In this article, we highlight the challenges facing research in the field of comparative historical dialectology, with a focus on morphological and syntactic phenomena. We define the term “historical comparative dialectology” and use four different studies from different varieties to illustrate how the aims of historical comparative dialectology can be approached. The studies consider different varieties and language stages (different dialects of Old and Modern Catalan, the dialectal continuum of Yucatec Maya, historical dialects of High German, different Occitan dialects) and discuss various phenomena from a historical comparative perspective: differential object marking, numeral classifiers, number distinctions of nouns, and inflectional patterns.

• Numeral classifiers in Yucatec Maya: Microvariation and syntactic change

  Vol. 8 No. 6 (2024)

  Barbara Blaha Pfeiler & Stavros Skopeteas. Part of the Special Collection "Towards a comparative historical dialectology". Numeral classifiers in Yucatec Maya are subject to two processes of language change that create variation in the contemporary language. The first process is the use of a general classifier instead of specific sortal classifiers. The second process is the use of the general classifier along with mensural classifiers. Our study examines the microvariation of the contemporary language in space and time, based on data from the Atlas of Yucatec Maya and draws inferences about the entity of change in diachronic perspective. Our findings show that these processes are partially interconnected, reflecting the emergence of a general marker of Cardinality (Krifka 1995, Bale & Coon 2014, Bale, Coon & Arcos 2019). The dispersion of these phenomena in geographical space shows that they only partially overlap, suggesting that the underlying processes may apply independently from each other. Furthermore, the use of the general classifier in expressions of measure does not apply equally to all mensural classifiers. Hence, a further source of variation comes from mensural classifiers: some of them lose their function as classifiers and are only used as measure nouns. Contemporary variation can thus be understood as the cumulative effect of these processes.