Discourse Driven V-to-C in Early Modern Romanian

Marc Olivier, 2022

This thesis explores the diachrony of clitic placement with infinitives (both in re- structuring and other infinitival clauses) in French in a corpus of legal texts from the mid-12th to the mid-19th century which was built as part of this research project. We find enclisis in non-restructuring clauses until ca. 1300, and clitic climbing (CC) in restructuring clauses until the late 18th century. The two orderings are subsequently replaced by proclisis. These findings challenge the view that enclisis and CC are necessarily found within the same system, as Middle French is a language with proclisis and CC. Furthermore, CC is the major ordering found in restructuring clauses in Old French, and its frequency tops 100% from the 14th to the early 17th century. This finding reveals that the construction was not optional in Middle French. This thesis develops a theory of cliticisation based on verb movement: we account for the shift from enclisis to proclisis in non-restructuring clauses with the loss of V-to-T movement with infinitives. Independent evidence for this hypothesis stems from the loss of infinitival suffix -r in early Middle French, which we show acted as a movement trigger in Old French. This proposal is further supported by the consideration of the crosslinguistic picture: Romance languages that have enclisis also have infinitival suffixes and V-movement to a high position (e.g. Standard Italian). Regarding CC, the analysis we propose is one of mono-clausal restructuring with cliticisation on the higher v-head. We argue that from the early 17th century on, the lower v-head is reanalysed as a cliticisation site, yielding proclisis. The diachrony of other Romance languages supports the view that cliticisation on the lower v is an innovation of late Medieval Romance. Unlike other canonical languages however, French did not retain the optionality of cliticisation on the higher v and proclisis generalised to all infinitival clauses.

This study analyzes the position occupied by pronominal clitics in the clause with respect to the verb in old Romanian (OR) on the basis of an extensive corpus analysis of 16th – 18th century texts. The corpus analysis shows that, from the earliest texts, OR pronominal clitics are attested in second, third, fourth, etc. position in the clause, and exceptionally also in first position. Therefore, they do not fully observe the Tobler-Mussafia Law, which was in function in other old Romance languages. OR pronominal clitics are IP-clitics, which can be placed both in pre- and in postverbal position (proclisis and enclisis). Due to the gradual reduction of V-to-C movement, pronominal proclisis generalizes.

Diachronica. International Journal for Historical Linguistics

Introduction Syntactic structures are complex objects, whose subtle properties have been highlighted and elucidated by half a century of formal syntactic studies, building on a much older tradition. Structures are interesting objects of their own, both in their internal constitution and in their interactions with various grammatical principles and processes. The cartography of syntactic structures is the line of research which addresses this topic: it is the attempt to draw maps as precise and detailed as possible of syntactic configurations. Broadly construed in this way, cartography is not an approach or a hypothesis: it is a research topic asking the question: what are the right structural maps for natural language syntax? Answers may differ, and very different maps may be, and have been, proposed, but the question as such inevitably arises as a legitimate and central question for syntactic theory. If it is a virtual truism that cartography can be construed as a topic and not as a framework, it is also the case that cartographic studies have often adopted certain methodological and heuristic guidelines, and also certain substantive hypotheses on the nature of syntactic structures, which form a coherent body of assumptions and a rather well-defined research direction; we will try to illustrate some ideas and results of this direction in the present chapter. If structures have, in a sense, always been central in generative grammar, the idea of focusing on structural maps arose around the early nineties, following a track parallel to and interacting with the Minimalist Program. Perhaps the main triggering factor was the explosion of functional heads identified and implied in syntactic analyses in the first ten years of the Principles and Parameters framework. One critical step was the full-fledged extension of X-bar theory to the functional elements of the clause (Chomsky 1986) as a CP-IP-VP structure; and the observation that other configurations, e.g. nominal expressions, were amenable to a hierarchical structure with a lexical projection embedded within a functional structure (such as Abney's DP hypothesis, Abney 1987). These advances provided a natural format for the study of the structure of phrases and clauses as hierarchical sequences of the same building block, the fundamental X-bar schema (or, later, elementary applications of Merge); the lowest occurrence of the building block typically is the projection of a lexical category, e.g. a noun or a verb, and this element is typically completed by a series of

2002

Revue roumaine de linguistique, 2020

By studying the grammaticalization of Romanian auxiliaries from a diachronic Romanian and a comparative Romance perspective, this paper argues that the output of grammaticalization is a predictable pattern in a given language, i.e. a language-specific parametric choice. Specifically, in the passage from old to modern Romanian we observe that a number of emergent periphrastic structures (innovations in contrast to Latin) died out, against the well-known transition from syntheticity to analyticity in the development of the Romance languages (i.e. the profusion of auxiliary structures in this particular situation). In order to account for what appears to be a diachronic paradox, we show that, under a rich cartographic structure of the IP, Romanian auxiliaries systematically grammaticalize as exponents of the category mood; the auxiliaries of the now-defunct periphrases have a richer feature matrix (and this accounts for their demise). The MoodP is also the target of synthetic (finite) verb movement, hence Romanian is, (micro)parametrically, a mood-oriented language, a hypothesis which accounts for the particular diachrony of periphrastic constructions in this language, as well as other properties.

The paper examines the interpolation and scrambling phenomena attested in Old Romanian and (modern) Istro-Romanian. By applying a coherent set of syntactic diagnostics, it is shown that these phenomena may be given the same analysis in both varieties: the discontiguity of the elements of the verbal cluster is the result of low verb movement of the lexical verb on the clausal spine. In a diachronic Romance comparative setting, the existence of low verb movement phenomena constitutes an important (yet overlooked) piece of evidence for the gradual emergence of V-to-I movement in the Latin-to-Romance transition. For the theory of diachronic linguistics in general, the preservation of interpolation and scrambling in Istro-Romanian shows that archaic phenomena may be preserved in isolated varieties, and that language contact (with Croatian in the case at hand) plays an important role in consolidating archaic features.

… and Parametric Variation, 2005

Sitaridou, I. (2012). ‘A comparative study of word order in Old Romance’. Folia Linguistica, 46/2, “The pace of Grammaticalisation in Romance”, (guest eds.) A. Carlier, B. Lamiroy & W. De Mulder). 553-604

The main objective of this article is to discuss word-order phenomena in Old French, Old Spanish, Old Portuguese, and Old Occitan from a comparative perspective. In particular, the following are examined: (i) the empirical evidence in order to assess the theoretical arguments for and against a verbsecond analysis of the Old Romance languages; (ii) verb movement in Old Romance; and (iii) whether the changes in the word-order patterns in the history of the Romance languages can be captured as yet another instance of grammaticalisation. By means of a uniform methodology and statistical analysis of novel data, it is claimed that: (a) Old Romance does not possess a Germanic V2; (b) there is variation in terms of word order among Old Romance languages, which is shown to be linked to the individual history of the languages; (c) V2 order is mostly an epiphenomenon of the discourse mechanisms and the nature of the left functional field in these languages.

Oxford University Press eBooks, 2005

From the perspective of language change, grammaticalization is generally viewed as the process whereby "a lexical item or construction in certain uses takes on grammatical characteristics" (Hopper & Traugott 2003:2). account for this type of change as resulting from the reanalysis of lexical heads into functional heads. The diachronic changes we discuss in this paper do not fall under this definition, but we view them as exemplifying a type of grammaticalization whereby illocutionary features come to be associated with distinct functional heads. We analyse the changes in the clausal organisation of Old French as following from the fact that the Topic/Focus functional head common to all clause types of the first stage gives way to a system with a number of separate illocutionary heads. We argue that the weakening of the Tobler-Mussafia (TM) constraint excluding object clitic pronouns from initial position in main clauses in Old French (OF) results from a gradual replacement of a common representation for V1 initial clauses by a new system where 1) satisfaction of a discourse-related [Top]/[Foc] feature by V is minimized, and 2) there is a reanalysis of the CP layer, with grammaticalization of illocutionary type features. 1. Old French as a V2 language. Old French is considered a V2 language (see e.g. , Roberts 1993), with the verb generally following a topic or focus element in main clauses. We account for that characteristic by adopting a split-CP approach (Rizzi 1997), and assume that the C system is composed of two functional projections ZP and FinP, as shown in (1). Z bears a discourse-related Topic or Focus feature, which we label [+D] (Discourse) for simplicity. Fin is the interface of the C * This work has been made possible thanks to a grant from the Social Sciences and Humanities Research Council of Canada (#410-mmmmmmm). We thank Ana Maria Martins and Itziar Laka for their comments and suggestions on a previous version of this work.

Despite a truly vast literature on Old French word order (Vance 1997; Rouveret 2004; Mathieu 2012; Labelle & Hirschbühler forthcoming) and a growing empirical understanding of word order in Old Occitan (Kunert 2003; Vance, Donaldson & Steiner 2009; Donaldson 2015, 2016) there remains as yet no formal account of the similarities and differences between these important early Romance varieties. This lacuna is significant as the correct analysis of core clausal word order in Old French and Old Romance in general is a hugely controversial area of Romance scholarship (see Rinke & Elsig 2010 and Benincà 2013 for opposing views). Based on detailed quantitative analysis of two 13th century texts, La Vie de Sainte Douceline in the case of Occitan and La Queste del Saint Graal in the case of French and a hand search of a range of 11th-13th century texts for each language, we propose a new account of the similarities and differences. Drawing on the insights developed in Rizzi (1997) and Benincà & Poletto (2004) that the left periphery consists of a rich set of hierarchically ordered functional projections, I assume that the V2 property, commonly conceptualised as entailing V-to-C movement, can vary in terms of which position in the left periphery the verb targets. In the case of Old Occitan, I hypothesise that the verb targets Fin°, the lowest functional head in the left periphery, and as such the whole range of left-peripheral positions are a priori active. In the case of 13th century French, I propose that the verb targets one of the highest positions in the clause, Force, which results in heavy restrictions in the left-peripheral structure available. This proposal permits a novel and revealing account of the differences between the relevant texts concerning V>3 structures, the licensing of null subjects and topics, the clitic pronominal system and the licensing of Verb Second structures in embedded clauses.