Linguistic Variation - Volume 19, Issue 1, 2019

The Verb-second (V2) phenomenon is one of the central issues of modern linguistic theory. This volume examines V2 from a micro-perspective, comparing various languages and periods. At the heart of the work presented here lies the clear assumption that Verb-second is by no means a unitary phenomenon, but rather a heterogenous and rich system that affects languages in various ways.

This article proposes a new typology of the V2 property, integrating new data from a corpus of Medieval Romance texts with data from Rhaeto-Romance, Early Germanic and Modern Germanic. The proposed analysis is that all V2 systems have a V-movement and phrasal movement trigger on the lowest left-peripheral head, Fin, and that in a subclass of V2 languages Force also has these properties. It is argued that the restrictions on and variation in licensing verb-initial and verb-third clauses within Romance and Germanic V2 systems fall out from the Fin/Force distinction.

In this article I will show that there is not a single diachronic path that leads to the loss of the verb second property so that different languages can have different ways towards a non V2 grammar. I will also show that the progressive loss of subject inversion contexts is not the only factor that played a role in the change, because there are V2 languages in which nominal subject inversion was never attested but they are still V2, i.e. they have I to C movement. I will identify various factors which may play a role and are related to the type of verb second found in different languages. I will consider various types of factors and show that they can play a role or not in the loss of V2 depending on the type of V2 that the language displays. I will also identify an additional factor that has not been considered in the literature namely the type of complementizer system: a change in this area can contribute to destabilizing the V2 grammar.

Through a focus on the properties of subject-finite verb inversion and XP fronting in three relaxed V2 languages, namely Cimbrian, Ladin and Mòcheno, this paper aims to widen and refine our understanding of relaxed V2 languages, i.e. languages in which the V2 property should be understood in a technical sense as obligatory V-to-C movement, not as a simple description referring to linearisation (Benincà 2006, 2013; Ledgeway 2016). It will be shown that inversion differs across relaxed V2 languages in two ways. In a first subtype, inversion is not associated with any marked pragmatic interpretation of the lexical subject and the subject appears in an A position in the IP area: this type is instantiated by Old Italian (Benincà 2006, Poletto 2014). A second option, instantiated by the languages considered in this paper, is that the lexical subject receives a pragmatically marked interpretation which is encoded in a Functional Projection (FP) in the vP periphery (Belletti 2004, Poletto 2006). This paper confirms that V3/V4 word orders involve the presence of a double articulation for foci and wh-elements, which appear in different positions in the CP layer in relaxed V2 languages (Poletto 2002, Wolfe 2015 a,b). It also contributes to our understanding of the syntax of topics in relaxed V2 languages by showing that (i) topics can be moved to CP and (ii) the movement option is not restricted to main clauses lacking an XP in the left periphery; it also occurs in interrogative clauses (unlike in the relaxed V2 varieties considered in Walkden 2014, 2015).

In this paper, I sketch the CP layer in main and embedded clauses in the history of English. The Modern English main clause is not as easily expandable as the Old English one, but the reverse is true in the subordinate clause, where Modern English has a more flexible embedded CP than Old English. I focus on the developments of the embedded CP. It has been claimed that Old English lacks an embedded split CP and therefore lacks embedded V2 and a host of other embedded root phenomena. I show this to be true for complements to both assertive and non-assertive verbs. In contrast, the Modern English matrix verb has an effect on the strength of the C-position. Assertive verbs in Modern English allow main clause phenomena in subordinate clauses whereas non-assertives typically do not. The main point of the paper is to chronicle the changes that ‘stretch’ the embedded clause and the changing role of main verbs. It is descriptive rather than explanatory, e.g., in terms of changes in phase-head status.

Old French is considered by many to have been a verb-second (V2) language. Furthermore, 13th century Old French featured a V2 system with strong restrictions on the prefield, meaning only a single constituent was generally accepted to the left of the finite verb. This bears a strong resemblance to the pattern found in the Modern Germanic V2 languages and has occasionally given rise to suggestions that V2 was a Germanic property inherited from the language of the Franks. In this paper, a concrete hypothesis is developed for the diachronic evolution of Old French V2 from Late Latin. It is argued that the hypothesis of Germanic influence is not necessarily incorrect, but too simplistic, as the two synchronic components of the Old French V2 construction -namely V-to-C movement and restrictions on the prefield – most likely have their own and independent diachronies as well. Comparative and historical evidence is presented to show that V-to-C movement is very unlikely to have been a product of Germanic influence and should rather be considered an internal development from Latin. As for the restricted prefield (so-called ‘linear V2’), the scarcity or even absence of evidence does not allow firm conclusions, but some general theoretical insights from the literature on language change and second language acquisition combine to make the idea of Germanic influence quite plausible.