The phenomenon of verb second (V2) has received considerable attention in the syntactic literature, especially in relation to the Germanic languages. In this book, Federica Cognola discusses the empirical and theoretical foundations of V2 in relation to the language of Mòcheno, a minority language spoken by around 580 people in the villages of Palù, Fierozzo and Roveda in the Fersina Valley in Trentino, Northern Italy.
Mòcheno has traditionally been analyzed as a German dialect, developed from the language spoken by German settlers in the 12th and 13th centuries. Compared to present-day German, however, Mòcheno displays a rather striking degree of variation and optionality at the syntactic level, including the presence of both OV and VO word orders, a non-strict V2 rule, and a complex distribution of NP and pronominal subjects.
Though not the first study on Mòcheno, this book (a revised version of Cognola's University of Padova 2010 Ph.D. thesis) is based on ‘the most extensive collection of syntactic data ever made for this language’ (p. 3). An impressive number of informants from distinct language environments, namely different families and farms, were interviewed for the study. The informants were asked to do both grammatical judgment tasks and free translations from Italian. As every question and answer for each informant is reproduced in the appendix, the empirical part of this study is fully transparent.
Previous non-syntactic accounts of Mòcheno (most notably Rowley Reference Rowley1986, Reference Rowley2003) have attributed the observed variation and optionality to language contact and bilingualism, as Mòcheno speakers have been bilingual for at least three centuries. Cognola refutes the view of ‘competing grammars’, and instead argues that the syntactic variation in Mòcheno can be fully accounted for within the confines of a single grammar. The main hypothesis is that Mòcheno has a V2 rule similar to that of the Old Romance languages (and not to that of German): the obligatory movement of the finite verb to CP co-exists with multiple access to the left periphery. Cognola further argues that the apparent syntactic variation is, in fact, systematic in relation to this ‘relaxed’ V2 rule. Her claim is both bold and fresh, and has potential bearing not only on our understanding of Mòcheno, but of the V2 phenomenon as such.
The theoretical discussion begins by defending the claim (originally made by Rowley Reference Rowley2003:281) that Mòcheno is a V2 language. This claim is by no means uncontroversial, at least if V2 is understood as a linear restriction. Consider the Mòcheno examples in (1) and (2) below.
(1)
(2)
As these examples show, Mòcheno does not display the ‘core properties’ of the V2 restriction, i.e. that the finite verb (i) has to be in second position, and (ii) must be preceded by exactly one arbitrary constituent. This fact is acknowledged already in the introduction to the book. Instead, it is argued that Mòcheno displays ‘correlated properties’ of V2. In the Germanic languages, these properties relate to (i) obligatory subject–verb inversion when non-subjects are fronted, (ii) structural correspondence between main declaratives and main wh-interrogatives, and (iii) structural asymmetry between main and embedded clauses. Mòcheno does not display these Germanic properties either, however: subject–verb inversion is possible but not obligatory (as shown in (1b) above), declaratives and wh-interrogatives are not necessarily structurally similar, and the asymmetry between main and embedded clauses does not concern word order, but the possibility of pro-drop. Following Benincà (Reference Benincà, Zanuttini, Campos, Herburger and Porter2006), Cognola argues that these are the correlated properties of V2 as found in the Old Romance languages. Thus, the V2 rule of Mòcheno is not similar to that of the Germanic languages, but rather to that of the Old Romance languages (in which the linear V2 restriction was also relaxed).
As this book is based on extensive new data from an understudied language, one would have liked the empirical presentation to be clearly separated from theory-dependent claims. But already the initial classification of Mòcheno as a V2 language is based more on theoretical reasoning than on undisputable empirical facts. This, in turn, makes the subsequent theoretical discussion somewhat unbalanced, as it serves both to defend and account for claims that, at times, seem to run counter to the (admittedly quite complex) data.
Set in the minimalist framework, Cognola's syntactic account of the V2 rule of Mòcheno builds on an elaborated C-domain in the sense of Rizzi (Reference Rizzi and Haegeman1997). According to her proposal, the finite verb must raise to the CP in all main clauses, more specifically to [Subj0] (assumed to be the lowest projection of the C-domain). SubjP is endowed with an EPP-feature that must be checked by movement of either a constituent or a silent pro to [Spec,SubjP]. The analysis hinges on the assumption that not all constituents ‘count’ for V2, which in theoretical terms translates to not being able to satisfy the EPP requirement. As long as one constituent is able to check the EPP-feature, there is nothing to prevent further fronting of constituents. This account thus captures the observation that Mòcheno allows several constituents in front of the finite verb.
Three different ways of promoting constituents to the C-domain are distinguished: Simple preposing (SP), Left dislocation (LD) and Hanging topics (HT). In Chapter 4, it is argued that only in case of SP, the fronted element establishes a Spec–head relation with the finite verb and thus satisfies the EPP requirement. The main empirical support for this claim is the observation, presented on page 127, that ‘no material can intervene between the fronted constituent and the finite verb’; see (3) below. In contrast, neither LD nor HT constructions require the fronted element to immediately precede the finite verb, as shown in (4) and (5).
(3)
(4)
(5)
The claim that LD and HD constructions do not satisfy the EPP requirement predicts that they should be able to co-occur with SP constructions; this prediction is also borne out.
The analysis proposed in Chapter 4 is subsequently challenged in Chapter 6, where it is shown that Mòcheno actually does allow intervening constituents also in SP constructions, as illustrated in (6):
(6)
This observation forces Cognola to assume that SP does not, after all, satisfy the EPP requirement, and that the previous conclusion was ‘wrong due to an incorrect interpretation of the data’ (p. 177). As the first conclusion (in Chapter 4) is presented as definitive, the revised conclusion in Chapter 6 comes across as entirely unexpected, and one is left to wonder why the revised conclusion was not presented already in Chapter 4.
The revision is not a mere presentational problem, however, as Cognola's entire syntactic analysis is founded on the assumption that the EPP-feature must be checked by the fronting of an element to [Spec-SubjP]. If the only fronted element fails to satisfy the EPP requirement, the analysis of Mòcheno as a V2 language ultimately becomes hard to maintain, especially in the light of the non-linear requirement noted above.
In order to uphold the original analysis, Cognola proposes that EPP can be satisfied by a silent pro in all cases in which no operator is fronted (i.e. a wh-word). While this analysis may be descriptively accurate, its explanatory value is not entirely obvious. First, the pro analysis does not extend to the Germanic V2 languages, as they are not pro-drop languages, as noted also by the author (p. 182). The pro-checking of EPP is thus an isolated Mòcheno trait, which makes the analysis slightly ad hoc. Second, the analysis builds on the assumption that Mòcheno is a pro-drop language, an assumption that is not unambiguously supported by empirical data.
Chapter 5 is devoted in its entirety to corroborating the hypothesis that Mòcheno is a partial pro-drop language. The problem is that Mòcheno does not fulfill the core property of consistent null-subject languages, i.e. the possibility of omitting definite subjects. In fact, Mòcheno subjects must be overtly realized in all syntactic positions. The only apparent exception to this rule involves the second person singular clitic in non-subject initial clauses in the Fierozzo and Roveda dialects. As was the case with the ‘relaxed’ V2 rule, Cognola instead argues that Mòcheno displays correlated properties of pro-drop (free inversion, that-trace violations and expletive null subjects). But not even these correlated properties are straightforwardly supported by the data; the discussion on expletive null subjects is especially unconvincing. Throughout this chapter, one cannot help but feel that the empirical data is sometimes made to fit the theory, rather than the other way around: the non-linear V2 rule of Mòcheno is supported by the existence of a pro-drop rule that, for the most part, does not allow actual subject omissions.
The claim that Mòcheno has a ‘relaxed’ (non-linear) V2 rule, in combination with the theoretical assumption that V-to-C movement does not exclude the fronting of several constituents, raises fundamental questions on the nature of the V2 phenomenon. If the proposed analysis is correct, the V2 requirement must clearly be regarded as an abstract rather than a concrete syntactic rule, without any necessary manifestation in ‘surface’ syntax. The obvious problem, from both typological and theoretical points of view, is then how to distinguish V2 languages from non-V2 languages – especially if the ‘correlated’ properties of V2 are relaxed as well. While some of these issues are briefly mentioned, Cognola does not address the greater theoretical implications of her theory in any detail.
The pronounced aim of this study is to provide a syntactic analysis of Mòcheno, in contrast to previous, sociolinguistically-oriented treatments of the language. However, the one-sided focus on syntax and theory-internal technicalities potentially blurs other aspects of relevance for our understanding of the grammar of Mòcheno. Throughout the entire study, Cognola frequently touches on the relevance of information structure, discussing different kinds of topics, focus and the distinction between pragmatic and syntactic subjects. As information structure is known to be of relevance for the ordering of constituents in the left periphery (mirrored also in the Rizzian C-domain), a more in depth discussions of these matters would probably have benefited the analysis.
In addition, this study raises a number of open questions related to the historical development of Mòcheno. Naturally, one cannot criticize a synchronic study for lack of historical data. But given that the main claim implicates that fundamental aspects of Mòcheno syntax have Romance origin, the diachronic development of this German language variety is highly relevant for the main hypothesis.
Overall, this study makes an important empirical contribution to our knowledge and understanding of an endangered minority language. It also challenges the ‘traditional’ notion of V2 as well as proposes a novel syntactic analysis of this phenomenon, the correctness and validity of which will have to stand up to scrutiny in future theoretical discussions.
