Locative inversion in Germanic and Romance

This chapter investigates locative inversion (LI) in Germanic and Romance, a subject inversion involving a preposed locative expression. LI appears in two primary types: overt LI and covert LI, where a locative reading obtains without a preposed locative. Covert LI occurs in null-subject languages, where a covert locative argument satisfies formal subject requirements. English LI is always overt, yet French only allows covert LI with verbs of appearance. Furthermore, Romance permits embedded and matrix LI, but English prohibits the former. This chapter proposes that cross-linguistic variation follows from varying conspiracies of syntactic and pragmatico-semantic factors. Firstly, verbs of appearance select a locative covert experiencer, which satisfies the French EPP. Secondly, multiple formal ingredients interact in different distributions to produce various instantiations of LI: an EPP in TP, and the ability of locatives to check: the EPP (Dutch), Subject of Predication (Italian), EPP and SoP (French), and EPP and Topichood (English).

The grammaticalization of SIC

The Old Romance continuations of Latin sic, such as Old French si and Old Italian sì, involve four different functions, all of which are referred to here as sic. The first one, which is closest to the original Latin usage, is that of a lexical adverbial, while the other three are functional elements introducing main clauses: the second sic follows elements preposed to the verb, the third one introduces clauses in a narrative sequence of events, while the fourth usage of sic has been described as a ‘weak consequential’ (Salvi 2002) . In this article, it is shown that these instantiations of sic in Old Romance, and in particular the third one, are parallel to the grammaticalized usages of svá in Modern Scandinavian. Furthermore, it is argued that the distribution of these functional elements in Old Romance, here represented by French and Italian, as well as Modern Scandinavian, represented by Swedish, can be successfully accounted for in a theory of syntax that incorporates certain notions of ‘narrative’, building on intuitions originating in Labov (1972) and subsequent work.

Issues in the left periphery of Old French and Old English

Topicalization in Old English and Old French is a poorly studied field. This chapter aims to identify and compare their different strategies for topicalization. We find that both Old English and Old French have evidence for high and low topics, which we argue are Hanging Topics and Left Dislocated Topics. In the case of Old English, we find evidence that both kinds of topics have been derived by movement into the left periphery. Old French displays a hybrid system where Hanging Topics have been base generated, while Left Dislocated Topics have been moved. Applying a Force-Fin-V2 analysis, we find that even in a Force-V2 language there may be Left Dislocated Topics. We propose that the Force-Fin-V2 model be refined, so that moved Left Dislocated Topics block further movement of the finite verb to Force, so that even in a Force-V2 language, the finite verb will be in Fin in the presence of moved Left Dislocated Topics.

Reflexive constructions in German, Spanish, and French as a product of cyclic interaction

German sich and Spanish se can have reflexive or anticausative interpretations but only Spanish se can have a passive interpretation. We argue that Spanish Passse is the result of interaction between the subject agreement cycle and the reflexive object cycle. We make two claims: i) pro merges in Spec-Voice in Passse, due to the subject agreement cycle; and ii) se heads Voice due to the reflexive cycle. The types of reflexive constructions a language has depends on the presence/absence of pro and the categorial status of the reflexive pronoun (head or DP). French appears problematic since it has Passse but lacks subject pro. However, Passse existed in Old French (Cennamo 1993), which was a null subject language (Vance 1997). Thus, French is consistent with this claim; i.e., it developed Passse when it had subject pro and se as a head. Passse survived into Modern French as a historical remnant.

Second positions

This paper is based on three central analytical ideas. First, following Roberts (2010), I assume that narrow-syntactic head-movement is a reflex of the Agree relation where the Goal is defective in relation to the Probe, in the sense that the Goal’s features are a subset of those of the Probe. Second, following Chomsky (2008), I assume that phase heads drive all narrow-syntactic operations in virtue of their uninterpretable formal features and their Edge Features (EF). Third, owing to the nature of head-movement as a purely Agree-based operation, head-movement to a phase head PH cannot satisfy PH’s EF. The third assumption has the important consequence that second-position (P2) effects arise where C’s uninterpretable formal features attract the verb or clitic under Agree/head-movement, and so C’s EF must attract some XP to its edge, subject to a discourse interpretation. The paper applies these ideas to the analysis of various cases of second-position effects across languages, mainly Romance and Germanic.

New prepositions in the house

In the literature on semantic and categorial change French chez and Mainland Scandinavian hos are often cited together as parallel examples of locative prepositions deriving from nouns referring to the concept ‘house’. In this paper we compare in detail the philological records and the more recent development of the two items as well as that of the cognate Insular Scandinavian hjá. We show that while there are similarities in the development of Latin CASA / French chez and hos, as frequently suggested in the literature, there are also significant divergences. We argue in favour of a reevaluation of the origin of hos aligning it with hjá rather than casa as suggested in Noreen (1892), and show that if so revised, the differences can be shown to arise from the different meanings of the source terms: Latin casa ‘hut, house’ and later ‘place’ as opposed to Old Swedish hos and Old Icelandic hjá ‘group of people, company’. We then go on to explore the consequences of these different diachronic trajectories for our general understanding of the connected semantic and syntactic developments and the time course of categorial change.

Topics in French and Norwegian

In this paper, we compare the structural and syntactic properties of different kinds of topics in French and Norwegian. We argue that both languages permit Hanging Topics and Left Dislocated Topics, but that the differences between the two types of topic are subtle. Properties that have been described for Italian and German cannot be extended to French and Norwegian, indicating that topicalization has language-specific characteristics. A fundamental difference between French and Norwegian is that while topics in French are base-generated, topics in Norwegian are moved.

Against complementizers

In this paper we investigate the syntax of complement clauses in some Romance and Germanic languages by focusing on word order asymmetries and extraction phenomena. We argue that complement clauses are relative clauses, as proposed in Manzini & Savoia (2003, 2011) and Kayne (2010). However, differently from the previous proposals, we claim that as in relative clauses (see Poletto & Sanfelici 2018a), the ‘complementizer’ partially spells out either the nominal element internal to the relative/complement clause, resulting thus into a raising derivation of the relative/complement clause, or the external nominal modified by the relative/complement clause itself, leading to a matching derivation. This difference in the raising vs. matching derivation accounts for a series of well-known asymmetries between some Romance and Germanic languages. In addition, we show that this proposal may be suitable to derive the different extraction patterns exhibited in ‘traditional’ relative clauses and complement clauses.

‘Have’ in English and Romance

This paper compares and contrasts the forms and usage of the verb ‘have’ in different varieties of English and Romance. ‘Have’ may serve as a verb of possession, a perfect auxiliary, a deontic modal, and a presentative in both English and Romance. However, its English uses as an epistemic modal, as a dynamic light verb, and as a causative have no counterparts in Romance, whilst the ‘double perfect’ of many Romance varieties is absent from English. Although restricted to certain varieties, parallels may be noted between English and some Romance languages (dissociation of auxiliary and possessive ‘have’; progressive loss of past participle agreement and of the competing auxiliary ‘be’ in the perfect; apparent ‘aoristic drift’). The paper concludes by discussing the significance of the data for the distinction between main verbs and auxiliaries; the status of ‘have’, particularly as a verb of possession; and the nature of grammaticalization.

Pro-drop in interrogatives across older Germanic and Romance languages

This chapter investigates the mechanisms of null subject licensing in direct interrogatives, an environment which is generally neglected in investigation into null subjects, using data from a range of early Romance and Germanic languages considered to be asymmetric pro-drop languages, i.e. languages in which null subjects are favoured in main clauses. We find that there is subtle variation between the languages in question, but that two factors in particular – interrogative type and person – are crucial in conditioning this variation, and we sketch analyses based on the differential availability of Agree relations with left-peripheral elements. Therefore, null subjects in main interrogative clauses are licensed in two slightly different manners in the two language families – a fact which we show follows from differences in the structure of their left periphery and in agreement morphology