Volume 31, Issue 2

This article presents evidence that, cross-linguistically or within the same language (family), there appears to be no morphosyntactic properties and/or structures specifically designated for the formation of middle voice constructions. What has been labeled a ‘middle voice construction’ is a semantic interpretation that, crucially, is blocked when an event variable is existentially closed by T. This article focuses on two ways of expressing a middle statement; namely (i) middle voice readings that occur with lexical-s passives, and (ii) adjectival middles – in Mainland Scandinavian, showing that properties such as the availability of an agent in middles pattern with whether an event variable is present (in the structure) or not. These are the result of two equally valid and productive grammatical structures: one where an event variable is present, an agent is projected and a modal operator blocks existential closure of the event variable, and another one where the event variable is not present in the structure, and therefore the operator is not necessary – hence impossible.

This article bears on three “verb cognate” (VC) constructions attested in Haitian Creole, where the verb and a homonymous replica of the verb cooccur within the same clausal domain. In the first two constructions – VC1 and VC2 – the verb's cognate is crucially unstressed and supports further lexical material, while in the third construction – VC3, which subsumes the subcase known as Predicate Cleft – the verb's cognate bears primary stress and is reduced to a lexical head. Haitian VC constructions differ from so-called “Cognate Object” constructions discussed for English and other languages by the strict-identity condition which – in Haitian – bears on the verb and its cognate, and by the arguably non “nominal” categorial specification of the verb's cognate. We describe the morphosyntactic properties and interpretations of VC1, VC2 and VC3, and pro-pose a syntactic derivation for each construction. We argue that the Cognate Phrase in all three cases instantiates a VP modifier – an adverbial – rather than an argument, but that the “cognateness” property should not be uniformly analysed across the three constructions: in VC1 and VC2, we argue that the verb's cognate is an expletive – a phonological filler whose function is to make a structural (head) position visible – while in VC3, we analyse it as a freely merged VP-modifier whose lexical identity with the verb triggers a contrastive-focus effect, which favours raising of the focused modifier to the clause periphery. Haitian verb cognates share some characteristics of “reduplicative” structures as presented by Stolz et al. (2011), but they do not seem to comply with these authors' view of “Total Reduplication”. Our results and assumptions are on the other hand consistent with Pereltsvaig's (1999a, b, 2002) claim that “Cognateness”, as such, has no relevance for syntactic theory.

The syntax of auxiliaries has given rise to much discussion in the generative literature (Akmajian and Wasow 1975; Emonds 1978; Akmajian et al. 1979; Pollock 1989; Chomsky 1993; Lasnik 1995b; Roberts 1998; Bjorkman 2011; Rouveret 2012). This paper explores the distribution of non-finite auxiliaries in Standard English, in particular the issue as to whether such auxiliaries raise for inflectional purposes or remain in their base positions and have their inflections lowered onto them. It is shown that auxiliary distribution is not determined by auxiliary type (passive, copular, progressive etc.) as the lowering accounts predict, but by the morphological form that the auxiliary takes. In particular, the auxiliaries be/been and being exhibit significantly different distributional properties across ellipsis, fronting and existential constructions in English that are difficult to capture under an affix lowering model, and lend themselves more easily to an auxiliary raising account. I therefore offer a syntactic account of auxiliary inflections which employs the theoretical uniformity of an Agree-based approach, with the empirical advantages that an auxiliary raising analysis affords. The auxiliary raising system that will be proposed essentially harkens back to Chomsky's (1993) and Lasnik's (1995b) approach to the auxiliary system, though with the utilisation of Bošković's (2007) notion of foot-driven movement.

In this paper I raise the following questions: (i) How appropriate/effective are traditional autosegmental representations? (ii) What insights do autosegmental representations help us express? (iii) Where do autosegmental representations fall short? (iv) What does this all mean for phonology? Although there is still general acceptance of the “basic autosegmental insight”, i.e. that certain features are semi-autonomous from each other and their anchors (e.g. tones vs. tonebearing units), there has been decreasing interest in representational issues in some current approaches to phonology, as well as proposed frameworks which abandon some of the key properties of autosegmental phonology. In this paper I will be particularly interested in the intersection of the following properties assumed in traditional autosegmental phonology: (i) assimilation = spreading; (ii) no line-crossing; (iii) branching elements ≠ successive linking of identical elements. I will first review these three autosegmental assumptions – and the good things they do for us – and then present some unexpected tonal data which are potentially problematic for one or another of these assumptions. After identifying what makes the cited cases problematic and the choices we have in accounting for them, I will conclude with a discussion of the implications such facts have concerning phonology in general.

When a verb takes two complements, the first complement, typically an accusative object, c-commands the second complement, and the second complement can host a relative clause containing antecedent contained deletion (ACD). This simple fact, which, upon careful considerations, reveals that English accusative objects raise to a higher position than previously thought. The theoretical claim is that objects raise to check Case to the outer vP-spec. We argue that the v consists of two separate heads, the higher v [acc] and the lower v θ ; objects raise to the outer spec of vP θ ; and that the verb moves to the head of the higher v [acc] .