Volume 31, Issue 3-4 - Glow Issue

In this paper we investigate the case split on the possessor in Udmurt. Traditionally, the choice between ablative and genitive possessor case is said to be driven by the grammatical function (GF) of the XP containing the possessor. We argue that the case split is not driven by GFs; rather, it is determined by the case value of the XP that contains the possessor. Importantly, there is no evidence which points to a possessor raising analysis in Udmurt. Instead, we present an analysis according to which the possessor always bears genitive but may be assigned another structural case by an external head. Due to a morphological constraint, stacked case features fuse into a single feature set in the postsyntactic morphological component. If accusative and genitive stack on the possessor, only the default semantic case marker, i.e., the ablative marker, can realize the resulting feature set. In any other context the genitive marker is chosen. We thus claim that there is no abstract ablative on the possessor; instead, the morphological ablative marker realizes a combination of two abstract structural cases.

Work on coordination has been concerned with the conditions in which elements can be “shared” between two conjuncts since at least Ross (1967). This paper aims to describe a curious case of apparently obligatory sharing, where coordinated clauses below a shared element in C – i.e. coordinated TP clauses – are required to “share” a single initial subject unless the subject of the second clause is focused. I argue that this restriction arises from properties of cyclic linearization (Fox and Pesetsky 2005, Richards 2010). Limitations of the linearization algorithm prevent it from distinguishing DP arguments from one another (Richards 2010). What distinguishes the subjects of TPs from arguments of other clausal conjuncts is that they are both visible to the linearization algorithm on a single cycle of Spell Out. Unable to distinguish the two subjects, this algorithm linearizes them in a single position, except in cases where subject-oriented focus requires an overt subject in the second clause. The final section of the paper extend the analysis to SLF coordination (Höhle, 1983, 1990), which shows a similar restriction on overt subjects in the second conjunct.

A well established assumption in the literature on null arguments is that so-called consistent pro-drop and radical (or discourse) pro-drop are distinct phenomena, characterizing different types of pro-drop languages. This paper argues that this dichotomy must be abandoned in favor of a unitary approach, and proposes to reduce both types of pro-drop to ellipsis of full-fledged argument DPs. It then shows that Fox' (2000) NP-Parallelism requirement, which accounts for the strict-sloppy ambiguities in VP-ellipsis, directly carries over to the same ambiguities in pro-drop (i.e. argument-ellipsis), and that the analysis explains a set of seemingly problematic data. The approach put forth here ultimately leads to characterizing the patterns of pro-drop observed across languages as epiphenomena.

In this paper, I propose that positional neutralization processes operate by deleting contrastive features from the bottom of language-specific contrastive hierarchies, rather than by restricting certain positions to one of the members of a neutralized opposition. This means that non-terminal nodes of the contrastive feature tree are licit lexical representations, and can be thought of as a principled implementation of the Prague School archiphoneme. I argue that such an analysis is preferable on conceptual grounds because the removal of a contrastive feature in a given position does not imply that the neutralized contrast is available in that position. I demonstrate that this has several empirical consequences: first, the intermediate realization of neutralized segments is expected because the deletion of a contrastive feature removes the phonetic boundary between the two segments that it distinguishes, and second, the set of available neutralizations in a single language is restricted by the set of possible configurations of features allowed within the contrastive hierarchical approach. Finally, I show how this model predicts limited (but attested) ternary feature contrasts, which arise when underlying non-terminal nodes are phonetically implemented.

What prevents pronominals from being locally bound? Does this a) reflect an intrinsic property of pronominals (Chomsky 1981), is it b) a relative (economy) effect, that only shows up where there is a more dedicated competitor (see from different perspectives, Safir 2004, Boeckx, Hornstein and Nunes 2007, Levinson 2000), does it c) have a semantic basis as in Schlenker (2005), or does it d) follow from general conditions on agree based chains, and reflexive predicates (Reuland 2011a)? To help resolve this issue, we investigate Khanty (Uralic, spoken in Northwest Siberia), a language that is reported to allow locally bound pronominals (Nikolaeva 1995), and assess whether it in fact does have them, and, which factors come into play when local binding obtains.

The article addresses the issue of Visser's Generalization (VG), holding that Subject Control verbs are incompatible with the passive, and proposes a solution based on a combination of the passive seen as smuggling and Obligatory Control seen as A-movement (Hornstein 2001, Hornstein & Polinsky 2010). The smuggling-type passive is perfectly compatible with the behavior of Object Control verbs, while also showing that in the case of passivized Subject Control constructions the crucial c-command link between the controller and PRO is not preserved, thus inducing the VG effect. This solution leads to three expectations confirmed in the literature: (a) lack of object promotion, meaning no smuggling of the infinitive and PRO out of the c-command domain of the implicit Agent, does not disturb Subject Control (van Urk 2011); (b) PRO requires c-command by its controller at LF, which holds true once the apparently problematic cases of intraposition and extraposition in Super Equi and the concept of logophoric extension from (Landau 2001, 2010) are reconsidered; (c) the smuggled constituent including the infinitive does not reconstruct to its original position. Alternative proposals are also considered to show the empirical adequacy of the presented solution.