Allomorph selection takes place on the basis of specificity, by means of the Elsewhere Principle. But it is also constrained by considerations of locality; in other words, the conditioning environment for an allomorph must be not only present, but also accessible in the relevant structure.

A crucial window onto the nature of the locality conditions is visibility: we can learn about the locality domains of allomorphy by examining when two elements X and Y can ‘see’ each other. But an equally important, and often underdiscussed, consideration is intervention, whereby X and Y stop ‘seeing’ each other when a third element Z intervenes.

Three case studies from Modern Greek exemplify the importance of intervention. I dub the phenomenon of interest default by intervention: when the context for a specific exponent is present but inaccessible due to intervening material, the insertion mechanism retreats to a default. I argue that this follows from a strict linear adjacency condition, and discuss root suppletion and the mechanism of Pruning in the process.

Publications:

The operation Agree has been invoked to explain Condition A of the binding theory, with mixed successes. Many Agree-based analyses of binding make a crucial, and hitherto underexplored, prediction: binding and agreement should go hand-in-hand, that is, there should be situations where you agree only if you bind.

We show that Turkish exemplifies a case of just this kind, providing crucial evidence in favor of the involvement of Agree in binding. At the same time, we argue that this involvement is more limited than sometimes argued: Agree is implicated in some, but not all instances of binding, with the dividing line perhaps involving an argument/adjunct distinction.

Publications:

Anaphora and agreement in the Turkish DP: Delimiting binding-through-Agree [Published in Natural Language & Linguistic Theory]

Analyses of clitic doubling abound, and empirically distinguishing between them for a given language often proves difficult. Modern Greek provides a crucial disambiguating data point: it is possible to clitic-double first conjuncts. This fact argues against movement-based analyses of clitic doubling, as these would incorrectly rule out first conjunct doubling as a violation of the Coordinate Structure Constraint. The only type of analysis compatible with the facts is one based solely on Agree, where the clitic is a probe copying the features of the doubled DP, without any movement. First conjunct doubling also provides insights into the mechanism of first conjunct agreement, and the analysis of the Person Case Constraint.

Publications:

First conjunct clitic doubling in Modern Greek: Evidence for Agree-based approaches to clitic doubling. [Published in Natural Language & Linguistic Theory]

First Conjunct Clitic Doubling, the Person Case Constraint, and First Conjunct Agreement: Insights from Modern Greek. [Published in Glossa]

Reflexivity is often associated with the presence of an overt anaphor; but in many languages, reflexive interpretations can also be achieved by means of a special morphosyntax, e.g. by different forms of ‘intransitivization’ yielding a reflexivized form.

Persistent questions on such forms concern what the precise link is between intransitivization and reflexivity; whether the relevant verbs are unaccusative or unergative; and how a single argument can be associated with two roles, including whether a dedicated reflexivizing operator is involved.

Using Modern Greek as a case study, I show how the hallmarks of reflexivization in this language — namely, nonactive morphology, unaccusative syntax, reflexive semantics, and the presence of the prefix afto- ‘self’ — all follow if reflexivity is tied to the presence of a reflexivizing Voice head.

Reflexivity and external argument introduction: Insights from Greek. [ms; comments welcome!]

Stative passives (e.g. the door appears opened) are often offered, in syntactic approaches to word-formation, as a prototypical case of phrasal layering: under many analyses, a stative passive amounts to a stativized eventive passive, where a large phrasal structure is embedded underneath a stativizing layer.

Greek has been the poster child of such analyses of stative passives, owing to the observation that the language arguably freely allows agent-oriented modifiers.

Identifying a range of novel generalizations, I show that Greek in fact widely counter-evidences a layering approach: stative passives do not contain eventive passives. The facts instead favor a complex head approach where stative passives are syntactically constructed but structurally ‘smaller’ than their eventive counterparts.

Phrasal layering versus complex heads in Greek stative passives. [NELS proceedings paper]

Discussions of blocking — i.e. cases where the ungrammaticality of one form is ostensibly related to the existence of another, competing one — are often narrowly circumscribed, with many focussing largely on derivational morphology. In this overview of approaches to blocking effects, we show how blocking bears on larger-scale questions on grammatical architecture, revolving around the question of what role competition plays in the grammar. We focus not just on cases of blocking, but also on the absence thereof, where apparently competing forms coexist, and conclude that what blocking, often discussed as a unitary effect, often follows from multiple interacting mechanisms.

Publications:

Blocking Effects [to appear in The Wiley Blackwell Companion to Morphology]

Implicit complements and the identity condition on ellipsis [CLS proceedings paper]

Indefinite null objects in Greek: Distinguishing between weakly equivalent ellipses [handout from a talk at NYU. Caveat lector: early work, wacky in parts]