Susana Bejar
“Phi-Syntax and PF Legibility”

If we take vocabulary insertion (VI) to be a step in the externalization of a syntactic structure, then the PF interface-legible elements of a syntactic structure must be understood to include at least the features and structural relations that are determinative of VI outputs. Contextual allomorphy, in this view, should be (at least in part) systematically correlated with differences in the specification/distribution of interface-legible elements of the syntactic structure, and is therefore an important kind of evidence to consider in connection with questions about grammatical architecture and modularity.

This talk examines cases of inflectional allomorphy that are indicative of PF legibility of various *abstract* derivational elements postulated within syntactic theories of Agree. These provide prima facie support for an architecture in which an independent syntax module feeds or is interpreted by externalization systems. However, the talk also considers empirical and abductive challenges to this position, and in particular the special methodological challenges posed by morphological evidence: idiosyncrasy and narrow distributions of patterns.

Nicole Dehé
“Parenthesis and the syntax-prosody interface”

Parentheticals have been defined as syntactic elements which are linearly integrated in the syntactic string, but are otherwise unrelated to the surrounding linguistic material, i.e. in terms of syntactic structure, semantic meaning, and/or intonation. It has also frequently been argued that parentheticals form intonational phrases (IPs) of their own, the main reason being their syntactic separateness.

In my work, I have looked at the prosodic phrasing of various types of parentheticals. In my talk, I will go back to that data and relate the results to more recent insights in a theory of the syntax-prosody interface.

Vera Gribanova
“Syntactic and postsyntactic operations: Consequences for ellipsis”

In this talk I examine paradigms of crosslinguistic variation concerning the verbal identity condition in verb-stranding ellipsis, building on a recent proposal about the mechanisms that yield head movement configurations (Harizanov and Gribanova, 2018).

When phrasal material is extracted from ellipsis sites (e.g. in sluicing), violations of lexical identity of the extracted material are permitted under focus of that material (Schuyler, 2001; Merchant, 2001). This is usually attributed to the licensing condition on ellipsis (Rooth, 1992; Heim, 1997; Merchant, 2001), which takes distinct variables inside the ellipsis domain and its antecedent to be identical. I focus on analogous paradigms with head movement out of ellipsis sites (yielding verb-stranding), which appear to lead to contradictory conclusions regarding the architectural status of head movement. Languages like Russian – among them Hungarian, European Portuguese, and Swahili – permit mismatches between extracted parts of the verbal complex and their corresponding antecedent components under focus, just as with phrasal extraction in sluicing. Languages like Irish do not permit such mismatches under any circumstances, pointing to a postsyntactic status for head movement: there is no genuine movement out of the ellipsis site, giving rise to a total identity requirement (Schoorlemmer and Temmerman, 2012; McCloskey, 2016).

A point of leverage into understanding these patterns comes from a proposal by Harizanov and Gribanova (2018), who argue in favor of a bifurcation, both empirical and theoretical, in head movement types. One type involves displacement of fully formed words to higher syntactic positions (e.g. verb second, long head movement). The other type constructs complex morphological words (e.g. affixation, compounding). They point out that the empirical properties of the two types are distinct, and justify a theoretical move in which they correspond to distinct operations, in distinct modules of the grammar. They propose that the operation responsible for upward displacement of heads is genuine syntactic movement (Internal Merge); on the other hand, word formation is the result of postsyntactic amalgamation, which has properties that are not associated with narrow syntax.

With this revised view in hand, we can revisit the paradoxical verbal identity patterns: we expect that mismatches in verb-stranding ellipsis will be permitted when head movement is syntactic, but not when it is postsyntactic. I present independently motivated analyses of Irish and Russian clause structure which support exactly this conclusion. Verb movement in Irish involves postsyntactic amalgamation only, predicting a strict lexical identity requirement. By contrast, verb movement in Russian involves both the syntactic and the postsyntactic head movement types, with one of the movement steps being syntactic and giving rise to the possibility of verbal mismatches in verb-stranding ellipsis.

Patricia Irwin
“The place of object positions in discourse”

This talk will show that an examination of the discourse properties of certain intransitive sentences can open up some useful ways of thinking about the relationship between syntax and discourse. I will show that analyzing sentences like “a lady danced in” as syntactically unaccusative and semantically as containing an existential proposition accounts for numerous syntactic and interface-level phenomena. I will also use the analysis to probe the division of labor between verbal roots and syntactic structure in order to shed light on current debates on argument structure and the nature of roots.

Peter Jenks (joint work with Hannah Sande)
“Cophonologies by Phase and the Moro verb”

In this talk I present a novel model of the syntax-phonology interface, Cophonologies by Phase (Sande & Jenks 2018) and show how it is able to derive a number of complex morphophonological properties of Moro verbs (Jenks & Rose 2011, 2015). I focus on three issues, all related to the distribution of tone: 1) the ability for lexical subclasses of verbs or morphological categories to condition specific tone patterns, 2) the restriction of these tone patterns to part of the verb, and 3) the ability of these tone patterns to occasionally interact with the tone of object noun phrases and to manipulate the position of object pronouns.

Cophonologies by Phase integrates Distributed Morphology with cophonology theory (Anttila 2002, Inkelas & Zoll 2007). Vocabulary Items are enriched to include three components, one of which is a cophonology, modeled as a constraint subranking. These cophonologies can only take scope in their phase or spell-out domain, and can interact with other morphological processes, including linearization. I demonstrate that this model allow us to account for all of the properties of Moro verbs listed above without the independent stipulations which were needed in earlier approaches to the same phenomenon.

Arsalan Kahnemuyipour
“Where phases have mattered”

In this talk, I will discuss several areas in my work where the notion of multiple spell-out and phases have played a role. I will start with a brief discussion of the sentential stress rule, second position clisis in Eastern Armenian (joint work with Karine Megerdoomian) and Turkish pre-stressing suffixes (joint work with Jaklin Kornfilt), but the focus of my talk will be Split noun phrase topicalization in Gilaki, and time-permitting Persian (joint work with Mansour Shabani). There are two types of split topicalization found in noun phrases in Gilaki. The first one splits a possessive noun phrase, placing the possessum in a clause-initial position, while leaving the possessor in the thematic position of the noun phrase followed by a gap. We call this construction possessor split. This construction is found only in the Eshkevarat dialect of Gilaki (hereafter EG). In the dialect spoken in the provincial capital Rasht (hereafter RG), a resumptive element has to be used in the position of the gap. (This is similar to Persian which also requires resumption in a similar construction.) To differentiate these strategies, I call the first one possessor split with a gap and the second one possessor split with resumption. In addition to this, there is another type of split topicalization found in EG, RG and Persian, which splits a noun phrase, placing the head noun in a clause-initial position, while leaving a numeral in the thematic position. We call this construction numeral split. Note that no resumption is needed/possible in any of these languages with numeral split. Using a battery of tests, I show that while possessor split with a gap is the result of syntactic movement, numeral split involves a base-generated structure. Meanwhile, similar tests indicate that possessor split with resumption also involves a base-generated structure. These proposals raise many questions. Here are a few. In EG, why is the movement strategy unavailable for numeral split and why is the base-generation strategy (without resumption) unavailable for possessor split? What is the difference between numeral split and possessor split, both base-generated constructions in RG, such that only possessor split requires a resumptive element? What is the difference between EG and RG such that the former allows the movement strategy in possessor split, while the latter uses base-generation (with resumption) only? I explore the idea that these questions may be addressed uniformly if we consider the properties of and the constraints on phases.

Jason Kandybowicz
“The anti-contiguity of wh- and C”

1. Introduction
Richards (2010, 2016) proposes a PF condition requiring wh- items to phrase prosodically with their scope-marking complementizers. For Richards, the two items must be prosodically contiguous (i.e. contained within a single Phonological Phrase). Kandybowicz (2017) develops an anti-contiguity proposal in which wh- items are barred from phrasing with overt C at the level of Intonation Phrase (iP).

(1) No iP may contain both overt C and wh-.

This talk presents new evidence in support of Kandybowicz’s anti-contiguity proposal on the basis of two surprising wh- asymmetries in Nupe, a Benue-Congo language of Nigeria. I show that both asymmetries can be directly explained in terms of (1).

2. Asymmetry 1
Embedded constituents can be focused in Nupe by moving to the left periphery of the embedded clause. Embedded wh- items, however, may not appear in left peripheral focus positions below obligatory C. That is, despite the availability of an embedded focus position, Nupe does not allow embedded questions or partial wh- movement.

3. Asymmetry 2
Nupe is not a wh- in-situ language, but like English it allows wh- in-situ in multiple questions. wh- in-situ in multiple questions, however, is limited to root contexts.

4. Analysis of the Two Asymmetries
I show that the two asymmetries above can be derived from (1). Prosodic analysis reveals that overt embedded C in Nupe does not induce iP phrasing of its clausal complement. The phonetic correlates of Intonational Phrasing in the language include the presence of low boundary tones at the right edges of iP, significant pauses/prosodic breaks, and the induction of F0 register reset. Phonological reflexes of Intonational Phrasing include the suspension of regular processes like regressive assimilation and glide formation. At the juncture of C and its clausal complement in Nupe, we find none of these indicators.

As a consequence of the fact that overt embedded C does not induce Intonational Phrasing of its complement, C and its clausal complement are prosodically contiguous within the iP in Nupe. Given (1), it follows that all wh- items in the language are restricted from appearing in embedded contexts, whether moved (i.e. Asymmetry 1) or in-situ (i.e. Asymmetry 2).

Ivona Kučerová
“Phi-features at the syntax-semantics interface”

Theoretical linguistics research has paid significant attention to possible transformations of syntactic features at the syntax-morphology interface. Much less attention has been paid to syntactic features at the syntax-semantics interface. I will argue that phasal transfer involves a translational process that creates feature bundles distinct from their narrow syntax representation. This newly created feature-bundle representation is readable both to the semantics module and to further narrow-syntax operations. Crucially, although readable to LF, the bundles themselves are created without any direct reference to compositional semantics. The presented argument is in line with recent work on argument structure (Wood & Marantz 2017) and non-compositional approaches to semantic representations (Zhai et al. 2014, Schlenker 2015).

Julie Legate
“VoiceP: Reevaluating parameters at the interface of syntax and morphology”

Forthcoming

Mark Norris
“The core and periphery of nominal concord”

In this talk, I discuss the formalization of nominal concord, whereby modifiers of a noun (e.g., demonstratives or adjectives) must match certain features of the noun (e.g., gender, case, or number). After discussing the current status of an ongoing typological survey of systems of concord, I propose a formal system whereby concord is split between syntax and morphology. I propose that the syntactic architecture for concord is uniform across languages that show concord in any form. Variation may occur with respect to (i) which categories participate in the concord system, and (ii) which features particular categories (or even particular lexical items) ultimately express. I propose these areas of variation are best captured in the morphological component. I illustrate with a case study from Ingush, a language with a relatively robust concord system that nevertheless shows a number of idiosyncrasies.

Maria Luisa Zubizaretta (joint work with Roumyana Pancheva)
“On the role of person-based features in the evidential-temporal connection”

At the forefront of syntactic theory is the question of what is the inventory of meaningful (i.e. interpretable) morpho-syntactic features that provide the semantically-relevant building blocks for natural language grammars. Novel insight has been provided to this question by the investigations of languages with the so-called direct-inverse system, such as the Algonquian languages. Paraguayan Guaraní (PG), a tense-less Tupian language, has been shown to also have a direct-inverse system (Zubizarreta & Pancheva 2017). Yet, unlike Algonquian, PG does not mark proximity in the nominal constituents that denote event-participants; therefore this notion plays no role in determining topichood or the form of the direct-inverse inflectional system in this language. Nonetheless, we argue in the present paper that the feature “proximate” and its related feature “distal” are active in other corners of the grammar in PG. They play a role in the characterization of the spatial relation between the speech participants and another individual in the case of nominal demonstratives, and more importantly for our present purposes, they play a role in the characterization of the spatial distance between events in the evidential domain, from which their temporal distance is inferred.