Byron Ahn
“Emphatic weak certainty in English: focus at the syntax-semantics/-phonology interfaces”

Joint work with Sunwoo Jeong and Craig Sailor

In the vast majority of cases across English, syllables marked with focus prosody must be the most prominent syllable within the semantic focus domain (Jackendoff 1972, inter alia). This essentially precludes lexically unstressed syllables from bearing prosodic focus. Despite this, particular linguistic contexts regularly allow lexically unstressed syllables to bear prosodic focus:
[A and B wait impatiently for their train, which is very late]
A: It looks like we’ll be late to the party.
B1: Yeah, appàrentLÝ.
B2: Yeah, pròbabLÝ.
B3: Yeah, màyBÉ.
B4: Yeah, lòoks LÍKE.
B5: Yeah, còuld BÉ.
B6: Yeah, sèems SÓ.
B7: # Yeah, dèfiniteLÝ.
Note that in all cases, the most prominent syllable could also be one where stress falls on a lexical stress (apPÁrently, PRÓbably, MÁYbe, LÓOKS like, CÓULD be, SÉEMS so), however, the cases where the stress is on the final syllable convey a particular meaning. Moreover, the availability of this pattern is constrained by syntactic and semantico-pragmatic factors. Across utterances where this focus phenomenon is acceptable, the following properties are shared:

• There is clausal ellipsis of a proposition, p;
• The speaker agrees with the interlocutor(s) that they cannot fully commit to adding p to the Common Ground (cf. B7).

This talk explores how to derive these facts, on the basis of:

• A clearer distinction made between syntactic, semantic, and prosodic notions of “Focus”;
• A silent morpheme under semantic focus;
• A process by which certain focus-marked silent morphemes can be realized as a prosodic enclitic.

This is part of a larger project that considers where focus pheonemona result in “surprising” prosodic patterns, typologically.

Artemis Alexiadou
“An indeterminate counter in Greek: plurality and imprecision”

In this talk, I will be concerned with the distribution and the morpho-syntax of the element kati ‘some’ in Greek. The peculiarity of kati is that it combines uniquely with plural nouns, although it bears no inflection and does not agree in gender, and case with the noun. Building on Tsoulas (2019), I will argue that kati is an indeterminate counter, favored in informal contexts by monolingual speakers. I will also attempt a syntactic analysis of kati that explains its combination with plural nouns as well as its semantic contribution.

Bronwyn Bjorkman
“When syncretism can (or can’t) fix feature mismatches”

This talk looks at the crosslinguistic profile of phenomena where syncretism—morphological forms that are neutral between different combinations of features—seems able to “rescue” syntactic structures that would otherwise be ungrammatical due to a conflict or mismatch in the features of a particular head or phrase.

Perhaps the most widely known pattern of this type involves free relatives in German, described by Groos and van Riemsdijk (1981). In German free relatives, the case imposed on a WH pronoun within the relative clause must match the case imposed by the higher clause on the free relative as a whole. Mismatches in case result in ungrammaticality—unless the relative pronoun happens to have the same morphological form for both case values, as was (‘what’) is for nominative and accusative. In exactly the case where the pronoun is syncretic for the conflicting case features, mismatches become grammatical.

Resolution via syncretism posses a challenge for many models of morphosyntax, including an interpretive theory like Distributed Morphology (DM, Halle and Marantz 1993 et seq.). For syncretism to ever resolve feature conflicts, structures with mismatching features on a single head must be syntactically licit—yet if they are syntactically licit, a theory like DM predicts that they should always have some grammatical morphological realization, if only by an elsewhere or default form, in which case syncretism would be irrelevant. If the theory is enriched so that feature conflicts can only be resolved in cases of morphological syncretism, however, the theory overgenerates in a different way, incorrectly predicting that all languages would exhibit interactions between syncretism and the availability of feature mismatches.

This talk reviews instances where morphological neutrality has been argued to resolve feature conflicts in Hungarian (Szamosi 1976), Norwegian (Taraldsen 1981), German (Groos and van Riemsdijk 1981; Pullum and Zwicky 1986), Finnish (Zaenen and Karttunen 1984), Russian (Asarina 2011), and Polish (Citko 2018). The profile of these phenomena is contrasted with instances where feature conflicts have been argued to instead be generally resolved via (syntactic or morphological) deletion of features, or via some forms of realizational hierarchy, with the aim of giving an explanatory account of when languages pursue each of these very different strategies.

Laura Kalin
“On the (non-)transparency of infixes that surface at a morpheme juncture”

Infixation is characterized by the intrusion of one form inside of another. Canonical examples involve an intramorphemic position for the infix, e.g., nominalization in Leti: k<ni>akri (<NOM>cry) ‘act of crying’ (Blevins 1999:386). Since infixal positioning is determined with respect to phonological elements (e.g., Yu 2007), and since infixes should in principle be able to combine with complex (multimorphemic) stems, it stands to reason that an infix could sometimes appear (incidentally) inside of an affix; this is indeed borne out, e.g., Leti again: v<ni>a-kini (<NOM>RECIP-kiss) ‘reciprocal kissing’ (Blevins 1999:400). The same reasoning also predicts that an infix could sometimes appear (incidentally) at a juncture between morphemes in its stem, intermorphemically. I ask, when an infix appears between two morphemes in its stem, does it disrupt relations between those two morphemes that we otherwise expect to be strictly local?

To address this question, I investigate the (non-)transparency of 10 infixes (from 9 languages; 8 language families) which can appear between two morphemes where (at least in the absence of the infix) there is some relationship or interaction between the two morphemes. The answer to the question is that infixes disrupt only a limited type of phonological intermorphemic interaction, but never interrupt semantic, syntactic, or morphological interactions/relationships.

These findings follow naturally under approaches where exponent choice (and thus infixation) are post-syntactic and cyclic (bottom-up), e.g., Bobaljik 2000, Embick 2010. The results also provide evidence for distinct types of phonological processes, with different timing, e.g., Kiparsky 1982, 2000, though that will not be the focus of this talk. Alongside other recent results from infixation and allomorphy (Kalin 2020), the present findings elucidate the fine timing of the morphosyntax-phonology interface.

Neil Myler
“Three approaches to Mirror Principle violations and their typological predictions”

Many otherwise divergent theoretical traditions have converged on the conclusion that Mirror-Principle-respecting morpheme orders are the default, and that this follows because by default word-formation proceeds in a semantically compositional fashion. However, no consensus has been reached on the proper account of exceptions to the Mirror Principle. This talk identifies three broad families of approach in the literature, which I dub the surface template approach, the word-internal phrasal movement approach, and the syntactic/semantic reconceptualization approach, and goes on to outline a program for testing these approaches by identifying their differing predictions regarding what sorts of Mirror Principle violations should and should not be attested in natural languages.

Becky Tollan
“Dependency paths and extraction constraints”

This paper examines the ways in which typological asymmetries between subjects and objects, in terms of wh-extraction possibilities, interact with how subjects and objects are case-marked. In particular, a number of languages which distinctly mark transitive subjects (as ergative) disallow extraction of those subjects, a hallmark of ‘syntactic ergativity’. Focusing on the Mayan language family, I present an account of syntactic ergativity based on the grammaticalization of a processing-based preference for nested as compared to crossing dependencies (Hays 1964; Kuno and Robinson 1972; Steedman 1985, Pesetsky 1982; see also Hankamer 1973; Kayne 1981; Kruijff and Vasishth, 2003; Levy and Manning, 2004). I propose that restrictions on the A’-movement of the ergative subject in Mayan languages arise because such movement would cross the prior A-movement path of the absolutive object, creating an illicit crossed dependency. In contrast, the A’-movement of an argument whose merge position is lower than that of the absolutive object creates a licit nested dependency, as mirrored in the A’-movement of accusative objects past A-moved nominative subjects in nominative-accusative languages (which are typically observed to exhibit a ‘subject advantage’; Holmes & O’Regan, 1981; et seq.). This account develops the ‘standard theory’ of syntactic ergativity (terminology from Deal 2016 for accounts based on absolutive inversion, e.g. Bittner and Hale 1996; Aldridge 2004; Coon et al. 2015, Coon et al. 2019) by strengthening its empirical foundation with insights from sentence processing and by increasing its empirical coverage within the Mayan language family and beyond.

Myriam Uribe-Etxebarris
“On polarity and negation in polar question-answer pairs”

The analysis of polar questions (PQs) and their corresponding response particles/answers, constitutes a research topic particularly suited to uncover the subtle interplay of the various components of grammar and the intricacies of the architecture of language. The variety of syntactic patterns, word orders and polar particles exhibited by PQs crosslinguistically and the impact they have on the interpretation of these interrogatives and their associated answers have led to very fruitful research on this question (a.o., Romero & Han 2004, Kramer & Rawlins 2011, Krifka 2013, 20015, 2017, Roelofsen & Farkas 2015, Holmberg 2016, Goodhue & Wagner 2018).

In this talk I will focus on PQs-answer pairs involving low and middle negation in English. In addition to English PQs, I will also discuss Spanish PQs, where answers to PQs license strings with more than one response particle, with well defined prosodic and morphological differences among them. My analysis builds on Wiltschko’s neoperformative approach and Holmberg’s (2016) analysis of PQS but, as will be shown, the resulting system differs substantially. More specifically, i) I pay particular attention to the different biases (epistemic vs. evidential) of the PQs (a.o. Domaneschi, Romero & Braun 2017, Goodhue 2018) and offer a new approach to the Polarity head. ii) As a result of this, I offer a direct derivation of the morphosyntax-pragmatic interface of the PQs under analysis and provide a homogeneous and coherent interpretation of negative features, where uninterpretable negative features ([uNeg] features) and negative concord operations are dispensable, avoiding some problems that Holmberg explicitly mentions regarding his own analysis. Further, iii) I reinterpret the split between the so-called truth/agreement answering systems (characteristic of languages like Japanese, Thai, Navajo, or Kashmiri) vs. polarity answering systems (Swedish, Finnish, etc.): under the approach advocated here in languages with exhibit both types of answering systems, like English (Holmberg 2016) or Spanish, polarity answering systems turn to be agreement answering systems, but at a higher level of the clausal spine. iv) I also point to the relevance of gestural features in the interpretation of answers to some PQs.

Susi Wurmbrand
“Voice restructuring cross-linguistically—evidence for a synthesis model of complementation”

Joint work with Iva Kovač and Magdalena Lohninger

A wide range of languages allows various forms of long object A-movement (LOAM)—configurations in which an embedded object is promoted to matrix subject. Cross-linguistically, there are four types of LOAM: Raising, Voice matching, Default Voice and Crossed Control. In this talk, I suggest a unified analysis of LOAM which is based on the concept of Voice restructuring. Voice restructuring involves the unification of matrix and embedded Voice due to the deficiency of one them. This Voice dependency also leads to an obligatory argument-sharing dependency, which may be semantic in nature, but is nevertheless sensitive to the syntactic structure. To account for certain argument structure restrictions (when the matrix predicate is thematic, the embedded predicate cannot be inchoative/unaccusative, whereas it cannot be causative when then the matrix predicate is non-thematic), a synthesis model of complementation is adopted, according to which the matrix and embedded predicates “select” each other and may affect the nature and properties of the other.