Location: OAK 112
9:30-10:00 Breakfast & coffee (provided)
10:00 Beccy Lewis (UConn) - A deficient indexical in British English
10:30 Thanos Iliadis (UConn) - The distribution of Modern Greek idhios
11:00 Satoru Ozaki (UMass) - Binding of degree pronouns
11:30 Katie Martin (MIT) - A new presuppositional account for slurs
12:00-13:00 Lunch (provided)
13:00 Keely New (MIT) - Island sensitivity of wh-the-hell in wh-in-situ languages
13:30 Daria Bikina (Harvard) - Why do some languages never relativize dative arguments? Evidence from Khanty
14:00 Aljosa Milenkovic (Harvard) - Symmetrical trade-offs are diachronically unstable: superlinear constraint cumulativity in Serbian
14:30 Anabelle Caso (Harvard) - Slay him naked: an interface approach to secondary predication
15:00-15:30 Coffee break
15:30 James Lee (Harvard) - Causivity in the Korean voice domain
16:00 Fedya Golosov (UMD) - Complex predicates with the light verb pɨr 'go' in Poshkart Chuvash: what counts as gradual change
16:30 Bergül Soykan (MIT) - The interaction between past and conditional in Turkish
A deficient indexical in British English
Beccy Lewis (UConn)
In many British English dialects (primarily those in Yorkshire, the North-East and Scotland, but there are others), a phonologically weak first person plural objective pronoun us is used where standard English would use the first person singular objective pronoun me. Evidence of singular us dates back to the 19th century and while it has been described by many authors (e.g. Wright 1905, Petyt 1985, Edwards 1993, Beal 2004, Snell 2007, a.o.), it has never received attention in the generative linguistic domain. In this talk, I provide novel empirical generalizations about singular us: it is an obligatorily de se pronoun and it can not be a bound variable. I show that these facts (coupled with its phonological deficiency) make singular us typologically anomalous; it doesn't behave as one might expect under the pronominal typologies of e.g. Cardinaletti and Starke 1999 or Dechaine and Wiltschko 2002. I offer a sample of potential analyses of singular us, all of which go some way to explaining these peculiar properties, but which fall short of explaining all the facts.
The distribution of Modern Greek idhios
Thanos Iliadis (UConn)
Previous literature has established a distinction w.r.t. the anaphoric properties of Modern Greek idhios, lit. ‘same’, depending on its syntactic position. When it occurs as an object, it is a long-distance anaphor: free in its local domain but, at the same time, obligatorily bound by a sentence-internal c-commanding antecedent. When it occurs as a subject, it emphatically refers to a discourse topic (Iatridou 1986, Anagnostopoulou & Everaert 2013). In this presentation, I reject this view and I argue that idhios is a strong personal pronoun competing with weaker forms, whose distribution is regulated by Discourse: in a nutshell, like Serbo-Croatian strong personal pronouns, idhios is sensitive to Focus and to the Information Structural status of its antecedent (Jovović 2022). At the same time, some of its otherwise puzzling properties (endophoricity, incompatibility with 1st and 2nd person antecedents) immediately follow from simple principles of economy regulating the internal setup of the MG pronominal system.
Binding of degree pronouns
Satoru Ozaki (UMass)
If the denotations of entity pronouns are indexed variables, and an entity pronoun in an entity abstraction may or may not be co-indexed with the binder of the abstraction, then we would expect both "free" and "bound" readings of pronouns. This seems true in English, where sentences like Rachel told every guy that she hated his outfit are ambiguous between a "free" and a "bound" reading of his. Do we expect the same thing to happen with degree pronouns? Is the co-indexation of a degree pronoun inside a degree abstraction with the binder possible, banned or obligatory? I will take some examples of constructions that are analyzed as making use of degree abstractions, put degree pronouns into these abstractions, and check whether the pronouns have "free" or "bound" readings, or both.
Why do some languages never relativize dative arguments? Evidence from Khanty
Daria Bikina (Harvard)
In Khanty (Ob-Ugric, Uralic), non-finite relativization can target subjects, direct objects, and most adjuncts, but never dative arguments. In order to relativize a dative argument, one must first promote it to a direct object position. This paper addresses this asymmetry in Khanty and shows that this occurs because of the nature of the relativizing probe: it has a complex composite probe that has a sole A'-feature REL and a set {TOP, φ} as its second feature. Datives never agree, so they must be topicalized in order to get fed to relativization.
Symmetric trade-offs are diachronically unstable: superlinear constraint cumulativity in Serbian
Aljiosa Milenkovic (Harvard)
In this talk, I present a case of superadditive ganging-up cumulativity in Serbian which features a symmetric trade-off between a higher-weighted faithfulness constraint and a single lower-weighted markedness constraint. Both degenerate feet and feet with a toneless head mora are tolerated in the language, but the combination of these two marked structures is strictly prohibited. I provide evidence that this superadditivity effect tends to be eliminated from the grammar. The survey thus lends empirical support to the recently advanced hypothesis that cumulativity effects featuring symmetric trade-offs between constraint violations are diachronically unstable (Smith 2022).
Slay him naked: an interface approach to secondary predication
Anabelle Caso (Harvard)
This paper examines the syntactic, semantic, and prosodic structure of secondary predicates, or nonverbal expressions which share an argument with (but which are distinct from) the finite matrix verb in a clause (Heidinger 2022). While resultatives signal an eventuality which is obtained as a result of the action of the primary predicate, depictives describe the state of their subject at the time when the action of the primary predicate occurs (Irimia 2012; Milway 2019). In finite metrical corpora, including the R̥gVeda and the Homeric poems, depictive and resultative secondary predicates are isolated prosodically and consistently demonstrate stage-level readings where individual-level predicates are otherwise expected (Carlson 1977; Casaretto 2020). These properties follow from analyses of secondary predicates as containing a small clause and an operator which is state-introducing (Kratzer 2005; Harley 2008), which in turn compose a recursive intonational boundary; such a description of secondary predicates as structurally and prosodically complex can be used to distinguish themfrom simple attributive entities across ancient metrical texts.
Complex predicates with the light verb pɨr 'go' in Poshkart Chuvash: what counts as gradual change
Fedya Golosov (UMD)
In Poshkart Chuvash (Turkic. Altaic), the verb pɨr ‘go, head towards some goal’ is grammaticalized into an aspectual light verb that forms complex predicates denoting a progressing gradual change of some state. What is especially interesting about this verb is its selectional restrictions. On the one hand, it can combine with any verbs if they form VPs denoting a gradual change. On the other hand, what counts as gradual change is surprising. Namely, only intransitive scalar predicates can combine with pɨr, while transitive scalar predicates must either have a non-agentive subject or be accompanied with a dynamic adverbial such as xolen ‘slowly, gradually’ or have a distributive plural object.
I argue that these selection restrictions are explained by the fact that the light verb pɨr requires the whole event to be gradual. Intransitive scalar predicates liketip ‘dry’, tol ‘get filled’, lajəklan ‘get better’ and their transitive counterparts with a non-agentive subject only denote a change of state and so are gradual, while transitive scalar predicates (tibët ‘dry (tr.)’, vula ‘read’, ɕi ‘eat’, əʐət ‘heat up’ etc.) with an agentive subject have a more complex event structure: there is also a causing subevent causing the change of state. Such predicates are no longer gradual in a sense that there are no salient plural stages of the event, and only addition of dynamic adverbials, as well as presence of a plural object “highlights” these stages.
I will come up with some ideas of potential semantic analysis of the light verb pɨr and discuss possible implications for the theory of the event structure.
The interaction between past and conditional morphemes in Turkish
Bergül Soykan (MIT)
Turkish has two verbal conditional morphemes, namely -sA and -ysA, which differ in their morphosyntax and semantics (Göksel & Kerslake, 2005). The -ysA marker attaches to inflected verbs and nominal predicates to generate what von Fintel and Iatridou (2022) call O-marked conditionals (O.COND). In contrast, the -sA morpheme only attaches to bare verb roots and creates a counterfactual interpretation as in X-marked conditionals (X.COND) proposed by von Fintel and Iatridou (2022). The past morpheme (PAST) can surface either before the O.COND (-ysA) and denote its conventional time information, or after the X.COND (-sA) and generate a counterfactual reading. In this study, I argue that the PAST in O.COND is interpreted as the evaluation time of either the antecedent proposition or the consequent prejacent of the conditional modal based on where it surfaces. However, the one with X.COND adds another level to X-marking by scoping over it and widens the domain of the modal so that it contains impossible/remote worlds. I agree with Kaufmann (2022) that X-marking can be split into two dimensions, but I stay agnostic about if the dimension that the PAST adds is temporal or modal in X.COND in Turkish.
Thanks to the UConn Department of Linguistics for funding this event.