about

The Circle is a weekly forum dedicated to discussion of the descriptive, experimental, and theoretical study of syntax and semantics, featuring presentations of ongoing research by members of the Berkeley Linguistics Department and other departments, as well as discussion of previously published works.

when

Fridays, 3:00-4:30pm

where

1303 Dwinelle Hall

coordinator

Emily Clem

Tessa Scott

 

University of California, Berkeley
Department of Linguistics



NEXT MEETING:

14 december
Jorge Hankamer (UC Santa Cruz) & Line Mikkelsen (UC Berkeley)
CP complements to D

Despite their apparent simplicity, the structure of DPs containing “complement” CPs (what we will call DCs) has long been obscure. One major strand of investigation has attempted to assimilate DCs to (close) nominal apposition, implying that N and CP form a structural unit which then combines with D.

Danish has two kinds of DCs, a bare DC with the superficial structure [D N CP] and a prepositional DC in which the CP is encased in a PP. Exploiting clues provided by the allomorphy of the definite morpheme, we argue that the bare and prepositional DCs have very different structures, neither of which can be assimilated to apposition between N and CP.

We show that the two kinds of DC have different semantic/pragmatic properties, the bare DCs being referent-establishing in the sense of Hawkins (1978) and the (definite) prepositional DC being anaphoric.

We then argue that English also has different structures for anaphoric and referent- establishing DCs, and that they are plausibly parallel to the structures we establish for Danish. We conclude by arguing that if the structure of any DCs in English is to be assimilated to apposition, it must be apposition between DP and CP.


PAST MEETINGS:

7 december
Noga Zaslavsky (Hebrew University of Jerusalem), Karee Garvin (UC Berkeley), Charles Kemp (University of Melbourne), Naftali Tishby (Hebrew University of Jerusalem), & Terry Regier (UC Berkeley)
Color-naming evolution and efficiency: The case of Nafaanra

Many theories hold that languages acquire color terms with time. Recently, it has also been claimed that this process is shaped by the need for efficient communication. However, most such research has been based on synchronic cross-language comparisons, rather than diachronic data. Here, we examine color naming evolution using diachronic data for a single language, Nafaanra. We find that the shift in color terms in Nafaanra is consistent with a recent proposal by Zaslavsky et al. (2018), grounded in an independently motivated information-theoretic principle of efficiency, providing the first direct support that color naming evolves under pressure for efficient communication.

30 november
Zachary O'Hagan (UC Berkeley)
Two sorts of contrastive topic in Caquinte

Recently accounts of contrastive topic (CT) have fallen into at least two camps. One camp, represented by Büring (2016) and others, analyzes CT as invoking a set of alternative questions (akin to the alternative propositions familiar from accounts of focus), the interpretation of which comes with conventional implicatures of pertinence, independence, and identifiability. The other camp, represented by Constant (2014) and others, analyzes CT as partially resolving a complex question in discourse. In this presentation I explore novel data from Caquinte (Arawak, Peru) that bears on this debate. In this language, CT is marked with one of two mutually exclusive second-position clitics, =mpani and =ga. The former attaches to demonstratives, topic pronouns, copulas, verbs, adverbs, and interjections in declarative clauses only; the latter attaches to demonstratives, topic pronouns, and adverbs in both declarative and interrogative clauses. When =ga appears in an interrogative clause, that clause is always a fragment question (e.g., English What about X?). I argue that these two markers encode a difference in whether the contrasted object is construed as part of interlocutors' attentional space: =mpani comes with the construal that it is; =ga that it is not. The unique appearance of =ga in fragment questions, as well as other differences in interpretation that I demonstrate, fall out as a result. This necessitates a modification to either of the accounts above. For Büring, I claim, it means that the pertinence implicature may or may not be active. For Constant it means that CT may involve discourse structures that do not initially consist of a complex question (i.e., a question with multiple subquestions).

The data that I present come from elicitation in Caquinte and Spanish, written narratives, audio and video recordings of spontaneous language use (which we will listen to and watch!), and notes of overheard speech. The distinction is subtle, and we will walk through a variety of contexts to get a firm grasp of the empirical facts. In order to situate the discussion and argue for the independence of CT as a linguistic phenomenon, I also briefly review issues of word order, continuation topic, switch topic (topicalization), information focus, identificational (exhaustive) focus, and the co-occurrences of =mpani and =ga with maximal elements (e.g., the universal quantifier maasano). Caquinte =ga is shown to resemble, though be far from identical to the Mandarin particle -ne described by Constant.

23 november
No Meeting (Thanksgiving)

16 november
CANCELLED DUE TO AIR QUALITY

9 november
Tom Roberts (UC Santa Cruz)
I can't believe what's not butter: Deriving distributed factivity

Given the assumption that selection is strictly local, we expect the ability of a head to take a particular argument to be insensitive to linguistic material above that head. The verb believe poses a puzzle under this view: while believe ordinarily only permits declarative clausal complements, interrogative complements are allowed when believe occurs under clausal negation and can or will. More puzzlingly still, can't believe also has a factive reading, similar to be surprised, where believe on its own does not:

  1. I (can't) believe that Elaine won the election.
  2. I *(can't) believe which candidate won the election.

Although believe is standardly given a Hintikkan proposition-embedding semantics, I argue that such an analysis is incompatible with the facts of can't believe and similar constructions cross-linguistically. Instead, I propose that believe is fundamentally question-embedding, à la Theiler, Aloni & Roelofsen (2018). I show that the right semantics for believe, including an excluded middle presupposition (Gajewski 2007), interact with the modal and negation to derive the factivity of can't believe and therefore its ability to embed questions (Spector & Egré 2015). I conclude that factivity need not be lexical: the right mix of semantic ingredients can conspire to yield a factive interpretation, even if those ingredients are distributed across multiple lexical items.

2 november
Amy Rose Deal (UC Berkeley)
Clausal complementation vs. “relative embedding”: on knowledge and happiness in Nez Perce

Traditional approaches to clausal complements and relative clauses make a sharp distinction between the two both syntactically and semantically. This distinction has been called into question by a series of recent arguments in favor of a more relative-like structure for complement clauses (Aboh 2005, Kayne 2008, Arsenijevic 2009, Krapova 2010, Legate 2010, Caponigro and Polinsky 2011, Moulton 2015, i.a.). A strong position, taken up by Kayne (2008) and Arsenijevic (2009) (and also to some extent Aboh 2005), holds that there is no variation in underlying structures: complementation universally features relativization as a matter of language design. In this talk I present challenges for this “all-complementation-is-relative” theory. Relative embedding, I argue, is neither a language universal nor a lexical fall-back strategy forced by the absence of complementizers like English THAT; rather, it is rather an independent option that languages may (or may not) make use of. My argument comes from a previously undescribed pattern of embedding in Nez Perce. In this language, certain apparent complements show morphosyntax distinctive to relative clauses; yet relative embedding exists alongside a pattern of ordinary clausal complementation of the English type. After discussing the syntax and semantics of the two clause types, I conclude with a consideration of the link, if any, between complement type and factivity.

22 october
Special day (Monday), time (11:00-12:30), and location (1229 Dwinelle Hall)
Ashwini Deo (The Ohio State University)
The emergence of split-oblique case systems: A view from the Bhili dialect continuum (Indo-Aryan)

Indo-Aryan ergativity is aspectually conditioned: the transitive subject, if marked, is marked only in perfective clauses, and verb agreement in most (but not all) such cases, is not with the subject but rather determined by case-marking on the direct object. Existing research has amply noted language-specific variability in overt marking of ergative case on the subject, overt marking of accusative case on the object (differential object marking (DOM)), and concomitant effects on verbal agreement. While Hindi-Urdu presents the best studied system, the systems obtaining in Marathi, Punjabi, Gujarati, Kutchi Gujarati, Nepali, and several dialects of Marathi have also been analyzed (Mahajan 1990, Mohanan 1994, Mistry 1997, Patel-Grosz 2012, a.o. for individual systems, with a comparative treatment in Deo & Sharma 2006).

The case-marking patterns found in several case-marking systems in the Bhili dialect continuum (a set of closely related Indo-Aryan dialects spoken in Western and Central India) present an interesting deviation. In contrast to the split-ergative pattern seen in much of Indo-Aryan (S = O, S ≠ A), these systems exhibit a split-oblique pattern (A = O, S ≠ A) in significant subparts of the case paradigms.

Three properties of the relevant paradigms are worth considering:
a. In several systems, there is syncretism between ergative and oblique marking in much of the pronominal and nominal inflectional paradigms (1pl, 2pl, 3sg, 3pl).
b. In some systems, the bare oblique is further used to mark possessors in lieu of a dedicated genitive case (with num-gen-case features) seen in standard languages like Hindi and Gujarati.
c. In other systems, the bare oblique is additionally used to mark direct objects (DOM) in parts of the pronominal paradigm.

In this talk, I investigate the hypothesis that the oblique form was recruited for marking agents in perfective, transitive clauses as well as patients with high animacy/referentiality properties for those cells in the paradigm that lacked distinct inflectional ergative and accusative marking. The emergence of a split-oblique system from an original split-ergative system can be thus tied to the reduction of the morphological case-inventories of particular languages. I take the first steps towards explaining these synchronic/diachronic patterns by appealing to a constrained interface between abstract and morphosyntactic case of the sort assumed in Kiparsky (2001). On this approach, abstract case features function as constraints on morphosyntactic case and the assignment of morphosyntactic case marking to abstract structural roles is determined by optimizing featural correspondence between the two.

19 october
Susan Steele (UC Berkeley)
The architecture of inflection

12 october
Round Robin

5 october
No Meeting (NELS)

28 september
NELS 49 practice talks
#1 Emily Clem (UC Berkeley)
Cyclic expansion in Agree: Maximal projections as probes
#2 Schuyler Laparle (UC Berkeley)
Locative inversion without inversion
#3 Tessa Scott (UC Berkeley)
Anti-clitic host requirement on second position clitic in SJA-Mam

21 september
Sabrina Grimberg (Stanford)
Between economy and recoverability: The case of subject doubling in Colloquial Finnish

In Colloquial Finnish (CF), there are three possible syntactic configurations for subjects: (i) the subject is focused and appears postverbally; (ii) the subject is non-focused and preverbal; (iii) the subject is focused and postverbal as in (i) but is also doubled by a preverbal pronoun. In order to derive the cases when the subject is pronounced only once (i.e. (i) and (ii)) and also the doubling instances (i.e. (iii)), I propose a Chain Resolution mechanism. The main intuition behind this operation is that the interplay between a general pressure to pronounce only the highest copy in a chain and the need to satisfy certain phonological requirements leads to the occurrence of a single or double subject copy in CF.

14 september
Ryan Bochnak (Universität Konstanz)
Combining coordination and focus: Towards an analysis of alternative questions in Washo

This talk explores some preliminary data on the interpretation of the coordinator -(i)ŋa in Washo (isolate; northern California and Nevada), which can be interpreted as either contrastive conjunction ('but') or disjunction ('or'). The semantic challenge is not only to uncover a single underlying meaning for -(i)ŋa (assuming that an ambiguity analysis is unattractive), but also to explain its interpretations in combination with other particles in the language. Specifically, when -(i)ŋa combines with the additive focus particle -saʔ 'also', the resulting interpretation is one of exclusive disjunction ('either...or'), while combining -(i)ŋa with -saʔ and the question particle -hé:š derives an alternative question. This latter observation is particularly intriguing, since current theories have yet to resolve the exact way in which focus is implicated in deriving the meaning of alternative questions compositionally.

7 september
No Meeting (SuB)

31 august
Sinn und Bedeutung 23 practice talks
#1 (3:10-3:55): Virginia Dawson & Amy Rose Deal (UC Berkeley)
Third readings by semantic scope lowering: Prolepsis in Tiwa
#2 (3:55-4:40): Rachel Rudolph (UC Berkeley)
A closer look at the perceptual source in copy raising constructions

24 august
Sinn und Bedeutung 23 practice talk
Agnes Bi (MIT) & Peter Jenks (UC Berkeley)
Pronouns, radical pro-drop, and ellipsis in Mandarin