university of california, berkeley


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.


Fridays, 3:00-4:30pm


Room 1303 (Dwinelle)

Recurring link. Passcode: sscircle .


Becky Jarvis

Phuong Khuu


University of California, Berkeley
Department of Linguistics

Fall 2021


3 december
Zachary O'Hagan (UC Berkeley)


10 december
RRR Week (No meeting)

17 december
Finals (No meeting)


19 november
Maksymilian Dąbkowski (UC Berkeley) and Hannah Sande (UC Berkeley)
Discontinuous vowel harmony in Guébie focus fronting

12 november
Ana Laura Arrieta-Zamudio (UBC)
Reflexives in San Pablo Güilá Zapotec: An approach on Discourse Representation Theory

Abtract is available here

5 november
No meeting

29 october
NELS (No meeting)

22 october

NELS practice talks

Amy Rose Deal:
Negative concord as downward Agree
Emily Drummond:
Abstract Case without morphological case
Madeline Bossi:
Negative bias and pragmatic reasoning in Kipsigis belief reports
Rebecca Jarvis:
Elided antecedents and exceptive sluices

15 october
Sherry Yong Chen (MIT)
Asymmetry in Presupposition Projection in If-Conditionals: An Acquisition Perspective

Presuppositions triggered in the consequent of conditionals are sometimes inherited wholesale by the complex sentence (1a), but sometimes, a weaker, conditional inference, as in (1b), is what projects:

(1a ) If Phil is traveling to London, he will bring his guitar on the trip.

(1b ) If Phil is a musician, he will bring his guitar to the party.

Under one family of theories (Karttunen 1974; Heim 1983; Beaver 2001; van Rooij 2007; George 2008; Fox 2008), the conditional inference is the “basic”, semantic presupposition of both (1a-b). It is pragmatically strengthened in (1a) due to plausibility considerations (it is unlikely that having a guitar depends on travel). Another class of theories (Gazdar 1979; van der Sandt 1992; Geurts 1996; Mandelkern 2016) takes the non-conditional presupposition to be basic, with weaker ones like (1b) resulting from local accommodation, which incorporates presuppositions into the asserted content. Under both approaches, triggers in the antecedent introduce a non-conditional presupposition:

(2 ) If Phil’s guitar is at the shop, he won’t attend the party. [ Phil has a guitar; presupposition inherited wholesale ]

The first approach thus holds an asymmetric view of presupposition projection in conditionals, modulated by position of the trigger; the second holds a symmetric view. Adjudicating between these approaches has been an empirical challenge, as speaker judgments are delicate and moreover, could reflect local accommodation processes in addition to the semantic presupposition triggered in these conditional sentences.

The present work approaches this challenge from the lens of child language acquisition. Previous developmental work (Zehr et al. 2016; Tieu et al. 2018) has shown that children do not apply local accommodation at the same rates as adults; the confound introduced by local accommodation can therefore be reduced in child language data. If there is genuine non-uniform projection in conditionals, as predicted on the asymmetric view, this population gives us an opportunity to see it transparently. Thus, in this ongoing work, we investigate the acquisition of presupposition projection in if-conditionals, finding developmental evidence which seems compatible with a family of theories that hold an asymmetric view of presupposition projection in conditionals.

8 october
Maša Bešlin (University of Maryland)
A Mayan temporal puzzle

Mayan languages have been claimed to lack tense morphology. Temporal interpretation is instead said to be guided by grammatical aspect (see e.g., Larsen 1988 for K’iche’, Vázquez Álvarez 2002 for Chol, Bohnemeyer 2002 for Yucatec Maya, Coon 2016 for an overview, a.o.). In this talk, I contribute to the discussion of how temporal information is encoded in Mayan languages by examining the distribution and interpretation of the tense/aspect (TA) markers x- and k- in K’iche’. Relying primarily on data from original fieldwork in Santa Lucía Utatlán (Sololá, Guatemala), I will conclude that, in K’iche’, these affixes mark (past and non-past) tense rather than grammatical aspect.

We will also see some indications that the same affixes in the closely related language Kaqchikel may indeed be aspectual. Although novel, the idea about the distinct functions of TA morphemes even in closely related Mayan languages should not be surprising; since the differences in the distribution of tense and aspect markers are relatively subtle, we could expect them to be prone to reanalysis. In addition to bettering our understanding of Mayan temporal systems, the findings I will report bear on phenomena such as TA concord in certain types of complement clauses, as well as on the ways we think about TA morphology (or lack thereof) in the derivation of deverbal elements such as participles and deverbal nouns.

1 october
No meeting

24 september
No meeting

17 september
Amy Rose Deal (UC Berkeley)
The logic of agreement: movement, morphology, and composite probes

In this talk I present new work on the interaction/satisfaction theory of Agree, the overall goal of which is to formally state an algorithm for the Agree operation. Toward that end, I discuss how the notions of interaction and satisfaction relate to the three primary observables for which Agree is responsible, namely movement, indexing (i.e. head marking), and flagging (i.e. dependent marking). I argue that all six options are indeed attested: move-what-interacts, move-what-satisfies, index-what-interacts, index-what-satisfies, flag-what-interacts, flag-what-satisfies. To make this case, I draw primarily on data from composite A/A' probing in Khanty (Finno-Ugric), recently analyzed by Colley and Privoznov (2020), along with the treatment of composite probes via conjunctive satisfaction in Scott (2021). I conclude with a presentation of a candidate Agree algorithm.

10 september
SuB (No meeting)

3 september
Erika Meyer (UMASS Amherst)
Likelihood, Predictability, and the Online Processing of Even

In this talk, I will discuss the online processing of the semantic operator even. Even is an expression that, when added to a sentence, introduces a presupposition that the sentence is less likely (and more surprising) than it otherwise might have been (Horn, 1969; Rooth, 1985, 1992; Greenberg, 2016; Francis, 2018). For example:

(1 ) Emily: I even hiked up Mount Tom this weekend!

In this sentence, the presence of even indicates that hiking up Mount Tom was in some way less likely or more notable than other things Emily might have done that weekend, such as go to the grocery store or take a nap.

Upon reading even or hearing it in speech, then, we must perform a rather complicated computation: we must comprehend that the sentence containing even is less likely than other possible sentences the speaker could have said. In this talk, I will explore how this computation might affect a comprehender’s prediction of upcoming material as they listen or read. The words and expressions we predict are, of course, the ones we consider to be the most likely to come next. Yet adding even to a sentence potentially disrupts this: even tells us that the sentence should be in some way less likely than it could have been. I will present the results of a cloze norm task and an eye-tracking while reading study investigating whether and how the presence of even and its likelihood presupposition influences the prediction of upcoming words.

27 august
Round Robin!