Whamit!

Speaker: Michelle Yuan (UCSD)
Title: Deriving VSO in San Juan Piñas Mixtec (and some puzzles along the way)
Time: Tuesday, March 8th, 1pm - 2pm

Abstract: (in collaboration with Gabriela Caballero and Claudia Juárez Chávez)

In this talk, I present in-progress research on the syntax of San Juan Piñas Mixtec (Tò’ōn Ndā’ví, though henceforth SJPM here), a previously undocumented variety of Mixtec spoken in Oaxaca, Mexico and in diaspora communities in California. This work is part of a larger collaborative project documenting and analyzing SJPM, as well as developing linguistic resources to be used for language reclamation purposes.

Like other languages of the Oto-Manguean language family, SJPM displays VSO word order; in SJPM, VOS may arise in very limited contexts (e.g. there is no pseudo noun incorporation that I am aware of). As such, SJPM is useful in evaluating approaches to verb-initiality cross-linguistically. I demonstrate that verb-initial word order in SJPM does not involve V0-movement (contra Macaulay 2005 and Ostrove 2020 on other Mixtec varieties)—rather, it is derived by object-stranding VP-movement. I then evaluate how the object comes to be “stranded” in VP-raising VSO configurations. Since Massam (2001), a common approach is to derive VSO from object shift followed by remnant movement of the VP (see also Lee 2006, Medeiros 2013, Collins 2017, a.o.). However, many recent proposals have instead sought to derive VSO/VOS without remnant movement, instead drawing on the prosodic status of different object types (Clemens 2014, 2019, Clemens & Coon 2018, Richards 2016, van Urk to appear). I argue that the prosodic approach is untenable for SJPM, and provide some evidence for the object shift/remnant movement approach, though residual issues still remain.

Speaker: Agnes Bi (MIT)
Title: Mandarin tone identification in whispered speech
Time: Monday, March 7th, 5pm - 6:30pm

Abstract: Phonetic contrasts are represented via multiple acoustic dimensions and signaled by multiple cues simultaneously. In the case of tones, it has been long established that fundamental frequency F0 serves as their primary cue (Gandour 1978, Yip 2002), and F0 information is sufficient for successful tonal identification (Abramson 1978). On the other hand, many studies have shown that secondary cues still contribute substantially to the identification process, and that listeners seem to be sensitive to a weighted combination of various acoustic dimensions (Di Paolo and Faber 1990, Wassink 2006, Zellou, Scarborough, and Kemp 2020, a.o.). In this project, I am interested in how much information listeners can extract from the secondary cues alone. More precisely, when listeners are deprived of a key dimension of cues, how do they utilize the remaining acoustic features? I will report findings from an experiment conducted over IAP that looks into the identification of Mandarin tones in whispered speech. The results seem a little mysterious, if not counterintuitive, and I would like to get some help with how to make sense of them.

Speaker: Sigwan Thivierge (Concordia University)
Date and time: Thursday March 10, 12:30-1:50pm
Location: 32-D461, https://mit.zoom.us/j/96158811570

Title: TBD

Abstract: TBD

Speaker: Maria Aloni (University of Amsterdam)
Title: Neglect-zero effects at the semantics-pragmatics interface
Time: Friday, March 11th, 2:30pm - 4pm

Abstract: In Free Choice (FC) inferences, conjunctive meanings are derived from dis- junctive modal sentences contrary to the prescriptions of classical logic [Kam73, Zim00, Fox07, Alo07]:

(1) Positive FC

a. You may go to the beach or to the cinema. —> You may go to the beach and you may go to the cinema.

b. ◊(a ∨ b) —> ◊a ∧ ◊b

[Alo21] presented a formal account of FC inferences in a Bilateral State- based Modal Logic (BSML). The novel hypothesis at the core of this proposal is that FC and related inferences are a straightforward consequence of a tendency in human cognition to neglect models that verify sentences by virtue of some empty configurations (neglect-zero) [BSK19]. There are two different ways to implement neglect-zero in BSML: (i) syntactically, using the [ ]+-enrichment function from [Alo21]; and (ii) model-theoretically, ruling out the empty set from the set of possible states (BSML∗). These two implementations both derive FC inferences and their cancellation under negation, but differ for example in their predictions with respect to the debated case of Negative FC (only derivable in BSML∗):

(2) FC cancellation under negation [AO06]

a. You are not allowed to eat the cake or the ice cream. —> You are not allowed to eat either one.

b. ¬◊(a ∨ b) —> ¬◊a ∧ ¬◊b

(3) Negative FC [MRSB21]

a. It is not required that Mia buys both apples and bananas. —> It is not required that Mia buys apples and that Mia buys bananas.

b. ¬◻(a ∧ b) —> ¬◻a ∧ ¬◻b

In the talk I will compare the predictions of these implementations with those of recent alternative accounts [Gol19, BLF20] and then argue in favor of a hybrid approach where neglect-zero can cause two kinds of effects:

(i) cancellable global effects (modelled by BSML∗);

(ii) more robust effects triggered by the conventional meaning of certain ex- pressions (modelled by obligatory local applications of [ ]+-enrichments)