The Weekly Newsletter of MIT Linguistics
Harrison Smith-Jaoudi will be presenting a paper by Kaplan reacting to the paper by Quine studied at LPRG two weeks ago. The meeting will take place on Monday 10/15 at 1pm in the 8th floor seminar room, i.e. the usual time & place.
Title : Opacity
Author(s) : Kaplan
Speaker: Stan Zompi (MIT)
Title: Stan presents Collins 2018 and De Clercq & Vanden Wyngaerd 2017
Date and time: Monday 10/15, 5-6pm
Location: 32-D831
Abstract:
“The discussion will revolve around two recent attempts to rethink contextual allomorphy. Collins (2018) outlines an approach to allomorphy that eschews reference to competition or blocking, and attempts to model allomorph distribution simply in terms of classic c-selection. De Clercq & Vanden Wyngaerd (2017), on the other hand, suggest that contextual allomorphy may just not exist: what are normally thought to be alternative allomorphs of one and the same morpheme actually are always portmanteau realizations of slightly different chunks of structure. We will compare both of these novel approaches to the standard treatment of contextual allomorphy in DM.”
(1) a. I hope to win the race.
b. Alex is hoped to know the answer to your question.
In this talk, I will show that the behaviour of hope is systematically variable, but not mixed. Present-oriented uses behave like wager verbs in allowing traces but not PRO and future-oriented uses behave like demand verbs in allowing PRO but not traces.
(2) *Alex is hoped to win the race.
I adopt Pesetsky’s proposal that empty categories in this position are (anti-)licensed cross-clausally by the matrix verb and develop an account of the behaviour of hope based on Wurmbrand’s (2014) proposal that future infinitives contain a syntactic realisation of a future operator. I propose that a doxastic (un)certainty presupposition of the verb determines whether or not the future operator is optional, obligatory, or banned in the complement of a given verb, and that this operator is a barrier to licensing of traces. Thus, as noted by Martin (2001), only future infinitives will allow PRO in their subjects. Such an account links the cross-clausal empty category (anti-)licensing abilities of the verb to its semantics, rather than relying on syntactic selection as in Pesetsky (1992). This facilitates a movement away from theories that posit purely syntactic restrictions on the derivation to ones that are more interface-oriented.
Speaker: Tatiana Bondarenko (MIT)
Title: Argument structure of thinking: a view from Buryat
Date and time: Wednesday, October 17, 1-2 pm
Location: 32-D461
Abstract:
Speaker: Lena Borise (Harvard University)
Title: Word stress and phrasal intonation in Georgian
Date/Time: Wednesday, October 17, 5:00-6:30pm
Location: 32-D831
Abstract:
Based on novel acoustic data, I show that Georgian (Kartvelian) has (i) word stress, which is fixed on the initial syllable and primarily cued by duration, and (ii) phrasal intonational pitch targets that are located in the right periphery of the prosodic word (penult and ultima).
While it has long been acknowledged that these loci (initial syllable and the right periphery of a prosodic word) carry most prosodic prominence in Georgian, its nature has been disputed. According to the existing literature, the initial syllable carries stress in di- and trisyllabic words, while in longer words there is another stress-like target on the (ante)penultimate syllable (Akhvlediani 1949; Robins & Waterson 1952; Tschenkeli 1958; Gudava 1969; Tevdoradze 1978). In such longer words either the initial syllable or the (ante)penult have been variably analyzed as carrying (primary) stress, with the other locus possibly carrying secondary stress. The uncertainty surrounding Georgian word stress led some authors to suggest that instead of word stress Georgian relies solely on phrasal intonational pitch targets (Zhghenti 1953; 1959; 1963; Alkhazishvili 1959), like French (Vaissière 1983; Féry 2001) or Korean (Jun 1993). This paper advocates for a mixed approach, whereby both word stress and phrasal intonational pitch targets are present in the language, though their interaction with each other is minimal.
Specifically, the instrumental results suggest the following:
1. Georgian has fixed initial stress that is primarily cued by duration, based on the consistently greater duration of the initial syllable as compared to the following ones in words of any syllable count.
2. The ultima carries a high boundary tone, as expected in broad-focus declarative contexts, according to the earlier literature.
3. The penult in words of any syllable count carries another pitch target, which is similar in nature to the phrase accent L that has been described as appearing on the penult of predicates in questions and focal contexts. Its theoretical status remains to be established.
Speaker: Reuben Harry Cohn-Gordon (Stanford University)
Title: Bayesian Pragmatic Models for Natural Language
Date and time: Thursday, 10/18, 5-6pm
Location: 46-5165
Abstract: The Rational Speech Acts model (RSA) formalizes Gricean reasoning through nested models of speakers and listeners. While this paradigm offers an elegant way to simulate pragmatic behavior in NLP tasks such as image captioning and translation, scaling from simple models to natural language presents several challenges. In particular, I discuss the problem of choosing alternative utterances among an unbounded set of sentences, including work on image captioning and on-going work on translation.