Archive for March 19th, 2012
As we wrote in last week’s issue, MIT was the host of the second-ever workshop on Formal Approaches to South Asian Linguistics (FASAL 2) last weekend, and it was a fantastic conference, with great speakers and a great atmosphere. Thank you, organizers - and special thanks to Pritty Patel-Grosz.for creating such a successful and exciting event!
photo credit mitcho
Speaker: Adam Albright
Title: Discovering and modeling cumulative markedness interactions with loglinear models
Date/Time: Monday Mar 19, 5:30pm
Location: 32-D461 (Note unusual location)
It is often observed that “phonology can’t count”. This principle rules out, among other things, languages in which a marked structure is tolerated once or twice within a word, but not three or more times. In this talk, I discuss a set of restrictions in Lakhota (Siouan) which are very similar to such a ‘threshold’ effect: roots often contain a single marked structure (fricative, aspirated or ejective stop, consonant cluster), but roots containing multiple marked structures are rarer than one would expect, based on the independent frequencies of those structures. This observation leads two questions: is the degree of underattestation significant, and if so how should it be accommodated in a grammatical model? I show that both questions can be addressed using log-linear (maximum entropy) models of constraint interaction. First, I present results of a series of statistical models of the Lakhota lexicon, attempting to predict the relative type frequency of root shapes based on their phonological properties. The results show that models with interaction terms, in which multiple simultaneous violations may be penalized more than expected based on the individual violations, do significantly better at predicting lexical counts. Furthermore, the effect is strongest for combinations of structures that are independently most strongly penalized. Thus, it appears that cumulative effects are real, and some form of ‘counting’ is indeed warranted. I argue that these statistical models are too powerful, however: they could, in principle, impose strong penalties on combinations that are independently penalized only weakly (or not at all), or they could even reverse the direction of the preference so that languages tolerate a marked structure only in the presence of another marked structure. I argue that we can avoid these predictions with a simpler model, in which markedness constraints interact with MParse (Prince and Smolensky 1993/2004). I present the results of a learning simulation, showing that the observed cumulative effects can be predicted using a small set of markedness constraints on simple structures. Finally, I consider some typological predictions of the proposed model.
Speaker: Hadas Kotek
Date/Time: Monday, Mar 19, 5:30p
Location: 32-D124 (Note unusual location)
I will present pilot results from two sentence processing studies that we ran in the department last month - one concerning multiple questions and the other concerning the quantificational nature of `any’. Both studies yield unexpected results that bear on the theories of questions and of free choice `any’, respectively. In particular, I will argue that covert wh-movement may not target the same position as overt wh-movement but rather a much lower position (if it occurs at all), and that any should be analyzed as an existential quantifier that QRs locally, targeting a lower position than the one targeted by the QR of `every’.
Speaker: Alex Silk (University of Michigan / MIT)
Title: Information-Sensitivity in Deontic ‘Ought’ and ‘Must’
Date/Time: Tuesday Mar 20, 10:00AM-11:30AM
(Note unusual time and location)
There is a growing literature on how deontic modals can be interpreted with respect to bodies of information or evidence. However, previous treatments of information-sensitivity in deontic modals focus exclusively on ‘ought’ and ignore important differences between weak necessity modals like ‘ought’ and strong necessity modals like ‘must’. In this paper I attempt to delineate and capture such differences in information-sensitivity between ‘ought’ and ‘must’. Drawing on and revising a suggestion by Aynat Rubinstein, I argue that ‘ought’ and ‘must’ exhibit different conventional signals vis-à-vis common ground assumptions: ‘ought’, unlike ‘must’, conventionally signals that the truth and acceptance of the necessity claim—currently and throughout the evolution of the conversation—relies on certain assumptions not currently established in the global discourse context. This hypothesis helps generate correct predictions concerning the contrasting felicity conditions of ‘ought’- and ‘must’- sentences and meanings of ‘ought’- and ‘must’- conditionals. It also correctly predicts that certain types of modus ponens violations can occur because of the presence of ‘ought’ but not because of the presence of ‘must’.
Speaker: Theresa Biberauer (University of Cambridge)
Title: 231 and the Final-over-Final Constraint
Date/Time: Tuesday, Mar 20, 1-2p
In terms of the Final-over-Final Constraint (FOFC), structures in which a head-final phrase dominates a head-initial phrase within the same extended projection should be ruled out (cf. Biberauer, Holmberg & Roberts 2007 et seq.). While 2-verb clusters at all stages of Germanic seem to reflect this constraint, the same is apparently not true of 3-verb clusters: as i.a. Wurmbrand (2005), Barbiers (2005), Schmid (2006), Biberauer & Walkden (2010), Biberauer (2010) and Salzmann (2011) observe, a number of West Germanic varieties – notably, West Flemish, certain Swiss German varieties and Afrikaans – feature structures in which 231 orders are either obligatorily or optionally available (3 here refers to the most and 1 to the least deeply embedded verb in the cluster). The purpose of this talk is, firstly, to give an overview of the data, highlighting in particular the extent to which 231 structures are available in Afrikaans, the least well studied of the troublesome Germanic varieties; secondly, to consider the data against the background of existing attempts to account for the FOFC phenomenon (Biberauer, Holmberg & Roberts 2011, Sheehan 2011, Cecchetto 2010, Hawkins 2012), all of which will be shown to fall short in different ways; and, finally, to consider the question of what 231 phenomena suggest about the nature of FOFC and, accordingly, what a successful analysis of this phenomenon might look like.
The MIT Linguistics homepage now has a link for a comprehensive (we’re getting there…) listing of papers and presentations by the members of our department over the past decade or so. (Click on the new Research tab.) Most of the papers currently included are by our graduate students, and links to faculty papers (and more links across the board) will be added in the coming months. All of our alumni are invited to inspect their listing and write in with corrections and additions.
The program for CLS 48 is out, and it looks as though much of MIT will be transplanted to Chicago for the meeting. Adam Albright and Kai von Fintel are invited speakers at the conference, and a total of eight other talks will be presented by a total of nine current graduate students in various solo and ensemble combinations that would stretch a semanticist’s imagination to represent succinctly: Marie-Christine Meyer, Yusuke Imanishi, Ayaka Sugawara, [Hadas Kotek, Yasutada Sudo and Martin Hackl], [Laura Kalin (UCLA) and Coppe van Urk], first-year student Paul Marty (félicitations!), Sasha Podobryaev and Mitcho Erlewine (in order of listing in the program). Recent alums Giorgio Magri and Maria Giavazzi will also be giving talks.
The program for Formal Approaches to Slavic Linguistics 21 (FASL) is also out, and features MIT-related talks by post-doc Erik Schoorlemmer and grad students Igor Yanovich and Liuda Nikolaeva - as well as an invited talk by Tania Ionin (2003 PhD from BCS, but a linguist at heart!). Поздравляем вас всех!
Speaker: Theresa Biberauer (University of Cambridge)
Title: One peculiarity leads to another: insights from Afrikaans analyticity
Date/Time: Thursday, Mar 22, 12:30-1:45p
This talk takes two superficially unconnected phenomena in Afrikaans – negation doubling by clause-final nie2 as in (1) and predicate doubling as in (2) – as its point of departure.
(1) Ek is nie1/nooit tevrede nie2 I am not/never satisfied POL “I am not/never satisfied”
(2) a. Sing SING daardie man! sing sing there.the man “That man really sings with gusto!”
(2) b. Gelukkig is ek nou regtig GELUKKIG! happy am I now really happy “As for being happy, I’m REALLY happy!”
Based on a combination of diachronic and synchronic considerations, I argue that these phenomena can in fact be connected and that doing so enables us to understand a third, from a Germanic perspective, very surprising fact, namely that Afrikaans readily permits embedded V2 wh-interrogatives, regardless of the nature of the selecting predicate. This latter property is shown in (3):
(3) a. Ek wonder watter eksamen skryf die studente vanaand I wonder which exam write the students tonight “I wonder what exam the students are writing tonight”
(3) b. Hulle vind net gou uit wie moet die kaartjies gaan koop they find just quickly out who must the tickets go buy “They’re just quickly finding out who needs to buy the tickets”
From a theoretical perspective, the significance of the data under discussion is argued to be i.a. the insight they deliver into the notion ‘acategorial element’ and the properties elements of this type exhibit in relation to the Final-over-Final Constraint (FOFC; Biberauer, Holmberg & Roberts 2007 et seq.), and also what they suggest about the underdiscussed question of the relationship between Minimalism’s putative phase heads and the more articulated functional domains assumed by cartographers.
Speaker: Mitcho Erlewine
Title: Association with traces and the copy theory of movement (practice talk for GLOW)
Date/Time: Thursday Mar 22, 10:00AM-11:30AM
In this talk I give a principled account for the observation that exclusive only must associate with a focus within its complement (Tancredi’s (1990) Principle of Lexical Association; PLA), utilizing the copy theory of movement and associated work on the interpretation of traces. Previous explanations for this fact come from the idea that traces cannot be F-marked. I argue contra Beaver and Clark (2008) that traces (lower copies of movement chains) can in fact be F-marked and that this is exemplified in F-marking contained in quantifiers which undergo QR. Instead, PLA effects arise through the interpretation of the predicate in both the higher and lower copies of of the moved constituent.
This special session of Syntax Square on Friday will be a practice talk for GLOW.
Speaker: Hadas Kotek
Title: WH-fronting in a two-probe system
Date/Time: Friday, Mar 23, 3:30-4:30p
(Note unusual time and location)
The study of wh-movement has distinguished among several types of wh-fronting languages that permit distinct patterns of overt and covert movement, instantiated for example by the Slavic languages, English and German. This talk extends the cross-linguistic typology of multiple questions by arguing that Hebrew instantiates a new kind of wh-fronting language, unlike any that are presently discussed in the literature. I will show that Hebrew distinguishes between two kinds of interrogative phrases: those that are headed by a wh-word (wh-headed phrases: what, who, [DP which X], where, how …) and those that contain a wh-word but are headed by some other element (wh-containing phrases: [NP N of wh], [PP P wh]). We observe the special status of wh-headed phrases when one occurs structurally lower in a question than a wh-containing phrase. In that case, the wh-headed phrase can be targeted by an Agree/Attract operation that ignores the presence of the c-commanding wh-containing phrase.
I develop an account of the sensitivity of interrogative probing operations to the head of the interrogative phrase within Q-particle theory. I proposes that the Hebrew Q has an EPP feature which can trigger head-movement of wh to Q and that a wh-probe exists alongside the more familiar Q-probe, and shows how these two modifications to the theory can account for the intricate data that will be presented in the talk. The emerging picture is one in which interrogative probing does not occur wholesale but rather can be sensitive to particular interrogative features on potential goals.