Proseminars are the special topics advanced graduate courses taught in our department, with course numbers in the 250 range. Typically these are prepared lecture courses given by faculty, but with strong student participation. They usually assume that enrollees will have already taken the core graduate courses in the relevant area.
This page reports proseminars being taught in the current academic year. We also include other graduate courses when their content has changed for the current offering.
For an archive of old proseminar topics, please visit the archive page.
On the Nature of Islands
I propose to begin this proseminar by laying out a buffet of topics that fall under the heading above and inviting the participants to select the ones that look most appetizing for us to pursue. I start with the disclaimer that I am by no means an expert on islands—far from it. I have a couple of little pockets of knowledge, but I hope everyone else will bring data and theoretical expertise to complement my own. Here is a sample of some of what you will find at the buffet table:
• What are the criteria for identifying something as an island, anyway?
Yes, we have definitions along the lines of “a constituent that is hard to extract from,” but in practice these constituents are often rather complex beasts in and of themselves, and we know from a couple of decades of processing literature that extracting (i.e., a long-distance dependency) is hard in and of itself, so people like Jon Sprouse have proposed that to warrant the label “island” (and hence the need for an island constraint), you need to be ‘harder to extract from than you would expect based on the nature of the constituent and the fact of doing extraction’. He operationalizes this definition as a statistical interaction in the acceptability ratings of a 2×2 set of sentence types, and by this measure, according to experiments conducted to date, not everything people have claimed is an island comes out as an island, and there are intriguing crosslinguistic differences (e.g. English vs. Italian). What to make of that?
• Relatedly, although island constraints are the poster children for UG constraints, there is massive (insufficiently acknowledged) cross-linguistic variation in which islands which languages enforce. (How) can you learn which islands your language tolerates violations of? Is there a notion of “parameter” that fits current theory that could capture such variation? Are ANY islands universally bad? The best candidate seems to be the Coordinate Structure Constraint, though even that has apparent exceptions (see (2) below), and there have been proposals (e.g. by Jacobson) that the CSC should be explained in terms of (roughly) semantic type mismatch rather than anything syntactic.
• To what extent does it makes sense in 2017 to keep talking about “islands” as if they are a unified phenomenon? There have been arguments (some recent, some dating back decades) that some “so-called” islands are syntactic, some are semantic/pragmatic, some are prosodic, some are processing-based, etc. To the extent these claims are right, could/should our theoretical architecture hook them in at a single locus? (Kyle Johnson shared some thoughts on this in his prosem this quarter.) If we were to “reduce” some islands to “processing considerations,” what sort of progress would this represent, as opposed to perhaps just kicking the can out of the syntacticians’ yard into the psycholinguists’?
• Does Minimalism have a theory of islands, really? We have phases (weak, strong, vanilla, chocolate, …), though we can’t seem to agree on what categories they are, and we have some notion that there are restrictions on how you can get out of them (via edges, and “edge features”, whose virtual conceptual necessity one might question), and to some people this feels like a warmed-over dish formerly known as Barriers served on a fancier platter whose shininess makes it harder to discern the predictions, while to others the dish now has a more principled flavor grounded in much deeper notions tied to the fundamental nature of the architecture, e.g. cyclic spell-out. Can we tell who is right (yet)?
• Language inaugurated a new “Perspectives” section recently with a target article entitled “Child language acquisition: Why universal grammar doesn’t help”, one of whose central arguments was that positing island constraints in UG is unnecessary (and actually counterproductive) because their consequences (to the extent they’re even empirically correct) fall out from discourse constraints that children must independently have/acquire anyway, drawing on work by i.a. Adele Goldberg. A response (2/3 of the authors of which have UCLA connections, including yours truly) argued that they completely failed to demonstrate this. Their reply to our response said we failed to make that case. Did anything useful come out of this exchange? Was it truly a scandal that Language published the target article in the first place (cf. Hornstein’s blog)?
• There is some recent intriguing work that seeks to wrangle and unify many notorious “exceptions” to island constraints by thinking more deeply about the meanings of the relevant sentences. One such example is Robert Truswell’s book/dissertation “Events, Phrases, and Questions,” whose central claim can be caricatured as ‘You can extract out of one event but not two’. (Why, one should of course wonder.) Some examples to give a taste of what he hopes to explain [judgments reflect (at least) one speaker of mongrel English]:
(1) a. Here is the influential professor that John went to college in order to impress.
b. ??the book that I went to college because I liked
(2) a. Which dress has she gone and ruined now?
b. *Which dress has she danced and worn?
(3) a. What did John drive Mary crazy whistling?
b. *What does John work whistling?
(4) a. Who did John go home without talking to?
b. ?*Who did John get upset despite talking to?
Feel free to email me suggestions for other items to place on the buffet table.
Scrambling and Clitics
Announcement: Mahajan/Sportiche Ling 252, Spring 2017, Mondays 2pm-5pm.
If you are interested in attending, please let us (Anoop and Dominique know).
Anoop and Dominique will co-teach a Proseminar on Scrambling and Clitics.
As early as in Mahajan 1991 ( Clitic doubling, object agreement and specificity. NELS 21:263-277.) the suggestion is made that (some) Hindi Scrambling shares properties with Clitic Left Dislocation in Italian and Romanian. In Sportiche (1992) (Clitic Constructions, ms., UCLA. published as: 1996. Clitic Constructions, in Phrase Structure and the Lexicon, L. Zaring and J. Rooryck, 213-276, Kluwer Academic Publishers, Dordrecht.), the claim is made is that cliticization in French (and Romance) is the counterpart of Germanic Scrambling.
The purpose of the seminar is to explore the connections between all these: Hindi Scramblings, Romance (and Greek) Clitics via Clitic Left and Right Dislocation, and Germanic Scrambling (German and Dutch at least) (and possibly Japanese and Korean Scrambling), with the aim of deciding whether a general theory unifying these movement types across languages can be formulated.
Since these suggestions were made, analytical tools used to establish the existence and properties of movement dependencies have been refined. We will describe these tools and apply them systematically to the three (or four) classes of languages: French (and Romance), Hindi, German and Dutch (possibly Japanese and Korean).
Linguistics 252, Hilda Koopman
Within current Minimalists approaches, we are looking for the syntactic framework that is most suitable for the interfaces with the semantics and the phonology, provides a likely path to acquisition, models the data from an individual speaker, extends to capture linguistic variation, and allows the development of SSWL properties to test hypotheses about (im)possible crosslinguistic variation. We will explore the question whether we can choose between (specific implementations of) different frameworks. (Not surprisingly) antisymmetry, and generalized U(niversal) 20 patterns, will play an important role throughout the quarter. We will look at left right asymmetries in scope interactions, and (further) test a U20 typology of morpheme orders, (attempt to test) a theory of the expected typology of second position phenomena, contrasting it with the Distributed Morphology framework. We will start off with a comparative study of the distribution of NegXPs in Germanic languages to determine what part of the syntax is stable and what minimally varies.
Meets Mondays 2-4.50 –(week1: in the syntax lounge).
Linguistics 254, Kyle Johnson
Tuesday from 9:00am-12:00pm
A Multidominant Theory of Movement
A common assumption is that there is a single operation, “movement,” that is responsible for certain types of long-distance dependencies. In its classic formulation, movement gives a moved term a new location and puts in the original location a silent variable that is bound by the term in its new position. Attempts have been made in this century to decompose movement into more elementary operations, and this seminar will trace one of those attempts — one that uses phrase markers that tolerate multidomance. The aim will be to explain some of the features that appear to define movement: semantic displacement, terseness and boundedness. The focus will be on Verb Movement, Wh Movement and Quantifier Raising. We will look at linearization schemes designed to flatten multidominant phrase markers into strings and how those schemes interact with the semantics of constituent questions, quantifiers and topicalized verbal projections. Key readings include Elisabet Engdahl’s 1985 UMass dissertation, Jairo Nunes’s 1995 University of Maryland dissertation and Hadas Kotek’s 2014 MIT dissertation.
Linguistics 252, Yael Sharvit
Indexicality and ‘de se’-ness, and the relations between them, in English and cross-linguistically (in the person as well as tense domains), has been the topic of much exciting research in recent years. We will explore some of the “old” and current literature on this topic, with the goal of understanding the important questions and some possible (and impossible) answers.