Research

The main focus of my research is the syntax of human languages and its interaction with representations of sound and meaning. I also work on the role of syntactic representations in sentence comprehension in normals and agrammatic patients.
My current research primarily focuses on relative clause formation and is funded by a European Research Council IR Grant, Grant Number: 224943, as part of the European Union's Seventh Framework Program.

1. Syntax-semantics interface

Relative Clauses
My work on relative clause formation has been focused on the interpretation of the head noun inside the relative clause. In Szczegielniak (2004), I have shown that there are two distinct types of relative clauses in Polish and Russian that are marked by two distinct relative markers: ‘co/cto’ (uninflected complementizer, similar to English ‘that’), and ‘ktory/kotoryi’ (inflected relative operator, similar to English ‘which’). Relative clauses derived via a Raising analysis (Sauerland 1998) have the relative marker ‘co/cto’, whereas relative clauses derived via a Matching analysis have the inflected marker ‘ktory/kotoryi’ (which I argue to be the operator). There is also evidence that there are no null relative operators in these languages. Support for two distinct mechanisms of relative clause formation comes from a battery of semantic and syntactic tests aimed at establishing whether the head noun is interpreted inside the relative clause. The tests include: the ability to break up idioms, the ability to have degree readings, the ability to have appositive readings, and Condition C and Condition A violations.
My current work explores the structure of the Left Periphery and its role in relative clause formation. I am examining whether the Left Periphery structure proposed by Rizzi (1997) is adequate for explaining the structure of relative clauses in Russian and Polish, as well as in other languages that allow multiple relative markers within one relative clause (Norwegian, Danish and Swedish, for example). This work is aimed at establishing what processes and structures are involved in relative clause formation.

Relevant Recent Publications/Presentations:

“VP Ellipsis and Topicalization” (2006) NELS 35 Proceedings. Leah Bateman, Cherlon Ussery (eds). GLSA UMass, BookSurge Publishing.Volume 2. 603-615.

“Relativization that you did…” MIT Occasional Papers in Linguistics. Vol. 24. 2005.

“Ellipsis and Relativization in Slavic” (2006). WECOL 2004 Proceedings, Vol 16. Michal Temkin Martínez, Asier Alcázar,  Roberto Mayoral Hernández (eds). University of Southern California. 373-385.

LingLunch, MIT, December 2007
“Relative Clause formation.”

WECOL, USC November 2004
“Ellipsis in Relative Clauses in Slavic.”


VP Ellipsis
My work explores the syntactic and semantic constraints on ellipsis. In Szczegielniak (2004), I have argued that VP ellipsis in Polish and Russian is licensed via overt VP topicalization. Following Rooth (1992), I assume that ellipsis has to be preceded by de-stressing. I show that de-stressing in Polish and Russian has to be preceded by the establishment of topic or focus relations in the syntax. In the case of VP ellipsis, this would be raising the VP to a topic position in the Left Periphery. This results in bare VP ellipsis structures, where everything but the subject can be elided. Support for this claim comes from the fact that constraints on forming VP topics are the same as constraints on bare VP ellipsis. Non-bare VP ellipsis (the type found in English where an auxiliary/modal/negation or ‘do’ has to be present) is derived by establishing what is focus in the syntax. The obligatory presence of an auxiliary/modal/negation or ‘do’ results from the need to have a focus ? head (Laka 1994) in the numeration.
My current research attempts to extend these findings for languages other then Polish, English and Russian. I am mainly working the properties of VP ellipsis in Swedish, German, Spanish, Tagalog.

Relevant Recent Publications/Presentations:

“Relativization that you did…” MIT Occasional Papers in Linguistics. Vol. 24. 2005.

“Ellipsis and Relativization in Slavic” (2006). WECOL 2004 Proceedings, Vol 16. Michal Temkin Martínez, Asier Alcázar,  Roberto Mayoral Hernández (eds). University of Southern California. 373-385.

“VP Ellipsis and Topicalization” (2006) NELS 35 Proceedings. Leah Bateman, Cherlon Ussery (eds). GLSA UMass, BookSurge Publishing.Volume 2. 603-615.

WECOL, USC November 2004
“Ellipsis in Relative Clauses in Slavic.”

Ling-Lunch, MIT November 2004
“Ellipsis and movement.”

NELS 35, UConn October 2004
“Ellipsis and Remnant Movement: Evidence from Slavic.”

Ling-Lunch, MIT March 2003
“ACD in Russian and Polish.”

Sluicing
In my work on sluicing, on the basis of data from Polish, German and Russian, I argue that contrary to Merchant (2001), Fox and Lasnik (2003) sluicing does not alleviate Island constraints. Instead I argue that the elided part of the sluice prefers to be syntactically parallel with the antecedent, but can be semantically parallel, or even discourse parallel (Hardt 2002, Culicover and Jackendoff 2005). In cases of apparent island alleviation, a grammatical continuation is possible that captures the necessary binding effects (Fox and Lasnik 2003).
Evidence supporting this claim comes from non preposition stranding languages like German and Polish, where it is shown that cleft continuations allow for the deletion of prepositions (in which case Polish is limited to d-linked (Pesetsky 1982) wh-sluices and German to nominative wh-endings, both required if the ending is a cleft). Further support comes from multiple wh-sluices in Polish, where island alleviation is not possible since multiple clefts are also not possible.

Relevant Recent Publications/Presentations:

"Islands in Sluicing in Polish". (2008) Proceedings of the 27th West Coast Conference on Formal Linguistics. Natasha Abner and Jason Bishop (eds). Cascadilla Proceedings Project, Somerville, MA, USA. 404-412.

WCCFL 27, UCLA, May 2008
“Islands in Sluicing in Polish.”

2. Syntax-phonology interface

Clitics
The focus of my research has been the properties of clitics in Slavic languages as compared to clitics in Romance languages. My work has led me to argue that Polish auxiliary clitics are Focus markers when not hosted by the verb and affixes when attached to the verb (see also Embick 1995, Borsley and Rivero 1994). This is in contrast to clitics in Slovak whose distribution is governed by their morpho-phonological properties. This work has been a part of a larger project of establishing the structure of Left Periphery in Slavic languages.
I also explore variation of clitics in other Slavic languages, mainly Polish versus Czech, Serbian and Croatian. The aim of this research is to establish what parametrical variation is responsible for the difference between Polish, where there are no second position clitics, and the other languages, where there are second position clitics.

Relevant Publications/Presentations:

“Clitic Position within the Left Periphery: evidence for a phonological buffer.” Clitics and Affix Combinations: Theoretical perspectives. L. Heggie, P. Ordoñez (eds). John Benjamins. 283-301. 2005.

“Deficient Heads, and Long Head Movement in Slovak.” Formal Approaches to Slavic Linguistics. The Indiana Meeting 1996. Steven Franks and Martina Lindseth (eds). Ann Arbor. Michigan Slavic Publications, 312 333. 1995.

“Certain Aspects of Cliticization in Polish.” Working Papers In Linguistics. Penn Review of Linguistics. Vol.2, No.2, 143 159. 1994.

LSA Summer Institute workshop entitled: “Perspectives on Clitic and Agreement Affix Combinations.” University of Illinois, Urbana Champaign, July 1999.
“An OT account of Slavic agreement clitics.”

30th. Poznan Linguistics Meeting, Adam Mickiewicz University, Poznan, Poland, July 1997.
“A Minimalist Account of Past Tense Constructions in Polish and Slovak.”

Formal Approaches to Slavic Linguistics Fifth Workshop, Wabash College and Indiana University, Indiana, April 1996.
“A Minimalist Account of Polish Auxiliary Clitics and of why there is no Long Head Movement in Slovak.”

3. Sentence comprehension

Comprehension of VP elided structures
(in collaboration with Evelina G. Fedorenko and Edward F. Gibson, BCS MIT)
We explore the extent to which different kinds of information are retrieved when processing VP elided structures. We developed a novel experimental paradigm (within-modality lexical priming) by combining a self-paced word-by-word reading paradigm with a visually presented lexical decision task. Subjects are asked to read sentences, which are interrupted at some point by a lexical decision task on a string of letters presented above the sentence. The time it takes subjects to make the decision whether a given string of letters is a word or not is taken to be a measure of ease of accessibility of that word. The general logic of the experiment is that processing an elided structure will result in some activation of the elided material, which should result in priming effects. This paradigm allows us to tap into different types of information about the antecedent of the elided structure that is available at the ellipsis site.
We assume that accessing semantic information is necessary for interpreting elided structures. However, phonological information may or may not be accessed. The results of this research will contribute to our understanding of whether only minimally-required processing takes place in sentence comprehension, or whether processing semantic information inevitably results in at least partial activation of phonological information, despite the fact that phonological information is not required for extracting the meaning from the sentences. Furthermore, at the word level, this research may help differentiate different between various models of lexical access (e.g., serial models vs. cascaded models).

Relevant Publications/Presentations:

“The processing of elided structures.” MS. Harvard/MIT 2006 with Evelina G. Fedorenko and Edward F. Gibson (MIT).
CUNY’05 Sentence Processing Conference, March 2005, Tucson Arizona

Activation of traces in long distance dependencies
(in collaboration with Edward F. Gibson, and Christopher Hirsh, Wexler Lab MIT)
We are also utilizing a similar paradigm to examine lexical activation in trace positions of wh-movement constructions. Specifically, we are exploring what information is being activated as part of a long distance relation formation and what information is being activated for semantic and pragmatic reasons, like for example, the establishment of coherence.

Comprehension of sluiced structures
(in collaboration with the Wexler Lab at MIT, Christopher Hirsh)
We are designing a set of experiments designed to explore the phenomena of island alleviation in Sluicing (Ross 1969). It has been reported that sluicing seems to alleviate the pronunciation of strings that otherwise pronounced would be ungrammatical. Consider the following example (italics indicates ellipsis):

1. John went to the store after calling someone, but I do not know who John went to the store after calling

The string 'who John went to the store after calling' is ungrammatical in English, and yet the sluiced structure is fine. Fox and Lasnik (2003) have proposed that deletion of such a string makes the structure grammatical. Following my research on sluicing, I argue that this is not the case. There is a perfectly viable continuation:

2. John went to the store after calling someone, but I do not know who it was that after calling him John went to the store

The continuation of the sluice is semantically parallel to the antecedent but not syntactically parallel. In order to establish whether comprehenders first process a syntactically parallel structure and only when it is ungrammatical do they come up with a semantically parallel one a self paced reading task is being designed to measure the reading times of sluices that alleviate islands and those that do not.

Relevant Publications/Presentations:

“All Sluiced up and nowhere to go.” MS. Harvard 2006.

LSA, January 2006 Annual Meeting
“Sluicing and P Stranding in Polish.”

Ling-Lunch, MIT 2005
“All Sluiced up and Nowhere to Go.”

4.Language disorders

(in collaboration, Beste Kamali, Clemens Mayr, Cedric Boeckx at Harvard University)
I received my training in studying language disorders in Alfonso Caramazza’s Lab at Harvard University. Our work explores the types of syntactic knowledge retained in agrammatic patients. Currently, we are examining the degree of knowledge constructions retained in patients who are diagnosed as Broca’s aphasics and who exhibit classical agrammatic patterns in production and comprehension in Turkish. Our experimental paradigm involves grammaticality judgment tasks, picture-sentence matching comprehension tasks, and elicited production tasks. This research will contribute to our understanding of the current theories of agrammatic production and comprehension.

Relevant Publications/Presentations:
3rd LPIA (Left Periphery in Aphasia)-Meeting: The Structure of the Left Periphery in Germanic languages – CP- and IP-related elements in normal and impaired speech. Venice 2006.

“Revisiting agrammatic comprehension: the case of Turkish.”
(With: Beste Kamali, Clemens Mayr and Cedric Boeckx, Harvard University)

HUMIT (Harvard-MIT Student Conference), September 2001
“Agrammatic Comprehension: The Case of WB.”
(With Xavier Alario and Alfonso Caramazza, Harvard University)

HUMIT (Harvard-MIT Student Conference), August 2000
“Phonological errors at the beginning of a word.”
(With Deborah Grossman and Alfonso Caramazza, Harvard University)

5. Bare Bones Syntax

This research explores a model where syntax is as simple as possible. My work takes current research aimed at simplifying syntax (for example: Chomsky 2001, Culicover and Jackendoff 2005, Brody 2000 Boeckx 2005) a step further. The model pursued in this research assumes that the syntactic component builds recursive merological sets. Syntax can read only lexical features directly involved in the construction of such sets: subcategorization/c-selection, and a long distance dependency feature. The sets are unordered. Other features that are semantic or morpho-phonological in nature (for example: case, agreement, wh-features, topic, etc.) the syntax ignores. The representation that syntax builds is a multidimensional set that has the topological property of not being altered by homeomorphisms, such as stretching or bending. In essence is it a set function that has a group of solutions constrained by the function’s geometrical properties. This function is read off by two real time processes. P1 linearizes the sets generated by syntax, it can read morpho-phonological features like agreement, tense, case. It cannot read the semantic aspects of case, or features like Force, Topic, or Focus, or scope features. The latter features are read off by P2 that processes semantic argument structure, and operator variable chains. P1 and P2 can communicate directly in cases when they encounter features in the syntactic representation that they cannot read. Furthermore P1, being a linearizer reads of the set representation from the inside out (top down in tree terms), whereas P2 reads off the syntactic representation form the inside out (top down in tree terms). In this model syntax is reduced to a recursive set building procedure. Movement is a by-product of P1 and P2 interpretation, phases are a result of the memory constraints imposed on P1 and P2. Islands are a result of miscommunication between P1 and P2.
The advantages of this model over current ones are: (i) integration of the real time processor with a static syntactic representation, the elimination of the stipulation that syntax has phase properties, has movement, can read features that are not syntactic in nature.

Relevant Publications/Presentations:

“Polish Optional Movement.” The Minimalist Parameter: Selected Papers from the Open Linguistics Forum, Ottawa, 21-23 March 1997. G. M. Alexandrova and O. Arnaudova (eds). (Current Issues in Linguistic Theory, Vol.192). John Benjamins, 137-159. 2001.

“'That t Effects' Cross Linguistically and Successive Cyclic Movement.” Papers on Morphology and Syntax, Cycle One MIT Working Papers in Linguistics, Vol 33. Karlos Arregi, Benjamin Bruening, Cornelia Krause, and Vivian Lin (eds). MITWPiL, 369 393. 1999.

“Why Move?” MS Harvard University. 2002.

Ling-Lunch, MIT 2006
“Bare bones syntax and how it interacts with processing. An investigation of the Strong Minimalist Hypothesis.”

Linguistics Theory Group, Harvard, November 2005
“The Topology of Syntactic Structures.”