The Weekly Newsletter of MIT Linguistics

Privoznov defends!

August saw several wonderful and successful dissertation defenses. On August 19, we were privileged to participate via Zoom in Dmitry Privoznov’s brilliant defense of his dissertation entitled “A theory of two strong islands”.
A syntactic island is a structural domain that blocks dependencies such as that between a wh-phrase and the gap that it binds from applying across its boundaries. The nature of islands and island phenomena have been a central topic of syntactic research for over a half-century — and Mitya’s research offers strong new evidence adjudicating among distinct approaches, along with some entirely surprising new results supporting his perspective.
With evidence from the Balkar (a Turkic language of the Caucasus), Russian, and English, Mitya’s dissertation supports the hypothesis that the island status of subjects and adjuncts reflects the schedule by which constituents are “spelled out” and frozen in the course of a syntactic derivation. Remarkably, he shows that the same regime of spell-out that blocks certain extractions acts to *permit* certain semantic connections between indefinite noun phrases and pronouns (that are blocked when islands are *absent*). Mitya ably presented and defended his results to an audience on two continents.

Great work — congratulations!!



For those who want to read the official abstract for his defense presentation:

“This thesis examines two strong island effects: the Adjunct Condition and the Subject Condition. It proposes that both are derived from the same basic principles that determine when and to which constituent the rule of Spell Out is applied over the course of the derivation. The proposed theory consists of two assumptions. First, between any two phrasal sisters at least one must be spelled out. Second, a spelled out phrase does not project its category. The immediate consequence of these is that all adjuncts and all specifiers must be spelled out, because all adjuncts and all specifiers are, by definition, maximal projections whose sister is a phrase. This theory predicts, first, that all adjuncts and all specifiers are opaque for extraction, and second, that all adjuncts and all specifiers are interpreted before their sister. The thesis examines these predictions and argues that they are indeed borne out, based on data from Balkar, Russian and English.”