ИСТИНА

Do languages that lack articles have an extended functional structure in the nominal domain, including a DP projection? This question has been a topic of a long-lasting debate in the linguistic literature. Previous studies focused on Slavic languages (but see Bošković and Şener 2012 on Turkish); some scholars (cf. Progovac 1998, Rutkowski 2002, inter alia) argued in favor of the DP projection, while Bošković (2005 and later work; Bošković and Şener 2012) argued against it. Pereltsvaig (2006, 2007) proposed that while some nominals in Russian are DPs, others are Small Nominals (SNs) of different sizes. In this paper, we provide novel evidence for the latter position based on another Turkic language, Tatar (spoken by over 5 million in Tatarstan, Russia). Drawing on our fieldwork on one subdialect of Tatar (spoken in the village of Kutlushkino), we show that different syntactic constructions call for nominals of different sizes. Moreover, we argue that Differential Object Marking (DOM) in Tatar—unlike in other Turkic languages such as Turkish or Sakha—can only be explained in terms of the amount of functional architecture in the object: DP objects receive structural (accusative) Case, as in (1a), while SNs (i.e. NPs or NumPs) remain Caseless, as in (1b). Thus, we rule out alternative analyses based on distinct positions of accusative and unmarked objects or on the semantic interpretation of the object. More generally, we propose that only DPs must receive structural case while SNs are not subject to such Case licensing requirements and may remain morphologically caseless. We further buttress this analysis by showing that DPs and SNs differ in their Case marking not only in the object position but inside nominals as well.