Non-nominative subjects in comparison

Abstract

Among languages with a sufficiently rich system of morphological Case we observe unmarked constituent orderings which deviate from the nominative preceding non-nominative pattern. This deviation, if one wants to call it that way, is to a large extent lexically and semantically predictable. Languages of this kind are classified as languages that permit nonnominative subjects. As is well known, however, they differ quite radically as to certain syntactic consequences which the non-nominative-first pattern may have. German and Icelandic are closely related Germanic languages which not surprisingly show strong similarities in their argument structures and syntax of Case. Nevertheless, they differ by the fact that non-nominative prominent DPs in Icelandic behave like genuine subjects while they do not (or do to a lesser extent) in German. The goal of the present article is to explore the possibility of deriving differences in subjecthood from the basic order of constituents. Icelandic has a head-initial VP which is separated from an external argument by a functional head F0 (or a number of functional heads), i.e., the order is SpecFP F0 [VP V ]. German has a head-final VP instead. There are strong indications that the order [VP V] F0 does not give rise to a functionally defined position SpecFP.Following work by others, I will argue that this is related to the nature of F0. Given that both German and Icelandic (and perhaps all languages) have a prominent argument that has certain syntactic and semantic properties one associates with the notion subject , my conjecture is that the difference derives from the fact that head-initial languages show a grammaticalization of the prominent argument that is missing in head-final languages. I will try to derive this grammaticalization from the nature of F0 and explore how far this can take us. Although the German/Icelandic contrasts will set the stage, datafrom other languages primarily from Bengali will be drawn into the argumentation. I will first turn to expletives, then point out the relevant similarities and differences between German and Icelandic. After this there will be a section on argument structure and unmarked ordering. The next section will turn to Minimalist assumptions about checking under c-command, the Extended Projection Principle (EPP) and reasons for displacement. We will then explore how in head-final languages checking can take place under m-command. Finally we will indicate how the different mechanisms for checking can yield the observed differences in subjecthood. I will throughout use the informal term subject for convenience. This usage should not imply that there is actually such a thing in the theory of grammar.