The case of PRO

Abstract

Icelandic case agreement suggests that nominative case is active in PRO infinitives in much the same way as in finite clauses, thus posing a difficult and a long-standing problem for generative (GB and minimalist) case theory and the PRO Theorem. In this article, I examine the Icelandic facts in detail, illustrating that the unmarked and common nominative morphology in Icelandic PRO infinitives is regular structural nominative morphology, suggesting that PRO cannot be reduced to a copy. What went wrong in the GB approach to PRO was not PRO itself but the binding theoretic and 'Case' theoretic conception of it. PRO is an empty category that is simultaneously a reference variable (like overt pronouns and anaphors) and a phi-feature variable (unlike overt expressions). Due to this unique combination of variable properties, PRO cannot be deduced from other traits of grammar, such as movement, nor can it possibly be lexicalized. Importantly, also, the facts studied here suggest that case is a post-syntactic category, assigned in morphology. In contrast, Person is evidently a syntactically active category, having some of the properties and effects that have commonly been attributed to 'Case'.

On EPP effects*

“…Second, Icelandic shows that nominative case is fully grammatical vP-internally, as in (15a) (this is a general phenomenon in Icelandic and many other languages, see the discussion in e.g. Sigurðsson 1989Sigurðsson , 2008Sigurðsson , 2009. Nonetheless, the definite NP in (15b,c) has to raise.…”

Icelandic Control Really Is A-Movement: Reply to Bobaljik and Landau

“…Contrary to what Bobaljik and Landau (2009) (and SigurLsson (2008)) claim, the MTC also has a very simple explanation for why control infinitives in Icelandic cannot license an overt subject, as illustrated in (23) (Jónsson 1996, reproduced in Bobaljik andLandau 2009:125) b. É g baL Maríu [aL (*hún/*Á sta) fara ein ÈangaL]. I asked Maria.ACC to she/Asta.NOM go alone.FEM.SG.NOM there 'I asked Maria (for her/Asta) to go there alone.'…”

“…The second type of purported challenge for the MTC from Icelandic involves control configurations in which embedded floating quantifiers and secondary predicates display Case agreement morphology that differs from that of the controller in the matrix clause (see Landau 2003, SigurLsson 2008, Bobaljik and Landau 2009). In the sentences in (12), for instance, the matrix subject bears (structural) nominative Case (see (12a)) and (quirky) accusative Case (see (12b)), but the secondary predicate in the embedded clause shows up with dative Case, which is the quirky Case assigned by the embedded verb.…”

“…SigurLsson (2008) offers a similar critique, which we address separately (seeBoeckx, Hornstein, and Nunes, in preparation).3 See, for example,Landau 2004 andSigurLsson 2008. …”

Case variation: Viruses and star wars

“…In a language where it has taken place, active Acc assignment in the v-system is based on the structure in (11), whereas the corresponding unaccusative and passive structures yield Nom, as sketched in (12); the arrows connecting v*-V and NP 1 indicate a matching relation that gets externalization process (suggesting that the PF derivation is NOT subject to the same locality and directional constraints as the syntactic derivation). In passing, it should be noted that PF case percolation does not bear on the (syntactic) Movement Theory of Control (MTC), contrary to common assumptions (see Sigurðsson 2008Sigurðsson , 2011c;…”