Empty categories revisited and revised cross-linguistically.

Ewa Bułat (University of Wrocław)

Working on the implementation of recent approaches to pro (Neeleman and Szendrõi 2005, Holmberg 2005, Hornstein 1999), I am going to propose a unified account of this category in English and Polish. Neeleman and Szendrõi (2005) notice that the traditional agreement-based theory of pro drop developed within GB by Rizzi (1986) faces difficulties with languages like Japanese and Chinese, which lack agreement, and yet allow pro subjects. They refer to this phenomenon as 'radical pro drop'. Holmberg (2005) advances this view, claiming that null pronouns in languages without Agr are in fact the only true instances of pro , that is to say, " pro exists, but (somewhat paradoxically, given the traditional view of pro ) only in languages which do not have agreement". Contrary to Rizzi (1986), Holmberg (2005) argues that languages like Polish (with subject agreement) cannot have a pro subject which is inherently unspecified for φ-features, and that pro cannot be identified by Agr (the φ-features of I) since Agr is uninterpretable in Minimalist theory.

Taking the above into account, there are at least two reasons for the need to re-examine the cross-linguistic occurrence of pro drop. First, if full lack of agreement allows subject deletion, then the lack of object agreement in English may be expected to license object pro in this language. Following Bhatt and Pancheva (2005) and Levin (1993), I would like to demonstrate that actually there are null objects in English, but this is in fact due to other than syntactic factors. In English, the possibility of omitting the object seems to be subject to idiosyncratic restrictions. In Polish, in contrast, null objects with arb interpretation seem to be generally available in generic contexts across a wide range of verbs. Therefore, the second issue that needs to be explored further is the role of semantic/pragmatic factor in pro licensing, the importance of which for a proper theory and typology of empty pronouns has been underestimated both in GBT and MP. Finally, I will try to check how the deleted subjects pro are related to PRO, the null subject of non-finite clauses. Hornstein (1999) assumes that PRO is small pro in cases where movement is prohibited. He refers to this phenomenon as non-obligatorily controlled (NOC) PRO:

(1) It was believed that [NOC PRO shaving was important].

However, following Bhatt and Pancheva (2005), 'uncontrolled' PRO does not always have an arbitrary interpretation. The arb interpretation is available only in generic environments; in episodic environments, the 'uncontrolled' PRO picks its interpretation from the local context, which is demonstrated in English and Polish examples below.

(2) It is difficult [PRO _arb to dance the tango].

Trudno jest [PRO _arb tańczyć tango].

(3) This morning it was difficult [PRO to dance the tango] since the floor was slippery and we were all tired].

Tego ranka trudno było [PRO tańczyć tango, ponieważ podłoga była œliska i wszyscy byliœmy zmęczeni].

This is part of the general context-sensitivity of null arguments, which once again proves the importance of pragmatic component in the investigation of empty categories.

All in all, I will propose that Polish and English partially behave according to the predictions made by Holmberg (2005). Namely, I will claim that agreement-based null subjects in Polish are not instances of pro , but regular, carrying a full set of phi-features pronouns that fail to be spelled out at PF. Nonetheless, if pro underlies NOC readings, and - ing and to can license pro in English, as suggested by Hornstein (1999), then it may be the case that gerundive -nie/-cie nominals and infinitival constructions in Polish could also license pro . Then, contrary to Holmberg's (2005) hypothesis, it may turn out that Polish actually has a pro subject of the type defined by Rizzi (1986), but in non-finite, and not finite clauses, as it has always been thought. What is more, another empty category that can be observed both in English and Polish is object pro , dependent on verb classes. There is merely a difference in the productivity of the null object option in the two languages, which is semantically/pragmatically determined. In fact, it seems highly probable that both subject and object pro are strongly influenced by some extra-syntactic factors.

References

Bhatt, R. and Pancheva, R. 2005. "Implicit Arguments", The Blackwell Companion to Syntax , ed. Martin Everaert and Henk van Riemsdijk, Volume II, Blackwell Publishing.

Holmberg, A. 2005. "Is There a Little Pro? Evidence from Finnish". Linguistic Inquiry 36, pp. 533-564. The MIT Press.

Hornstein, N. 1999. "Movement and Control", Linguistic Inquiry 30, pp. 69-96. The MIT Press.

Levin, B. 1993. English Verb Classes and Alternations: a Preliminary Investigation , The University of Chicago Press.

Neeleman, A. and Szendrõi, K. 2005. "Pro Drop and Pronouns", in Proceedings of the 24th West Coast Conference on Formal Linguistics , ed. John Alderete et al., pp. 299-307. Somerville, MA: Cascadilla Proceedings Project.

Rizzi, L. 1986. "Null Objects in Italian and the Theory of pro ", Linguistic Inquiry 17, pp. 501-557. The MIT Press.