Rivero’s (1991, 1998, 2000, 2003) analysis of verb fronting has met with a fair amount of criticism in the literature. Among other things, critics take issue with her stance that clitics are distributed across the clausal domain. For example, Ćavar & Wilder (1994) argue that auxiliaries cannot be in I since—if they were—we would expect to find the SpecIP subject position filled; Bošković (1995) maintains that verb-movement cannot be to C since li and participles cannot cooccur. These objections lose their force however when viewed in light of evidence from other Slavic languages. For example, in Czech pronominal subjects occupy a position between a comp and aux (i.e. C and I) (1a) and participles cooccur with li (1b).
(1) a. Kam tys chodila do školy? (Müllerová et al. 1992)
‘Where did you go to school?’
b. Zlobil-li jsem, proč jsem nebyl vyhozen? (Vaculík, 1998)
‘If I misbehaved, why wasn’t I tossed out?’
This is just what one would expect if the distributed clitic account were correct. Are we then to accept Rivero’s original position? In this paper, while rejecting Long Head Movement, I argue that the distributed view is essentially correct. Specifically, I generate clitics in their respective categorial projection—complementizers in C, tense auxiliaries in T, pronominals in argument position with subsequent movement to AgrO, etc., without any further movement to C.
Such an analysis has both empirical and conceptual advantages over existing accounts. Empirically, only a structural analysis can account for the kind of subject-object asymmetries we observe in (1a-2a).
(2) a. *Kam tos dala? (Fried, p.c.)
b. Kams to dala?
‘Where did you put it?’
As a comparison of (1a-2a) shows, a subject pronoun ty stands to the left of aux (= SpecIP) in (1a), but in the very same position an object pronoun to is disallowed (2a). Rather an object must occupy a position lower than aux (2b).
Such an analysis suggests a solution to long-standing empirical problems, as well. For example, we know that the copula is not a clitic in Czech, as evidenced by its ability to occupy the initial position (4a). On the other hand, it is commonly accepted that clitics occupy 2P. The fact that nonclitic je should precede clitic ti in (4b) is a mystery if the traditional assumption is correct. But the mystery disappears as soon as the situation is viewed in structural terms: if the copula is located in T and object clitics occupy a position lower than T (cf. 2b), then the unexpected position of ti in (4b) suddenly makes sense.
(4) a. Je mi jednaadvacet. (Mills 2003)
‘I’m twenty-one’
b. Kolik je ti let?
‘How old are you?’
c. *Kolik ti je let?
Conceptually, all things being equal, the simpler account is to be preferred. The proposed account offers an attractive alternative to existing accounts by moving the debate from ad hoc properties such as prosodic deficiency and stipulative mechanisms such as movement to and from C to accounting for residual clitic effects.
References
Bošković, Ž. “Participle Movement and Second Position Cliticization in Serbo-Croatian.” Lingua 96 (1995): 245-266.
Ćavar, D. & Ch. Wilder. “‘Clitic Third’ in Croatian.” Clitics: Their Origin, Status and Position. Eds. van Riemsdijk H. and Hellan L. Eurotyp (1994): 19-61.
Rivero, M. L. “Long Head Movement and Negation: Serbo-Croatian vs. Slovak and Czech.” The Linguistic Review 8 (1991): 319-351.
Rivero, M. L. “Verb Movement and Economy: Last Resort.” Topics in South Slavic Syntax and Semantics. Eds. Dimitrova-Vulchanova M. and Hellan L. Amsterdam: Benjamins (1998): 1-23.
Rivero, M. L. “Finiteness and Second Position in Long Head Movement Languages: Breton and Slavic.” Syntax and Semantics 32. Ed. Borsley R. New York: Academic Press (2000): 295-321.
Rivero, M. L. (ed.). Comparative Balkan Syntax. Oxford University Press, 2003.