In many Slavic languages all wh-words are moved to clause initial position at S-structure. It is standardly assumed that there are two distinct, parameterized types of languages based on the landing site of the multiple fronted wh-phrases (Rudin 1988, Norvins 2002): the Bulgarian type where all wh-phrases form a ‘group’, and the Serbo-Croatian type, where wh-phrases can and sometimes must be split.
Previous analyses have presented the split as the interaction of the first wh-word in the cluster and a parenthetical, an adverb, a particle, or a second-position clitic (Cichocki 1983, Rudin 1988, Bošković 1996, Citko 1997). I am focusing on the fact that in Russian and Czech, several functional heads in a row can split the wh-cluster (NP+ConjP, AdvP+PP, etc.). The crucial data are exemplified in (1) - (3), which show that the wh-cluster can be split by multiple constituents and the constituents represent several types of functional heads (verb, noun, conjunction). I will not discuss long movement of a single wh-word across a clause boundary that results in a clause splitting the wh-cluster mainly because this phenomena pertains particularly to Serbo-Croatian and does not occur in the discussed languages.
Taking a generative approach, my primary goal is twofold: a) to define the landing site for the elements that split the fronted wh-cluster and b) to draw conclusions about placing Czech and Russian within the typology of multiple wh-fronting.
I argue that the constituent distribution within the split (allowing different syntactic elements, in addition to clitics) supports the proposal put forward by Uriagereka (1995), where clitics are said to occupy a specific functional projection associated with such notions as theme and rheme. I extend this specific functional projection to being capable of hosting additional types of elements as a Discourse-related phrase. The ability of the splitting elements to represent the discourse-related phrase is triggered by the syntax of specificity and, in the case of pronominal clitics, referentiality. Thus this discourse-related phrase carries a load of a site where point of view is encoded (specificity) and also the association of a referent to a semantic denotation (referentiality).
I also demonstrate the restrictions on the possibilities of wh-words ordering combinations (elaborating on superiority, focus, etc.) (cf. also discussion in Billings and Rudin 1996) and I develop a hierarchy of the wh-words with regard to their order in the wh-cluster.
Data:
1. a) Russian:
Он рассказывал
о том, где они и как познакомились с Марией.
He was-telling about that where
they and how got-acquainted with Mary
‘He was telling about how and where Mary and
he got acquainted.’
b) Czech:
Povídal o tom, kde
se tam s kým seznámil.
He-was-telling about that where
‘clitic’ there with whom he-got-acquainted
‘He was telling about where he got acquainted
and with whom (he got acquainted).’
2. Czech:
Kdo píše a co o Kanadě?
Who writes and what about Canada
3. Czech:
Kdo a co komu koupil?
Who and what for-whom bought