Hilda Koopman (UCLA): Are objects scattered in all languages?
In many languages, direct objects show different syntactic distributions ("scramble" ) (and/or formal encodings), depending on a number of variables: whether they are weak pronouns, definite DPs, quantified DPs, specific indefinites, indefinites, focused DPs etc, etc. . (cf Diesing 1977, Aissen 2003, etc .) We can diagnose where in the hierarchy these types of objects appear ("scramble" to) by examining the interactions (order and scope) with elements on their right and left: i.e. objects visibly occur at different height in the structure along universally stable patterns (definite > specific indefinite > indefinite.. etc).
Why would this be the case? What variation do we expect to find?
I would like to explore the hypothesis that there is in fact no variation between languages in this respect.
(1) In all languages objects are scattered at spell out.
To do so,
(i) I will spell out the theory, and the grammatical architecture that makes this hypothesis plausible, and the predictions testable;
(ii) discuss why OV languages are all expected to show this quite transparently (as they do), but VO languages are expected to be quite opaque (which seems to be the case):
(iii) apply the theory: a comparative study of Yiddish, Dutch and English, showing the great underlying hierarchical stability of scattered objects (and why this is so difficult to see in English);
(iv) finally I will discuss an African OV language (Bambara, Mande) which seems to be a direct counterexample to (1). Yet, as I will suggest and support, even here scattering and height will be seen to be operative.
In conclusion, if the research direction taken here is on the right track, it tells us what specific grammatical theories should be pursued. In some implementations of the Minimalist Program (parametric variation as to low spell out or high spell out), there is no expectation that all languages have the different types of objects occupy different positions at spell out. Indeed, given copy theory, copies can be spelled out low but interpreted high, or spelled out high and deleted low.
Two papers will be helpful as preparation for the talk:
Cinque, G. 2005. Deriving Greenberg's universal 20 and its exceptions, it Linguistic inquiry, 36. 3. 315-332.
Diesing, Molly 1997. Yiddish VP order, and the Typology of Object movement in Germanic, NLLT 15: 369···427, 1997
An exploration of the SSWL database http://sswl.railsplayground.net/ which I will try to use a bit in the talk will be useful as well.