Licenser under cover: The genitive of negation in Slovenian
- Matic Pavlic, University of Ljubljana
- Glendinning Lecture Theatre 2D67, UWE Bristol
In Slovenian, syntactic environments can be found in which noun phrases in negative sentences do not take the nominative or accusative, as one would expect from the parallel positive sentences. Instead, they are inflected for genitive (GenNeg). The phenomenon is widespread among Slavic languages. Consequently, it is well described (Babby 1980, Pesetsky 1982, Brown 1999, Pereltsvaig 1999, Harves 2002) – yet not satisfactory explained. Most of the theories assume that GenNeg is assigned by the NegP head, but on the basis of pleonastic negation Brown&Franks (1995) provide counter-evidence to that claim. Now, if it is not NegP who assigns GenNeg, it remains unexplained why GenNeg is found only in negated sentences. Also it is not clear, why the relevant NPs change their case precisely to genitive.
Four characteristics of GenNeg should explain: (i) as mentioned above, GenNeg appears only in the scope of sentence negation; (ii) cross-linguistically, GenNeg can target different sets of syntactic environments, yet they all have to originate from the position of the underlying direct objects (Bailyn 1997: 86); (iii) the switch from nominative or accusative to genitive is accompanied by „some difference in syntactic structure and/or in semantics or pragmatics” (Partee&Borschev 2002a: 181); (iv) the distribution of the GenNeg is strikingly similar to the distribution of negative quantifiers (Pesetsky 1982, Pereltsvaig 1997).
Negative quantifier nič and GenNeg. It is known that certain quantifiers have the same effect on the verb as Slovenian GenNeg, stated in (iv). These quantifiers also require their complement to be genitive. It thus seems reasonable to assume that GenNeg is licensed by a phonologically null quantifier. This is the main contribution of my paper. I observe that the distribution of GenNeg is very similar to the distribution of the n-numeral (a numeral which is a negative word in the sense of Zeijlstra 2004) nič `nothing´. Thus I put forth the hypothesis that GenNeg is licensed by the covert version of this word.
Four characteristics of GenNeg should explain: (i) as mentioned above, GenNeg appears only in the scope of sentence negation; (ii) cross-linguistically, GenNeg can target different sets of syntactic environments, yet they all have to originate from the position of the underlying direct objects (Bailyn 1997: 86); (iii) the switch from nominative or accusative to genitive is accompanied by „some difference in syntactic structure and/or in semantics or pragmatics” (Partee&Borschev 2002a: 181); (iv) the distribution of the GenNeg is strikingly similar to the distribution of negative quantifiers (Pesetsky 1982, Pereltsvaig 1997).
Negative quantifier nič and GenNeg. It is known that certain quantifiers have the same effect on the verb as Slovenian GenNeg, stated in (iv). These quantifiers also require their complement to be genitive. It thus seems reasonable to assume that GenNeg is licensed by a phonologically null quantifier. This is the main contribution of my paper. I observe that the distribution of GenNeg is very similar to the distribution of the n-numeral (a numeral which is a negative word in the sense of Zeijlstra 2004) nič `nothing´. Thus I put forth the hypothesis that GenNeg is licensed by the covert version of this word.