Linguistics Colloquium: Natasha Fitzgibbons

“Negative Concord Items are not Negative?”
Natasha Fitzgibbons, Concordia University

In logic, two negatives cancel each other out leading to an affirmative proposition. Many natural languages, however, make use of the option not provided by logic – negative concord. The lexicon of negative concord languages contains a special class of words, n-words, that have to cooccur with sentential negation:

(1)	Ja nikogo *(ne) znaju.	Russian
	I    n-who     not  know
	‘I do not know anybody.’

There is no consensus in the literature on the locus of negative meaning in negative concord languages: is it sentential negation head or the n-word that is semantically negative? The majority of researchers agree that n-words can supply negative meaning at least in some special circumstances.

In this talk, I explore some evidence that has been used to argue that n-words can contribute negative meaning. I discuss three languages (Russian, Serbo-Croatian and Spanish) and two arguments: (i) acceptability of n-words in fragment answers (2) and (ii) some cases where double negation (DN) interpretation is available in a negative concord language (3). Both of these empirical observations appear to force the conclusion that n-words can contribute a negative meaning:

(2) a.		Serbo-Croatian
i.	Šta   si    kupio?
	what are bought
	‘What did you buy?’
ii.	Ništa
	nothing
	‘Nothing.’ (intended: I bought nothing.)
b.		Russian
i.	Kogo ty    videl?
	Who  you saw
	Who did you see?’
ii.	Nikogo.
	n-who
	‘Nobody.’ (intended: I saw nobody.)

(3)  a.		Spanish
	Nadie   miraba a  nadie.
	n-body looked  at n-body
	‘Nobody looked at anybody.’
	‘Somebody looked at somebody.’
b.		Russian
	Konečno  nikto       ne   znal!
	Surely      n-who Neg knew
	‘Like hell nobody knew!’
	‘Somebody knew.’

With respect to fragment answers, I argue for an ellipsis account of licensing of n-words: there is sentential negation present in the logical form of (1a,b). I also explain how recoverability of deletion is satisfied in these cases. With respect to availability of double negation readings in some negative concord contexts, I will argue for an additional negative element in the CP domain in Russian and Spanish. In Spanish, it is signaled by intonation only, but in Russian, in addition to intonation, it has to contain phonologically overt material. The conclusion that n-words can contribute negative meaning is thus not necessary.