Cecilia Poletto - The Internal Syntax of Negation
In this course I apply a cartographic approach to negation and propose that what is commonly defined as NegP is a cover term (much like IP or CP) for a complex set of projections that have each their own specific value and none of which has the semantic value of NegP (see Breitbarth (2012) for the proposal that NegP does not exist as such). Hence, Zanuttini's proposal, illustrated in (1) has to be restated as (2), where all negative markers in the clause are actually generated inside a big XP:
(1) [NegP1 non [TP1 V+Agr [NegP2 mica [ TP2 [AdvP already] [NegP3 niente [ Asp perf. Vpast part [Asp gen/progr [AdvP always] . [NegP4 NO]]]]]]]]
(2) [ Focus/Operator NO [ScalarP non [MinQ mica [ExistentialP (ni)-ente ]]]]
I claim that this complex XP is not the real semantic operator, but is only indirectly connected to it. As the syntactic realization of Tense differs from its semantics (see on this Giorgi and Pianesi (1997)), I think that all the elements that seem to encode negation in Romance are not the expression of this negative operator. Other authors have proposed that this is the case: Manzini and Savoia (2005) give a similar explanation assuming that negative adverbs have a nominal nature and are generated inside the predicate layer, Bayer (2009) proposes that negative adverbs start out in the object position. I will propose that this complex XP is generated at the border of the vP and then each of its internal elements can move to a different position in the clausal spine of FPs for feature checking. Hence Zanuttini's proposal in (1) can be reinterpreted as the result of the splitting of the structure in (2) after checking operations have applied.
Arguments in favor of the approach:
1. Such an approach straightforwardly accounts for cases of 'discontinuous negation' (as for instance standard French) in a way similar to Belletti's proposal of DP-doubling as stemming from a unique (big) DP: the two (or more) negative markers originate inside the same projection NegP and are then split to reach different positions where each checks its features (notice that this is already implicit in Pollock (1989) treatment of French negation).
If negation is marked through a quantifier, this will simply target the position of quantifiers in the sentence, which is presumably in the low IP space where also quantifiers like the universal one (tutto, 'everything') is located in Cinque's (1999) hierarchy. This straightforwardly derives why the presence of more than one negative marker does not give rise to a double negation interpretation, but to negative concord. It also explains why the various 'negators' are placed where they are in the structure of the clause: they just move to the FPs where they can check their features. In a sense, negation doubling is similar to DP doubling because the two elements actually behave like one (in terms of sharing a thematic role in the case of DP doubling and in terms of sharing the licensing of a unique negative interpretation). Such an account also captures rare but existing cases of tripling where the three elements are all morphologically different and cannot be the spell-out of the trace of the highest one.
2. One further argument to analyze negation doubling as originating from a single constituent (much as DP doubling in Belletti's (2004) analysis) is the fact that the two negative elements actually occur as a single constituent: for instance in dialects like Paduan, where constituent negation is marked by two elements together.
3. A third phenomenon that a cartographic approach to negation provides an account for is the different, though recurring etymology of elements which are reanalyzed to become the negative marker. As shown by Zanuttini (1997), Northern Italian dialects display several negative markers, but they are all derived from a restricted set of elements a) a minimizer b) the n-word corresponding to 'nothing' c) the item corresponding to the polarity particle. I will show that there exist at least two further etymological sources in Southern Italian: a) the negative marker neca found in Sicilian (see Cruschina (2011)), which is derived from a cleft clause, and Basilicatan manco (related to the adverb/verb 'lack'). The very same etymological processes are described for other language groups like Bantu or Austronesian by van der Auwera (2009). I will show that there is a correspondence between micro- and macrovariation in the etymological types forming negation markers. The fact that the same etymological sources are associated to negation in different language groups can be explained if we admit that each of these types spells out the same internal 'pieces' of the complex XP, but none of them is the actual semantic negation. If we adopt an analysis as the one in (2), Jespersen cycle receives a natural structural explanation in terms of lexicalization of different elements inside the big NegP.
4. The hypothesis also captures cases of movement of lower elements to higher position: minimizers are generally located higher than T2 but lower than T1, as Zanuttini (1997) shows, while negators deriving from n-words are generally lower. However, Manzini and Savoia (2005) show that this is not always the case, although the majority of the dialects still conform to Zanuttini's schema. If we want to keep Zanuttini's original generalization, and at the same time explain Manzini and Savoia's exceptions, we can assume that a first set of the exceptions in terms of movement of the 'negator' from the position it usually reaches to check its features to a higher one (as it happens with colloquial Italian preverbal mica). Other counterexamples where a given 'negator' occurs lower than the position where Zanuttini places it in the structure can also be explained by admitting that the element remains in its original merge position inside XP.
After having justified the labels in (2) and tried to see what the indirect link between all these elements and the actual semantic operator is, I will try to extend this approach to cases of negative concord where the two items are either two n-words or an n-word and a 'negator'.
The complete number and type of projections included in the complex XP triggering a negative interpretation of the clause is probably more complex than then one illustrated in (2). However, although (2) still has to be refined, this is the correct way to account for the wealth of negative markers and for their apparently rather wild distribution across the clause.
Van der Awuera, Johan (2009) The Jespersen Cycle. In Cyclical Change. Elly van Gelderen (hrg.) John Benjamins, Berlin.
Bayer, J. (2009), Nominal negative quantifiers as adjuncts. The Journal of Comparative Germanic Linguistics 12. 5 - 30
Breitbarth A. (2012) The development of negation in historical low German. Ms. University of Ghent.
Cinque, G. (1999). Adverbs and Functional Heads. Oxford University Press, Oxford.
Giorgi A. and F. Pianesi (1997) Tense and Aspect: from Semantics to Morphosyntax. Oxford University Press, New York.
Manzini and Savoia (2005)
Pollock, C. (1989). Verb Movement, Universal Grammar and the Structure of IP. Linguistic Inquiry, 20: 365-424.
Zanuttini R. (1997). Negation and Clausal Structure: A Comparative Study of Romance Languages. Oxford University Press, New York.
 This means that when double negation occurs the two negative elements are not generated together.