Some notes on DP-internal negative doubling

Liliane Haegeman (Université Charles de Gaulle LILLE III)


In the literature on negation, attention has been paid to the encoding negation at the sentence level. Relations such as those between a negative head and a negative constituent or that between multiple negative constituents have, for instance, been examined mainly in terms of sentence structure (see, for instance, Horn 1989, van der Wouden 1994, Bayer 1990, Corblin and Tovena 2000, Déprez 1999, etc). The realisation and interaction of negation markers at sub-sentential levels such as DP, PP etc has, to the best of my knowledge, received little and only sporadic attention. I suspect that this is because in the core case there is one negative marker per DP, PP etc. and negative concord (NC) has not been signalled at those sub-sentential levels. However, the Flemish data in (1), described by Vanacker in the late 1970s in a paper written in Dutch, reveal that co-occurrence of negative constituents is also attested at the sub-sentential level. The examples in (1) were collected in the Flemish speaking area in northern France (Vanacker 1975: 127):

(1) a. Bij de jonge gasten en-es er nie vele geen Vlaams mee(r) gesproken.
with the young people en-is there not much no Flemish more spoken
'Among the younger people, not much Flemish is spoken.'
b. T'es daarvoren da kik nie vele geen beesten en-oude.
it is therefore that I not many no cattle en-keep
'That's why I don't keep many animals.'

In these examples, the negative DPs, nie vele geen Vlaams ('not much no Flemish') and nie vele geen beesten ('not many no animals'), contain two expressions of negation: the negative marker nie which negates the quantifier vele ('many'), and the negative quantifier geen, the Flemish/Dutch equivalent of English no, or German kein.

The DP-internal co-occurrence of what seems to be two negative quantifiers is interesting and raises a number of questions. First the data suggest that at least in the Flemish dialects, the surface position of the quantifier veel ('many') can (or perhaps must) be different from that of the negative quantifier geen ('no'), allowing them to co-occur. Moreover, the order in (1) suggests that the quantifier veel is spelt out in a position higher than geen.

That quantifiers and articles may not occupy the same position is not a novel idea. As early as 1977 Jackendoff (1977: 105) indicated that quantificational elements need not all be spelt out at the same position, and specifically he assigned a different position to the English quantifiers no and many (see also Giusti (1997) for a recent discussion of the position of prenominal quantifiers). However, in Jackendoff’s proposal English many would actually be spelt out lower than the negative quantifier no:

  Since some quantifiers [some, each, all , no, any, lh] are now Art[icle]s and some [many, few, several, lh] are Q[uantifier]s, the phrase structure component will generate structures in which two quantifiers appear, one in each position, e.g. *no many men, *all several men, *any much wine. (Jackendoff 1977: 105)

Jackendoff (1977:105) rules out such co-occurring quantifiers on semantic grounds:

  These are ruled out semantically, however, by the Specifier Constraint (5.1.), which forbids two (semantic) quantifiers in the same NP specifier. ' (Jackendoff 1977: 105, my italics)

One might in fact expect that if there are two quantifiers in the Flemish constructions, geen ('no') and nie vele ('not many'), the construction will crash because one of the quantifiers will quantify vacuously. Obviously, this is not the case since such data are attested

Vanacker's data in (1) are mainly drawn from Flemish dialects in Northern France, but he signals that the phenomenon is also to be found in the West Flemish coastal areas (1975: 132). My own WF dialect (Haegeman 1992), which is spoken in the rural area inland of Knokke-Heist, also exhibits such DP-internal negative doubling. The relevant data have already been briefly discussed in Haegeman and Zanuttini (1996). The main purpose of the present paper is to render the DP-internal negative doubling data accessible to a wider audience by offering a detailed description in English. As far as I can tell, the empirical facts of the WF dialect which I will be describing parallel those described by Vanacker. I hope that this description may encourage other researchers to look at the pattern. In a more speculative second part of the paper, I will also offer some proposals for an analysis.

The paper is organised as follows: section 2 sets the background and describes the properties of DPs containing negative markers in WF. Section 3 provides a detailed description of the syntactic properties of WF DPs with negative doubling and shows that these seem, to all intents and purposes, to share the external syntactic properties of non-doubled DPs. Section 4 is added to complete the survey and deals with internal negative doubling and NP ellipsis. Section 5 offers an analysis of the internal structure of DPs with negative doubling. Based on additional data from English, an articulated DP is elaborated to accommodate the patterns observed. Section 6 introduces additional data involving degree markers. Section 7 summarises the paper.

Back to Syntactic Microvariation