The quantifiers veelmany and weinigfew in (256) are not only existential, but in addition express that the cardinality of the boys walking in the street is higher/lower than some contextually determined norm. In terms of Figure 1, repeated below, this means that they express that the intersection of the set of boys (set A) and the set of entities walking in the street (set B) is non-empty: A ∩ B ≠ ∅.

The contextually determined norm is not an absolute number, but has a certain range; in the semantic representations in (257), n and n' refer to the lower and upper bounds of this range, respectively.

Degree quantifiers differ from purely existential quantifiers in that they modify not only plural count nouns but also non-count nouns, such as the substance noun water in (258). In this case, of course, the notion of cardinality is not applicable; instead, the degree quantifier expresses that the quantity of the substance denoted by the noun is greater/smaller than some contextually determined norm.

The quantified noun phrases in (257) and (258) function as the subject of an expletive er construction and are therefore clearly weak. However, the examples in (259) show that it is also possible to use degree quantifiers in strong noun phrases.

Like the existential quantifier enkele, the degree modifiers veel and weinig can either quantify over a pre-established set of entities in domain D or express a more “generic” meaning, i.e. quantify over all relevant entities in the speaker’s conception of the universe. In the first reading, an example such as (259a) asserts that, out of a contextually determined set of books, the number of books containing misprints is higher than some tacitly assumed norm. In the second (“generic”) reading, the speaker claims that a relatively large proportion of all existing books contain misprints. Example (259b) exhibits the same kind of ambiguity.

The degree quantifiers veel and weinig are adjectival in nature: they can themselves be modified by degree modifiers like ergvery or tetoo and can be used as input for comparative and superlative formation, although the superlative form (het) minste gives rise to a marked result.

	veel	weinig
degree modification	erg/te veel boeken ‘very/too many books’	erg/te weinig boeken ‘very/too few books’
comparative formation	meer boeken ‘more books’	minder boeken ‘fewer books’
superlative formation	de meeste boeken ‘most books’	??de minste boeken ‘fewest books’

Given its adjectival properties, it is not surprising that quantificational veel can be found in the same position as attributive adjectives, i.e. in a position between the article dethe and the noun. The primeless examples in (261) further show that veel must be inflected with the attributive –e ending when the article is present, but that the inflection is optional when there is no phonetically realized article. The primed examples show that although inflected weinig is possible when the article is present, it leads to a degraded result without an overt article.

The alternation found in (261b) suggests that we are dealing with a structurally ambiguous example. A plausible analysis of this case would be that the presence of the attributive –e ending signals that the quantifier veel functions as an adjectival modifier of the head noun boeken, while its absence signals that we are dealing with a quantifier located in the specifier position of NumP. The representation in (262a) is comparable to attributive constructions such as (de) mooie boekenthe beautiful books; the representation in (262b) is comparable to run-of-the-mill quantificational constructions such as (*de) sommige/alle boekensome/all books; cf. Section 20.2.1, sub V, for a discussion of the silent noun number.

The judgments on the primed examples in (261) suggest that weinig cannot easily be used as an attributive modifier of the head noun, as in (263a), and strongly favors a quantificational use, as in (263b).

There are a number of potential problems with the proposed analysis. The first concerns N-ellipsis. Example (264a) shows that N-ellipsis is normally possible in the context of an attributive adjective, provided that the context provides sufficient information to identify the content of the empty noun; cf. Section A28.4. It is therefore surprising that N-ellipsis is excluded with the inflected quantifier vele in (264b). Although the reason for the unacceptability of (264b) is not clear, it should be noted that its superlative counterpart in (264c) is perfectly acceptable; this can be taken as support for the proposed analysis.-

A second potential problem concerns the quantitative er construction discussed in Section 20.3. This construction can arise with indefinite noun phrases when the nominal gap e follows a cardinal numeral or a quantifier, but produces an unacceptable result when the gap follows an attributive modifier of the empty noun; cf. (265).

The analysis in (262) would therefore predict that quantitative er can only occur with bare veel, but this is not fully borne out; the examples in (265) show that the bare and inflected forms are both possible, although the inflected form seems somewhat outdated, as evidenced by the fact that it appears less frequently on the internet and mainly in formal texts.

We leave these potentially problematic issues for future research, but we think that the analysis in (262) deserves serious consideration.

Like most existential quantifiers, the degree quantifiers veel and weinig are not normally used as independent arguments: example (274b) is acceptable, due to the presence of [NP e] licensed by quantitative er, but example (274c) with a truly independent quantifier is unacceptable.

Things are different, however, when we are dealing with non-count nouns. Since the quantitative er construction requires the empty NP to be plural, it is not really surprising that example (275b) is excluded. However, unlike (274c), (275c) is perfectly acceptable on a mass interpretation: veel/weinig can refer to a certain amount of liquid (e.g. wine).

Note that the judgments on the examples in (274) and (275) remain the same when veel and weinig are modified by degree modifiers like ergvery or tetoo or replaced by their comparative forms meermore and minderless. Replacement by the superlative forms (het) meestthe most and (het) minstthe least is excluded in these examples, but this is for the independent reason that definite noun phrases are not possible in quantitative/expletive er constructions.

We can capture the differences between the examples in (274) and (275) by means of the structural representations in (276), which are based on the general format of our earlier structural analysis of quantified nominal phrases of the type veel/weinig N. The (a)-representations assume that the degree quantifiers are modifiers of the silent nouns number and amount (cf. Kayne 2002), which head the complex quantifier in the specifier position of NumP. The feature values [+plural] and [Ø] of the Num-head further indicate that mass nouns differ from count nouns in that they are not specified for number (cf. Mattens 1970). The (b)-representations express that the empty NP in the quantitative er construction must be licensed by a Num-head specified for number, thus accounting for the acceptability contrast between (274b) and (275b). The (c)-examples assume the presence of the silent noun substance, which is incompatible with a Num-head specified for number.

The (c)-representations in (276) suggest that the independently used quantifiers can only alternate with noun phrases headed by a [-count] noun, but the fact that the sentences in (277b&c) can be used in the same context shows that this is not correct. The two examples differ in that the subject in (277b) is used to refer to an indeterminate number of books (say about 100), while the subject in (277c) is used to refer to an indeterminate amount of books (e.g. a large pile). This suggests that it is not so much the feature [±count] of the noun that is relevant, but the question of whether the noun heading the quantificational phrase in the specifier of NumP is number or amount. If so, the (c)-representations in (276) correctly predict that the difference in interpretation is reflected in the number agreement on the verb: the subject in (277b) triggers plural agreement on the verb while the subject in (277c) triggers singular agreement.

There is another difference between quantitative er constructions and independently used degree quantifiers: the empty NPs in quantitative er constructions are anaphoric in the sense that their interpretation depends on the preceding context; in (274b) the empty NP must be interpreted as “bikes” and in (277b) as “books’. The independently used degree quantifiers may or may not be anaphoric: veel/weinig may refer to a certain quantity of wine in (275c) or to a certain number of books in (277c), but in both cases the independently used degree quantifier may also refer to a certain number of different (possibly unidentified) things.

So far we have only considered examples with inanimate subjects. If we compare the examples in (277) with the count noun boekenbooks with the examples in (278) with the count noun kinderenchildren, we see that they behave in a similar way with regard to the quantitative er construction, but differently when it comes to the independent use of the degree quantifier: the independent use is impossible with inanimate NPs, but possible with NPs referring to people (as well as certain animals), provided that the quantifier is inflected with –en. Furthermore, the inflected quantifiers velen and weinigen in (278c) differ from the bare quantifier veel in (275c) and (277c) in that they trigger plural agreement on the finite verb. Note that some people consider weinigen in (277c) to be obsolete.

The two differences can be accounted for if we adopt our earlier assumption from Section 20.1.1.6, sub IIIC2, that plural forms of the quantifiers in (278c) can arise thanks to the presence of the silent noun persoon. If so, the subjects in the examples in (278) can be given the structural representations in (279).

The difference in acceptability between (279c) and (276c) implies that the silent noun persoon is different from the silent noun substance in that it is a count noun; it also brings with it that the degree quantifiers are inflected with –en, which is the result of merging it with the plural suffix of the silent noun persoon.

Note in passing that the analysis in (279) confirms the suggestion in the 2012 edition of this work that there is no direct relation between the independent use of the degree quantifiers in (280c) and their use as attributive modifiers of the noun phrases in (280a&b). However, this still leaves the acceptability of vele in (280b) as a problem, because attributive modifiers are generally not possible in quantitative er constructions; cf. *Er lopen er [vrolijke e] op straat (intended reading: “There are happy people walking in the street”).

The fact that the independently used quantifiers veel and weinig in the previous examples function as the subject of expletive er constructions shows that they can be used as weak noun phrases. However, they can also be used as strong noun phrases, as shown for veel in example (281a); for completeness’ sake, this example also shows that the superlative het meeste can be used as an independent argument (note that het is not an article in this case, but part of the superlative). The remaining examples in (281) show that such independently used quantifiers can be used in all regular argument positions accessible to inanimate noun phrases, i.e. as direct object or as complement of a preposition.

The same is true for the independently used quantifiers velen and weinigen. Example (282a) first shows that these quantifiers can be used not only as weak but also as strong noun phrases. The remaining examples show that these quantifiers can be used in all regular argument positions accessible to [+human] noun phrases, i.e. as (in)direct object or as complement of a preposition.

The examples in (283) and (284) show that veel/weinig can also be used as the predicate in a copular construction or as a measure phrase with verbs like kosten. In such cases, veel and weinig can also be replaced by their comparative or superlative form; the latter can optionally take an -e ending.

The degree modifiers expressing that the cardinality of the intersection satisfies a contextually determined norm behave more or less like uninflected veelmany and weinigfew. If genoeg, voldoende and zat trigger plural agreement on the finite verb, they must be accompanied by quantitative er; this implies that these examples show that genoeg, voldoende and zat cannot be used as independent arguments referring to a collection of entities of a contextually determined kind.

However, if these elements trigger singular agreement, quantitative er cannot be realized. Like veel/weinig in (275c), the quantifiers in (286c) can then be construed with a mass interpretation; they refer e.g. to a certain quantity of wine, or to a set of discrete entities of a disparate or unidentified kind.

This suggests that the examples in (285) can be analyzed along the lines of the silent noun number and those in (286) along the lines of the silent noun amount; we are dealing with the representations in (276) with genoeg/voldoende/zat instead of weinig/veel.