Four case studies
In order to account for the degree modifying uses of massa’s, duizend , een parti] en tig, we will trace and document their development from their purely lexical uses to the currently attested instances of modification. It will be argued that, despite the different origins of these elements (a plural size noun, a numeral, a singular size noun and a suffix, respectively) they all go through similar stages in their development from quantifier to degree modifier. Massa’s and een partij go through a similar shift from binominal construction to quantifying construction: lexical uses tick over into quantifying uses which in turn lead to subsequent degree modifying functions. The first part of this development, i.e. from lexical to quantifying is a well-documented process of grammaticalization, which, especially in the case of size nouns, has been attested in many a language (see Keizer 2001; Brems 2003, 2007a, 2007b, 2010; Denison 2005; De Smedt, Brems, and Davidse 2007; Traugott 2008a, 2008b; Langacker 2009: 6080; De Clerck and Colleman 2013, to name a few). In these cases, a semantic extension or delexicalization motivates changes in the distribution which can eventually lead to a complete syntactic reanalysis, involving rebracketing (reversal of head positions), functional shifts of the first noun in the binominal construction (in what follows: N1) into modifier, host-class expansion from concrete to abstract second nouns (in what follows: N2s), synchronic layering and cross-linguistic replication (cf. Traugott 2008a). Within this context Brems (2011) distinguishes two major functions in English of these non-lexical uses: a quantifier use (as in loads of people) and a valuing(-quantifying) use in which the referent is evaluated rather than quantified (as in a load of crap or a bunch of liars). In Dutch, too, similar uses and similar processes can be attested. Doetjes (1997: 99), for instance, observes a process in which the size noun, e.g. een hoop (‘a heap’), een berg (‘a mountain’), tonnen (‘tonnes’), een paar (‘a pair’), etc.) “turns from an expression indicating a specific amount only [...] into an expression which can also be used to indicate a non-specific quantity, which is either relatively big (‘a lot’) or small (‘a bit’)” (see also Joosten 2003; Joosten et al. 2007). In addition to purely hyperbolic quantifying uses, valuing quantifying uses are attested as well: non-lexical, diminutive uses of stelletje (originally ‘couple’) and zoo(i)tje (originally ‘stew’), for instance, are subject to “functional crystallization” (Brems 2007a: 215) and only function as valuing-quantifiers with a negative semantic prosody in binominal constructions, e.g. een stelletje amateurs (‘a bunch of amateurs’), een zootje flauwe moppen (‘a bunch of lame jokes’). Since all of our cases involve quantifiers, each of the sub-sections below will first of all briefly sketch this development from lexical to (valuing)-quantifying uses.
Most of the attention, however, will be devoted to the second stage in the development, i.e. the further development from quantifying to degree modifying uses. Actual frequencies and contexts of use (e.g. possible host class expansion from adjective to adverb and verb, or vice versa) of the attested degree modifying uses of massa’s, een partij, tig en duizend will be examined more closely and subjected to individual comparison. This general trend in which quantifiers develop into degree modifiers (a trend which can also be observed in colloquial English, e.g. heaps funny, loads better as shown in De Clerck and Brems in press) will be captured within a construction grammar framework. Following De Clerck and Brems (in press), who show that the degree of expansion of modifying uses is partially influenced by the degree of grammaticalization of quantifying uses (cf. piles vs. loads as degree modifiers), individual differences will be explained on the micro-constructional level resulting from differences in grammatical con- structionalization (see section 3 for a more elaborate discussion).