Three’s a Charm – On Negational Modification as a Third Type of Quantity Modification of English all and French tout

In 2017, Njende et al. investigated the conceptual analogy between degree modification (i.e. quality modification, as in very hot or almost full) and quantity modification (i.e. the modification of quantifiers, as in very few or almost all), building on Paradis and her study of (un)boundedness (1997, 2000, 2001). In contrast to the vast body of research on quantifiers (e.g. Barwise & Cooper 1981; Gärdenfors 1987; Langacker 1991, 2008, 2016, 2017; Doetjes 1997; Benninger 1999; Radden & Dirven 2007) the literature on quantity modification has been quite scant up to then (e.g. Njende et al. 2017: 34).
In their paper, Njende et al. apply the categorization of degree modification into proportional and scalar modifiers to quantity modification. They explain that relative quantifiers (e.g. all, most, some), which compare the size of a predicated mass to that of a reference mass, take proportional modifiers, which comment on the coincidence (be it partial or total) between said predicated and reference masses (Njende et al. 2017: 55). Proportional modifiers in turn subdivide into two categories: approximating modifiers, as in almost all, and totality modifiers, as in absolutely all (Njende et al. 2017: 55).
It is argued here that, in addition to these two types of modification, a third type has to be recognized, i.e. negational quantity modification expressed by not in English and (ne) pas in French. We will explore the scope of negation as well as relevant syntactic tests, which proved crucial for the data analysis, especially for French.
Interestingly, we found that negational modification also co-occurs regularly with a semantic specification phenomenon (SSP). SSPs build on the addition of another element – typically a second quantifier, either preceding or following not all – to generate contrast in the sentence. This phenomenon presumably aims to clarify (even just partly) the intended meaning, as in (1) below.
(1) Not all atheists believe in evolution. Many do, but not all. (YCCQA_uk)
The English and French data used for this study were extracted from a lower-register written comparable corpus called Yahoo-based Contrastive Corpus of Questions and Answers (YCCQA) (De Smet 2009). The YCCQA corpus totals 29 million words, is available in English, French, German, and Spanish, and covers the time period 2006 to 2009 (De Smet 2009). Provided that sufficient data was available, datasets of 150 occurrences were compiled using AntConc (Anthony 2010).
The contribution made by this study is thus threefold: (i) it complements Njende et al.’s model with a third subtype of proportional quantity modification of the relative quantifiers all and tout; (ii) it discusses the many problems linked to the scope of negation, and suggests a test for the scope of negation with quantity modification; (iii) it investigates semantic specification and offers a description of SSPs with English all and French tout.