Collins (and Elbourne) on free pragmatic processes

Abstract

The debate between literalism and contextualism bears on the (in-)existence of ‘free' pragmatic processes, i.e. pragmatic processes of interpretation which contribute to shaping intuitive truth-conditional content without being mandated by anything in the sentence itself. In his new book John Collins defends the contextualist position. He focusses on so-called ‘unarticulated constituents' (e.g. the unmentioned location of rain in a statement like ‘It is raining’) and argues against the idea that the existence of certain bound readings for the implicit component entails the presence of a covert variable in the syntax to which the alleged unarticulated constituent is assigned as semantic value. In the absence of properly syntactic evidence, Collins says, 'judgments about semantic readings are invariably amenable to a pragmatic explanation'. In a recent paper, however, Paul Elbourne offers the sort of evidence which Collins says would be needed to support the literalist position. On Collins' behalf, I discuss, and rebut, the evidence offered by Elbourne. In the last section, I argue against Collins that optionality is a better criterion than unarticulatedness for establishing that a given meaning component results from a free pragmatic process.

KEYWORDS:

• Contextualism
• binding argument
• unarticulated constituents

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 Linguistic Pragmatism and Weather Reporting, Oxford University Press, 2020.

2 See Recanati (2004). Collins speaks of ‘pragmatism’ rather than contextualism but the position spoken of is the same.

3 There is a minor debate regarding the extent of the phenomenon; the so-called ‘minimalists’ hold that there are significantly less context-sensitive expressions than people tend to think. See e.g. Cappelen and Lepore (2005).

4 In Literal Meaning I give the following example to illustrate the distinction between what is said (the statement made) and what is meant: If, being asked whether I can cook, I answer ‘I am French’, what I say is that I am French, and what I mean is, presumably, that I can cook (Recanati 2004, 5).

5 Collins cites Stephen Neale with approval : ‘We cannot dogmatically assume that there must be isomorphism [between content (what is said) and syntax], and we should recoil from the unargued goal of attaining isomorphism by freely adding aphonics to LFs as if adorning some garish Christmas tree with a new light wherever it seems too dark’ (Neale 2005, 187).

6 In the plain ‘it is raining’, the variable (for a location) remains free and must be assigned a particular value in context. In ‘everywhere I go, it rains’, the variable is bound by the quantifier: for every place l such that I go to l it rains in l.

7 Even though he endorses the over-generation objection, Collins seems to think that the verb ‘eat’ in this example might be construed as an ‘object-deleted transitive’ (Collins 2020, 143).

8 Thus ‘wherever I go, it rains’ has a reading in which the place of rain does not depend upon the destination of my travels: for example, wherever I go, it rains [here]. Unfortunately, Collins does not explicitly address the issue, whether this consideration also applies to everywhere I go, it rains.

9 To put it more strictly, this is the only bound reading Elbourne takes to be available. He mentions other possible readings, where no binding occurs: the location of the rain that is seen is fixed and does not covary with the donkey owners. See Elbourne (2022, 65, footnote 9).

10 Punkt-readings of weather sentences are readings in which no location is tacitly mentioned.

11 As Collins points out, ‘(ia) is not ambiguous in the way (ib) is:

(i) a With binoculars, Mary saw Bill

 b Mary saw Bill with binoculars

In (ia), with binoculars can only modify the VP see Bill, and so is unambiguous. In (ib), with binoculars can modify either Bill or see Bill. Just so, in most of Bill’s classes cannot modify three Frenchmen alone’ (Collins 2020, 148, footnote 20).

12 The resulting sentence is: In most of Bill’s classes, he telephones three Frenchmen.

13 Collins imagines a set-up in which the Frenchmen are not located within the classes where the failings occur: ‘imagine (…) that Bill teaches English to Frenchmen as a home tutor, but marks their work during his day job as a maths teacher at a high school. In this scenario the Frenchmen are not in any of Bill’s classes’ (Collins 2020, 147).

14 See what Collins says of objectless uses of ‘kiss’ on p. 66: if such uses were acceptable – as objectless uses of ‘notice’ undoubtedly are – they would involve a covert or elided object, he claims.

15 Recanati (2007 and 2010, chapter 3).

16 On the distinction between argument roles and argument places, see Perry (2001, 47), cited in Collins (2020, 85, footnote 3).

17 The parameter is lexically foregrounded in the case of simple expressions like ‘enemy’ or ‘ready’, but for complex phrases like noun-noun compounds it is the meaning of the construction, not the meaning of any particular lexical item, that introduces the free variable in need of contextual instantiation. In a nutshell: Linguistic foregrounding that is not syntactic is semantic foregrounding, which may or may not be lexical.