Language and World

A Defence of Linguistic Idealism

Richard Gaskin

• Publisher: Routledge
• Year: 2020
• eBook ISBN: 9781000167214
• Edition: 1
• Topic: Philosophy
• Subtopic: Linguistics

This book defends a version of linguistic idealism, the thesis that the world is a product of language. In the course of defending this radical thesis, Gaskin addresses a wide range of topics in contemporary metaphysics, philosophy of language, philosophical logic, and syntax theory.

Starting from the context and compositionality principles, and the idea of a systematic theory of meaning in the Tarski–Davidson tradition, Gaskin argues that the sentence is the primary unit of linguistic meaning, and that the main aspects of meaning, sense and reference, are themselves theoretical posits. Ontology, which is correlative with reference, emerges as language-driven. This linguistic idealism is combined with a realism that accepts the objectivity of science, and it is accordingly distinguished from empirical pragmatism. Gaskin contends that there is a basic metaphysical level at which everything is expressible in language; but the vindication of linguistic idealism is nuanced inasmuch as there is also a derived level, asymmetrically dependant on the basic level, at which reality can break free of language and reach into the realms of the unnameable and indescribable.

Language and World will be of interest to scholars and advanced students working in metaphysics, philosophy of language, and linguistics.

1
Context and Compositionality

1 The Context Principle

This is a philosophical inquiry into the relation between language and the world. Like all inquiry, it must start somewhere: I start—as a number of philosophers do, both classic and contemporary¹—with the existence of linguistic communication, the conveying and understanding of meaningful items of discourse. Meaningful sentences have meaningful subsentential components, and the meaningfulness of these components is governed, as I take it (an argument will be provided shortly), by what is generally known as the context principle. This principle has been formulated in a variety of ways, but in its most familiar, Fregean version it states that words mean (bedeuten) something only in the context of a sentence.² We might gloss this as saying that words are made for sentences; sentences are what give words their function and point. Like Frege, I shall focus here on the declarative sentence, making the assumption that other kinds of sentence can have their functions either reduced to or explained in terms of declaratives.³ Since Aristotle the declarative sentence has been defined to be a group of words that is true or false⁴—assuming that contextual parameters are fixed, we must add, and waiving a few complications (Aristotle himself may have wanted to except future contingents from the rule).⁵ I follow that definition here, modulo the proviso about the role played by context (but without making an exception of future contingents). The focus on declarative sentences is not just a matter of convenience: more importantly, the connection with truth and falsity, mediated by declaratives, is both essential to and distinctive of language.

I do not insist that a group of suitable words must actually be uttered before it can take a truth-value.⁶ Any group of words generated by the language and suitable to discharge the semantic function of a sentence counts as a sentence, whether it is uttered (inscribed, entertained) or not. I thus work with a semantic, not a syntactic, definition of the sentence: this is standard practice in modern linguistic theory, and indeed it is hard—actually, I think, impossible—to see how a syntactic, or purely structural, definition could have the requisite generality to ensure complete capture of the relevant entities.⁷ Sentences are syntactic entities, but the criterion for whether a given syntactic construction can serve as a sentence must be semantic—a matter of whether the construction can perform the communicative function of a sentence. By ‘words’, the components of sentences that are governed by the context principle, I shall understand morphemes in general, so including word roots and semantically significant affixes (such as ‘un-’ and ‘-able’).⁸

Frege mentions the context principle several times in The Foundations of Arithmetic and clearly regards it as significant, but he says very little about it either there or elsewhere. For more detail we must look back to Jeremy Bentham’s Essay on Language and forwards to Wittgenstein’s Tractatus and Philosophical Investigations. Bentham not only anticipated Frege’s version of the context principle; he also defended it by pointing to the absurdity of supposing that words can be anything other than an abstraction from sentences. In effect, for Bentham, sentences are the given in linguistic communication; words and their meanings are derived, that is, theoretically secondary. The mistake in the Aristotelian tradition, according to Bentham, was to treat terms (words) as conceptually prior to sentences. Once this fateful step has been taken, we are then naturally led to speculate that ‘finding these terms endowed, each of them, somehow or other, with a signification of its own, at a subsequent period some ingenious persons took them in hand, and formed them into propositions’⁹—that is, into (as we would say now) sentences. But the speculation is outrageous: words, after all, do not come neat; they come with an implicit syntax. They are, as used to be said, ‘parts of speech’: the English word ‘cat’ is a noun, the French word ‘naître’ a verb, the German word ‘wenn’ a conjunction, and so on. Words come essentially packaged with functions, and these are sentential functions. Bentham’s ‘ingenious persons’ would have had nothing to do; that is his point. The words they found would already be governed by an implied syntax; they would already be geared up to figure in sentences. Thus Bentham against the atomism of the tradition.¹⁰

The sentence enjoys conceptual priority over words because it is the smallest unit of discourse with which one can, as Wittgenstein said, ‘make a move in the language game’; above all, it is the smallest unit of discourse that can bear a truth-value.¹¹ No kind of word is exempt from the context principle: in particular, proper names are not exempt. This was a point on which Wittgenstein insisted in the opening sections of the Philosophical Investigations:

Only someone who already knows how to do something with it can significantly ask after a name. … Naming is so far not a move in the language-game—any more than putting a piece in its place on the board is a move in chess. We may say: nothing has so far been done, when a thing has been named. It has not even got a name except in the language-game. This was what Frege meant too, when he said that a word had meaning (Bedeutung) only as part of a sentence.¹²

Further, as Wittgenstein hints, the context principle applies not merely to words in respect of sentences but also to sentences in respect of languages (a point I shall revisit in §3).¹³

Note the structure of the dialectic that I have presented so far: the phenomenon of linguistic communication is taken as a given; the context principle, by contrast, I do not assume but have argued for, following Bentham and Wittgenstein, on the basis of the fact that a word comes together with grammatical properties as a matter of its essence. Note, too, the connection that has been drawn between sentences and truth-values: this connection, I have said, is what makes the declarative sentence fundamental to discourse. Since the advent of the Fregean context principle, sentences have enjoyed a privileged status in theorizing about language and world. Partly that is because sentences, as Michael Morris puts it, ‘simply strike us as units’ (2017, p. 10). But there is a reason for their so striking us; there is a reason for the theoretical centrality of the (declarative) sentence in metaphysical investigations. That reason is given by the fundamental connection between sentences and truth-values.

I have emphasized, following Wittgenstein, that names are fully as subject to the context principle as other kinds of word. This point was denied by Peter Geach:

An act of naming is of course not an assertion; it may be correct or incorrect, but not, strictly speaking, true or false; it does, however, ‘express a complete thought’, as grammarians say about sentences—it has a sort of independent sense. Naturally the sense of a name used this way is not independent of the language, or of the situation that makes such use appropriate; but it is independent of any verbal context expressed or understood—it is not like the sense of a fragmentary expression that answers a spoken or unspoken question.

(1950, p. 462)

But how can an act of naming be correct or incorrect, if it is independent of any verbal context, expressed or understood? What would make a bare utterance or inscription of ‘Socrates’ correct or incorrect? It seems that there is no possibility of answering this question unless we can supply a suitable sentential context, reconstructed from explicit or tacit verbal features of the occasion of utterance or inscription, into which the name ‘Socrates’, as uttered or inscribed on that occasion, slots. In fact Geach seems to concede this very point in the quoted passage when he allows that the sense of a name ‘is not independent … of the situation that makes [its] use appropriate’. Elsewhere, indeed, Geach agrees that some one-word utterances abbreviate sentences, but he claims that the following are independent uses of names: ‘nouns in the vocative case used as greetings; labels stuck on things, e.g. “poison” on a bottle or the name-labels sometimes worn at conferences; ejaculations like “Wolf!” and “Fire!”’ (1962, p. 26). But it is surely obvious that these uses also abbreviate sentences—indeed abbreviate declarative sentences. Is there always a declarative sentence lurking behind such one-word utterances? Ian Rumfitt suggests not: ‘when Rodolfo exclaims “Mimì” just before the final curtain of La Bohème, he performs the complete speech act of invoking the late seamstress, even though he expresses nothing capable of assessment as true or as false’ (1995, p. 857). But, as I have said, non-declaratives can be reduced to, or explained in terms of, declaratives. In many cases this reduction goes via performatives, such as ‘I promise’, ‘I banish’, ‘I baptize’, or, as here, ‘I invoke’. Performative statements are special inasmuch as their truth is guaranteed by their being uttered (in the right context, by an authorized person, etc.); regular declaratives, by contrast, require co-operation from the extra-linguistic world for their truth. But that subtlety does not spoil the point, which is that Rodolfo does (perhaps indirectly) express something capable of assessment as true or false.

The conceptual priority of the sentence holds notwithstanding the likelihood that words came first in the historical development of language. For, even assuming the temporal priority of single words, those items of ‘protolanguage’ must have had an implicit syntax.¹⁴ This would be so even when such words were uttered in isolation;¹⁵ a fortiori when they were uttered in combinations. Wolfram Hinzen has argued that complex noun phrases could have evolved independently and in the absence of sentences, and even that such independent noun phrases could have been employed in assertoric speech acts.¹⁶ I agree that that could have happened—perhaps did happen. But the possibility that language evolved in that way does not show, as Hinzen supposes, that sentence structure and a correlative notion of truth-value are not essential to assertion, or that it is wrong to accord declarative sentences a fundamental status in semantics: for, to put it compendiously and to anticipate somewhat, the contents of assertions, even if they are effected by the use of isolated noun phrases, are true or false propositions, propositions being constitutively the referents of declarative sentences. (I shall examine the purport of these terms of art in more detail in Chapters 3–5.) In fact, as I and others have argued, the meanings of complex noun phrases are propositionally structured.¹⁷ So the apparent syntactic independence of the complex noun phrase from the declarative sentence is superficial: semantically, the former depends on the latter.

James Hurford has suggested, in the manner of Geach, that interjections such as ‘Ouch!’ and ‘Damn!’ have no descriptive content, do not abbreviate sentences, and are mere relics of animal cries.¹⁸ But the plausible conjecture that such cries antedated the utterance of complete sentences in the evolution of language does not entail that we should exempt them from the context principle. On any occasion when such a cry is emitted, there is no difficulty in assigning to it descriptive content—content which is given by a complete sentence. Of course, it is often the case that those who emit such cries cannot themselves articulate that sentential content. So, for example, human infants and non-human animals cannot articulate that content; and we are here presuming that our ancestors could not do so either, at least if we go back to a sufficiently remote stage of human development, when (as we are hypothesizing) only isolated words of protolanguage were uttered. But these facts, assuming that they are facts, do not derogate from the conceptual priority of the sentence. The meanings of one-word utterances are and always have been conceptually dependent on the meanings of whole sentences: these sentences we can express, even if other beings, who—as we are forced to allow, at least in the most favourable cases—entertain those contents, cannot.¹⁹ Compare one-word commands: these stand alone, and in that sense are ‘syntactically unintegrated’, as Hurford puts it (2012, p. 221), just like one-word interjections; but one would not call them syntactically unintegrated in any deeper sense than that.²⁰ Such imperatives presuppose a sentential syntax. You might as well say that any stretch of stand-alone discourse was syntactically unintegrated; you could say that the whole of uttered human language, from beginning to end, was syntactically unintegrated. If you make the unit of discourse large enough, you will find that it is unintegrated. But it is no strike against the context principle to observe that, if you widen the domain far enou...

Table of contents

• Cover
• Half Title
• Series Page
• Title Page
• Copyright Page
• Contents
• Preface
• 1 Context and Compositionality
• 2 The Theoreticity of Meaning
• 3 Reference and Ontology
• 4 Reference and Sense
• 5 Propositions
• 6 Truth, Falsity, and the World
• 7 Realism, Pragmatism, and Linguistic Idealism
• 8 Linguistic Idealism: Problems and Solutions
• Bibliography
• Index