Necessity and Language

Morris Lazerowitz, Alice Ambrose

• Publisher: Routledge
• Year: 2016
• ISBN:
  9781315523910
• Edition: 1
• Topic: Philosophy
• Subtopic: Language in Philosophy

The problem of necessity remains one of the central issues in modern philosophy. The authors of this volume, originally published in 1985, developed a new approach to the problem, which focusses on the logical grammar of necessary propositions. This volume gathers their seminal essays on the problem of necessity, together with new material at the original time publication.

1 NECESSITY AND LANGUAGE

Wann man sich von der Wahrheit fürchtet (wie ich jetzt), so ahnt man nie die volle Wahrheit

(Wittgenstein)

Underlying much of Wittgenstein’s later thinking was the wish to reach a correct understanding of the nature of philosophical utterances, and this wish is also discernible in his Tractates.¹ His later investigations led him to some iconoclastic ideas about what a philosophical theory is and what a philosopher does who supports his theory with an argument. Wittgenstein saw more deeply into philosophy than anyone before him, but, for the most part, he seemed to prefer to express his perceptions in metaphorical language rather than in the language of straightforward reportage. Part of the reason for this may have been the wish to soften the hard things he saw. Remarks like ‘philosophical problems arise when language goes on holiday’² and philosophical language is ‘like an engine idling, not when it is doing work’³ give expression to disturbing perceptions into the nature of technical philosophy, but use a form of the mechanism of sotto voce to deflect them. Where their translations into prosaic language would tend to stir up anxiety, these words can be accepted as colourful jibes which need not be taken seriously.

It is important to notice that he stated in a number of places that philosophical propositions are not empirical. This insight into the nature of philosophical propositions (and into the modes of investigation employed in philosophy) made it fundamentally important for him to understand the logical difference between empirical statements and statements which have a priori necessity, and especially to understand the nature of necessity, to obtain, so to speak, an inside look at it. He appears to have emerged with a conventionalist view, which on the surface at least he seems never to have given up. This is the view, generally speaking, that necessary propositions are about the literal use of terminology in a language. A number of writers have described him as a conventionalist, and it must be allowed that there is considerable justification for this description. One of his frequently cited expressions, ‘rule of grammar’, which he used to characterise necessary propositions, unquestionably lends some substance to the claim that he took one of the traditional positions about logical necessitation. On one occasion G. E. Moore, who was puzzled by the term ‘rule of grammar’, remarked to me that he thought Wittgenstein meant by it what is meant by the more familiar term ‘necessary proposition’. My impression at the time was that Moore thought Wittgenstein was so using ‘rule of grammar’ that in his use of the term a rule of grammar was not verbal. Moore’s line of reasoning was perhaps the following: a rule of grammar in Wittgenstein’s sense is a necessary proposition, and since a necessary proposition says nothing about usage, a rule of grammar says nothing about usage. There can be no doubt, however, that Wittgenstein wished by his special use of the word ‘grammar’ to say that in some way necessary propositions are verbal.

I

Conventionalism is open to a number of obvious objections which Wittgenstein must have known. It is worth remarking that conventionalists who are aware of these objections are not moved to give up their position. This is mystifying and certainly calls for an examination, for if conventionalism is taken at face value as making a factual claim about the nature of necessary propositions, the objections are as conclusive as any objections could possibly be. One frequently repeated criticism is that to suppose a necessary proposition to be one which makes a declaration about verbal usage, or one which ‘records usage’, is to imply that a necessary proposition is not necessary. The negation of a true verbal proposition is a false verbal proposition, but not a proposition which could not, in principle, be true. Put roughly, the negation of a true verbal proposition is not a self-contradiction, and precisely the same kind of investigation which establishes the truth of a verbal proposition, such as recourse to dictionaries and the like, could theoretically establish its denial. To use an expression of Wittgenstein’s, we know what it would be like for a verbal proposition, which happens to be true, to be false, and for one which is false to be true. By contrast, we do not know what it would be like for a false arithmetical proposition to be true — for example, for 4 + 3 to be less than 7. Taken literally, the philosophical claim that necessary propositions are about usage is refuted with complete finality by the objection that the view that they are implies that they are not necessary.

Another well-known objection is that a necessary proposition does not say anything about terminology, because it says nothing about what language it is expressed in or about any words occurring in it. The two sentences ‘Red is a colour’ and ‘Rot ist eine Farbe’ have the same meaning, which would not be the case if the propositions expressed in the English sentence made a declaration about words occurring in the sentence, and the proposition expressed in the German sentence made a declaration about words occurring in it. Wittgenstein certainly was aware of these objections, and there is reason to think that his conventionalism, which undoubtedly was the usual philosophical article at first, was transformed by his growing insight into the way language works.

Verbal usage and logical necessity are in the same way bound up with each other, and it is not going too far to think that part of Wittgenstein’s investigation into language and necessity was directed to bringing to the surface in what way they are bound up. Thus, in more than one place Wittgenstein remarks that a philosopher rejects a notation under the illusion that he is upsetting a proposition about things.⁴ This observation shows a recognition of the difference between an explicitly verbal statement and its semantic counterpart formulated in a different idiom; an idiom which easily gives rise to the illusion that the statement is about things. The difference between a verbal proposition and a necessary proposition may only be a difference in the form of speech in which they are expressed. But the difference in the form of speech may be of great importance, and seeing this difference can lead to an understanding of the way in which language and necessity are connected. To put the matter more concretely, seeking the unlikeness, without losing sight of the likeness, between, say, the proposition that being an uncle entails being male and the proposition that being male is part of the meaning of the word ‘uncle’ can lead to a correct understanding of how the sentence which expresses the entailment proposition is related to the proposition it expresses.

Consider for a moment the following sentences:

(1) A camel is a herbivore.

(2) Ein Kamel ist ein Pflanzenfresser.

(3) A camel is an animal.

(4) Ein Kamel ist ein Tier.

(5) The word ‘animal’ applies, as a matter of usage, to whatever ‘camel’ applies to.

(6) The word ‘Tier’ applies, as a matter of usage, to whatever ‘Kamel’ applies to.

Having a clear notion of necessity requires seeing how (3) and (5) are related to each other, i.e. in what way ‘A camel is an animal’ is like The word “animal” applies as a matter of usage, to whatever “camel” applies to’; and in what way they are unlike. To see this it is necessary to see also what (1), ‘A camel is a herbivore’, has in common with (3), and in what way it is different from (3). Furthermore, it is important to see how the fact that (1) and (2) have the same meaning is both like and unlike the fact that (3) and (4) have the same meaning. A clear grasp of these features of likeness and unlikeness requires seeing how (3) is related to (5), and (4) to (6).

Getting a proper view of these matters will help us understand what it is about the philosophical view that necessary propositions are verbal, or that they state facts of usage, which makes it possible for a philosopher to hold it despite being aware of conclusive objections to it. Seeing what makes this possible will help dispel the mystery surrounding a long-standing dispute in which able philosophers with a well-preserved sense of reality can, to all appearances, debate the truth-value of a view which is known to be false without having strange ideas about each others’ psychology.

Some philosophers, for one reason or another, have denied that there is a difference between logically necessary and empirical propositions, a direct consequence of which is that there is no difference in kind between the propositions expressed by ‘ A camel is a herbivore’ and ‘A camel is an animal’. Without going into the reasons for the philosophical claim that there is no difference, it can be seen that the mode of verification relevant to the proposition expressed by the first sentence is different from the mode of verification relevant to the proposition expressed by the second: observation is relevant to the first out not to the second.⁵ Both sentences, equally, can be expressed as general statements of the form ‘All a’s are b’s’, i.e. as ‘All camels are herbivores’ and ‘All camels are animals’, which makes it tempting to think that both are generalisations. Their grammatical similarity appears to blind some philosophers to an important semantic difference between them. The proposition expressed by the first sentence, unlike the proposition expressed by the second, does not, to use Kant’s phrase, have ‘strict universality’. The first is an inductive generalisation from observed instances and could in principle be upset by future instances: no number of confirming cases, however large, removes the theoretical possibility of there being a camel that is not a herbivore. By contrast, the second proposition has strict universality, which is to say that, unlike the first, it does not carry with it the theoretical possibility of being upset by a counter-instance. This means that it is not an inductive generalisation.

C.I. Lewis has stated that a logically necessary proposition might, in addition to having an a priori demonstration, be established by ‘generalisation from observed instances’,⁶ that is, be established in the way in which a law of nature is established in science. Undoubtedly what Lewis was impressed by — and perhaps wished to highlight — is the similarity between the sentences expressing the two. But putting aside considerations of this sort, it will be clear that taken at face value his claim implies both that a logically necessary proposition of the form ‘ All a’s are b’s’ has an associated theoretical disconfirming instance and that it does not have one. The difference between ‘A camel is an animal’ and ‘A camel is a herbivore’ may be brought into focus by noting that the first can also be expressed as an entailment, ’being a camel entails being an animal’, and the second can not—being a camel does not entail being a herbivore. Nothing is more plain than that being a camel is logically consistent with being a herbivore and also with not being a herbivore, and that experience alone, not penetration into the meanings of the words ‘camel’ and ‘herbivore’, will show whether it is a herbivore or not.

To come back to the philosophical claim that a necessary proposition is verbal, it can easily be seen that even though it is expressible in the form ‘All necessary propositions are verbal’, it is not put forward as a generalisation which issues from an examination of instances. Instead, it is put forward as a statement to which there can, in principle, be no exception, or as one whose universality is ‘strict’. Construed in this way it can be restated as an entailment: being logically necessary entails being about the use of terminology. But looked on as making an entailment claim we are puzzled to understand the continued disagreement which revolves around it. There is no debate over whether being a camel entails being an animal; and if philosophical conventionalism did actually come down to a straightforward, elementary entailment claim, to the effect that being logically necessary entails being verbal, there is no question but that the debate over it would have been brought to an end long ago. But if the conventionalist thesis is not to be taken as an entailment statement, correct or incorrect, then truly a familiar view is turned into an enigma. We do not know what the conventionalist is asserting, nor do we know what we are disputing. There can, of course, be no doubt that in some way we do understand the view and the arguments for and against it; and the conclusion would thus seem to be that our understanding of the view, like our understanding of our dreams, is hidden from us. No one who lets himself become a curious observer of the philosophical scene can fail to entertain the idea that philosophy is an activity which takes place in one of the obscurer parts of the mind. Conscious understanding of the apparent entailment statement should help us understand the nature of philosophical views in general. For if, as Wittgenstein has declared, philosophical propositions are not empirical, then it is natural to suppose them to be a priori and to be making entailment claims. Again, as in the case of the conventionalist position, what needs to be seen is what makes possible the continued disagreement centring on them. For example, if the philosophical statement ‘A sense datum is private’ is an entailment statement, it is one which is turned into a mystery by the continuing disputation over it. The only hope of dispelling the mystery and arriving at an undistorted perception of the philosophical theory lies in clarifying how a logically necessary proposition is related to the sentence which expresses it; or, to put it generally, how logical necessity is related to language.

In the Tractatus Wittgenstein makes a number of remarks about tautologies which throw light not only on tautologies but also on all statements having logical necessity, whether analytic or synthetic a priori. Proposition 6.1 states that The propositions of logic are tautologies’, and 6.11 that Therefore the propositions of logic say nothing’. The view which comes out of these two propositions is that tautologies say nothing. This view has been linked with the idea that they say nothing about things, that is, with the notion that they have no ‘factual content’. Thus in 4.462 Wittgenstein states that tautologies are not ‘pictures of reality’, the implication being that they give no ontological information. The statement, ‘It is either raining or not raining’ says nothing about the weather; ‘A plant is either an oak or not an oak’ says nothing about what a plant is; ‘An object is either a plant or not a plant’ says nothing about what an object is. This can perhaps be made more perspicuous by considering the negations of these statements. The expressions, ‘not both an oak tree and not an oak tree’, ‘not both a plant and not a plant’, do not function as predicates which tell us what a plant or an object is not, unlike ‘not both a camel and not herbivorous’, which does function to deny what a creature is. To say with regard to anything that it is not both a camel and not a herbivore is to say what the thing is not, and this is because the predicate ‘both a camel and not a herbivore’ presents a possible ‘picture’ of the thing. But to say with regard to a plant that it is not both an oak tree and not an oak tree is not to say what the plant is not, in as much as ‘both an oak and not an oak’ does not have a use to describe any plant, actual or hypothetical. Tautologies say nothing about what there is and what things are like, and contradictions say nothing about what there is not and what things are not like. Predicates of the form ‘ϕ or not ϕ’ equally with ‘ϕ and not ϕ’ have no descriptive content.

These considerations apply to all analytic propositions, and to synthetic a priori propositions as well. Kant, and many philosophers after him, have held that synthetic a priori propositions, the predicates of which are connected by ‘inner necessity’ to their subjects, but are not components of them, have factual content; that is, they delineate features of the world. But it will be clear that a true proposition, and hence a logically necessary proposition of whatever kind, will tell us something about what there is only if its negation states something to be the case which in fact is not he case. Kant’s claim that the proposition that every change has a cause is a pr...

Table of contents

• Cover
• Half Title
• Title Page
• Copyright Page
• Original Title Page
• Original Copyright Page
• Table of Contents
• Acknowledgements
• Editor’s Introduction
• 1. Necessity and Language
• 2. Factual, Mathematical and Metaphysical Inventories
• 3. A Priori Truths and Empirical Confirmation
• 4. Assuming the Logically Impossible
• 5. Invention and Discovery
• 6. Mathematical Generality
• 7. The Infinite in Mathematics
• 8. The Metaphysical Concept of Space
• 9. The Passing of an Illusion
• 10. On Making a Philosophical Problem Disappear
• Publications
• Index of Names