Frank Ramsey and the Realistic Spirit

Steven Methven

• Publisher: Palgrave Macmillan
• Year: 2015
• Print ISBN: 9781137351074
• eBook ISBN: 9781137351081
• Topic: Philosophy
• Subtopic: History & Philosophy of Mathematics

This book attempts to explicate and expand upon Frank Ramsey's notion of the realistic spirit. In so doing, it provides a systematic reading of his work, and demonstrates the extent of Ramsey's genius as evinced by both his responses to the Tractatus Logico-Philosophicus, and the impact he had on Wittgenstein's later philosophical insights.

Part I

The Realistic Spirit

1

The Realistic Spirit

The aim of this chapter is to begin the task of characterising the realistic spirit. This is not a trivial undertaking; Ramsey uses the expression only once, in ‘General Propositions and Causality’ (GP&C), one of the last papers that he wrote. Nonetheless, we may find thoughts expressed in other late papers which will assist in the elucidation of the view. Note first that when a philosopher nowadays uses the word ‘realistic’, she can expect her audience to make a series of wholly predictable associations. The expression is nowadays connected with the general thesis of realism, an umbrella term under which any number of local theories shelter. When Ramsey uses the term ‘realistic’, he does so in explicit opposition to many (though not necessarily all) theses that might be labelled realist. That is how I shall use the word also.

Variable hypotheticals have formal analogies to other propositions which make us take them sometimes as facts about universals, sometimes as infinite conjunctions. The analogies are misleading, difficult though they are to escape, and emotionally satisfactory as they prove to different types of mind. But these forms of ‘realism’ must be rejected by the realistic spirit. (GP&C, p. 252)

What constitutes the content of either the general thesis of realism or of particular local realist theories is no easy matter to articulate, though realism about subject matter X is traditionally thought to be a metaphysical thesis which is committed to the existence of the entities implicated by X and to the independence of those entities from minds, linguistic practices, epistemic limitations and so forth.¹ Anti-realism about X is then the denial of either or both of these commitments for X.

It is not, however, immediately clear what Ramsey means by the term ‘realism’. Why, for instance, does he take both a theory of variable hypotheticals that implicates facts about universals and one that involves infinite conjunctions to be forms of realism? While the former view may be described as involving positive claims about the existence and mind-independence of universals, thus satisfying the traditional description, the latter view is not so readily amenable to such characterisation. What Ramsey finds problematic in the second view is its apparent commitment to an ‘infinite collection’, such a notion being ‘really nonsense’ (loc. cit.). Ramsey is not denying the coherence of there being, for instance, infinitely many numbers, propositions or objects (‘there may be an infinite totality’) but rather the idea that there could be such a thing as a completed yet infinite process, such as collecting together infinitely many numbers, propositions or objects (this topic is picked up again in Chapter 8 and Chapter 9). The ‘realism’ he detects in the infinite conjunction account is that it is a theory which seeks to explicate our everyday use of variable hypotheticals in terms of a process – the conjoining together of infinitely many conjuncts – of which we can make no real sense.

It is not, thus, realism tout court that the realistic spirit rejects, but only ‘forms of “realism”’ such as the two mentioned here. What these forms of realism have in common is that they are dependent for their putative explanatory worth – in their accounting for our ordinary practices of saying, inferring and doing this or that – upon entities (such as relations between universals), activities (such as completed infinite processes) or, as I shall argue in Chapter 2, perspectives which, upon closer inspection, turn out to be such that we ‘can understand nothing’ of them (loc. cit.).

The range and extent of the realisms to which Ramsey might have objected is not readily determinable. Rather than attempting to provide such a determination, I shall instead, in the course of this chapter, try to spell out what I take to be certain methodological commitments expressed in GP&C, as well as others of the Late Papers, that may be taken to characterise the realistic spirit. In the next sections, I shall argue that while Ramsey’s account of variable hypotheticals may appear to reveal a connection between philosophising realistically and a sui generis commitment to ontological parsimony, that is not the correct way to understand him. After all, for Ramsey, a philosopher who embraces the realistic spirit must come to see various forms of realism not as false, but as incomprehensible. For this reason, I shall argue that his commitment to parsimony, to whatever degree it ultimately amounts, stems rather from a particular conception of philosophy on which its aim is usefulness. What is meant by ‘useful’ is made clearer by looking at three contrasts that Ramsey makes in ‘Philosophy’ (Ph.) and in his criticism of the Tractarian conception of logical inference. It is the aspiration that philosophy be useful which puts it at odds, for Ramsey, with a number of manifestations of realism.

1.1 Realism and the realistic

In GP&C, Ramsey aims to account for a species of generalisation that he calls a variable hypothetical. These are universal generalisations for which no domain restriction, implicit or explicit, is present. In particular, Ramsey is interested in law-like generalisations, such as ‘All men are mortal’ and ‘Arsenic is poisonous’, which are of this kind, but which also appear to posit a connection between the properties in question that goes beyond the merely accidental. Compare these examples with the kind of universal generalisation that Ramsey thinks can be accounted for by a conjunction of particular propositions, such as ‘Everyone in Cambridge voted’.

In the latter case, once we have settled what is to count as Cambridge (i.e. delineated a border and specified a particular time), the generalisation will be equivalent to, but not synonymous with, the conjunction of a finite series of particular propositions: for example ‘Lorna voted and Rob voted and ... ’. ² This is, more or less, the view of TLP, except that there Wittgenstein extends the treatment to all generalisations, independently of considerations of domain size.³ But, as we shall see, it is this thought – the extension of a treatment of a class of sentence from one case to a superficially similar but radically different case – that Ramsey holds to be incompatible with the realistic spirit.

Ramsey’s view is that statements of the former kind are not truth-functional; indeed, they are not truth-apt at all. Rather, they constitute ‘rules for judging: “If I meet a Φ, I shall regard it as a Ψ.”’ (GP&C, p. 241). That is, to assert that such-and-such is a law is to express that one has adopted a certain habit of judging. Apparent disagreement as to which law-like statements are true is disagreement about which rules of judgement that the parties endorse. As such, disagreement over which generalisation ought to be endorsed is never a disagreement about a matter of fact but instead disagreement about the ways in which the antagonists order their cognitive lives on the basis of their singular experience.⁴

Towards the end of the paper, Ramsey considers two alternative views that see variable hypotheticals as statements of fact: the Tractarian view, in which such generalisations are infinite conjunctions,⁵ and what I shall call the Platonic view, in which statements of law have as their truth-conditions the existence of relations between universals. Ramsey describes both of these views as meaningless, and not merely false.

But may there not be something which might be called real connections of universals? I cannot deny it, for I can understand nothing by such a phrase; what we call causal laws I find to be nothing of the sort. So too there may be an infinite totality, but what seem to be propositions about it are again variable hypotheticals, and ‘infinite collection’ is really nonsense. (GP&C, p. 252)

Thus, writes Ramsey,

both these forms of ‘realism’ must be rejected by the realistic spirit.

Now Ramsey has an explanation for why these views are sometimes seen to be satisfactory: variable hypotheticals beguile by their surface similarity to other propositions, and our desire for univocal treatments tempts us into views with maximal generality. But

the analogies are misleading, difficult though they are to escape, and emotionally satisfactory as they prove to different types of mind.

So here is a starting point from which we can begin to characterise the realistic spirit, namely that someone who does philosophy realistically will be constantly on guard against easy or beguiling pictures, analogies and metaphors.

Ramsey goes on to give an example that is supposed to explain why one should wish to take a realist (note: not realistic) view of some range of phenomena.⁶ We are asked to consider a society of humans who have never eaten strawberries, because they have always held that strawberries cause stomach-ache. Since, ex hypothesi, none has ever eaten a strawberry, nor ever will, each of their beliefs ‘If I eat a strawberry, I get ill’ is, taken as a material conditional, vacuously true. Clearly, that does not correctly characterise the content of their belief, and so an alternative account is required.

In attempting to provide that account, the realist temptation amongst us strawberry-eaters is to say that there is something wrong with what that society thinks because, or so we shall say, we know that if they had eaten a strawberry, they would not have been ill. It is, so we say, a fact that if they had eaten a strawberry, they would not have been ill.

But this is where we, and our invocation of facts, and not they, are wrong, claims Ramsey. What is a fact is that we have in the past eaten strawberries without becoming ill, but that no more entails that had they eaten them they would not have become ill than it entails that we shall not become ill if we eat them in the future. This is not an argument that culminates in scepticism about induction or counterfactual reasoning: Ramsey’s claim is not that our having enjoyed strawberries in the past is insufficient evidence for inferring that we will continue to enjoy them in the future. Our so inferring is not an epistemic failing, and it may, indeed, be just the thing that we ought to infer given the laws governing the consumption of strawberries that we have come to accept. Nor is it incorrect, given those laws, to infer that the strawberry abstainers wouldn’t have become ill had they eaten the strawberries. Rather, the failure arises from the involvement of a certain kind of pretence, namely that of holding that what didn’t happen but could have happened, or what will happen but hasn’t, is a fact which either succeeds or fails in rendering our statements true or false and which we could somehow come to know.

The question concerns the relation between counterfactuals and statements of law. If a statement of law such as ‘Strawberries do not make humans ill’ is considered to be a statement of fact, analysed in accordance with either of the realist theories that Ramsey rejects, then so too must be the counterfactuals which attend them. On either view, a statement such as ‘If they had eaten strawberries, they would not have been ill’ is furnished with truth-conditions according to the analysis of the law statement. On the Tractarian view, it would come down to its being the case that the conjunction which is the analysans of the law statement contains a conjunct that expresses the non-toxicity to humans of each strawberry in the domain.⁷ The counterfactual claim is then just a claim about strawberries, in particular, about the individual strawberries that humans in that society were ever in a position to eat, its truth-conditions being the conjunction of the relevant conjuncts occurring already in the conjunction which is the analysans of the law statement.

On the Platonic view, a law statement is true if there holds a relation between the universals implicated in the statement – the universal being a strawberry and the universal being non-toxic to humans. The truth-value of the counterfactual claim would then be inherited from the truth-value of the relevant claim about those entities and the nature of the relation that binds them. ‘If they had eaten the strawberries, they would not have been ill’ is true in virtue of the fact that there is a relation between the relevant universals, such a fact being responsible for the truth of the related law statement.

Ramsey, however, urges an alternative explanation by way of a new example.

If we regard the unfulfilled conditional [that is, the counterfactual regarding strawberries] as a fact, we should have to suppose that any such statement as ‘If he had shuffled the cards, he would have dealt himself an ace’ has a clear sense true or false, which is absurd. We only regard it as sense if it, or its contradictory, can be deduced from our system. Otherwise, we say ‘You can’t say what would have happened’, which sounds like a confession of ignorance, and is so indeed, because it means we can’t tell what will happen in a similar case, but not because ‘what would have happened’ is a reality of which we are ignorant. (GP&C, p. 253)

Ramsey is pointing out an interesting and prevalent feature of the conditions under which we are, in general, willing to assert a counterfactual. In the card case, Ramsey’s point is that we would regard the counterfactual as correct were it deducible from the system of variable hypotheticals by which we ‘meet the future’ (GP&C, p. 241). That is, if there were some statement of law upon which we agreed (or even disagreed) regarding the distribution of aces in a well-shuffled pack, which settled ‘our expectation, as to the outcome of any state of affairs whenever or wherever it may occur’ (GP&C, p. 247), then we would have something to say, either in agreement or disagreement, about how things would have turned out had he shuffled and about how things will turn out if he shuffles on the next round.

On Ramsey’s view, a counterfactual is (correctly) assertible just in case a future-looking conditional either is assertible under similar conditions or was assertible prior to the state of affairs that the counterfactual concerns. And both are assertible only if there is a related variable hypothetical to which the speaker would assent. That is, assent to a statement of law explains why one would be prepared to assert any of this range of conditional statements. These connections are plausible. Consider, for instance, a case where S says to her audience ‘If you shuffle that pack, the first card you draw will be an ace’. The audience is likely to ask her why she thinks what she thinks, to which the only adequate response is that she offer up a law-like statement which justifies the prediction. ‘I have a hunch’ or ‘Because I want it to’ will not succeed in furnishing her statement with a rational justification. Of course, there could be a disagreement about the status of the law statement, but on Ramsey’s view, that is not a disagreement about the truth of the statement but about whether or not one should adopt it as a rule of judgement, and it is therefore also a disagreement about which counter...

Table of contents

• Cover
• Title
• Introduction
• Part I  The Realistic Spirit
• Part II  Meaning
• Part III  Mathematics
• Part IV  Influence
• Notes
• References
• Index