Colin McGinn

  • The Secret Connection: Causation, Realism and David Hume by Galen Strawson
    Oxford, 291 pp, £32.50, August 1989, ISBN 0 19 824853 9
  • J.L. Austin by G.J. Warnock
    Routledge, 165 pp, £30.00, August 1989, ISBN 0 415 02962 7

Philosophical reputations come and go – they surge and gutter – according largely to the prevailing intellectual climate, and are only tenuously tied to the actual merits of the views put forward by the reputand in question. To have a reputation is to have something perishable and fleeting, an imposition from without, no sooner bestowed than withdrawn.

Take the case of David Hume. In the dark days of logical (sic) positivism Hume’s reputation ran high as the philosopher who first did away with causal necessity; he was thought to have shown that causation consists in nothing, objectively, but constant conjunction: things happen in regular sequences but nothing makes them happen that way. In reality, the cement of the universe consists in nothing over and above the dependable concatenation of separable events. But when positivism quietly expired, and natural necessity regained its lost respectability, Hume’s standing correspondingly dipped. The neglected Locke began to seem like the philosopher with the better eye for metaphysical truth, while Hume started to look guilty of trying to deduce metaphysical conclusions from epistemological premises: ‘if no ideas then no reality’.

Now here comes Galen Strawson to argue that Hume has been grievously misrepresented all along: for the real David Hume never denied the objective reality of causal necessity. He firmly believed in it. And so Hume’s reputation is set to rise high again. He did not, after all, commit the mistake of letting the ideational contents of our minds determine what the world might really contain – though he did indeed think there was a problem about our achieving an adequate grasp of the nature of objective necessary causal relations. Hume, then, is a sceptical realist about causal necessity, contrary to the widely-received idealist interpretation; and sceptical realism is a view much favoured in this post-positivist era. The positivists were right in their high estimate of Hume, but for exactly the wrong reasons.

J.L. Austin was a philosopher with a legendary reputation. Although he published little, he is revered, especially in Oxford, for his critical acumen, withering good sense, originality, and talent for hitting the nail on the head. He was made White’s Professor in Oxford at the tender age of 40. His intellectual powers are said to have struck terror into the hearts of his contemporaries, to the point of deterring some of them from daring to put pen to paper, or mouth to thought. Indeed, it might fairly be said that Austin’s reputation depends largely upon his reputation: one tends to hear more about his philosophical reputation than about his philosophical ideas. It therefore comes as a bit of a shock to read Geoffrey Warnock’s study. The impression here conveyed is that Austin was almost pathologically incapable of getting anything right. Time and again Warnock has to correct obvious mistakes, apologise for unclarities, expose ground-floor misconceptions. It is all very puzzling. Even as Warnock attempts to celebrate his subject we see the man’s reputation sink wanly over the horizon. He may have initiated some fruitful lines of enquiry, later developed by others, but he himself seems to have been unable to pursue these lines with any surefootedness or perspicacity. You begin to understand why he wrote so little. Funny things, reputations. Steer clear of them if you can.

Attend now to a typical causal sequence – say, Mike Tyson’s fist colliding with his opponent’s jaw and the opponent dropping to the canvas. The blow, we say, caused the fall. Now we can distinguish three views about what this causal connection involves. One claims that there is no kind of necessity relating the events to each other: all that occurs in reality is that one event is succeeded by another. A second view insists that a species of necessity underlies the savagery of the nexus: the opponent had to fall, given that his jaw was subject to the force unleashed on it (and the circumstantial conditions were as they were). However, this second view concedes, we cannot know or perceive the nature of this binding necessity: we can assert that it exists but we can have no proper conception of what it ultimately involves. A third view agrees that causal relations carry objective necessitation, but this view is more sanguine about our capacity to understand such necessitation: science can tell us what the nexus depends on, if it is not already clear to common sense. These three views of causation and our access to it may be labelled anti-realist, sceptical realist and naive realist, respectively.

Strawson contends, against the common anti-realist interpretation, that Hume believes something like the second view. His main ground for attributing this view to Hume is that Hume repeatedly asserts the view, especially in the Enquiry. Thus: ‘experience only teaches us, how one event constantly follows another; without instructing us in the secret connection, which binds them together, and renders them inseparable’; ‘we are ignorant of those powers and forces, on which [the] regular course and succession of objects totally depends.’ Strawson adduces many such quotations, and disposes of rival interpretations of their purport: they are to be taken at face value, not as ironic or as occurring in suppressed oratio obliqua. He further contends that this agnostic position chimes better with Hume’s strictly non-committal scepticism about the world beyond our ideas: for such scepticism does not permit him actually to deny that there is necessity in nature. Similarly, Strawson argues, for the self and external objects: all we really know of them is contained in our ideas, which fall short of what we routinely take ourselves to know, and which fail to supply the basis for the kind of understanding claimed by certain rationalist philosophers of the period; but that does not imply that there is nothing more to these things than what is thus contained – quite the opposite. Causal necessity is something in which we do and may continue to believe: it is just that our ideas do not penetrate to its underlying real nature. What Hume objects to, on this interpretation, is not the objective existence of causal necessity: his objection is rather to the epistemological thesis, held by many philosophers of his day, that our minds furnish us with a full grasp of the nature of this necessity. We can reasonably assume that there is such a thing – Hume never doubts it – but we cannot arrive at an understanding of its inner reality.

And the reason we cannot embrace causal necessity in thought, for Hume, is that our ideas are derived from our impressions, and we have no impression from which we could read off the inner workings of objective causation. This thesis of Hume’s creates an initial problem for Strawson’s interpretation, to which he is acutely sensitive, since it is prima facie hard to see how Hume could consistently believe that something exists and yet deny that we can form any idea of it: how is it possible to formulate this existential thought if its components are not available to the thinking mind? Strawson registers the tension but argues that it can be relieved. The key is to distinguish merely referring to something from having a ‘positively contentful conception’ of it: Hume allows that we have a ‘relative idea’ of causation, which enables us to refer to it; what he denies is that we have any impression-based revelatory conception of the nature of that to which we refer. In this respect, his position mirrors that of Locke and Berkeley and Kant, who also had need of a category of concepts which by their own lights fall short of everything a proper hard-working concept should be: dummy concepts, as it were.

I find Strawson’s case for the sceptical realist interpretation thoroughly convincing. The textual evidence for it is well-nigh overwhelming; its consonance with other elements in Hume’s philosophy is striking; and the apparent clash with the theory of ideas is satisfactorily deflected. Hume emerges as a common-sense British Kantian. What is surprising is that a reader of the Enquiry could ever have run away with the anti-realist interpretation. (It should be noted that Strawson does not claim to be alone in interpreting Hume correctly. As he remarks in his preface, others are onto the same interpretation, notably John Wright in his The Sceptical Realism of David Hume.) But, as I observed above, a philosopher’s actual words are seldom sufficient to deter a reading that fits contemporary orthodoxy (c.f. Wittgenstein). I would make only two criticisms of Strawson’s otherwise admirable book. First, it is rather repetitive, as if the author feels that it is not enough simply to make his case once and well. I found that my level of credence had stabilised after a couple of restatements (or is it that I, like other philosophers, am easily persuaded that my intellectual heroes think the same things as I do?). Second, he does not appreciate a further tension in Hume’s overall position – namely, the tension between his tolerance of our ‘natural beliefs’ and his radical scepticism. It is really not consistent to grant us permission to believe what we naturally do believe and at the same time insist that we do not know any of the things we commonly take ourselves to know, since one cannot consistently continue to believe what one believes one cannot know. To believe is to hold oneself to know, so one cannot believe what one holds oneself not to know. One can, of course, combine belief in something with an admission that one does not know the nature of that thing, and this is clearly one part of Hume’s general thesis: but it is another matter to try to hang on to one’s beliefs while acknowledging scepticism with respect to what one claims to know. I have no right to believe in what I know I cannot know.

As to Hume himself, the obvious point of weakness, identified by Strawson, lies in his general theory of ideas. In effect, this theory takes perceptual confrontation as the model of what a good concept ought to be. Hume’s concept police discriminate against any putative citizen of the mind that cannot produce sensuous credentials. This theory is doubly mistaken. In the first place, it dogmatically banishes concepts that don’t enjoy a perceptual prototype, thus repudiating those of a more ‘intellective’ kind. Secondly, and more damagingly, the theory is wrong even about those concepts for which it was expressly designed – namely, sensuous concepts. As Berkeley noticed, and Wittgenstein rammed home, this picture of concept possession by immediate ostensive confrontation is multiply flawed: no concept can be generated by mere confrontation with what it is a concept of. In fact, all concepts are much more like the kinds of concept Hume officially found defective. From this perspective, then, the concept of causal necessity is as healthy as any concept we have. And so there is nothing in what Hume says to prevent us from going one step further than him and embracing naive realism about causality: there is causal necessity in the world and we can form an adequate conception of it. I therefore see no warrant for Strawson’s making the following pessimistic concession to Hume: ‘It seems that there will always be a sense in which the nature of even the simplest causal interaction is entirely unintelligible to us.’ Which sense is that, once we have rejected, as Strawson does, Hume’s restrictive and discriminatory theory of ideas? Some causal relations may well be unintelligible to us in principle, but why suppose that the unintelligibility is ubiquitous? Is snookerball causation really ‘entirely unintelligible’? Indeed, if every causal nexus is said to be unintelligible, then the point of declaring some to be so is blunted. One wonders what intelligibility would be if we could get it.

Geoffrey Warnock begins his study of Austin by remarking that ‘his reputation owed much to his certainly formidable personality,’ and that ‘the impression that he made as a philosopher upon those who knew him may be difficult to fully appreciate for those not included in that now diminishing number.’ Not being one of that number, I can only say that for me the difficulty is real. A certain jaunty contempt is never very far from the surface of his prose, a quality I can imagine intimidating some, but for the most part his arguments lack force and his doctrines are shallowly obscure. His studied casualness too often lapses into mere slapdashery. Warnock lists the defects Austin detected in the work of other philosophers: ‘carelessness; haste; a persistent tendency to invent and to rely on ill-defined and slippery technical terms; oversimplification; reckless and premature generalisation; and, perhaps above all, a predilection for ambitious either-or dichotomies’. I am sure that Warnock intended no irony here, but the rest of his book is almost a case-study in the diagnosis and correction of such faults in Austin himself. Was Austin peculiarly prone to these occupational hazards – and by Freudian projection tended to see them all around him? In any case, the following chapters consist largely of Warnock accusing Austin, evidently correctly, of precisely these failings. Did nobody dare venture these critical points at the time? Did Warnock himself not step in with the objections he now so effectively marshals? Did Austin listen? We are told that he advocated a co-operative approach in philosophy, in which patient criticism would lead to agreement and truth, but it is hard to believe that his own papers were the upshot of such collective efforts: there are just too many things wrong with them.

Take his suggestion that the wrongness of saying,‘I know it is so, but I may be wrong,’ is parallel to the wrongness of saying, ‘I promise I will, but I may fail’: that is, the suggestion that ‘I know’ is, or is akin to, a performative verb. Calling this suggestion ‘really unprofitable and misguided’, Warnock makes a number of simple objections to it. You can sincerely say, ‘I know,’ and not know, but you can’t do the same with promising. You can say, ‘He promised to do it, but he won’t,’ but you can’t say, ‘He knows it is so, but it isn’t.’ You can know something without saying, ‘I know,’ but you can’t do the same with promising. There is in general no conventional or ritualistic setting in which you say, ‘I know,’ unlike promising. You do not, as a rule, in saying, ‘I know,’ do anything beyond saying so, unlike promising. The explanation of the original datum is just that knowledge implies truth and has nothing specifically to do with speech acts and what they lead audiences to expect. Contrary to Austin’s thesis, ‘I know’, unlike ‘I promise’, is as descriptive as any first-person attribution. And, I would add, knowing is not an act at all, which precludes its being effected by the utterance of a performative verb. These objections are a. elementary and b. definitive. Ten minutes’ reflection should have made it clear that the assimilation is simply a mistake, prompted by the most superficial of similarities between the two verbs as they (sometimes) occur in the first person

Austin’s paper ‘Truth’ defines truth as follows: ‘A statement is said to be true when the historic state of affairs to which it is correlated by the demonstrative conventions (the one to which it “refers”) is of a type with which the sentence used in making it is correlated by the descriptive conventions.’ Warnock struggles to clarify what Austin might have meant by the two kinds of ‘convention’, but it remains unclear whether this is just a confused way of talking about indexicality in natural language, having little to do with truth in general. Certainly the account is hard to extend beyond simple indexical subject-predicate sentences: general statements, hypotheticals, mathematical statements and analytic truths cannot be forced into the Austinian mould. Isn’t this the very kind of over-generalisation on which he heaped scorn? In comparison with Tarski’s semantic theory of truth, available at the time he was writing his paper, Austin’s version of the correspondence theory looks at best quaint and at worst mired in obscurity and intractable difficulty.

The two chapters on action and ability find Austin frequently unclear, careless of important distinctions, and far too ready to dismiss defensible ideas for inadequate reasons. I mention two examples: his conflation of the question whether it is normally superfluous to append ‘intentionally’ after a verb of action with the question whether it is true to append that adverb; and his not noticing that you can have an ability which you do not successfully exercise every time you try to. Not very difficult points, really.

We turn then, hopefully, to the final long chapter ‘Words and Deeds’, which addresses itself to what is commonly regarded as Austin’s most important and enduring work. And indeed his treatment of the performative aspect of speech has been fertile enough, giving rise to what has come to be called ‘speech act theory’. The central idea to begin with is that uses of language are not exclusively ‘fact-stating’: some utterances also enable us to perform actions of various sorts – promising, betting, bequeathing, naming, acquitting and so forth. We do these things by uttering appropriate indicative sentences, but the sentences (Austin claimed) do not describe us as doing what we thereby do. (Why we cannot do something with language at the same time as describing ourselves as doing just that is never made clear.) So, it initially seems, Austin is directing us to distinguish the ‘constative’ use of language from the performative use: there are two kinds of speech act to consider.

However, as Warnock is at pains to point out, this alleged dichotomy subsequently evaporates into the insistence that all uses of language have a performative aspect. It turns out, on close examination of Austin’s text, that he has been roundly conflating at least three different definitions of ‘performative’, and their demonstrable inequivalence ends up pulling the notion in opposite directions, eventually causing its disintegration. There is the notion of a speech act uttered in a conventional setting, such as a marriage ceremony; there is the notion of a speech act which makes its own character explicit; and there is the notion of a speech act in which something is done, which threatens trivially to include every speech act. It is thus quite unclear what distinction Austin was endeavouring to capture with his original constative/performative dichotomy. Not surprisingly, therefore, he abandons in mid-stream the attempt to characterise the nature of the distinction and proceeds to analyse the structure of speech acts in general – distinguishing the locutionary, illocutionary and perlocutionary aspects of an utterance. Here again Warnock is obliged to correct exaggerations, inconsistencies, slips, confusions – but at least we are now engaged upon an adequately conceived project.

I have not yet mentioned Austin’s noted critique of Ayer in Sense and Sensibilia. This work is almost entirely negative in intention, consisting in generally convincing demonstrations that Ayer says many false and confused things about the verb ‘to see’. But what ought now to strike us is Austin’s own propensity, when engaged upon more constructive work, to fall into comparable traps. As he himself acerbically remarks, discussing Ayer, ‘there is nothing so plain boring as the constant repetition of assertions that are not true, and sometimes not even faintly sensible; if we can reduce this a bit, it will be all to the good.’ Boring, yes, and irritating too – though at least Ayer was trying to tackle hard and deep philosophical questions that resist ready formulations. It seems to me that Austin, while for the most part eschewing the traditional questions of philosophy, shows an equal proneness to falsehood and confusion, and with less excuse.

His personal charisma must have been powerful indeed, because he wrote little of lasting value. Perhaps his greatest legacy was his early translation of Frege’s Foundations of Arithmetic (from which he seems to have learned little). Warnock’s book has the merit of providing us with a sober and not unsympathetic dismantling of a reputation that has long seemed inflated. Scrupulously courteous as he is to Austin, I cannot help feeling that he is well aware of the perlocutionary effect his illocutionary acts are likely to have.