Hume and Scepticism

Hume has had much to complain about in his readers, besides the (perhaps legendary) early lack of them. The first reviewers accused him of obscurity, egotism and ‘evil intentions’. Warburton charged him with framing a system of human nature on the principle of ‘necessity, in opposition to liberty and freedom’. A French reviewer complained: ‘Never has there been a Pyrrhonian more dogmatic.’ And as if Warburton’s ill-natured and uncomprehending critique were not enough, the London Review (no relation) some years later carried the no doubt fabricated story of Hume, in a fit of ‘violent rage’, demanding satisfaction at sword-point from the terrified printer of Warburton’s piece.

More recently the range of Hume interpretation, and the heat of dispute, have hardly diminished. The massive misinterpretation that must exist (given the massive disagreement among interpreters) is partly the result of Hume’s style – where beautifully clear sentences combine to make puzzling paragraphs and even more baffling books – and partly the result of a philosophy that seems to turn back on itself – containing a scepticism and a theory of belief each of which seems to put the other in question. One can easily get a sense of vertigo faced with these epistemological puzzles, which Quine has more recently made vivid without as yet inspiring much agreement on how to resolve them. Matters are not helped by the fact that on the one occasion when Hume replied in public to criticisms of his work, he had his eye on evading accusations jeopardising his candidacy for the Chair of Pneumatical Philosophy at Edinburgh. His ‘Letter from a Gentleman’ contains a response to the charge of denying the immortality of the soul, for example, that sounds less than totally ingenuous.

Robert Fogelin, who has written extensively on Wittgenstein and more recently on Hume, is mainly concerned in his latest book to correct an imbalance in Hume interpretation.* In the earlier part of this century a common view was that Hume’s main achievement was to work out the extreme sceptical consequences of British empiricism. Hume had shown that, on the empiricist principles of Locke and Berkeley, belief in the external world, in induction, and even in the self, was ungrounded and unreasonable. The conclusion was that our knowledge could reach no further than our present and perhaps past mental states. Reason had demonstrated the falsity of some of our deepest, even instinctive, beliefs. A rival interpretation subsequently gained ground, however, which insisted instead on the power of instinct over reason. Hume’s supposed sceptical conclusions were ones that human nature made it impossible for anyone, Hume included, to accept. Human nature was too strong for us to abandon belief that the sun would rise tomorrow, or that independent material objects exist. Far from being essentially destructive, the main aim of the Treatise was a constructive one: to develop a Science of Man (what we would call Psychology), and ‘to introduce the experimental method of reasoning into Moral Subjects’. The keyword of this interpretation, principally developed by Norman Kemp Smith in his magisterial book of 1941, was ‘naturalism’ – which covered the related views, that mental processes, including belief, can be the subject of empirical investigation, that one must or should acquiesce in one’s ‘natural’ beliefs, that empirical beliefs cannot be assessed rationally, and that reason is and ought to be the slave of instinct. From the point of view of the naturalistic interpretation, the scepticism attributed to Hume by the first interpretation is an unattainable state. If, on the first interpretation, reason proved the poverty of instinct and natural belief, on the second, reason was portrayed as their slave.

Fogelin’s main point is that the sceptical aspect of Hume’s philosophy is seriously undervalued by his naturalistic interpreters. Book I of the Treatise contains sceptical arguments concerning, among other things, the understanding (in Part III), reason (in Part IV s.1) and the senses (Part IV s.2). Hume’s scepticism about induction and about the senses is ‘wholly unmitigated within [its] domain’; his scepticism about reason is even more wide-ranging, having a ‘wholly unrestricted’ range of application. Hume’s scepticism is typically theoretical rather than prescriptive: he challenges the supposed grounds for a system of beliefs, holding ‘that we have no rational grounds’ for them, but he does not in general ‘call for a suspension of belief’ or even ‘for more caution in giving assent’. And according to Fogelin, this unmitigated ‘theoretical epistemological skepticism’ and Hume’s ‘naturalism’ (the attribution of which he accepts from the interpretations which he is rejecting) are ‘mutually supportive’.

Scepticism has the star part in Fogelin’s book, which discusses in turn all the main topics of Book I, and some key issues in the rest of the Treatise. When sceptical arguments are on stage, their role is to press an ‘epistemological skepticism that is wholly unmitigated’ – a recurrent phrase. Even with non-sceptical issues, such as belief, and the nature of our ideas, Fogelin has scepticism working out of sight, setting the stage for Hume’s views and clearing away any rivals.

There are several reasons to worry about Fogelin’s picture of Hume. First, it is doubtful whether Hume’s arguments are as sceptical as Fogelin thinks. I shall concentrate on one central example, Hume’s argument about induction. ‘If reason determin’d us,’ Hume says, then all causal reasoning would proceed on the principle that the unobserved is like the observed (e.g. that the future will be like the past). This principle of the uniformity of nature can be proved, however, neither by deduction a priori, since its negation is conceivable (the future could conceivably be quite different from the past), nor by empirical argument or causal reasoning, since that would be arguing in a circle (given that all causal reasoning itself presupposes the principle at issue).

On Fogelin’s interpretation, Hume’s conclusion is that ‘every inference from what is present to what is not present is on a par in having no rational foundation.’ This suggests that – as far as Hume the philosopher is concerned – our belief that the sun will rise is on a par with the lunatic’s belief that the next time he tries he’11 be able to fly. All beliefs about the unobserved would be equally unreasonable – those of common sense no more, and no less, than those of the lunatic. Indeed Hume the naturalist would explain the former as more natural (and perhaps inescapable) than the latter – but that would be just a matter of instinct and the nature of the human mind.

I am sure that this misrepresents Hume. Hume’s argument was that if reason determined us, it would proceed upon a uniformity principle for which it was unable to find a foundation. The conclusion may simply be: reason does not determine us. And here ‘reason’ means, I think, a faculty of conscious deductive reasoning, or – perhaps more precisely and less restrictedly – a faculty of understanding, grasping, interpreting and calculating. Hume is actually running together two views, the first more psychological, the second more philosophical. First, our belief that the sun will rise tomorrow has not, in fact, been produced by a process of conscious deductive reasoning: hence Hume’s references to the fact that ‘the most ignorant and stupid peasants – nay infants, nay even brute beasts –’ learn by experience. Secondly, that belief could not be reached by deductive reasoning from independently establishable premises. Neither of these views, however, implies that the belief that the sun will rise tomorrow is illegitimate or unreasonable – unless we add the highly dubious premise that every belief which has no independent deductive foundation is unreasonable. Hume’s argument neither concludes nor implies that all inductive beliefs are really on a par in being equally unreasonable.

Is there no inductive scepticism in Hume then? Hume would certainly have seemed sceptical to contemporary supporters of a priori or rational science, in rejecting any intelligible connection between cause and effect. In addition, there are passages where sceptical doubts come to the surface. The ‘secret nature [of bodies], and consequently all their effects and influence, may change, without any change in their sensible qualities. This happens sometimes, and with regard to some objects: Why may it not happen always, and with regard to all objects?’ Hume here and elsewhere undoubtedly sees the possibility of pressing an inductive scepticism. One might use the word ‘sceptic’ for anyone who so much as raises such doubts without being sure he can quell them – and in this sense Hume is certainly a sceptic. But it is more common to restrict the term to someone who rates all beliefs of a certain type as equally (or perhaps unequally, but all hopelessly) unreasonable. Nothing in Hume’s argument, I have suggested, commits Hume to being an inductive sceptic in this stronger sense, even while wearing his philosopher’s (as opposed to psychologist’s) hat – and Fogelin’s ascription of a ‘wholly unmitigated skepticism’ must be wrong.

My first objection – that Fogelin exaggerates the sceptical content of Hume’s arguments about induction – runs against prevailing views of Hume. Both sceptical and naturalistic interpreters have usually agreed that Hume’s arguments have the sceptical conclusion that induction is unreasonable, though the naturalistic interpreters add that Hume cannot have finally espoused that conclusion. So though it is bad that Fogelin fails even to consider that this agreed view may be wrong, it is perhaps not surprising.

My second objection is to a more culpable shortcoming: Fogelin fails to take seriously the important non-sceptical passages in Hume. There are plenty of occasions where Hume distinguishes what might be called good and bad causal beliefs. There is the argument that belief in miracles is unfounded – that is, counts as bad causal reasoning. There is the chapter on ‘unphilosophical probability’ – on beliefs which are said to lack ‘reasonable foundations’ – and then the chapter on the ‘rules, by which we ought to regulate our judgment concerning causes and effects’ (my italics). There is the crucial passage where Hume says that a person who hears a voice in the dark and forms the belief that someone is there ‘reasons justly and naturally; tho’ that conclusion be deriv’d from nothing but custom’. By contrast, people frightened by ghosts ‘reason naturally’ only in ‘the same sense that a malady is said to be natural; as arising from natural causes, tho’ it be contrary to health’. This would be strange talk from someone committed to the view that all causal beliefs are equally unreasonable.

Perhaps these passages are inconsistencies, but they are not incidental inconsistencies: they recur frequently and they are backed up with a kind of theory for distinguishing better and worse causal inferences. The former arise from principles ‘in the imagination’ which are ‘permanent, irresistable, and universal’; the latter from principles which are ‘changeable, weak, and irregular’. General rules are distinguished as being ‘capricious and uncertain’ or else ‘more extensive and constant’. ‘The vulgar are commonly guided by the first, and wise men by the second.’

Hume sometimes pretends that these distinctions simply describe variations in psychological activity, without evaluating them. But his practice – of criticising superstition and ill-founded belief – clearly allies him with one side of the distinction. Hume is not just saying that there are two kinds of causal inferences, which are commonly labelled ‘good’ and ‘bad’, though really each is as good as the other from the philosophical point of view: he clearly thinks ‘bad’ patterns of reasoning are actually bad.

Fogelin does consider some of these passages – but he apparently never sees how they challenge his attribution of unmitigated epistemological scepticism to Hume. Perhaps he thinks they can all be put down to ‘naturalism’ – the side of Hume already overemphasised by commentators. Maybe Hume is simply indulging his natural instincts – speaking as a man, rather than a philosopher – when he criticises superstitions and advances rules ‘to distinguish the accidental circumstances from the efficacious causes’. But this would surely be a desperate suggestion – excluding from Hume’s philosophy some of its most characteristic features.

This is closely related to my third worry about Fogelin’s book: that it invokes ‘naturalism’ far too readily. This slippery term covers at least three views, the second of which is only in part attributable to Hume, while the first does nothing to help Fogelin’s sceptical interpretation, and the third actually works against it. On the first use, ‘naturalism’ simply covers the ‘empirical inquiry into the causes of mental phenomena’ – an innocent enough enterprise which, despite Fogelin’s Chapter 11, seems to me neither to support ‘an unmitigated epistemological skepticism’ nor to be supported by it. (Has Fogelin ever tried to persuade a psychologist of the sceptical tendencies in his work?) At other times the term covers the view that certain beliefs are naturally, indeed inexorably, forced upon us – so it would be idle to recommend suspension of these beliefs. This is certainly Hume’s attitude to some beliefs (for example, in the existence of body): ‘Nature has not left this to [our] choice.’ But it is not his attitude to beliefs in general, and certainly not to all causal beliefs: he plainly thinks human beings can be weaned off superstition, and can be taught better principles of assessing evidence. Belief is a natural causal process, but not just that: it may be a natural causal phenomenon and still be a rational one. (Hume can be a compatibilist here, as with free will.) People often point out how Hume says we cannot believe whatever we like, simply at will. But this does not show that our beliefs are essentially irrational, not up for rational assessment and possible correction. The fact that we cannot ‘decide to believe’ does not imply that we cannot perform those paradigmatically rational acts of deciding, of making up our mind, and even changing it, as a result of evidence and argument. (I here use ‘rational’ in our modern sense – rather than simply for the products of an 18th-century faculty of reason.) Hume is, for the most part, not a naturalist at all in this second sense. A third use of ‘naturalism’ would be for Hume’s view that philosophy itself is – at least for him – the result of certain natural inclinations. In this sense, indeed, Hume’s distinctions between good and bad causal reasoning are the outcome of naturalism. But this is now only to say that they are part of his philosophy, on a par in this respect with his sceptical arguments: all are the outcome of careful thought, of natural capacities indeed, in a person with a natural inclination towards philosophy, but certainly not the result of ‘mere’ instinct. Once Hume’s recommendations of good inductive method take their place as part of his philosophy, indeed of his epistemology, it becomes obvious that what epistemological scepticism there is in him is not ‘unmitigated’.

My fourth worry concerns Fogelin’s treatment of moderate scepticism in Hume. In a curious way, Fogelin might almost claim to have accommodated some of these criticisms – to have left a place for the Hume whose scepticism is certainly not unmitigated. At the beginning and end of the book, Fogelin makes a distinction between a ‘moderate – probabilistic, academic – skepticism’ and ‘a more radical version of skepticism that is Pyrrhonian rather than Carneadean’. Moderate scepticism includes the injunction to proportion one’s belief to the strength of the evidence, and to maintain a suitable modesty in one’s convictions. Fogelin might therefore claim that Hume’s strictures against superstition and bad causal reasoning are taken care of in his interpretation – as part of this moderate scepticism.

There are two things wrong with this, however. First, the point would remain, that Hume’s epistemological scepticism is certainly not ‘unmitigated’ as long as his distinctions between good and bad causal reasoning form part of his philosophy, even if they gain the additional label of moderate scepticism. Secondly, this ‘moderate scepticism’ is far too important to be mentioned at the beginning and end of the book and ignored in the middle. Fogelin admits in the Introduction that moderate scepticism is Hume’s ‘general posture’, but then announces that the more radical, Pyrrhonian scepticism will be the main topic of the book. Moderate scepticism reappears in the Conclusion as the ‘standpoint from which [Hume] normally addresses his reader’. But the intervening ten chapters concentrate instead on what radical scepticism can be found in the Treatise. This gives the book a curious feeling of being continually out of focus – attending to the more extreme aspects of Hume, when they are not central, even according to the author. This might be permissible in a book which simply aimed to illuminate a neglected area, but it is decidedly odd in one that is trying to correct an imbalance in earlier commentators.

I have not meant to suggest that there are no grounds for a sceptical interpretation of Hume. But these grounds fall short of justifying the attribution to him of a ‘wholly unmitigated’ scepticism. Hume himself describes ‘the philosophy contain’d’ in the Treatise as ‘very sceptical’. But what counted as sceptical in the 18th century does not always do so now – take the rejection of necessary connection, for example. Some of our beliefs – in space and time, physical objects, and the self – are indeed presented by Hume as palpably irrational. But his attitude to causal beliefs seems rather different. And though he distinguishes good and bad causal reasoning, it may be said that he hardly makes out his entitlement to do so. But this, too, falls short of wholly unmitigated scepticism. An unmitigated sceptic would surely say: ‘All beliefs are equally unreasonable. As a philosopher I ought just as much to work for the abandonment of the belief that the sun will rise tomorrow as for the abandonment of superstition and illusion – however hard, even impossible, for a man that may be.’ But Hume does not say this – and the passages where he comes closest to doing so need to be weighed against others where what he says is clearly incompatible with it.

My four main criticisms could be directed at much else that has been written on the Treatise. Fogelin is not the first to exaggerate its sceptical content, to ignore contrary evidence, to make uncritical use of the term ‘naturalism’, or to neglect Hume’s moderate scepticism. There are welcome features of his book that make it stand out from the crowd. He pays attention to several neglected passages in Book I of the Treatise – on space and time in Part II, on probability in Part III, and on ‘antient’ and ‘modern’ philosophy in Part IV. There is a good discussion of the ‘communication of vivacity’ from impressions to ideas, and a notably clear survey of interpretations of Hume’s worries about personal identity. Fogelin sometimes brings in useful historical background – for example, Bayle’s article on Zeno to illuminate Hume’s work on space and time. But it is misleading to say that ‘Hume does not mention Bayle by name in this context.’ For though the Treatise does not mention him, one of Hume’s letters specifically recommends Bayle’s articles on Zeno and Spinoza, together with three works by Descartes, Malebranche and Berkeley, as an introduction to his own work. Sadly, neither of the other two Continental philosophers recommended by Hume receives even a mention in Fogelin’s book – though an understanding of Hume’s Cartesian inheritance is essential to understanding his views on causation and power. This is not just a pedantic grumble – it is a sign of the general lack of historical sense in Fogelin and in many other Hume commentators, which contributes to misunderstandings of the content of his views, almost as much as of their context.

If there are general morals to be drawn from these criticisms, they are these. We need a much better understanding of the slippery terms ‘scepticism’ and ‘naturalism’ if discussions of whether Hume is ‘really a sceptic’ or ‘really a naturalist’ are to make much progress. If forced to use the terms, I should say that Hume was both less of a sceptic and less of a naturalist than Fogelin, and many others, suppose. But in the absence of a clear definition of the issues, even that means very little. Secondly, we need a better understanding of Hume’s Cartesian inheritance, which is so widely neglected. Hume wrote much of the Treatise while staying at La Flèche, Descartes’s Jesuit College – and an extensive, significant and sometimes surprising influence is visible, for those with eyes to see. Thirdly, any study which confines itself to the Treatise, as Fogelin’s does for the most part, runs a serious risk of misunderstanding even that text. Progress has been made in relatively specialised studies in all of these areas: it is therefore bad that general works on Hume should continue to show so little awareness of this work.

A final point is that Hume’s work is immensely knotty and complex, and any general and sympathetic interpretation will have to take account of opposing tendencies that produce tensions between his views. Though I have traced the relatively clear lines in Hume while discussing the equally clear lines of Fogelin’s interpretation, it is important to draw attention to obscurer lines that often run in different directions.