Before its 1997 relocation to St Pancras, the main reading room of the British Library sat under a soaring Victorian dome in the heart of the British Museum in Bloomsbury. Researchers in possession of a coveted reader’s ticket could work in one of the most atmospheric reading rooms in the world, inspired by the knowledge that such luminaries as Karl Marx, Virginia Woolf, George Bernard Shaw, and Mohandas Gandhi had laboured there before them. There were, it must be admitted, minor inconveniences. Ordering books was a time-consuming process that required the completion of call slips on triplicate carbon paper. These would be conveyed to the bookstacks in Perspex capsules, propelled by compressed air through a labyrinthine system of tubes. This process was preceded by perusal of the printed catalogue, which consisted of hundreds of large, blue, hard-bound volumes arranged in circular cases at the centre of the room. Individual entries gave the appearance of having almost been pasted in, giving the whole catalogue the appearance of massive multi-volume scrapbook. The system had its compensations, however. The arrangement of entries often revealed relationships between sources that might otherwise have gone undiscovered. The way in which authors were characterised was also revealing. The catalogue entry for David Hume, for example, reads: ‘HUME (David) the Historian’.Footnote ¹

Twenty-first-century readers may find this categorisation puzzling. We now think of Hume as primarily a philosopher and, indeed, one of the progenitors of philosophical naturalism.Footnote ² Hume is also something of a role model for many contemporary analytic philosophers, although few share his enthusiasm for history. Survey data reveal that more philosophers identify with Hume than with any other non-living philosopher and by a significant margin. (For those interested in the rankings, Aristotle comes a distant second, followed by Immanuel Kant.)Footnote ³ But in his day, and for a considerable period after, Hume was known as a historian. The subsequent change in the disciplinary identification of Hume is noteworthy, for at least some of his celebrated philosophical positions turn out to rely upon covert historical commitments. This is especially so for one of Hume’s most admired arguments: his case against believing reports of supernatural activity in the form of miracles.

Hume’s famous treatment of miracles is set out in section 10 of the Essay concerning Human Understanding (1748). The section has two parts. In the first, Hume outlines his logical case against lending credence to miracle reports. The second part seems to consist in ancillary historical and anthropological considerations that lend support in various ways to the core philosophical argument that precedes them. Most scholarly attention has accordingly focused on the argument of the first part, which a good number of present-day philosophers still regard as having dealt a telling blow against the rationality of believing in reports of supernatural occurrences. My suggestion will be that a key premise of Hume’s argument lurks among the historical claims made in the second part of the chapter and that it is here that the real force of the argument is to be found. This goes to a more general thesis that the true foundations of modern naturalism lie not in philosophy or the logic of the natural sciences, but in tacit assumptions about historical progress and an accompanying hierarchy of cultures.

Hume recognises at the outset that the business of taking on reports of supernatural interventions is likely to be an endless task if each case has to be evaluated on its own merits. As he puts it: ‘Does a man of sense run after every silly tale of witches or hobgoblins or fairies, and canvass particularly the evidence?’Footnote ⁴ The situation called for an approach that could in principle cover all instances. Hume believed that he had discovered just such an argument, one that could serve, for the wise and learned, as ‘an everlasting check to all kinds of superstitious delusion’.Footnote ⁵ His goal was thus to fashion a ‘silver bullet’ that would establish a presumption of guilt for all miracle reports.Footnote ⁶ In setting up his argument, Hume articulates three considerations. First, ‘A wise man proportions his belief to the evidence.’ Second is his definition of the miraculous: ‘a transgression of a law of nature by a particular volition of the Deity, or by the interposition of some invisible agent’. Third, laws of nature are said to be established by the weight of unvarying testimony to some uniformity in the natural world. What Hume thinks necessarily follows from this is that ‘as a firm and unalterable experience has established these laws, the proof against a miracle, from the very nature of the fact, is as entire as any argument from experience can possibly be imagined’. This is because, by definition, testimony to the breach of a law of nature will always be outweighed by testimony to the law of nature which is supposedly being breached. Hence, the wise man, who weighs up competing testimonies, will always land on the side of the inviolability of laws of nature. For Hume, two further things follow: ‘a miracle can never be proved, so as to be the foundation of a system of religion’ and, because the various religions are incompatible with each other, the putative miracles of one religion necessarily cancel out those of another.

Allocating a key role to weight of testimony leaves Hume with a dilemma, however, although not one that he explicitly acknowledges. The framing of laws of nature might well rest upon cumulative testimony. But there was also cumulative testimony to the reality of miracles and wonders – and this from a variety of cultures past and present. Indeed, this was the very problem that Hume sought to address. Weight of testimony alone, therefore, was insufficient to settle the question. The resolution of this dilemma lay in the appeal to history and anthropology that is quietly introduced in the second part of the chapter. In essence, Hume needed to move from the issue of the quantity of testimony to its quality.Footnote ⁷ Apparently some testimonies are more equal than others:

It turns out, then, that it is this ‘strong presumption’ that relieves the investigator of the obligation to pursue any serious enquiry into the veracity of particular relations of supernatural events. Typical purveyors of tales of the supernatural include, for Hume, ‘monkish historians’, ‘the vulgar’, ‘ignorant people’, ‘barbarous Arabians’, ‘the ignorant and stupid’, and so on.Footnote ⁹ The main thrust of Hume’s argument, then, actually turns on a historical thesis about the process of civilisation, in which some cultures and races (civilised ones) are more advanced and trustworthy than the rest (ignorant, barbarous, and so on). It is this presumption that enables the testimony of certain groups to be discounted without further investigation. Stated in this uncompromising way and extracted from the matrix of philosophical argumentation in which Hume had embedded it, this stark assertion looks very much like an unsupported prejudice, couched in terms likely to be at least mildly offensive to present-day sensibilities.

Unhappily, in his essay ‘Of National Characters’ Hume would go still further and remark, albeit in a footnote, on the intellectual inferiority of non-white peoples. Only among civilised white nations, Hume remarked, do we encounter excellence in action and speculation.Footnote ¹⁰ It is this superiority in speculative ability that enables judgement to be passed by the white and civilised on the beliefs and practices of the non-white and barbarous. My interest here is not in a posthumous prosecution of Hume on this issue, nor in mounting a guilt-by-association case against his modern admirers. Rather it is to wonder about the possibility of a lingering, covert influence of some of Hume’s more dubious anthropological assumptions on our present naturalistic outlook. It is safe to say that few today would stand by the questionable cultural rankings expressed in ‘Of Miracles’ and we have long dispensed with the descriptor ‘barbarous nations’ to characterise peoples given to supernatural beliefs. Yet we might still enquire whether, protestations to the contrary, some implicit commitment to these sentiments sits beneath the surface of at least some of our modern, naturalistic forms of intellectual enquiry. To be sure, the prejudicial condescension of this stance will be less explicit, and the ostensible justifications for disbelief are far more likely be expressed in terms of references to science, or laws of nature, or, for philosophers, an adoption of the scepticism about the supernatural of the kind that Hume elaborates. There is little talk of barbarians or backward savages in the Western academy: on the contrary, and all to the good, there is an increasing effort to acknowledge the value of indigenous perspectives, notwithstanding their prima facie incompatibility with the ruling naturalistic assumptions of the modern natural and social sciences.Footnote ¹¹ It is worth reflecting in more detail, then, on the ways in which unspoken historical assumptions might continue to inform contemporary disciplinary commitments and whether these might conflict with other values that are increasingly regarded as important.

In Chapter 6 we will consider how the ideas of historical progress that Hume casually adverts to were formalised and written into the foundational narratives of the modern social sciences. Naturalism would play a prominent role in these progress stories as a marker of civilisational advance. For now, though, I want to shift the focus of attention to the conceptual framing of Hume’s argument, which is equally revealing about some of the assumptions that underpin modern naturalism. These relate to the fact that certain concepts – ‘belief’, ‘laws of nature’, ‘supernatural’, ‘religion’ – are often treated as if they were unproblematic, self-evident, transhistorical, and universal.

As already noted, contemporary philosophers have focused most of their attention on the first part of Hume’s ‘Of Miracles’, seeing in it Hume’s most original contribution to the discussion.Footnote ¹² These elements of the argument are among the most overworked of all in the philosophical literature, but one reason for revisiting them here is that the assumptions and concepts that underlie virtually every step in this chain of reasoning are illustrative of some general feature of our modern approach to the question of belief in the supernatural. These turn out to be problematic, to varying degrees, when viewed in historical perspective.

First is Hume’s contention that a ‘wise man’ should proportion his belief to the evidence. Taken at face value, this seems an eminently sensible recommendation. On closer examination, however, the idea that our beliefs should be determined solely on the weight of evidence encounters some difficulties. There is a long-standing discussion among philosophers on ‘the ethics of belief’ – whether we have a moral obligation to believe only those things for which we have convincing evidence. A number of acute thinkers – Blaise Pascal, Immanuel Kant, Søren Kierkegaard, and William James among them – have argued that there are occasions on which for prudential, moral, or religious reasons we actually have an obligation to believe things without sufficient evidence. Indeed, it has been proposed that the most important things in life require conviction in the absence of adequate evidence. Added to this, most would allow that there are certain ‘first principles’ that we need to accept, without proof, if we are to know anything at all. As will become apparent in Chapter 4, it was long held that theistic belief was one such principle. Indeed, contra Hume, Cicero had maintained that lack of religious belief was a mark of barbarity.Footnote ¹³ My concern at this point is not to come down on one side or the other of these discussions, but to indicate that the issue is not quite as straightforward as Hume implies. More importantly, and looking ahead to Chapter 3, the emergence of an ethics of belief in the early modern period signals a major shift in Western understandings of what it is that faith and belief consist in. These early modern discussions, initially occurring in a religious context, established the conditions for the subsequent epistemological preoccupations of modern philosophy. They also point us in the direction of a potentially different way of assessing the merits of cultures that seem prone to the advocacy of supernatural beliefs.

Second, is Hume’s definition of a miracle: ‘a transgression of a law of nature by a particular volition of the Deity’. Again, this may seem relatively straightforward and uncontroversial. But it is important to note that the formal conception of a ‘law of nature’, at least in the sense that Hume and early modern natural philosophers used it, did not come into existence until the seventeenth century.Footnote ¹⁴ It follows that earlier historical actors who either observed ‘miracles’ or gave credence to miracle accounts, could not have been operating with the same conception of miracle that Hume is urging upon us, since the idea of a law of nature was not available to them. Hume is thus engaging here in a kind of ‘bait-and-switch’ strategy – not uncommon in philosophy of religion – offering a stipulative definition of some notion or doctrine, and then proceeding to a critical analysis of it without being overly scrupulous about whether this version of the notion corresponds to the way in which it operates in the wild. In this case, there is a significant mismatch between the philosophical conception, cleaned up and abstracted for the purposes of philosophical argumentation, and the variety of ways in which native users had spoken about the miraculous.

Of course, it could be argued that it is precisely the lack of a conception of laws of nature that is the problem. This lack would become the mark of the intellectual immaturity or ‘barbarism’ of the cultures in question. But a retreat to this line of reasoning must contend with the fact that the early modern conception of laws of nature was strongly underpinned by theological considerations. From the seventeenth to the nineteenth century, laws formulated within the sciences were consistently aligned with theological readings of nature and offered as evidence for the divine superintendence of the natural world. The prominence of this view and its decline in the nineteenth century will be discussed in more detail in Chapter 5. For now, we can say that few of the theological champions of laws of nature understood them in ways that generated major difficulties for miracles. More generally, this historical grounding of laws of nature in the divine will is illustrative of one of the central themes of this book – a persistent pattern of the indebtedness of naturalism to covert theological premises. There might seem to be an irresolvable tension between conceiving of God as the source of both the regularities of nature and miraculous ‘interventions’. But, leading on to the next point, this tension was largely the creature of a particular view of the natural/supernatural divide.

If the notion of ‘laws’ operating in Hume’s argument cannot be accepted uncritically, this is even more so with respect to ideas of ‘nature’, ‘supernatural’, and ‘transgression’. Hume’s conception of the miraculous requires something like a natural/supernatural distinction as, more generally, does modern naturalism. But a two-tiered natural/supernatural understanding of reality was a relatively late historical development, as will become apparent in Chapter 5. Suffice it to say for now that earlier thinkers did not subscribe to a notion of divine transgressions, interpositions, or interventions into some relatively independent natural order. For Augustine, to take a single pre-modern example, miracles were not contrary to nature but contrary to our knowledge of nature.Footnote ¹⁵ Again, the suggestion might be that Hume is working with a more sophisticated conception of nature than his medieval predecessors – one based on the sciences, for example. But this is not something he argues for directly, and neither was this reading one that was shared by the scientific practitioners of the period. It might also be the case that Hume’s argument works in the terms in which it is expressed. However, if it fails to match the historical instances it was intended to target, we are just back to the ‘bait-and-switch’ move. Admittedly, matters are complicated by the fact that many of his religious contemporaries shared some of the contents of this conceptual toolkit, and some at least were willing to conduct the debate on those terms. In this case the relevant point goes to an important change in how religious belief was conceptualised that was shared by advocates and critics alike.

Third is Hume’s conclusion that ‘A miracle can never be proved to be the foundation of a system of religion.’ Here the unexamined assumption is that religion is a ‘system’ consisting of propositions for which a particular form of justification is required. Miracles, in this scheme of things, are meant to act as ‘proofs’ for the propositions that constitute the substance of a religion. Like Hume’s stipulative conception of ‘laws of nature’, however, the very idea that Christianity was a ‘religion’ constituted by its propositional content was a product of the early modern period, and not a notion to which pre-modern individuals subscribed.Footnote ¹⁶ So again, this line of argument operates with a rather abstract understanding of the phenomenon that it purports to be addressing. It is certainly true that during this period we encounter religiously motivated defences of Christianity that exhibit a similar understanding of the issues. Apologetic appeals to miracles, along these lines, were responsible for eliciting Hume’s critique. But again, the more general point here is that the new understanding of the nature of miracles and the evidential role ascribed to them tells us something important about new understandings of religious belief. These will be discussed in more detail in the next chapter.

The final element of Hume’ case is this: miracles are supposed to serve as evidence for the truth of propositional beliefs that constitute religion, and all religions claim to be true based on their own proprietary miracles. However, the various religions, understood in this propositional sense, posit conflicting truth claims. It would then follow that either the miracles appealed to as evidence for competing religions did not occur, or even if they did, that they could not logically serve to guarantee the truth of incompatible systems of religious truths. The difficulty here, related to the point above, is the assumption that there are plural ‘religions’, modelled on Christianity, that can be understood in terms of mutually incompatible belief systems. Yet again, this represented a new way of understanding religious phenomena that first arose in the early modern period. What is significant about Hume’s point is that it highlights how the appearance of multiple Christian confessions in the wake of the Protestant Reformation led to a reconceptualising of the religious life and its justifications.

There is also an element of the pot calling the kettle black in Hume’s reference to the embarrassments of religious pluralism. René Descartes observed of the state of philosophy that ‘it had been cultivated for many ages by the most distinguished men, and that yet there is not a single matter within its sphere which is not still in dispute’.Footnote ¹⁷ The ancient schools of Scepticism were a direct response to this ‘known problem’ within philosophy. For their part, modern philosophers have belatedly realised that the state of their own field is itself worthy of philosophical analysis and the philosophy of disagreement has become a lively topic in the last two decades. Needless to say, perhaps, philosophers have been reluctant to conclude, on the basis of the fact that many of them hold mutually exclusive positions, that philosophy is an irrational activity. The so-called ‘steadfast view’, for example, provides reasons for thinking that it is rational to stick to your guns, even in the face of strong peer disagreement.Footnote ¹⁸ In the comparable case of religion, even if we do regard religions as systems of propositional beliefs, as Hume seems to, the implications of disagreement among them might not be as destructive as he seems to think. Some of the problems associated with philosophical disagreements go away if we think of philosophy more as an activity than a set of theoretical commitments. The potential parallels with religion need not be laboured. The way in which religion was reconceptualised in the early modern period, coming to be understood in terms of beliefs to be supported by particular kinds of evidence, was at least partly responsible for generating the pluralistic predicament to which Hume drew attention in this final line of criticism.

In sum, Hume’s celebrated argument helpfully exemplifies a number of the key issues that arise in attempts to understand the historical roots of modern naturalism. These determine the structure of what follows. Restating that order in terms of the discussion above, the historical fortunes of the ideas of faith and belief will be treated in the next two chapters. The question of how rational proofs relate to the religious beliefs, and how that relation has changed over time will be dealt with in Chapter 4. The fifth chapter will focus on the historical origins of the now familiar natural/supernatural distinction, along with the emergence of the accompanying notions of naturalism and supernaturalism. Chapter 6 will provide an account of how naturalism came to be written into accounts of historical progress.