1. Introduction
Paul Russell’s The Riddle of Hume’s Treatise: Skepticism, Naturalism, and Irreligion (Reference Russell2008) is an important and impressive work that makes for compelling reading. Apart from its interest to Hume scholars, Russell provides extensive and detailed information about a fascinating set of debates in the early Modern period between ‘speculative atheists’ or free thinkers, like Thomas Hobbes and Benedict Spinoza, and religious philosophers, like Samuel Clarke. Russell shows that Hume made significant contributions to this broad range of subjects. It turns out that even in a seemingly nonreligious text like A Treatise of Human Nature, Hume’s thoughts were largely motivated by topics in the philosophy of religion. Russell claims:
The primary aim of Hume’s series of skeptical arguments, as developed and distributed throughout the Treatise, is to discredit the doctrines and dogmas of Christian philosophy and theology with a view toward redirecting our philosophical investigations to areas of ‘common life, ’ with the particular aim of advancing ‘the science of man.’ (2008, 290)
It is not too far off the mark to suppose that the primary aim of the Treatise was to replace religious philosophy with Hume’s ethical naturalism.
For Hume scholars the book is important in other ways because Russell sheds light on the ‘ultimate riddle’ of the Treatise: ‘is it possible to reconcile Hume’s (extreme) skeptical principles and conclusions with his aim to advance the “science of man”;’ (2008, 3)? Or more particularly: Does Hume’s skepticism undermine his ‘secular, scientific account of the foundations of moral life in human nature’ (290)? Which brings us to Russell’s controversial thesis: ‘the irreligious nature of Hume’s fundamental intentions in the Treatise’ is essential to solving the riddle (11). Connecting all of these strands together is Hume’s naturalism. Russell makes a compelling case for Hume’s irreligion as well as his atheism. It is difficult to find areas of disagreement between us, although one is that I think Hume is an irreligious theist and not an atheist. Perhaps this is only a verbal dispute, as Russell (289) and Hume (Reference Hume and Smith1947, 216–19) suggest. My hope is that my case for Hume’s theism sheds even more light on his naturalism and ultimately the riddle of the Treatise.
2. An argument for atheism in the Treatise?
A good place to begin is with Hume’s alleged argument for atheism. In Chapter 13, an argument for atheism is offered that, although never explicitly stated, Russell claims is found in the penumbra of the Treatise. The argument concerns the problem of skepticism about the external world. For Descartes, our knowledge of the existence of the external world is founded on our knowledge that God exists and that he is not a deceiver (2008, 168–9). Several philosophers in the Modern era endorsed this view. Within this Cartesian justification lies the claim that God would be a deceiver were our views about the external world incorrect (169). Yet God, by his very nature, cannot be a deceiver. From this, we get the following argument for atheism:
1. We naturally and inescapably believe in the existence of body (i.e., usually and primarily in the vulgar form).
2. Our belief in the existence of body is false and based on an illusion (i.e., we are deceived about this).
3. If God exists, and we are naturally deceived about the existence of body, then God is a deceiver.
4. God cannot be a deceiver.
5. If we are deceived in our natural belief about body, then God does not exist.
6. Therefore, God does not exist. (177)
First, consider premise (3): If God exists, and we are naturally deceived about the existence of body, then God is a deceiver. One way that I might be deceived is if I believe that an ordinary object is in front of me – a chair or a table – yet there is no ordinary object. Perhaps I am in the Matrix world, and robots are systematically making me falsely believe that I am sitting in a chair. Similarly, God might be an evil genius. Is this the way in which God might be a deceiver? It seems not.
The vulgar believe that ordinary objects are the direct objects of perception: We see chairs, tables, etc. directly. Yet ordinary objects are physical objects that exist independent of our experience. This combination of beliefs is problematic given the argument from illusion. Take your finger and press it against your eye. You might see two chairs when previously you saw just one. Because there is only one chair yet you see two, ordinary objects are not always the direct objects of perception and direct realism – which holds that the direct and only object of perception is a mind-independent, physical object – is wrong. The criticism appears decisive. Likely it would be decisive were it not for the fact that the other theories of perception – indirect realism and idealism – are equally problematic. Idealism is also prone to the argument from illusion and indirect realism ultimately leads to skepticism (2008, 175; see Hume Reference Hume, Selby-Bigge and Nidditch1975, section 12, part 1).
Premise (2) says: Our belief in the existence of body is incorrect and based on an illusion – that is, we are deceived about this. Perhaps what we are deceived about is the nature of the objects of perception. I see a chair and come to believe that the chair exists. Adopting the vulgar view, I judge that the chair is the immediate object of my perception and here is where I go wrong. Contained within these puzzlements are two distinct sets of questions.
a. What is the nature of ordinary objects (e.g., chairs, tables)? Are ordinary objects mind-independent or mind-dependent; are they physical objects or mere bundles of ideas?
b. What is the nature of objects of perception? Do we perceive ordinary objects directly or are they mediated by ideas?
Both of these more fundamental questions presuppose the existence of ordinary objects. That ordinary objects exist is a starting point, an assumption about which all parties agree. Hume appears to regard question (a) as a non-issue. Given my understanding of Hume (and Russell agrees), Hume is a neutral monist: there is just one kind of stuff and Hume remains neutral about whether it is physical, mental, or whatever. On the other hand, question (b) is never given a satisfactory answer, for all theories of perception are problematic. Nonetheless, Hume believes that ordinary objects exist along with the vulgar and most other parties of the various philosophical debates.
On many issues Hume remains a skeptic. He holds that there are no convincing responses to a range of philosophical problems, no way for reason to determine which among a set of philosophical theories is correct. Nonetheless, we believe in spite of our skepticism. Skeptical arguments are idle and only influence our beliefs and behavior for a limited period of time: ‘an hour hence he will be persuaded there is both an external and an internal world’ (Reference Hume, Selby-Bigge and Nidditch1978, 218). We are compelled by ‘action, and employment, and the occupations of common life’ to hold beliefs despite our skepticism (1975, 159). Skepticism only goes so far. Skeptical arguments not withstanding, we still believe in the existence of ordinary objects. We bicker about their nature and how we might come to know about their nature but we accept as a provisional starting point that ordinary objects exist. This is a shared assumption among all parties in the debate. Philosophical disputes remain but because solutions to these problems have little practical significance, it is unclear that this is a meaningful ‘deception.’
3. Skepticism, naturalism, and positivism
Russell is concerned with ‘the fundamental skepticism/naturalism dichotomy’ (2008, 4) in Hume’s Treatise. The short version of the riddle of the Treatise is: Can we reconcile Hume’s skepticism with his naturalism? For if the skeptical arguments are adequate enough to demolish the speculations of religious philosophers, why aren’t they sufficient to undermine Hume’s naturalism? Assuming by Hume’s naturalism something like the Norman Kemp Smith (1941) interpretation of Hume, then really there are three Humes: skeptic, naturalist, and positivist. If one is looking for a streamlined, positivist analysis of some philosophical concept –causation, free will, identity, morality – what better place to start than with the Treatise? As an undergraduate I imagined that logical positivism began when Moritz Schlick read the first three sections of the Enquiry. He stopped, thinking he could figure out the rest. Had he only finished the Enquiry – or read the Treatise – then logical positivism would have been avoided.
Russell recognizes these three interpretations of Hume: skeptic, naturalist, and positivist. He notes what Kemp Smith calls ‘the Reid-Beattie interpretation, ’ which is ‘destructive’ in character (Russell Reference Russell2008, 4; see also 14–15). According to this view, ‘Hume is read as a systematic skeptic, whose principle aim is to show that our “common sense beliefs”; (e.g., beliefs in causality, independent existence of bodies, in the self, etc.) lack any foundation in reason’ (4). Russell doesn’t spend much time discussing this interpretation and for good reason. It portrays Hume as an excessive skeptic in contrast to the more mitigated skepticism that he endorses (1978, book 1, part 4; 1975, section 12). Of course, positivism is a kind of naturalism, so this is precisely where things get confusing. What kind of naturalism does Hume endorse and what kind does he reject?
P.F. Strawson distinguishes between reductive and nonreductive naturalism (Reference Strawson1985, 1). There are at least two forms of reductive naturalism.Footnote 1 On the one hand there is reductive empiricism, such as the positivism that Hume puts forth in the first three sections of the Enquiry (1975) and elsewhere. Second, there is metaphysical reductionism, such as reductive physicalism or eliminative materialism. As I understand Strawson’s distinction, these are all kinds of reductive naturalism. Thus, reductive naturalism is often coupled with eliminativism, for it eliminates whatever it cannot reduce. Many reductive naturalists are atheists. They cannot reduce the concept of God to a set of experiences or beliefs, so they reject God’s existence. But God is the least of their worries. Reductive naturalism does not allow for the existence of anything ‘which is not ultimately reducible to or explicable in terms of the natural sciences’ (Strawson Reference Strawson, Hahn and Hahn1998a, 168). Thus, reductive naturalism is also often aligned with skepticism. Strawson writes:
An exponent of some subvariety of reductive naturalism in some particular area of debate may sometimes be seen, or represented, as a kind of skeptic in that area: say, a moral skeptic or a skeptic about the mental or about abstract entities or about what are called ‘intensions’ (1985, 2).
In contrast, nonreductive naturalism ‘provides for a richer conception of the real, making room, for example, for morality and moral responsibility, for sensible qualities as genuinely characterizing physical things, for determinate meanings, meaning-rules, and universals – all as we ordinarily conceive them’ (Strawson Reference Strawson, Hahn and Hahn1998a, 168). The central point of nonreductive naturalism is that it proclaims naturalism without epistemological or metaphysical reduction of any sort, and in the face of compelling skeptical arguments. Of course, the nonreductive naturalist has some explaining to do. Why is it that we are invited to believe in, say, moral responsibility given the skeptical attacks against it? The reductive naturalist has a response: elimination. What is the nonreductive naturalist response to the problem of skepticism?
According to nonreductive naturalism, the problem of skepticism is ‘not to be met with argument’ but is ‘simply to be neglected’ (Strawson Reference Strawson1985, 13). Strawson continues:
To try to meet the skeptic’s challenge, in whatever way, by whatever style of argument, is to try to go further back. If one is to begin at the beginning, one must refuse the challenge as our naturalist refuses it (1985, 24–5).
Implicit in this nonresponse to the skeptic is the suggestion that there are no adequate responses to the problem of skepticism. This applies to broad range of problems, both epistemological and metaphysical (Campbell Reference Campbell2011, chapter 5). Strawson writes:
the point has been, not to offer a rational justification of the belief in external objects and other minds or the practice of induction, but to represent skeptical arguments and rational counter-arguments as equally idle—not senseless, but idle—since what we have here are original, natural, inescapable commitments which we neither choose nor could give up (1985, 28; cf. Reference Strawson and Hahn1998b, 242).
Strawson’s concession is only a partial victory for the skeptic because skeptical arguments have no lasting impact. We continue to act as if our ordinary beliefs were true, regardless of our inability to provide either conclusive support for those beliefs or an adequate response to the skeptical arguments that undermine them. Skeptical arguments are ‘idle.’ In Hume’s Dialogues Concerning Natural Religion, Cleanthes makes this same point when he says to Philo:
Whether your scepticism be as absolute and sincere as you pretend, we shall learn by and by, when the company breaks up: We shall then see, whether you go out at the door or the window; and whether you really doubt if your body has gravity, or can be injured by its fall; according to popular opinion, derived from our fallacious senses, and more fallacious experience (1947, 132).
This is not a refutation of skepticism, for ‘arguments on both sides are idle’ (1985, 29). Were our beliefs based on reason, skeptical arguments would undermine those beliefs. The beliefs are not undermined, so they are not based on reason.
Return to the three interpretations presented at the beginning of this section. Is Hume best understood as a skeptic, naturalist, or positivist? Nonreductive naturalism is sympathetic to both the naturalist interpretation of Hume as well as the skeptic interpretation. It allows us to keep the best of both interpretations: the skeptic wins but it is all for naught. Hume is a reductive naturalist on the positivist interpretation. That view cannot be correct. For one thing, it renders Hume’s rejection of excessive skepticism and acceptance of mitigated skepticism a complete mystery.
Hume’s reductive analyses of various philosophical concepts are an important part of his philosophy but it is best to see them as versions of the very skeptical worries that are dismissed by his nonreductive naturalism. Given the dictates of empiricism (1978, book 1, part 1; 1975, sections 1–3), our concepts of causation, external object, and personal identity should be limited to Hume’s positivist analyses. Nonetheless, our natural beliefs about causes, ordinary objects, and the self extend beyond the empirically given. Hume is a naturalist but not a reductive naturalist; he is a skeptic but not an extreme or destructive skeptic.
4. Atheisms and theisms
Russell discusses J.C.A. Gaskin’s distinction between absolute and relative atheism (1988). The absolute atheist denies the existence of all Gods whereas the relative atheist believes in ‘a more contracted or radically different idea of god from that which prevails in their society’ (2008, 47; Gaskin Reference Gaskin1988, 219–23). For Gaskin, Hume is an ‘attenuated deist’ because he assents to the proposition that God exists yet does not commit to any specific set of religious views. Russell notes: ‘this does not commit him to belief in a Christian god’ (2008, 47).
Later in his book, Russell distinguishes between thin and thick theism. According to thin theism, ‘there exists some “supreme”; intelligence that is the origin, creator, and governor of this world’ (2008, 282). On the other hand, ‘thick theism presupposes a richer set of attributes, such as infinity, omniscience, omnipotence, and moral perfection’ (282). Russell writes: ‘whether we are judged an atheist or not may depend not only on whether the standard of theism is thick or thin but also on what particular set of thick attributes are considered as essential for belief in God’ (282).
Russell notes that Hume ‘leaves the door open regarding the possible truth of thin theism, ’ and Hume is better regarded as a ‘skeptic’ or ‘agnostic’ (2008, 284) than an ‘atheist.’ But Russell adds:
As Hume presents it, religion based on thin theism is refined into nothing. Although he does not actually deny the hypothesis of thin theism, he leaves us with nothing to believe in—much less any doctrine or teaching we can guide our lives by… Hume embraces and endorses no kind of theism—thick or thin (283-4).
Russell’s worry is a genuine one. The question arises: Is Hume’s conception of God too thin? Consider, for instance, Philo’s reversal.
If the whole of natural theology, as some people seem to maintain, resolves itself into one simple, though somewhat ambiguous, at least undefined proposition, that the cause or causes of order in the universe probably bear some remote analogy to human intelligence: If this proposition be not capable of extension, variation, or more particular explication: If it affords no inference that affects human life, or can be the source of any action or forbearance: And if the analogy, imperfect as it is, cannot be transferred, with any appearance of probability, to the other qualities of the mind: If this really be the case, what can the most inquisitive, contemplative, and religious man do more than give a plain, philosophical assent to the proposition, as often as it occurs; and believe that the arguments on which it is established, exceed the objections which lie against it? (1947, 227)
Philo suggests that his conclusion is ‘not capable of extension, variation, or more particular explication’ and explicitly says ‘the analogy, imperfect as it is, can be carried no further than to the human intelligence.’ For this reason, some think that Philo’s conclusion is that a God without moral attributes exists (Tweyman Reference Tweyman1987).
Philo’s conclusion is ‘that the cause or causes of order in the universe probably bear some remote analogy to human intelligence.’ The conclusion is not that the cause of order is an intelligent being but that it is remotely analogous to intelligence. As Graham Priest (Reference Priest1985) indicates, even an atheist could accept that such an entity exists. One might say, for instance, that the laws of nature are analogous to intelligence in the sense that they are orderly and systematic. A rotting turnip is, in this way, analogous to human intelligence (Hume Reference Hume and Smith1947, 218). Does this lend support to Hume’s endorsement of an amoral, impersonal God? No. There is nothing in Philo’s argument that would rule out God’s having either moral or personal qualities. At most we cannot conclude that he has them. Philo’s argument proves the existence of something but we have no idea whether that something has or lacks any of the classical attributes, for we know nothing about its nature other than that it resembles human intelligence.
A closer look at thin theism reveals that it has parts: (a) God is a supreme intelligence, (b) God is the origin and creator of the universe, and (c) God is the governor of the universe. A deist might accept (a) and (b) but not (c). Hume might endorse an even thinner conception of theism if he is limiting his belief to the dictates of reason. As Russell puts it: ‘All we can understand by the word God, therefore, is “the cause of the World”;’ (2008, 51). On this understanding of Hume, he could accept (b) yet reject both (a) and (c). As he expressed in a letter to William Mure, Hume objected to ‘every thing we commonly call Religion, except the Practice of Morality & the Assent of the Understanding to the Proposition that God exists’ (Reference Hume, Klibansky and Mossner1954, 12–3). This seems like a proclamation of theism, for Hume assents to the proposition that God exists. This is Hume’s minimal theism.
The concept of the relative thinness is helpful but by classifying Hume as a thin theist, we are extending the content of his actual claims. Philo’s conclusion is that the cause of the universe is analogous to intelligence, not that the cause of the universe is intelligent. Hume assents to this and little else besides minimal theism. Hume is not a thin theist but for different reasons than Russell supposes.
Another set of distinctions might be helpful. Transcendental theism is the view that God exists at least in part outside of the natural order whereas natural theism claims that God exists completely within the natural order (see Reich Reference Reich1998). There is also the contrast between classical and neoclassical theism (see Hartshorne Reference Hartshorne1962). According to the former, God exists and has all of the classical attributes: immutability, eternality, omnipotence, omniscience, benevolence, etc. According to the latter, God exists but lacks at least some of the classical attributes. If God’s transcendence is a classical attribute, then classical theism entails transcendental theism. As I understand it, the transcendental/natural theism debate is a debate about whether or not transcendence is a divine attribute. Often transcendental theism is associated with classical theism even though the logical link between the two is questionable. In any event, Hume is best understood as a neoclassical naturalist.
Gaskin makes a strong case for Hume’s minimal theism by providing several quotes, written throughout Hume’s lifetime, as well as a few historical anecdotes (Reference Gaskin1988, 219–21). Perhaps the most persuasive evidence comes from The Natural History of Religion, which offers multiple versions of the design argument.
• The whole frame of nature bespeaks an intelligent author; and no rational enquirer can, after serious reflection, suspend his belief a moment with regard to the primary principles of genuine Theism and Religion. But the other question, concerning the origin of religion in human nature, is exposed to some more difficulty. (Reference Hume and Root1956, 21)
• … I allow that the order and frame of the universe, when accurately examined, affords such an argument [for God’s existence] … (1956, 24)
• Were men led into the apprehension of invisible, intelligent power by a contemplation of the works of nature, they could never possibly entertain any conception but of one single being, who bestowed existence and order on this vast machine, and adjusted all its parts, according to one regular plan or connected system. (1956, 26)
• Whoever learns by argument, the existence of invisible intelligent power, must reason from the admirable contrivance of natural objects, and must suppose the world to be the workmanship of that divine being, the original cause of all things. (1956, 38)
• But being taught, by more reflection, that this very regularity and uniformity is the strongest proof of design and of a supreme intelligence, they return to that belief, which they had deserted; and they are now able to establish it on a firmer and more durable foundation. (1956, 42)
• A purpose, an intention, a design is evident in every thing; and when our comprehension is so far enlarged as to contemplate the first rise of this visible system, we must adopt, with the strongest conviction, the idea of some intelligent cause or author. (1956, 74)
Some of these passages suggest that Hume was a thin theist, for instance, when he writes: ‘nature bespeaks an intelligent author.’ There are two responses to make.
First, one can read the conclusions of these arguments as variants of Philo’s conclusion. These conclusions claim the existence of an ‘invisible intelligent power, ’ that is, a power analogous to human intelligence. God is intelligent in the sense that he is analogous to human intelligence.
Second, more is made of the issue than is warranted. Suppose we agree that Philo’s conclusion is merely minimal theism but that Hume actually endorses thin theism. What contradictions arise for Hume? None. We already think that Hume believes in the existence of external objects even though he accepts that that belief is both unsupported and (apparently) disproven by reason. Belief may extend the dictates of reason. That is the whole point of Hume’s naturalism. It would be odd to find that suddenly, when we are talking about belief in God’s existence, that things would be different. Of course, we must be careful to stay as close to the dictates of reason as possible, even if we can’t stay as close as the reductive naturalist desires.
Is the conjunction of minimal theism together with Philo’s conclusion too thin a theism to be any genuine theism at all? Think of the belief that there is something outside your window. You might not know anything about the details: what it looks like, whether it is benevolent or malicious, whether it is even a person. Yet you might believe there is something outside your window nonetheless. This is analogous to belief in God combined with a thoroughgoing skepticism about God’s nature. No doubt, Hume was a skeptic about God’s nature and attributes; he was irreligious, a staunch critic of popular religion and popular Christianity in particular. My only additional claim is that Hume was a theist.
5. Concluding remarks
In closing, let me emphasize my strong enthusiasm for The Riddle of Hume’s Treatise. Besides its important philosophical content – some of which is discussed above – Russell’s book provides a wealth of information for future scholars about the debate between religious philosophers and free thinkers: arguments, criticisms, and who held what view when. This is not the place to tell that story, just another reason to read the book. Yet it also gets at one final possible argument for Hume’s atheism. By the standards of Hume’s Christian contemporaries, all free thinkers are atheists.
According to religious philosophers of the early modern period, atheists adopt the following theses: ‘materialism, necessitarianism, ethical relativism, and skepticism about natural and revealed religion’ (Russell Reference Russell2008, 26). Choosing just one item on this list, all skeptics are atheists. This holds even for mitigated skeptics like Hume. Similarly, Clarke links atheism with naturalism. In order to secure his own theism, he endeavors to establish ‘five connected points’ (29).
• God necessarily exists.
• The laws of nature are contingent, for nature is a result of God’s free choices.
• Space and time are infinite; matter is finite.
• God is present in nature; space and time are divine attributes.
• The soul is immaterial and endowed by God with free will.
If we use the acceptance of the above theses and claims as a determiner for whether someone is a theist, then Hobbes, Spinoza, and Hume are all atheists. Spinoza accepts the first claim, of course, and Hobbes and Hume accept that persons have free will. Nonetheless, if a theist must accept all the claims on the list, then Hobbes, Spinoza, and Hume are all atheists.
In response, consider a hypothetical free will debate between Hume and Spinoza. Hume wants it all: free will, determinism, and the compatibility between them. He is a soft determinist. Spinoza is a hard determinist: determinism – worse necessitarianism – is true, so no one has free will. Spinoza accepts a kind of freedom –blessedness – that is compatible with determinism but blessedness is not free will, according to Spinoza.
Spinoza says to Hume:
In order to have free will, one must be the ultimate cause of one’s actions. But given determinism – which you accept – no one is the ultimate cause of anyone’s actions. Each of our actions traces back to events over which we had no control. We may have blessedness but blessedness is not the kind of ultimate agency and responsibility we need for free will. Given determinism, no one has free will.
In spite of arguments like this, Hume accepts that some persons have free will. That Spinoza doesn’t accept Hume’s conception of free will cannot be relevant to the issue of whether or not Hume accepts that some persons have free will. Perhaps Spinoza regards Hume’s conception of free will as too thin but Hume does not. Likely, Hume regards Spinoza’s conception of free will as incoherent, so Hume’s conception is as thick as you are going to get. Besides, Spinoza’s argument is just another argument for free will skepticism, and we know what Hume’s response would be.
In a similar way, one might defend the claim that Hume is a theist. The relevant issue is whether or not Hume assents to minimal theism: the proposition that God exists. The relative thinness of Hume’s theism does not matter. The moral of the story is: Don’t let your opponents set the taxonomy.Footnote 2
Disclosure statement
No potential conflict of interest was reported by the author.
