Alexander, we are told,4 refused to fight the battle of Gaugamela at night, lest he be accused of having stolen a victory. Can philosophers, who claim to champion truth, afford to be less scrupulous? Even if they deem arguments they attack to be unworthy of respect, should they not provide sound, rather than inadequate or spurious, reasons for rejecting them? This is exactly what D. C. Stove has argued that Hume fails to do in Part IX of the Dialogues, assuming, as is generally done, that the speeches of Cleanthes and Philo present reasoning that Hume would personally endorse.5 Stove maintains that most of the criticisms of Demea's argument are "extremely defective, and that they even include an inconsistency" (Stove, 300). This contention has been challenged by Donald Stahl, who has published a succinct point by point rebuttal of Stove's contentions.6 The Stove-Stahl controversy only deals with the reasoning presented in paragraphs 5, 6, 7, and 10 of Part IX because those are the contentions Stove attacks (see p. 3 below, at the beginning of §2, for a summary of the theses of each paragraph of Part IX). He has "nothing to say against" the criticisms presented in paragraphs 8 and 9, which he assumes will be generally agreed to be less important than the others (Stove, 308). He totally ignores the argument of paragraph 11.
However, there are excellent reasons for believing the criticisms of paragraphs 8 and 9 to be crucially important to Cleanthes' case, and much can be said against his arguments. Further, if, as many suppose, Philo most nearly represents Hume's own view, careful attention to paragraph 11, which is Philo's last word on Demea's argument, may be especially rewarding. Consequently, this paper concentrates on these still-neglected portions of the traditionally neglected Part IX, although it occasionally wanders onto the terrain disputed by Professors Stove and Stahl. I shall (1) review Demea's argument, (2) argue for the significance of the criticisms of paragraphs 8-9, (3) evaluate those criticisms, and (4) assess the role of Philo's concluding remarks in paragraphs 10-11.
In earlier parts of the Dialogues Hume has Demea utter such blatant inconsistencies as to make the reader wonder whether this character is supposed to be at all thoughtful about the implications of his assertions. At the beginning of Part II, he proclaims that God's existence is surely not at issue, since no one "of common sense . . . ever entertained a serious doubt with regard to a truth so certain and self-evident."7 The only proper object of theological inquiry is the nature of God, but that is "altogether incomprehensible and unknown to us" (ibid.)! So, God's existence is certain beyond any reasonable question, and the legitimate question about the divine nature is unanswerable. Such a radically apophatic position (Cleanthes calls it "mystical") leaves so little scope for natural theology that it is hardly surprising that the sceptical Philo readily concurs, once he specifies that he takes 'God' to mean just "the original cause of the universe (whatever it be)" (D 142). But does Demea really mean it? He immediately buttresses this allegedly self-evident truth by appealing to the authority of Father Malebranche and the whole tradition of Christian theology, apparently oblivious to the impropriety, to say nothing of the futility, of supporting an indubitable proposition by citing authorities (D 141-2). Further, scarcely more than a page later, after Cleanthes sketches out the design argument, this same Demea is horrified that Cleanthes gives no abstract arguments or a priori proofs and calls for a "demonstration of the being of a God" (D 143).8
When, in Part IX, Demea himself undertakes to provide the "simple and sublime argument a priori" he had expected from Cleanthes, he characterizes the argument as "the common one," which suggests he is paraphrasing authority rather than thoughtfully arguing. The argument, in fact, as has been generally recognized, closely follows one given by Samuel Clarke in the Boyle Lectures of 1704.9 Demea begins by assuming that "whatever exists must have a cause or reason of its existence," since nothing can cause itself to exist. He then states a dilemma: in tracing the causal ancestry of present existences, either one traverses an infinite series of causes without an ultimate cause or else one ends with an ultimate cause which exists necessarily.10 Next he claims that he can prove that the first horn of this dilemma is absurd. But the subsequent reasoning does not attack the idea that the causal succession is infinite, as the statement of the dilemma leads one to expect. Quoting the text,
In the infinite chain or succession of causes and effects, each single effect is determined to exist by the power and efficacy of that cause which immediately preceded; but the whole eternal chain or succession, taken together, is not determined or caused by any thing: And yet it is evident that it requires a cause or reason, as much as any particular object, which begins to exist in time. The question is still reasonable, why this particular succession of causes existed from eternity, and not any other succession, or no succession at all. If there be no necessarily existent Being, any supposition, which can be formed, is equally possible; nor is there any more absurdity in nothing's having existed from eternity, than there is in that succession of causes, which constitutes the universe. What was it, then, which determined something to exist rather than nothing, and bestowed being on a particular possibility, exclusive of the rest? External causes, there are supposed to be none. Chance is a word without a meaning. Was it nothing? But that can never produce any thing. We must, therefore, have recourse to a necessarily existent Being, who carries the REASON of his existence in himself; and who cannot be supposed not to exist without an express contradiction. There is consequently such a Being, that is, there is a Deity (D 188-9).
Demea does not arrive at the Deity by tracing the chain of causes from present existences to a primordial cause. Rather than terminating the causal series temporally, he instead consents to an infinite causal succession and argues that there must be a reason for the existence of the whole series. Clarke makes much the same move but is more explicit about his motives. Although he believes that the causal series actually did begin to exist some finite time ago, he declines to argue "from the supposed Impossibility of Infinite Succession" because he thinks there are no sound arguments against the possibility of the infinite. That the world is temporally finite can be known only by revelation (Clarke, pp. 24, 71-4). Given how Demea initially phrases his dilemma, and given the good Christian character with which Hume endows him, surely we must suppose that Demea, too, would have been perfectly happy to offer a proof whose conclusion closely conformed to the orthodox doctrine of creation, had it been possible to do so. This supposition will be recalled shortly.
It will be easier to assess the relative importance of the criticisms proposed by Demea's interlocutors if we first orient ourselves by surveying all their objections. Cleanthes' are found in paragraphs 5-9 and Philo adds his in paragraphs 10-11. They are:
¶5 The non-existence of anything is conceivable and therefore not demonstrable.This summary makes it easy to apprehend a pattern in these objections. Paragraph 5 is a general argument against all a priori demonstrations of "matters of fact," which Cleanthes proclaims as "entirely decisive" reasoning upon which he is "willing to rest the whole controversy." Ironically, he immediately proceeds, with typical inconsistency, to give four more arguments against Demea's a priori proof. These latter four criticisms are of two sorts. Those proposed in paragraphs 6 and 7 attack the concept of a necessarily existent deity, first on intensional, then on extensional grounds.11 They are criticisms of the object, or end, of Demea's proof. Those of paragraphs 8 and 9 attack Demea's use of causal reasoning to move from any particular entity to an ultimate explanation of that entity and its causal antecedents. They are criticisms of the means Demea's proof employs to achieve its end. Philo's remarks, contained in paragraphs 10 and 11, are of an entirely different order.¶6 Since we can conceive anything not to exist, 'necessary existence' is a meaningless or contradictory expression.
¶7 The material universe may be the necessarily-existent being.
¶8 The causal relation implies both temporal priority and coming-to-be and therefore cannot pertain to something eternal.
¶9 The sum of the causal accounts of the parts of a collection of entities is the complete causal account of the whole collection because the whole is an arbitrary mental construct, not another entity.
¶10 That the digits composing any product of 9 always add up to some lesser product of 9 at first seems to evidence either design or a remarkable stroke of chance, but it can be shown to be entirely due to numerical necessity. Similarly, things may be disposed as they are simply because that arrangement is necessitated by the nature of matter.
¶11 A priori arguments convince only a select few habituated to abstract reasoning and are thus ineffectual foundations for religion.
This pattern suggests that the argument of paragraph 5 (and 6, which simply applies the general rule to "necessarily existent being") may be especially significant, as Cleanthes' claims intimate. True, if any object might not exist, then in no case could a priori reasoning validly show the existence of an object to be necessary. Cleanthes deduces this universal contingency from the alleged fact that "whatever we conceive as existent we can also conceive as non-existent." But that is precisely what is at issue, since Demea thinks that there is, and must be, one entity which cannot be conceived not to exist. He supports this belief by a strongly intuitive principle, the principle of sufficient reason, according to which everything contingent must have a complete cause. Thus, although paragraphs 5-6 seem pivotal, they amount to little more than a dogmatic assertion that we can conceive anything not to exist.
If importance is determined by argumentative effectiveness, the pattern simply does not support the contention that paragraphs 5-7 are more important than paragraphs 8-9. Clearly a causal argument for a necessarily existent being could be as decisively derailed by a successful attack against the propriety of its use of causal reasoning as by a successful attack against the intelligibility or extension of the concept 'necessarily existent being.' Indeed, in this instance a case can be made for the greater importance of objections to the use of causal reasoning.
Admittedly, the most obvious--perhaps the only--way to attack the ontological argument is to challenge the concept of 'necessarily existent being.' However, Demea's a priori proof is not a version of the ontological argument. The ontological argument begins with the concept of a necessarily existent being and proceeds to deduce the actual existence of that being. Demea's argument instead concludes with a necessarily existent being because it finds that causal explanation either terminates (i. e., causation initiates) with such an entity or fails to be complete. This difference is crucial for assessing the relative significance of objections to the argument. Should one be able to show that causal explanation does not require a single first cause, the entire motivation for Demea's invocation of a necessarily existent being would be undercut. Although, considered abstractly, an alleged demonstration of the impossibility of a necessarily existent being might seem equally decisive, in practice it would not be.
First, as I have already intimated and as Stove himself points out (Stove, pp. 304-5), it simply begs the question. Cleanthes' argument in paragraphs 5-6 rests on the self-evident status of his claim that whatever can be thought to exist can be thought not to exist. He attempts to rule out Demea's argument by a psychological or logical (which is not clear) decree about what can be thought. Appeals to what can be thought or imagined notoriously have force only for those who already share the same intuitions about the matter. Demea is certainly not alone in thinking 'necessary existence' to be an intelligible idea, since many thinkers, including such post-Humean philosophers as Charles Hartshorne and Norman Malcolm, share his conviction. If Demea and others think 'necessary existence' an intelligible idea, does that annul Cleanthes' decree? The answer is far from obvious, for when two abstract a priori claims contradict one another, no crucial experiment is readily available to determine which, if either, is true. On the other hand, should Demea's use of causal inference to establish the divine existence be shown to be improper or flawed, his argument fails even if the concept of necessary existence is coherent.12
Further, were the argument against the concept of necessary existence entirely successful, it would have consequences Cleanthes clearly does not anticipate. If God does not exist necessarily, then presumably God's existence is contingent. If God's existence is contingent then it causally depends on something else, and God cannot be 'God' even in the sense of "original cause" which Cleanthes wishes to retain. If his criticism of Demea succeeds, he decisively undercuts his own brand of theism. Cleanthes does phrase his argument entirely in terms of our conceptual abilities. In basing the arguments of ¶¶5-6 on the nature of our faculties, is he trying to leave open the possibility that God might really be necessary, although human beings can never know that? But then in thinking "God does not exist" the concept 'God' would not refer to God.13
A third reason for preferring to attack the means rather than the end of Demea's proof is that our tendency to seek explanatory completeness is normally stronger than our confidence in allegations of absurdity.14 The history of knowledge is replete with inferences to the existence (or non-existence) of entities, from the atoms of the ancients to the black holes of today, whose antecedent probability was vanishingly low but whose existence (or non-existence) seemed requisite for a complete account of the phenomena. Zeno, and more recently, F. H. Bradley, thought that our ordinary concepts of space and time were self-contradictory and gave very powerful arguments to support their claims; but their point of view did not prevail in the face of the general usefulness of these concepts. Aristotle thought that the idea of an atom (an indivisible magnitude--whether a Platonic plane or a Democritean solid) was, on unimpeachable logical and mathematical grounds, incoherent; nor is it clear that we have even yet definitively settled its status. What is clear is that Aristotle's arguments did not deter subsequent atomists, from Epicurus onward. In the Treatise, Hume himself argues for the existence of colored points, the concept of which seems as absurd to most people as the concept of necessary existence does to Cleanthes (T, I. Part II, §§ 3-5, pp. 33-65).15 He asserts the existence of colored points nevertheless because, within his conceptual framework, he requires them to make his account of visual experience complete. These cases show that an idea which is prima facie incoherent may still be accepted if it is the best available candidate for completing an explanatory account. Since Demea proposes the existence of a necessary being precisely as a solution to a problem of causal incompleteness, it is important that Cleanthes' arguments in paragraphs 8 and 9 succeed lest his previous a priori strictures on 'necessary existence' be overridden by the exigencies of explanatory completeness.
Do they succeed? In a word, no. Cleanthes' first complaint about Demea's inference to a "general cause or first author" is that nothing which exists eternally can have a cause, since the concept of cause "implies a priority in time and a beginning of existence" (D 190). He provides no further defense of this dictum. Were Demea more a philosopher and less a disciple, he could reply to this criticism in two ways.
First, he could point out that his argument does not require that the causal series be eternal; and, as I have suggested above, it is likely that he would be happier if he could show that it was not. After all, a temporally finite causal sequence intuitively suggests a first cause, whereas a temporally infinite sequence does not. Demea addresses the harder possibility. His point is that even if the causal series is temporally infinite, each member of it, being determined by its antecedents, is unable to serve as the ultimate reason for the existence and peculiar character of the series. Rather than attacking the possibility of there being an infinite series of intermediate causes and effects, his argument is a reductio directed at the inadequacy of a non-terminating causal account. Causal explanation, and hence the causal efficacy to which it refers, must come to an end, even if the number of causes and effects is infinite. If the causal series is temporally endless then this explanatory termination must lie outside time, i.e. the series' very unendingness entails that it have "a cause or reason, as much as any particular object, which begins to exist in time" (D 188).
Here Demea faithfully echoes Clarke, who makes a similar pronouncement, "because Duration in this Case makes no Difference" (Clarke, p. 26). However, where Demea rather abruptly jumps from talk of causes for particular entities to talk of a cause of the whole succession, Clarke is more forthcoming about why the whole requires a cause. He assumes that whatever exists does so either necessarily or by virtue of some cause. He then observes that "no One Being in this Infinite Succession is supposed to be Self-existent or Necessary . . . but every one Dependent on the foregoing." Clarke's point is that the whole series requires some "Original Independent Cause" precisely because "where no Part is necessary, 'tis manifest the whole cannot be necessary" (Clarke, p. 25).16 But if not necessary, it must be caused; and the only way to avoid a viciously infinite regress is to terminate explanation in an uncaused cause. The essential element in Clarke's, and Demea's, notion of 'cause' is not temporal priority but that determinate "power and efficacy" (D 188) which produces any effect. If the causal series happens to be temporally unending, so that every cause depends on another antecedent cause, then there could be no determinate causal efficacy anywhere in the series unless the whole were determined by some atemporal condition.17
In effect, Demea and Cleanthes confront us with different intuitions about what is more important in causal explanation. They agree that causes typically both temporally precede and produce their effects.18 Confronting the limiting case of accounting for the entire temporally infinite history of causal production forces them to sacrifice one of these conditions. Demea is willing to give up temporal precedence to obtain a complete account of production, whereas Cleanthes is willing to truncate productive explanation to retain temporal precedence of cause to effect. Their disagreement echoes an ancient dispute about the appropriateness of seeking a cause for that which is eternal. Aristotle complains that Empedocles failed to provide a reason for the eternal cosmic alternation between attractive and repulsive forces and that Democritus claimed that showing that something always happens suffices to explain it. That is a mistake, he contends, because "it is always true that the angles of a triangle equal two right angles, but there is another cause (or 'reason,' aitia) for this eternal fact" (Physics 8.1 252a26-252b5). Aristotle's point is that this eternally true theorem still requires explanation, because its truth depends on its deducibility from self-evidently true axioms. If using a mathematical case seems a tad sneaky, a physical one can be substituted (although finding possibly eternal physical states of affairs is not easy). Suppose all matter is eternally in motion. From that eternal fact alone it doesn't follow that there is no cause for its actually being the case. Democritus says there is not, but Aristotle assures us that there is, indeed, must be. Demea's point is similar to Aristotle's: the causal series may have always existed, but since it didn't have to exist, there must be some reason why it does. An eternal effect of course requires an eternal cause.19
If the essential feature of causation is productive power, temporal precedence of cause to effect is not required. It is required only if causation is essentially regular succession. Doubtless that is Hume's view; but it is not argued for in the Dialogues, although it may be insinuated in Philo's criticism of Cleanthes' empirical arguments. But if it is to be the basis for rejecting Demea's argument, it does require justification, because temporal precedence of cause to effect is not a datum of experience but a thesis of a particular theory of causation. That is sufficiently well shown by the fact that Aristotle argues that causes must be simultaneous with their effects (Post. Ana. B.12), countless philosophers have supposed that they may or must be simultaneous with their effects, and a few have even supposed that effects may precede causes.20 Cleanthes does not provide any argument for causes necessarily preceding their effects, nor does Hume elsewhere provide a sound argument for their doing so.21
Demea could also avail himself of an et tu quoique response, which circumstance may have been planned by the ironic Hume for the amusement of perspicacious readers. Cleanthes, too, argues for an "Author" (D 143, 155, 163) or "original cause" (D 153) of the world,22 albeit on the basis of apparent design rather than ontological insufficiency. Now he objects that there can be no "general cause or first author" of an infinite causal sequence. So his own argument is secure from his own criticism only to the degree that the universe can be shown to be temporally finite (in other words, he must try to justify the very belief that Clarke thought could be known only by revelation). Indeed, Cleanthes attempts to establish precisely that in Part VI; but the best case he can make is that if the world were eternal, cherry trees, vines, sheep, swine, dogs, and grain would have arrived in various parts of the globe earlier than they in fact did (D 173). Clarke's frank admission that he knows of no sound refutation of temporal infinity seems immensely dignified beside this pathetic argument.
Cleanthes' second complaint about Demea's use of causal inference, in paragraph 9, has both an ontic and an epistemic aspect. On the ontic side, it is that the whole chain of beings is only a product of "an arbitrary act of the mind" and does not independently exist in nature. The epistemic side is that there is no explanatory utility in inquiry into the cause of a whole if the causes of its component parts are known. The epistemic aspect is elicited by an example:
Did I show you the particular causes of each individual in a collection of twenty particles of matter, I should think it very unreasonable, should you afterwards ask me, what was the cause of the whole twenty. This is sufficiently explained in explaining the cause of the parts.
Would the sort of separate causal accounting envisaged by Cleanthes be adequate, or even possible? His claim seems plausible for twenty grains of sand on the beach; but does it work for twenty particles of shell in the body of a mollusk on that same beach? Do shells and shell particles exist in nature or is there only calcium carbonate? Suppose the twenty particles are bits of graphite constituting the tip of a pencil used to line out an infelicitous expression in a manuscript. An account which traced either 'pencil lead particles' or 'line particles' from their present disposition to their first formation would leave out a great deal, notably everything to do with pencils and lines. Have we explained the line on the page when we have collected the causal genealogies of each of the graphite particles of which it is composed? The idea seems patently absurd, as does the notion that those particles are united into a whole by an "arbitrary act of the mind," unless we are to suppose that neither animals nor their artifacts are found in "the nature of things." Although that degree of scepticism may be found in the Treatise, neither Cleanthes nor the sceptical Philo proposes it in the Dialogues for the very good reason that its truth would be as damaging to their views and to common sense as to Demea's argument. Even in the inorganic realm, what is to count as a non-arbitrary "particle of matter"? A chunk? A molecule? But these are wholes. In the end Cleanthes' criticism relies on the truth of material atomism, which also is not argued for in the Dialogues. Further, even if atomism be true, Cleanthes' argument is still not salvaged. Let us take as our atoms a hundred ballots in a ballot box. We might ask why this ballot is in the box, to which the answer might be "because John Smith placed it there." This question could be repeated for every ballot in the box. But surely it is also appropriate to ask why all these ballots, or any ballot, is in the box, to which the answer is "because there is a special referendum today." This explanation is not contained in the explanations of how each ballot got there, indeed those explanations provoke, rather than answer, the question as to why all these individuals are performing similar acts. On the other hand, a satisfactory answer to the question about the whole collection also explains why John Smith et alia placed their ballots there.
Besides, Cleanthes' example is only tangentially relevant to Demea's argument. Demea is talking about a "chain or succession" of causes and effects in which each member causally depends on its predecessors, but Cleanthes instances a "collection" of material particles. He substitutes a primarily spatial whole in which the connection of the parts is arbitrary for a primarily temporal one whose parts are linked together by a single causal history. This move lends his claim what little plausibility it may have, because sometimes it really is "unreasonable" to ask for an explanation of a whole whose spatially contiguous parts have already been explained. But it is not unreasonable to ask for some further explanation of a causally connected series of particulars. That is just what we do when we search for the explanation of an entire series of chemical or physical reactions or, in historical explanation, for a series of human actions. There doubtless is some arbitrariness both in the level of generality and in the temporal point at which we stop, since those depend on what question we have posed. Why should the question as to the reason for the whole universe, assuming the whole to be causally interconnected, be declared out of bounds? If the world is bound together with causal glue, the "arbitrary act of the mind" lies in the inquirer's prying this whole apart into apparently independent parts rather than in constructing it from those parts. Demea concludes there are no independent parts, and Cleanthes gives no argument to prove him mistaken. In presuming causal atomism, he is at least equally guilty of the arbitrariness of which he accuses Demea.
Cleanthes is also oblivious to some reflexive consequences of his criticism. As previously noted, Cleanthes himself argues, not that God is the direct cause of the present state of the universe, but that he is its ultimate cause. Consider his use of the propagating books example in Part III. The particular causes of each volume are its animate or vegetative progenitors, yet Cleanthes wishes to infer the existence of an "original cause" or author (D 153).23 Yet, were the universe "sufficiently explained" by particular causal accounts, there could not be an original cause.24 Hume seems to be ironically toying with his characters again. If Cleanthes' point is sound, the goal of his own theology is logically impossible. We may imagine that he would try to evade this consequence by claiming that whereas explaining purposeful order requires tracing causes until one arrives at an intelligence, explaining sheer genesis need not go beyond the first efficacious existent. But Demea is saying precisely that if every finite existent is conditioned, then no one of them is self-sufficiently efficacious.25 To assume otherwise is, again, simply to presuppose an unjustified ontological atomism.
I have said that the criticisms proposed by Philo are of a different order. In fact, the first one, presented in paragraph 10, does not attack Demea's proof at all! Philo maintains that just as the remarkable properties shared by all integers having a factor of 9 are not due either to design or chance but to mathematical necessity, most probably "the whole oeconomy of the universe is conducted by a like necessity" (D 191). Stove seizes on this passage to contend that Philo takes the material world to exist necessarily, contradicting Cleanthes' arguments, in paragraphs 5 and 6, respectively, that 'necessary existence' is either impossible or meaningless (Stove, pp. 306-7). Stahl correctly notes that Philo does not attribute necessity to the existence of the world but only to its "order," "disposition," or "oeconomy" (Stahl, p. 507). However, he does not draw the obvious conclusion that Philo's remarks, for that very reason, have no bearing whatsoever on the cosmological argument, although they weigh heavily against Cleanthes' design argument. If the order of the world is necessary, an external cosmic designer is not required, nor even permitted.
Kemp Smith tells us that Philo "approves Cleanthes' reasonings" (D 116). That seems to be the common interpretation; but its truth is far from evident. Why does Cleanthes, a fellow theist, strive so hard to refute Demea? Apparently because the cosmological argument leads only to a first cause of the universe, and remains silent with respect to the attribute of purposive intelligence which Cleanthes sees behind the phenomena of nature. Demea's argument relies entirely on causal efficacy; Cleanthes' on purposive agency. Cleanthes wants a God like us. It is this thesis which both Philo and Demea, for quite different reasons, oppose. In the third paragraph of Part II Philo grants that nothing exists without a cause and that the original cause of the universe, whatever it may be, we rightly call 'God.' In the first paragraph of Part IV Cleanthes claims that Demea's negative theology is indiscernible from the beliefs of atheists and sceptics. Cleanthes attacks Demea's argument because he thinks we can know more about God than Demea's conclusion permits. If God is a wholly different kind of cause from us--a necessarily existent, eternally active, cause--then Cleanthes is cut off from interpreting divine activity in terms of human rationality. Obviously Philo does not share this concern; nor does the cosmological argument pose much of a threat to him. It is instructive to remind ourselves that the best known early proponent of this type of argument, Aristotle, found little use for purposiveness in explaining nature.26
Moreover, there are a number of clues which suggest that Philo is only feigning agreement as entrée to renewing his attack on Cleanthes' arguments. Philo's opening remarks in paragraph 10 are quite equivocal:
Though the reasonings, which you have urged, Cleanthes, may well excuse me, said Philo, from starting any farther difficulties; yet I cannot forbear insisting still upon another topic (D 191).
Then he proceeds to offer the argument whose real target, as we have just seen, must be Cleanthes, not Demea. Paragraph 11 begins with "But dropping all these abstractions, continued Philo; and confining ourselves to more familiar topics . . . ," which also could indicate something less than full confidence in all the preceding arguments, Cleanthes' as well as Demea's. True, some of Cleanthes' criticisms echo familiar Humean positions, notably the proclamation in paragraph 5 of what Stove calls Hume's "favorite maxim," the principle that matters of fact are indemonstrable (Stove, p. 303). But does Cleanthes successfully use the maxim to refute Demea? Stove reasonably contends that it is irrelevant against Demea, who regards God's existence as a necessary truth rather than a matter of fact. Stahl defends Cleanthes by observing that "Hume's and Demea's conceptual frameworks are so disparate that any comparison of views is likely to involve begging questions," coupled with the taunt that Demea begged the first question by defining God as existent (Stahl, p. 506).27 Is this not a roundabout admission that Demea is not refuted, as declaring oneself the winner and leaving the battlefield is one way to lose?
Whether or not Hume is signaling that some of these arguments are faulty, it is noncontroversial that he distrusts any a priori argument, or "abstract reasoning," which is not "concerning quantity or number."28 Now, all the arguments of Part IX, prior to the final paragraph, seem to be non-quantitative a priori arguments: Demea's argument, Cleanthes' criticisms of it, and Philo's oblique contribution in the penultimate paragraph. The first of Cleanthes' criticisms is just the enunciation of Hume's "favorite maxim," that matters of fact or existence cannot be demonstrated. Were this principle known to be true, it should be "entirely decisive," as Cleanthes claims. That he proposes further arguments presupposes the persuasive inadequacy of this maxim. In the closing paragraph, Philo reverts to this methodological level to proffer his only observation on Demea's proof. His point, however, unlike Cleanthes', is empirical rather than a priori, more psychological than epistemological, about failure to convince rather than failure to prove.
He claims that a priori theistic arguments typically appeal only to those whose success with mathematical discovery has led them to expect similar success in applying those same deductive methods to metaphysics and theology.29 Most people, however inclined to religious belief they may be, are not convinced by such arguments. This majority is presumably constituted by those whose existential judgments are based on experience rather than abstract reasoning. Thus his remark is, in effect, a psychological explanation of how a few are led to violate Hume's favorite maxim: they have been seduced by formal science. His observation about ordinary folk being suspicious of a priori theological arguments seems accurate (as anyone who has introduced the ontological argument to undergraduates is well aware). Yet a rationalistic metaphysician might retort that the most trustworthy judge of truth is the expert in the relevant discipline, not the masses. But if the results of philosophical theology are as negative as Philo believes them to be, and its very practice a misguided hybris about the powers of the intellect erected on a category mistake, that discipline is bogus and no experts exist.
That these metaphysically inclined reasoners are few and unable to influence many persons suggests a fuller explanation of why Hume does not lavish attention on Demea's argument comparable to that he gives to Cleanthes'. The design argument is more important, not because of any logical superiority, but because it is more seductive to most minds. Its greater persuasiveness rests on its being a response to the question of why natural things are as they are, a question which seems to be a routine, almost inevitable, extension of ordinary causal reasoning. In focusing on accounting for the apparent properties of objects the design argument resembles useful explanations sufficiently to acquire a patina of "scientific" respectability, which certainly contributes to making it more popular, and hence more dangerous. The cosmological argument would never be mistaken for a scientific argument, but the design argument not only could, but has been, so taken. More importantly, the design argument's question is one to which imagination readily suggests answers based on analogy to artefacts. The cosmological argument, by contrast, poses the question of the logical implications of natural things being causally dependent in the manner presupposed in our causal explanations of them, a question incapable of exciting the imagination. The abstruseness which Philo attributes to the argument stems from its being the product of second-order reflection on causality rather than a first-order reflection on things. Its failure to arouse imagination renders the cosmological argument emotively deficient in comparison to the design argument. It should be recalled that the capacity to motivate belief via the "irresistible influence" of feeling and imagination is precisely Cleanthes' last-ditch apology for the design argument in Part III (D 154-5).30 Philo does not there challenge Cleanthes' claims;31 but now he makes clear that he does not think the cosmological argument shares that sort of power.32 I suggest that Hume proportions his discussion to his perception of the arguments' relative potentiality for intellectual mischief.
Interestingly, Philo's thesis that accepting reasoning as recondite as that of a priori theological demonstrations depends on special uncommon psychological factors is the one clearly empirical claim in the whole set of objections (despite its being called a "certain proof" in the final sentence). It also is the only criticism which even Demea is depicted as accepting, confessing in the very next paragraph (the beginning of Part X) that one seeks God "from a consciousness of his imbecility and misery, rather than from any reasoning" (D 193). It echoes the doctrine of The Natural History of Religion, which traces the quest for the divine to anxious "hopes and fears" rather than any "speculative curiosity."33 Although admittedly not the sort of rigorous logical demonstration most philosophers prefer, it may well have been the one Hume found most convincing. Hume, after all, makes reason the slave of the passions; and paragraph 11 goes to the heart of the passional deficiency of the a priori proof--its failure to move one to belief.
If Philo's criticisms fail either to rely on or to support those of Cleanthes, what overall picture can we draw of Hume's purpose in Part IX? If arguments are weak in a treatise one is entitled to accuse the author of careless reasoning; but the case of a dialogue is not so clear-cut. Yet Hume's intentions are not as obscure as Plato's, who could claim that he had authored no treatise to betray his personal viewpoint. Cleanthes unquestionably espouses some identifiably Humean doctrines; but he does so without sufficient justification and in spite of inconsistencies with his overall position. Hume may indeed support the arguments given by Cleanthes, perhaps being unusually careless in his reasoning because he believes Demea's argument to be of a generally discredited sort. On the other hand, there is a subtler option, to which the foregoing discussion lends some support. Perhaps Hume intends Cleanthes' criticisms to recoil on their author. The quality of Cleanthes' reasoning against Demea is little better than that he earlier uses in defense of his own thesis; and, to whatever extent it is successful, it damages his own case. Philo is content to observe a battle in which he has nothing to lose. He then deftly appropriates the idea of 'necessary existence,' introduced by Demea, to further damage Cleanthes' case for tracing cosmic order to a transcendent intelligence and recalls the idea of the emotive forcefulness of argument, introduced by Cleanthes in Part III, to expose the persuasive deficiency of Demea's a priori proof. Philo alone emerges unscathed from the contest of Part IX. Perhaps here the victor is the one who has lost least.
1. A revision of "A Word on Behalf of Demea," Hume Studies XV, 1 (April, 1989), pp. 120-140.. Return
2. Although he does sometimes seem to speak for Hume through the beliefs he shares with Philo and through part of his criticism of Cleanthes' argument in Part III. There is also a sense in which all of the characters speak for Hume, not just because he wrote all their speeches, but because each represents some tendency to belief formation found in us all.Return
3. One of the many ironies in the history of ideas is the fact that, while the design argument no longer evokes the sort of respect it enjoyed in the days of Hume and Kant (despite such modest rehabilitation attempts as that of Richard Swinburne in The Existence of God [Oxford: Oxford University Press, 1979]), the cosmological argument has, of late, enjoyed something of a revival. See, e.g., Bruce Reichenbach, The Cosmological Argument: A Reassessment (Springfield, IL: Thomas, 1972), William Rowe, The Cosmological Argument (Princeton: Princeton University Press, 1975), Keith Ward, Rational Theology and the Creativity of God (New York: Pilgrim, 1982), William Lane Craig, The Kalam Cosmological Argument (London: Macmillan, 1979). Return
4. Plutarch, Vit. Alex. 31; Arrian, Anab. 3.10. Return
5. D. C. Stove, "Part IX of Hume's Dialogues," Philosophical Quarterly 28 (1978), pp. 300-309. In §4 below I register my dissent from the common view that both Cleanthes and Philo express Hume's personal sentiment. Return
6. Donald E. Stahl, "Hume's Dialogue IX Defended," Philosophical Quarterly 34 (1984), pp. 505-507. Return Return
7. Hume, Dialogues concerning Natural Religion, ed. Norman Kemp Smith (London, 1947), p. 141. Hereafter cited as D. It is not inconceivable that Demea does, although rather obliquely, speak for Hume here. Insofar as we are "of common sense" we do not doubt; the "indolence and indifference" of the vulgar prevent their pursuing philosophical critique of their beliefs. See A Treatise of Human Nature, ed. L. A. Selby-Bigge (London, 1888), Bk. I, Part IV, Sect. III, p. 223; hereafter cited as T. It is notable that Cleanthes, too, in Part III, under the pressure of Philo's criticisms, insists that the arguments of natural religion have the same force as "common sense and the plain instincts of nature" (D 154). Return
8. Perhaps this ambivalence between a non-argumentative "mystical" view and demonstrative knowledge was Hume's way of representing the inconsistent methods followed by orthodox religious apology. Return
9. Published as A Demonstration of the Being and Attributes of God: More Particularly in Answer to Mr. Hobbs, Spinoza, and Their Followers (London, 1705). J. C. A. Gaskin suggests that there might have been some influence from Leibniz (Hume's Philosophy of Religion [London, 1978], p. 177, n. 2). Mossner writes that Demea "may be taken historically" as Clarke and Cleanthes as Bishop Butler (The Life of David Hume [Oxford, 1980], p. 319). I am uncertain what being "taken historically" means, but unqualified identification of the doctrines of Hume's characters with those of historical personages seems extreme. For example, Demea sometimes cites, and more often reflects, the views of Malebranche. See John Wright's comment in The Sceptical Realism of David Hume (Minneapolis, 1983), n. 68, pp. 185-6. Return
10. In speaking of an "ultimate" cause Demea clearly means a first cause, which is the ultimate object attained by tracing backwards from effect to cause. Return
11. Together they form an inconsistent set, since if a necessarily existent being is impossible, the material world could hardly instantiate it. Cleanthes here, as in Part III, argues as rhetorician, not as philosopher; he uses a variety of arguments without troubling much about their mutual compatibility. Return
12. Stove provides an example to show that it is coherent: "there necessarily exists a prime number greater than ten and less than twenty" (Stove, p. 305). This example suffers from the weakness pointed out by Stahl (p. 506), that it only asserts the necessary existence of abstract entities. But Demea's argument does not even require unhypothetical necessity; he contends only that a conditioning entity necessarily exists if a conditioned entity or set of entities exists. He does not invoke necessary existence as part of the definition of 'God' but as part—the final step—of a causal explanation of what is. Return
13. Hartshorne has repeatedly argued, I think correctly, that the idea of a "contingently-existing God" is both religiously and metaphysically unintelligible (see, e.g., The Logic of Perfection and Other Essays in Neoclassical Metaphysics [LaSalle, 1962], pp. 41, 89-117). God is either necessary or impossible. If all objects are contingent, then no object is God. I doubt that Hume saw this, even if he would have enjoyed the irony of Cleanthes yet again damaging his own case. Return
14. This point is akin to one that Philo makes in paragraph 11. A priori arguments, whatever their logical merits, often simply fail to convince. Philo's observation is about Demea's proof, but Cleanthes' contention about necessary existence is equally a priori. Return
15. Whether 'necessary existence' is coherent is at least disputed by philosophers; but would anyone champion 'colored point'? Points are the unextended, imperceptible, limits of one-dimensional extension. Color is a perceived quality of extension in two or three dimensions. Need more be said? Return
16. He is far from clearly right in this. Formally, the argument commits the fallacy of composition. We cannot infer that a "whole," say, mankind, is perishable because all men are perishable. Metaphysically, it would seem crucial whether the parts were finite or infinite, and whether the whole is merely the collection of the parts or has characteristics of its own. If the whole is only a finite collection of parts each of which is perishable, then it would seem the whole must be perishable as well. But Clarke (and Demea) allows that the succession may be infinite, which would seem to permit, or even require, that the succession be analogous to Aristotle's view of time, according to which the whole is necessary (in at least one sense; see the following note). Fortunately, the issue of the ontological status of the actual infinite need not be settled to deal with Cleanthes' criticism, since he does not entertain considerations such as these as part of his argument. That he does not is an even more surprising omission in the context of ¶7, where he suggests that the world itself might be necessary, since that supposition directly challenges Clarke's analysis. Return
17. If power and efficacy are subjugated to temporal succession, the argument seems more tenuous. Consider the relationship of part to whole in an Aristotelian view of time. For Aristotle, every part of time must perish but the whole of time is necessary precisely because each part must have an antecedent and a successor. But this analogy points up an ambiguity in "the whole," since the totality of temporal sequences is never simultaneously actual. In Aristotle's own argument for an unmoved mover cause and effect are assumed to be simultaneous. It is the manifest actuality of the effects in this simultaneous sequence of "power and efficacy" which is taken to imply an actual first efficient cause operative at this moment (and so at any moment). The introduction of temporal succession into the concept of causation weakens the argument, since then the cause of a present actuality would reside in some immediately past actuality, and so on, ad infinitum. If an infinite past is necessary and is the source of causal power, then the totality of previous events could be the sought-for necessary being. God would be the past. Return
18. Inasmuch as both accept the second of these two conditions, neither espouses a "Humean" view of causality. Return
19. The argument as such is not specific enough to identify that cause. In Part VIII, Philo revives a version of Epicurean cosmology as an alternative to Cleanthes' teleological theory. In the course of expounding that view, he responds to Demea's suggestion that motion implies a first mover by claiming that it is just as likely, a priori, that it is inherent in matter (D 182-3). Philo's suggestion clearly applies better to Cleanthes than Demea; for although it would not strictly rule out a first cause, since the first cause might be matter itself, it would rule out a transcendent, purposive or intelligent first cause. Philo touches on this theme again in ¶10 of Part IX, suggesting that the present disposition of things might be necessitated by something in the "intimate nature of bodies." If being the cause of motion and all the forms brought about thereby qualifies it to be "cause of this universe," then, by Philo's definition of 'God' in Part II, matter would be God. Return
20. This latter position is typically prompted by consideration of Newcomb's Problem and similar paradoxes. Return
21. Hume does give an argument at T 76, but it is not sound. Even though he subsequently adopts a position which is in spirit Humean, Robert Fogelin shows this argument for the temporal precedence of cause to effect to be quite defective ("Kant and Hume on Simultaneity of Causes and Effects," Kant-Studien 67 (1976), pp. 51-59). Return
22. The latter term is introduced by Philo in Part II. Return
23. In Part III Cleanthes makes clear that he is seeking a single intelligent cause of apparently purposive phenomena, despite having proximately causal explanations of each phenomenon; see Dye, "Superhuman Speech and Biological Books," HPQ 5:3 (1988), pp. 263-4. If this is permissable for explanations of teleological order but not for explanations of existence, Cleanthes neglects to provide any justification for this disparity between the two cases. Return
24. There could be a first particular cause, i.e. the cause of the first effect, if one could prove that the series of causes and effects was temporally finite. We have already seen that Cleanthes doesn't do very well at that. Return
25. Cleanthes' criticism in ¶7, that were something to exist necessarily that something might be the material world itself, would be harder for Demea to answer. Surely the contingency of parts need not imply the contingency of the whole, as Demea assumes, any more than the movement of the parts of a clock implies the movement of the clock. But since Cleanthes rejects the idea of necessary existence, and in ¶9 rejects the world (the whole), he must assert the ¶7 criticism only counterfactually. Even more ironically, this criticism of Demea's argument resembles the earlier criticism Philo made of him in Part VIII (see note 0 above). Cleanthes could hardly be the sincere spokesman for a thesis more obviously destructive of his own view than Demea's. Which is too bad, since it is the best he has to offer. Return
26. There are exceptions, notably the movements of the planets and stars. His natural philosophy remains basically mechanistic. The teleological explanation he uses is non-rational, unless the phenomena being explained are the acts of rational animals. However, there seems no reason to believe that Hume knew much about Aristotle. Return
27. The taunt is false. Demea does not define God at all and seems to think that doing so exceeds our limited intellectual abilities. He gives a cosmological, not an ontological, argument. The attribute of necessary existence is determined by the exigencies of explaining the existence of the causal series, not from definition. Return
28. Enquiry concerning Human Understanding, ed. Selby-Bigge, 3rd Ed. (Oxford, 1975), Sect. XII, Part III, pp. 163-5. Return
29. Here it is difficult to suppress the idea that Hume must be thinking of Leibniz as well as Clarke. Return
30. One is reminded of Kant's assessment of the design argument as the one most in agreement with "the common reason of mankind" and able to overcome "doubts suggested by subtle and abstruse speculation" (Critique of Pure Reason, A624, B652). Return
31. The task of responding falls to Demea, who points out the religious deficiencies of Cleanthes' anthropomorphism. Perhaps part of Hume's dramatic intention is that Cleanthes, by leading the assault on Demea in Part IX, is "getting back" at Demea for his criticism in Part III. Return
32. The text does not suggest that Philo requires persuasiveness as a sufficient, rather than only as a necessary, condition of a good theological argument. So his position here does not support those who infer from his failure to respond in Part III (as Demea is silent here) that he accepts Cleanthes' "irregular argument," which is rhetorically persuasive but not cogent. Return
33. The Natural History of Religion, II. Return
Last Updated 6 September, 1998