I. The Text. The two examples are both highly imaginative, even bizarre, thought experiments. The first is "that an articulate voice were heard in the clouds, much louder and more melodious than any which human art could ever reach" which "extended in the same instant over all nations, and spoke to each nation in its own language and dialect" in words which conveyed "some instruction altogether worthy of a benevolent Being, superior to mankind" (D 152). The second is "that there is a natural, universal, invariable language, common to every individual of human race; and that books are natural productions, which perpetuate themselves in the same manner with animals and vegetables, by descent and propagation" (D 153).
Cleanthes insinuates that no one, not even a sceptic, could fail to suppose that the causes of such events were intelligent and purposeful. He then argues that natural phenomena, especially anatomical structures, such as the the eye, manifest design still more strongly than do these imaginary instances. In effect, phenomena are God's utterances or writings. Finally, he adds that the theistic inference from the observation of natural "contrivance" is "universal" and "irresistible," i.e., presumably, that it has the same sort of force he has attributed to his two examples.3 This alleged universal and irresistible influence of the theistic argument is cited to show that the inference is legitimate. However, he admits that the argument may not convince "an ignorant savage and barbarian" because of his "stupidity." As Philo's problem is not stupidity, but "too luxuriant a fertility" in inventing objections, he supposedly cannot escape the force of the argument. At this point, Philo is said to be "a little embarrassed and confounded" (D 154-5).
II. The Commentators. The principal common feature of most commentary on this passage is agreement that Cleanthes has presented some form of inference which Hume would take to be legitimate, although there is disagreement about just what portion of his case is sound and about the effectiveness of the argument as a whole. A brief survey of some representative views will help delineate the range of previous interpretation.
Kemp Smith claims that Hume has Cleanthes propose the superhuman voice "example" principally to show Cleanthes evading his obligation to give a cogent response to the criticisms of the design argument put forth by Philo in Part II. Philo does not respond to the illustration because it is "both irrelevant and misleading"; but, considered as reasoning, it is, "under the supposed circumstances, . . . a legitimate inference by analogy" (D 101). He is less charitable in assessing the breeding books example, seeing it as "very bewildering" and as illustrative principally of the difficulties involved in asserting the similarity of organisms to artefacts. "Organization" (i.e. biological or organic process) is a different principle from that "project and forethougt" required to produce articulate speech or a book. Cleanthes confounds these incompatible principles—a real book must be produced by intelligent purpose, so to call a biologically propagated entity a 'book' is "to insist on a resemblance while refusing the conditions which can alone make it relevant" (D 103).
Nelson Pike also takes the supposed voice from the clouds to be the basis of a legitimate inference. But he does note that Cleanthes himself observes that Philo's objections to the design argument would hold equally against the attribution of an intelligent cause to the voice, given that "its loudness, extent, and flexibility to all languages, bears so little analogy to any human voice" (D 153).4 He accordingly sees Cleanthes as retracting his claim that the design argument rests on inferences like the "arguments from experience" found in the sciences. Rather the inference is, as Cleanthes says, "irregular," and this irregularity consists in the conclusion's being "drawn directly from the data" instead of from "an empirically established correlation between classes" (Pike, 229). Pike emphasizes that the inference is nevertheless "rational," even if it is not the same kind of inference found in scientific inductions (ibid.). Pike does not find in Hume any more details about what other characteristics this form of inference may have beyond its being "of a direct and immediate sort," nor does he offer any additional characterization of his own (Pike, 230-3). He nevertheless goes on to claim that Philo accepts this argument as legitimate and that this acceptance provides the basis for Philo's so-called "confession" in Part XII (Pike, 233-4). So, presumably Hume is supposed to have thought the argument as a whole to be successful.
P. S. Wadia takes both illustrations as presentations of legitimate inferences. He also takes them to be at least partially relevant to Philo's critique. Philo has contended that the belief that the world is a product of intelligent design could be justified only by prior observation of the processes of world production, but Cleanthes here successfully produces species of experience "within our world which would force even a sceptic like Philo to posit an intelligent Source who controls (at least some of) the laws of nature."5 Wadia agrees with Kemp Smith's assessment of the superhuman voice illustration as "a legitimate inference by analogy." What he adds beyond anything said by Kemp Smith or Pike is the claim that such events would be miraculous, although they are logically possible. Were they to occur, they would provide evidence for design which Philo could not reject (Wadia, 288-9).
Stanley Tweyman contends that Wadia is mistaken in supposing Cleanthes to be appealing to the miraculous, because Cleanthes classifies the inference to cause in these imaginary instances as of the same sort as the inference to an intelligent cause of the world, which he, in Part I, has characterized as scientific. So the inference to a cause of the "Vegetable Library" is also of a "scientific character."6 Yet, in another article, Tweyman stresses that Cleanthes presents two kinds of argument in these illustrations—"an inference by analogy, and an immediate inference or natural argument . . . ."7 If Tweyman is consistent in these two articles, he must be referring solely to the analogical argument when he maintains, against Wadia, that Cleanthes' arguments are supposed to be of a scientific nature. He allows that the articulate voice from the clouds may sufficiently resemble human voices to permit our inferring an intelligent cause of it, although Philo does not try to resolve that question. Whatever Philo's views about the soundness of an analogical argument to an intelligent cause of the voice, he consistently rejects the analogical argument to an intelligent cause of the world (Tw 270). However, Cleanthes also claims that we would immediately, instinctively, (i.e., for Tweyman, non-analogically) affirm that the voice was purposeful. Here, and in the following illustration, Cleanthes uses "natural arguments" which are not "rational arguments." These arguments attain their conclusion, not by logical constraint but by the force of "our natural tendencies." Philo is "far more sympathetic" to this latter kind of argument (Tw 268). Tweyman goes on to speculate as to whether Cleanthes' appeal to natural arguments "accords with" Hume's view of "natural belief in connection with causality, physical objects, and the self." He concludes that it is "far from obvious" that this is so (Tw 273-4).
III. Proposed Interpretation. Contrary to the prevailing view, I think Cleanthes presents no legitimate inferences or reasoning towards which Hume would be kindly disposed. Therefore, (1) I shall attempt to show that these examples are incoherent and that Cleanthes' argument is entirely rhetorical. It would be unfortunate if Hume, in the work he may have regarded as his finest, should have either mistaken a bad argument for a good one or inserted such a remarkable passage for the relatively trivial purpose of portraying Cleanthes reacting irrelevantly and evasively to Philo's previous criticism. Accordingly, to avoid these disagreeable conclusions, I shall (2) present evidence that Hume does not accept the sort of reasoning presented by Cleanthes and (3) suggest an explanation of the purpose of the passage which rescues it from being nothing more than a dramatic interlude depicting the confusion produced by Philo's sceptical critique.
It is difficult to understand why so little objection has been raised against Cleanthes' voice in the clouds illustration. If the voice is, as claimed, in the clouds, and not just in the imaginations of the auditors, it is a real sound, a physical and not merely a psychological event. However, sound takes time to travel, whereas Cleanthes' voice is propagated instantaneously over the entire globe. Perhaps Cleanthes does not literally mean "in the clouds" but just 'from overhead'; and perhaps the entire atmosphere, rather than a specific area, is the source of the voice. But there is more. Cleanthes does not say, as the Acts of the Apostles says of the apostles' speech at Pentecost, that each auditor hears in his own language, but rather that the voice simultaneously speaks to every nation in its own language.8 That change takes us from the realm of the marvelous to the realm of the absurd. Voices utter words, not bare meanings; and the word 'green' is not the word 'vert.' Apparently we are being asked to imagine that the same voice, at the same time, says "aimez vous ton prochain" to a Frenchman and "du sollst deinen Nächsten lieben" to a German standing across the border, only a few feet away. Can this be consistently done?
Suppose our German and French auditors were both recording bird songs when the event occurred. What would be registered by their two machines? (Anyone bothered by the anachronistic character of this thought experiment is free to substitute the vibratory movements of membranes, tiny bones, and nerve fibers for the analogous electrical and magnetic events.) The whole of our knowledge of electronics and acoustics requires us to believe that two properly functioning recorders fed by microphones of similar sensitivity located only a few feet apart, with no physical barriers between them, will record the same distant sounds, save for predictable phase shift. If they do, there is no affront to logic, but the recordings are of the same language. At least one listener was certainly only "hearing things." If they do not, matters are more complicated. The only coherent assumptions would be that at least one piece of equipment had malfunctioned, or that at least one tape had been altered since the simultaneous recording. For if we suppose the equipment to have been in good order and the tapes to be unaltered, the very criteria for the identity of 'a voice' will have to be sacrificed. A voice is a variety of sound. 'A sound' (in the physical, not the psychological sense) means vibrations of an appropriate medium, at the appropriate frequencies and amplitude to be detected by some organism's auditory apparatus, propagated in the manner specified by the acoustical laws. Should voices regularly manifest a "translation effect" when crossing national frontiers, the nature of air and the acoustical laws would have to be such as to provide for that effect. In the end, the whole fabric of nature would be radically different from the way it in fact is, and 'a voice' would not mean what it in fact does. There may be a possible world in which there are phenomena, 'noices,' which are just like voices except that they manifest the translation effect.9 However, this cannot be the case in the actual world, since the physical conditions for 'voices' and 'noices' are incompatible. 'A voice' names a phenomenon which does not and could not conceivably behave as Cleanthes suggests. Cleanthes could with equal propriety have asked us to imagine a pink voice as to imagine an omnilingual voice.10
Therefore, even Wadia's assessment of the event as miraculous seems overly charitable, for that assumes it to be conceivable, although in violation of known laws of nature. Had Cleanthes asked us to suppose a celestial voice speaking a single language, or many celestial voices each speaking a different language, that could be done (provided our imagining did not extend very far into the details of the event's causal structure); but to conceive a voice simultaneously expressing the same sentiment in myriad tongues cannot be done. Unless I am missing something, the illustration is logically impossible because the supposed event would have contradictory acoustical properties.
A philosophically sophisticated reader might suggest that what I have missed is that Cleanthes intends to assert only a single voice-type, individuated from other types of voice by properties, such as tonality and resonance, which normally serve to identify an individual voice, regardless of the language spoken. There are tokens of this voice-type for every language and dialect, each expressing the same sentiment and seeming to emanate from the sky. On this view, neighboring French and German auditors both hear the same (type of) voice simultaneously speaking in two languages.
Although this involves no obvious physical impossibility, it seems an unnatural reading of the text, which hypothesizes the existence of "an articulate voice . . . in the clouds." The indefinite article permits English, unlike languages with no equivalent device, to refer unambiguously to individuals rather than types.11 Even cases like "That's a Mercedes" or "That's a voice, not rustling leaves" identify an instance as of a certain type. Cleanthes continues by referring to "this voice," which also normally picks out instances rather than types (which is why Aristotle calls individual things 'thises'). But since "this voice" sometimes means the voice of a certain person rather than any particular instance of that person's speaking, let us suppose, momentarily, that that is what Cleanthes intends. Both logical and epistemic difficulties remain.
If "this voice" is one as a person's voice is one, it cannot enunciate several distinct words at once. Call types which cannot have several simultaneous tokens "sequentially instantiated." For example, 'sunrise' is sequentially instantiated. The sun cannot rise in the same instant over all nations; and were it reported to have done so, it would be certain that the experiences of all observers other than those located within a small longitudinal range were non-veridical. A person's voice is also sequentially instantiated. To bring Cleanthes' case down to earth: suppose one were to hear, in a darkened room, several simultaneous utterances, completely indistinguishable in vocal quality, expressing the same proposition in different languages. This would be a most remarkable and puzzling circumstance, but surely one would not be tempted to explain it by saying that the utterances were by one and the same voice.12 Relocating the sounds to the clouds and increasing their loudness and melodiousness only weakens our identification of them as of the type 'voice,' as Cleanthes candidly admits.
If the tokens are not of one person's voice, but are type-identical only in having the same meaning and audible qualities, the logical problem gives way to an epistemic one. The circumstances must entitle one to infer a single intelligent cause. The conditions for inferring a single speaker being violated, is it nevertheless obvious that all the multilingual speakings have one intelligent cause? No. Extra-terrestrial beings (or 'gods' in the ancient Greek sense) may have decided to instruct us in concert. Further, intelligible utterances need not imply intelligent causes at all. If a crowd "speaks with one voice" its unanimity is ultimately due to non-intelligent psychological or environmental conditions.
Given an event so extraordinary and unprecedented, it would be unreasonable to jump to any explanation. Cleanthes' jump to a supernatural explanation additionally involves an insuperable logical difficulty. Philo would agree that every existence has a cause (D 142), and, despite Cleanthes' insinuation, would never accept an "accidental whistling of the winds" as a final explanation any more than he would "divine reason or intelligence" (D 153). To be a cause, however, is to be an antecedent existence of a sort universally experienced as the apparent source of a given effect. Cleanthes would extend this concept, which applies only between objects of possible experience, to a principle which cannot possibly be an object of experience. His inconsistency can be expressed as a version of the Liar Paradox. The reason for supposing that a supernatural intelligence is the source of the voice is that like events have like causes. But if the voice is caused, it has a natural explanation as the effect of some antecedent existence. So, for a supernatural intelligence to be the source of the event, the event must be uncaused. But if the voice is uncaused, there is no reason to suppose a supernatural intelligence as its source. The causal tie which justifies Cleanthes' inference is annihilated by that inference, so if the inference succeeds it fails.
All told, I think it best to take Cleanthes au pied de la lettre about "a voice" and "speaks" because Hume is not normally a careless writer, and it would have been easy to make the example explicitly refer to many simultaneous voices or to the simultaneous hearing of the same message by people of all nations. That the text does not explicitly propose one of these apparently miraculous13 but possible alternatives suggests that Hume is trying to say something about the nature of popular arguments for theism. Might not his intention be to illustrate that their force is so independent of logical consistency, so much a matter of imagination and emotion, that they can seduce one even to accept purposive explanations for impossible events? If so, he has succeeded all too well, given the many otherwise erudite readers who have systematically overestimated the cogency of this example.
What of the bizarre business of the biological books? This example is introduced with the comment that it will "bring the case still nearer the present one of the universe" (D 153). The commentators do not explain this remark. Cleanthes' point must be that the books are regularly produced through natural processes, unlike the singular event of the voice. As we have seen, the voice example is certainly paradoxical and probably physically impossible. In a world where books propagated, there would be some natural law of books, for example, book offspring could replicate or resemble their parents. An Iliad would beget Iliads or something Iliad-like, and an Aeneid would beget other Aeneids.14 So understood, the example contains no exceptional or miraculous events. Is it coherent in other respects? This question can only be answered by considering how the illustration is used.
Cleanthes says that it supports an inference, not to the immediate cause of any given biological volume, but to "its original cause" or "the first formation of this volume in the loins of its original parent" (ibid.). Therefore, the illustration is a special application of one form of cosmological argument—the argument to a first cause. Because he fails to attend to the distinction between proximate and first cause, Kemp Smith's criticism—that Cleanthes commits a category mistake in calling something propagated a 'book'—goes somewhat astray. Cleanthes is not confusing propagation with the thought processes of an author, he is merely substituting propagation for printing or book manufacture. The illustration is supposed to force us to admit that there would have to be an intelligent author of propagated volumes as surely as there are intelligent authors of printed volumes. In other words, whether the proximate causes are natural (and secret) or artificial (and known) should have no bearing on the inference to an intelligent original cause.
But why are we sure that printed volumes have intelligent authors? We are confident that the causal history of a book can be traced back through the manufacturing process to an author because the publication process is our own artifice and we know that it is unable to create the "refined reason and most exquisite beauty" found in books. Were books propagated naturally, this inference to an author would be unwarranted, since, for all we could know, literary production might be a power present in biological generation. Even if one felt impelled to say that the books manifest intelligence, it would be more plausible to locate that intelligence in the books themselves (if they are more like animals) or in their parent plants (if they are more like fruit) than in a remote first cause. "Like effects, like causes" is the principle of Cleanthes' entire argument. Now he supposes an effect radically different from those standardly observed to arise solely from biological processes. Were that effect a customary outcome of those processes, as it would be according to his supposition, we would not hold our present conception of the capabilities of those processes. The observed effects being radically different from present experience, their observed, proximate, causes would be conceived to have causal powers proportionate to those observed effects. So, this example, too, is ultimately paradoxical. The inference to an intelligent first cause is superfluous and without foundation unless propagation alone is unable to create intelligent discourse; but that is known to be so only if books are not bred. Thus the example presupposes both that books are bred and that they are not bred.15
In addition to this internal incoherence, the example is inconsistent with other views Cleanthes espouses. In Part IX he makes three objections to Demea's use of the argument to a first cause. Although the first objection, against the concept of a necessary being, is not relevant to the argument of Part III, the other two objections are relevant. The second is that, given our ignorance of the qualities of matter, it may, for all we know, exist necessarily (D 190). To broach this topic calls to mind Locke's famous argument that "God can, if he pleases, superadd to matter a faculty of thinking . . ."16 If we are ignorant of the qualities of matter, might not the biological books be expressing their own thought in the marks on their pages as we express our own thought in our speech? The third objection is that "it is absurd to inquire for a general cause or first Author," if "the particular causes of each individual" have been identified (D 190-1). But that is precisely what Cleanthes is demanding here, since he has stipulated the particular causes of each volume, namely its procreation by its proximate parent. Either Cleanthes is unwittingly inconsistent, or he does not himself regard the example as representing a logically sound inference and is using rhetoric insincerely.
If these examples furnish no rational support for theism, does his argument as a whole provide some such support, despite the difficulties with the illustrations used? The form of the argument is a fortiori. The tie between the examples and natural phenomena is just that the latter are said to display "many stronger instances of design." Even among informal arguments, a fortiori arguments are odd. One is inclined to wonder how they can affect belief at all. If a given set of circumstances provides insufficient grounds to necessitate a certain conclusion, how can the inference be strengthened by comparison to analogous situations in which the same conclusion is less strongly supported? At best, the parallelism must provide only rhetorical, not logical, reinforcement.
But Cleanthes seems uninterested in logical inference. When he turns from his examples to the "argument for theism" itself, he asserts that "the arguments of natural religion" have the same force as "common sense and the plain instincts of nature." He speaks of "the most obvious conclusion," of "natural and . . . convincing arguments"; and he asks Philo "if the idea of a contriver does not immediately flow in upon you with a force like that of sensation" upon anatomizing the eye (D 154). In addition to this talk of instinct and feeling, he explicitly concedes that the theistic argument may be "contradictory to the principles of logic." He maintains that "its universal, its irresistible influence proves clearly, that there may be arguments of a like irregular nature" to literary effects which work despite their seeming to break the rules of good style (D 155). This corroborates my suggestion that Hume has put logically incoherent examples into Cleanthes' mouth in order to show that theistic arguments have a curious force to persuade even if they are completely irrational. In fact, Cleanthes' final appeal to universal consent to justify the theistic inference does not depend on the examples and could as well have been made without them. Moreover, it is no sooner made than qualified. The argument may not work on "an ignorant savage and barbarian," because "he never asks himself any question" about the original cause of experienced objects. If one is an ignorant savage if one does not believe that there is a terminus to the sequence of causes and effects, then many of the best natural philosophers, from Democritus onward, must be so classified.
In Part III Cleanthes finds no additional fault with Philo's denial that a similarity between artefacts and natural phenomena sufficient to license belief in a cosmic designer can be derived from valid causal reasoning. Therefore, it seems that his claim that "similarity of the works of nature to those of art . . . is self-evident and undeniable" (D 152) must signal some alternative way to establish belief in a cosmic designer. Were Cleanthes' new tack classifiable as a kind of reasoning Hume has previously recognized, surely that would shed some light both on his purpose and on Hume's likely assessment of his achievement. I think it can be classified in terms of distinctions made in the Treatise.
For Hume, all beliefs arise from custom, but custom gives rise to belief in different ways. The "natural" way is through a uniform past experience of the conjunction of two objects, so that the present experience of one readily brings the other to mind. Another way is through education, a repetitive exposure to ideas alone, without any direct experience of the objects those ideas represent. Ideas frequently presented may so impress themselves on the imagination as to produce beliefs as strong as those derived from abstract reasoning or experience, even though that method is not "recogniz'd by philosophers."17 He also thinks there are swifter ways to enliven a idea sufficiently for it to become a belief, namely by exciting the passions or impressing the imagination. Here is Hume's description of how the passions do this:
Admiration and surprize have the same effect as the other passions; and accordingly we may observe, that among the vulgar, quacks and projectors meet with a more easy faith upon account of their magnificent pretensions . . . The first astonishment, which naturally attends their miraculous relations, spreads itself over the whole soul, and so vivifies and enlivens the idea, that it resembles the inferences we draw from experience (T 120).
His comment on imagination is even more interesting:
a vigorous and strong imagination is of all talents the most proper to procure belief and authority. 'Tis difficult for us to withold our assent from what is painted out to us in all the colours of eloquence; and the vivacity produc'd by the fancy is in many cases greater than that which arises from custom and experience. We are hurried away by the lively imagination of our author or companion; and even he himself is often a victim to his own fire and genius (T 123).
Is this what Cleanthes is intimating when he says he will use "illustrations, examples, and instances, rather than . . . serious argument and philosophy"? I think so.
Consider the evidence. Contrary to some scholarly opinion, Cleanthes presents no good arguments. He does paint striking images "in all the colours of eloquence." He compares his arguments to literary devices which "animate the imagination, in opposition to all the precepts of criticism" (D 155). He ultimately appeals to universal consent, but excludes the "ignorant savage," i.e. he tacitly acknowledges that a certain education is required to reach his conclusion. An essential part of that presupposed education is the belief that there must be an "original cause," a belief which Cleanthes later, in a more philosophical mood, argues against. All this strongly suggests that he is here speaking with the vulgar.
The treatment in the Treatise of what Hume calls "unphilosophical probability" provides the key to his attitude towards Cleanthes' remarkable illustrations. In discussing unphilosophical probability "deriv'd from general rules" Hume makes a distinction between the ways in which vulgar and philosophical thinkers are apt to be affected by partially resembling circumstances. He writes,
When an object appears, that resembles any cause in very considerable circumstances, the imagination naturally carries us to a lively conception of the usual effect, tho' the object be different in the most material and most efficacious circumstances from that cause. Here is the first influence of general rules (T 150).
Exchanging the terms 'cause' and 'effect' in this passage to reflect the fact that Cleanthes is arguing inversely from effect to cause yields a perfect description of how his examples work. Voices and books normally require intelligent causes, and since the hypothesized phenomena are like ordinary voices and books in many respects, one cannot help imagining an intelligent cause. Hume continues,
But when we take a review of this act of the mind, and compare it with the more general and authentic operations of the understanding, we find it to be of an irregular nature, and destructive of all the most establish'd principles of reasonings; which is the cause of our rejecting it. This is a second influence of general rules, and implies the condemnation of the former. Sometimes the one, sometimes the other prevails, according to the disposition and character of the person. The vulgar are commonly guided by the first, and wise men by the second (T 150).
There are three points of significant similarity between what the Treatise calls "the first influence of general rules" and Cleanthes' procedure:
1. Hume contrasts the "first influence" with "more general and authentic" reasoning, and Cleanthes emphasizes that the theistic inference is immediate, while objections to it require "time, reflection and study" (D 154). The Treatise makes clear that Hume does not take the first idea to come to mind to be the best.
2. This "first influence of general rules" is described in exactly the same terms Cleanthes applies to his arguments—both are "of an irregular nature."
3. Cleanthes' opening disavowal of philosophy and closing invocation of the popular acceptance of the theistic inference clearly indicate that his discourse should be ranked as the sort which, according to the Treatise, most appeals to "the vulgar."
Given these parallels, that Cleanthes' reasoning is supposed to be of this kind seems certain.
Consequently, I cannot agree with Tweyman that Philo would be "more sympathetic" to these "natural arguments" than to the earlier pseudo-scientific arguments. In the Treatise the sceptics' reaction to this dialectic of rule-governed inference is only to "have the pleasure of observing a new and signal contradiction in our reason" (T 150). Of course, Philo, like Hume himself, seems to be a "moderate," rather than a radical, sceptic. This is doubtless why Cleanthes tries to put his arguments on a par with "common sense and the plain instincts of nature," towards which moderate sceptics have a sort of natural piety. But just saying that arguments are 'natural and convincing' doesn't make them so. In fact, Hume accepts some natural or instinctive inferences but rejects others. He countenances those that are universal and necessary. It should be remembered that reason itself is "nothing other than a wonderful and unintelligible instinct" (T 179). Our beliefs in causation, external existence, and personal identity do not admit of rational proof but are necessary for all subsequent reasoning in science and for the practical activities of life. Tweyman, too, observes that Cleanthes fails to show that theistic belief is one of these so-called 'natural beliefs.' I agree; but then whence this belief?
Hume does write of other natural inferences which, though universal, are avoidable, because they are required neither for science nor for common life. The "irregular" and "destructive" operation of the imagination we have just been examining in the Treatise is said to occur "naturally"; but whether it leads to the formation of a belief depends "upon the disposition and character of the person." The Natural History of Religion recognizes a "universal tendency amongst mankind to conceive all beings like themselves, and to transfer to every object those qualities, with which they are familiarly acquainted," which can nevertheless be corrected by "experience and reflection."18 Cleanthes' argument trades on both of these natural, but undesirable, tendencies of human thought.
Therefore, Hume would not allow that there are "legitimate inferences" in Cleanthes' arguments, as most commentators assert. Cleanthes himself notes that Philo could raise his usual objections against the voice supposition. Hume regards irregular inferences as "destructive of all the most establish'd principles of reasonings" and thinks that only the vulgar, rather than the wise, are "commonly guided" by such thought (T 150). In the oft-quoted letter to Gilbert Elliot, Hume confesses that he has not been able to make Cleanthes' argument "formal and regular" and that "the propensity of the Mind towards it, unless that Propensity were as strong and universal as that to believe in our Senses and Experience, will still, I am afraid, be esteem'd a suspicious Foundation."19
In this same letter to Elliot, Hume also writes, "The Instances I have chosen for Cleanthes are, I hope, tolerably happy, & the Confusion in which I represent the Sceptic seems natural." If the foregoing contentions are sound, the instances are not "tolerably happy" because they are part of a sound argument. Consequently, the naturalness of Philo's confusion cannot reasonably be supposed to stem from his being impressed with that argument. Understanding the point of Part III comes down to finding a good reason for Hume's being happy with Cleanthes' bad argument.
There is such a reason. In Part II Cleanthes had presented arguments of a pseudo-scientific, quasi-Newtonian, sort. Philo asserted the Humean theory of causation and showed that, on that understanding of causes, no legitimate inference to the nature of an unexperienced cause of a singular phenomenon (the world) was possible. Thus Cleanthes' pretensions to scientific procedure in his theistic reasoning were undermined. In Part III, then, he is falling back on the one alternative left to him. He is appealing directly to our natural tendency to imagine anthropomorphic and familiar causal machinery behind any phenomenon whose causes are hidden to us.
This agrees with Hume's view about the nature of religion. Religion arises out of just this tendency, which indeed is the point of his introducing that notion in the Natural History of Religion. Rational religion is always an ad hoc attempt to furnish additional justification for beliefs previously arrived at by irrational means. Cleanthes has already admitted as much at the end of Part I, where he claims that it is "natural . . . for men to embrace those principles by which they can best defend their doctrines" and, further, that the truth of our principles should be judged by whether they lead to "the confirmation of true religion, and serve to confound the cavils of atheists, libertines, and freethinkers of all denominations" (D 140). Following this program, I suggest that he first tries "scientific" theology; and when that fails he then tries rhetorical, popular, theology. Hume is happy with the instances because they are good examples of the kind of argument religionists commonly, and successfully, use to persuade the vulgar.
But why is Philo, who is certainly not among the vulgar, "a little embarrassed and confounded"? There are a couple of plausible explanations. Philo holds that "nothing exists without a cause" and is perfectly willing to call the world's cause 'God.' He even thinks we "justly ascribe" intelligence to this cause because this is the highest perfection we know (D 142). Both ascriptions are vacuous as we are entirely ignorant of the true nature of this cause and can give no reasoned justification for so characterizing it. If Tweyman is correct in supposing that Cleanthes abandons the argument from analogy in appealing to common sense and instinct, Philo may simply be dumbfounded because this seems like an unwitting capitulation to his own position that such attribution cannot be justified. This suggestion is weakened if we understand the appeal to instinct still to affirm the analogy (as seems plausible, contrary to Tweyman), despite the insufficiency of the analogy alone to support the theistic inference.
I prefer to think that the explanation lies in Cleanthes' thinking that instinct can establish the nature of that rather scholastic object—an "original" or first cause. Philo had already suggested that Cleanthes was bringing into "the schools" what should be left "in the temple" (D 142). Now Cleanthes proposes to attain his pre-determined goal by following the first promptings of the imagination rather than considered judgment, rejecting rational critique as irrelevant and unnecessary. Since all philosophers can do to undermine a belief is to show that belief to be theoretically or practically irrational, Philo may be confounded just because he is a philosopher. If the savage is too ignorant and stupid to appreciate Cleanthes' argument, Philo is too learned to respond to it. Neither rhetoric nor dogmatic theology is his vocation.
Accordingly—and the commentators seem to have overlooked this—, Hume has the only appropriate character, Demea, respond to Cleanthes on theological, rather than philosophical, grounds. What he maintains, namely that human attributes can be applied to God only negatively, was asserted both by him and Philo in Part II (D 141-2). The point is reintroduced here to remind us that Cleanthes' arguments fall far short of satisfying sophisticated (Neo-Platonic and Christian) religious sentiment, which demands a God who is Wholly Other. By appealing to the primitive source of religion in imagination, Cleanthes is committed to the anthropomorphic deity that imagination posits. In effect, Demea is also too learned, as a theologian, to accept Cleanthes' argument; and its status as popular theology is made all the more obvious.
Philo does recover his loquacity when philosophical critique is again possible, and in a way which makes clear that he has not accepted Cleanthes' argument. In Part IV, after Cleanthes has argued that the via negativa espoused by Demea is equivalent to atheism, Philo recommences by ironically pointing out the theological corner into which Cleanthes has painted himself. "But if idolators be atheists, as, I think may justly be asserted, and Christian theologians the same; what becomes of the argument, so much celebrated, derived from the universal consent of mankind?" (D 160) Is it merely a coincidence that Cleanthes had invoked the universal consent to theism as his last-ditch defense of the legitimacy of his "irregular" arguments? I think not.
James Dye
Northern Illinois University
1. Dialogues concerning Natural Religion, ed. Norman Kemp Smith (London, 1947), p. 152. Hereafter cited as D. Return
2. Arguing for a self-evident truth would seem to be a superfluous and paradoxical undertaking; but that has not deterred many philosophers from doing so. Return
3. Natural phenomena are said to evidence design more strongly than the examples, but the examples supposedly suffice to make an inference to an intelligent cause "irresistible." Return
4. Hume, Dialogues concerning Natural Religion, ed. with commentary by Nelson Pike (Indianapolis, 1970), pp.225-6. Return
5. P. S. Wadia, "Philo Confounded," McGill Hume Studies, ed. David Fate Norton, Nicholas Capaldi, and Wade L. Robison (San Diego, 1976), p. 286. Return
6. Stanley Tweyman, "Remarks on P. S. Wadia's 'Philo Confounded'," Hume Studies, VI, 2 (November 1980), p. 158. Return
7. Tweyman, "The Articulate Voice and God," Southern Journal of Philosophy, XX, 2 (Summer 1982), p. 268. Hereafter cited as Tw. Return
8. Is this not the model for Hume's illustration? Both involve a sound or voice from heaven, a comparison to the noise of the wind, and auditors of every nation. Of course, in Acts 2:1-13, men, rather than clouds, are heard to speak in the languages of the auditors and some auditors only hear unintelligible speech, such as intoxicated persons might produce. More recently, one of our pop divines, Pat Robinson, spoke in English to a Shanghai crowd, which heard him in their native Chinese dialects (Newsweek, 14 Oct. 1985, p. 77). However unlikely such events are, they can be consistently imagined, provided our imaginings ignore the actual causal mechanisms connecting speech and hearing. Return
9. Assuming that the phenomenon would have some consistent physical description, that it could occur in any possible world is far from obvious. Could the linguistic habits of a nation causally affect the acoustical properties of the atmosphere above it? Return
10. Simultaneous transmission of different information is a commonplace fact of contemporary life, as in electrical or optical telecommunication. But modulated high-frequency signals are not voices, although voices may be translated into them and reconstructed out of them. If an omniliqual translation machine were spliced into the transmission line, one could get many-voice output from one-voice input, but no single voice could conceivably utter different words at the same time. Return
11. The article may refer to types in classificatory statements where it has the same meaning as 'every,' as in "a lion is a four-footed animal." Cleanthes is clearly asserting existence, not classifying. Return
12. Actually, we, in the twentieth century, might say that very thing, because we are familiar with recording technologies and could expect that we were hearing reproductions of the same voice. It seems safe to rule out this line of thought for Cleanthes, not only because it is anachronistic but also because he takes "flexibility to all languages" to be a property of the voice. Return
13. It is only apparently miraculous, because previously unknown but natural causes could be imagined. For example, as noted above, the voices could be the work of technologically-advanced extraterrestrial beings. Return
14. Since Cleanthes specifies that there is but one universal language, there could not be both an Iliad and an Aeneid if those names are only properly applied to works in two different languages. Return
15. It might be supposed that it would make a difference if there were also humanly-authored books as well as generated ones. Knowing that human books required authors, could we not infer that the generated books also had authors? No, not unless one is willing to call the biological books themselves 'authors'. The situation would be the same as with other natural processes man has been able to replicate, such as fermentation, flight, and even the production of specially adapted organisms. That beer, airplanes, and frost-preventing bacteria have designers does not provide any good grounds for assuming similar causal histories for the natural things on which these products are modeled. Return
16. Locke, An Essay concerning Human Understanding Bk.IV, Ch.3, 6. Return
17. Hume, A Treatise of Human Nature, ed. L. A. Selby-Bigge (London, 1888), Bk. I, Part III, Sect. IX, pp. 116-7. Hereafter cited as T. Return
18. Hume, The Natural History of Religion, in Four Dissertations (London, 1757). Facsimile edition (New York, 1970), p. 27. Hereafter cited as N. Return
19. Hume, The Letters of David Hume, ed. J. Y. T. Greig (Oxford, 1932), I, Letter 72, p. 155.Return
Last Updated 6 September, 1998