Swinburne's Probabilistic Argument for God's Existence: A Response to Hume's Scepticism, Schemes and Mind Maps of Design

Download Swinburne's Probabilistic Argument for God's Existence: A Response to Hume's Scepticism and more Schemes and Mind Maps Design in PDF only on Docsity! 4 H U M E ’ S S C E P T I C I S M V S . S W I N - B U R N E ’ S I N F E R E N C E How Hume’s Scepticism Regarding the Design Inference is still Applica- ble Today Peter Casurella,1 McMaster University ￿ . ￿￿￿￿￿￿￿￿￿￿￿￿ I recently logged into a social networking site to find that a friend had posted a humorous picture for my consideration. Before me, a young man was gleefully engaged in an act of nasal exploration with his index finger. The caption read: “Your finger fits perfectly into your nostril. Checkmate, atheists”. This satirical picture is referring to the Argument from Design, one of the classic approaches to arguing for God’s existence via an analogous comparison of intentional and ordered objects created by man, to the apparent intentionality and order of the universe; which, it is argued, must likewise have an intelligent creator. In the example of our friend above, the assumption is that it is no accident that one’s finger conforms to the size of one’s nostril, but rather, that the convenience of this arrangement was thought out beforehand by a god who made man in such a way that objects could be removed with ease from one’s nose. In his Enquiry Concerning the Human Understanding, the Scottish philoso- pher David Hume began what many consider to be the most devastating philosophical attack on the argument from design.2 His remarks here, combined with his post-humously published attack on the design argu- ment in Dialogues Concerning Natural Religion, are considered by many 1 Peter Casurella is a second year Masters student in philosophy at McMaster Uni- versity, in Hamilton, Canada. His primary area of research is in the analytic phi- losophy of religion; specifically focusing on arguments and digressions of interest surrounding proofs for the existence of god. 2 Hume published this paper in 1757. 34 ￿￿￿￿ ’￿ ￿￿￿￿￿￿￿￿￿￿ ￿￿ . ￿￿￿￿￿￿￿￿￿ ’￿ ￿￿￿￿￿￿￿￿￿ 35 to have greatly reduced its philosophical importance.3 However, despite Hume’s attacks, the argument has persisted and evolved over the cen- turies and arguably remains the single most compelling argument for god’s existence. In his 2004 edition of The Existence of God, the esteemed Oxford philosopher and theologian Richard Swinburne advances a prob- abilistic form of the Argument from Design, building a case that the de- sign inference leads us to believe that the existence of god is more likely than the alternative. His argument connects with human intuitions in a compelling way. However, he does not succeed in his project. This paper will argue that Hume’s 250 year old scepticism regarding the de- sign inference cannot be bypassed by Swinburne’s probabilistic attempt at dodging its conclusions. I will begin by reviewing how Hume’s epis- temological commitments lead him to conclude the irrationality of the design argument. Next, we will examine how Swinburne goes about trying to dodge the problems presented by Hume. Finally, if Hume’s epistemology, or something like it, is correct, then I will show that Swin- burne’s dodge fails to carry his conclusion clear of Hume’s scepticism. ￿ . ￿￿￿￿ ’￿ ￿￿￿￿￿￿￿￿￿ ￿￿￿￿￿￿￿￿￿￿ ￿￿￿ ￿￿￿ ￿￿￿￿￿￿￿￿ ￿￿￿￿ ￿￿￿￿￿￿ The design argument is certainly not new. In the 4th century BC, Plato wrote that one of the main reasons for men’s belief in the gods was “the order of the motion of the stars, and of all things under the dominion of the mind which ordered the universe” (Plato, 12.966e). Aristotle too believed that the beauty of nature was what first caused men to wonder about how the universe had come to be. He argues in On Philosophy that anyone who observes the scope and beauty of nature would “judge both that there exist gods and that all these marvelous works are the handiwork of the gods” (Aristotle). In the middle ages, Thomas Aquinas propelled the argument to lasting fame through his inclusion of it in his five ways of knowing that God exists. Aquinas argues that any non- conscious object, which exhibits purpose in its design, must be under the direction of a conscious and intelligent being. In the same way that arrows do not head towards a bulls-eye without intelligent direction, neither do acorns grow into oaks without intelligent direction. Therefore, everything in nature must be directed towards its goal by someone with intelligence; e.g. God (Aquinas). 3 This paper was published twenty-two years later, in 1779. 38 ￿￿￿￿ ’￿ ￿￿￿￿￿￿￿￿￿￿ ￿￿ . ￿￿￿￿￿￿￿￿￿ ’￿ ￿￿￿￿￿￿￿￿￿ to be a complex object analogous to a house or a watch. Therefore, the universe itself must have a designer. In his later work, Dialogues Concerning Natural Religion, Hume allows the assumption that God really is the designer of the universe and goes on to level five arguments against the soundness of the design inference, based on the logical fallout which such an inference leads to. Firstly, he says that the analogy is weak. It is not at all clear that the universe so closely resembles a man-made machine that it always appears or- dered or designed. Many inexplicable and seemingly random events occur around us all the time. Next, Hume argues that the argument is also underdetermined. We have no reason to believe that order and complexity only arise from intelligence. “For all we know”, says Hume, “matter may have a source of order within it, just as mind does, hav- ing it inherently, basically, not acquired from somewhere” (1779, DCNR: 2).5 Further, Hume argues that if we posit an intelligent designer, that designer must also possess the properties of order and complexity, be- cause causes must contain sufficient properties to produce their effects, and must be alike in enough ways such that the correlation is rational. If God is comprised of order and complexity, then we are simply pushing the query back one stage, and are then, left to wonder what caused an ordered god? Hume’s final two critiques in the Dialogues deal with the properties of God which men have ascribed to him. It seems contradictory that a perfect god should create a manifestly imperfect world. It also seems impossible that we could reasonably infer from the universe’s creation, any other quality about god, beyond those qualities absolutely necessary to instantiate the universe. However, Hume also shows that the design inference itself should never be allowed off the ground in the first place. In Hume, Newton, and the Design Argument, Robert Hurlbutt nicely summarises this argument: The design argument is not scientific in that it offers no ev- idence for the causes of the world order. The world is one particular, not a member of a species, whose members have been observed. In order to demonstrate a cause for any ef- fect it is necessary to have observed the cause and effect in conjunction, indeed, in constant conjunction. And no one has seen the origin of one world, let alone “worlds” (Hurl- butt, 1965: 151). 5 DCNR will refer to Dialogues Concerning Natural Religion. Hume’s comments on this point are almost prophetic. The later development of the theory of evolution, both biological and stellar, would later reinforce this point of his very strongly. In many ways, matter does seem to organise itself, at least sufficiently, and for long enough, to produce beings with the kind of complexity which we exhibit. ￿￿￿￿ ’￿ ￿￿￿￿￿￿￿￿￿￿ ￿￿ . ￿￿￿￿￿￿￿￿￿ ’￿ ￿￿￿￿￿￿￿￿￿ 39 On Hume’s understanding of cognition, causal inferences only arise from cases where we have observed a constant conjunction between ob- jects, such that the mind becomes accustomed to expect one object to follow upon observation of the first. Every time we see a house being built, we have also observed people actively building such things. The conjunction of the two objects is very firmly fixed in our minds: intelli- gent purpose and designed object. This is the case with all the objects of human contrivance. Since our minds come to expect this conjunction, whenever we see what appear to be objects, or systems, which seem to exhibit order, complexity, and intent, we are naturally drawn to infer that a designer must be the cause. However, while it may be psychologically understandable why we make the inference, this does not justify our application of this inference to the universe as a whole. It is only “when two species of objects are found to be constantly conjoined, that we can infer the one from the other” (1778, EHU: 11.30). In the case of the instantiation of the universe, we have only observed one instance of an effect without even observing its cause. Therefore, we cannot make any rational conclusions about what the cause of the universe must be. ￿ . ￿￿￿￿￿￿￿￿￿ ’￿ ￿￿￿￿￿￿￿￿ ￿￿￿￿￿￿￿￿ In his influential book Inference to the Best Explanation, Peter Lipton de- scribes how these inferences work. When we find ourselves wanting to explain a certain phenomenon, but insufficient data exists to deductively describe the event, “given our data and our background beliefs, we infer what would, if true, provide the best of the competing explanations we can generate of those data”, so long as the best is good enough for us to make any inference at all (1999: 58). Given competing explanations for some event a, we compare and contrast rival explanations and deter- mine which one is more likely to be true based on its ability to account for the phenomena.6 Since Hume’s 18th century attack on the design argument, and in light of the purchase his thought has gained in the intellectual commu- nity, many of the more respected attempts to salvage the design infer- ence have adapted by making IBE claims.7 In 2004, the distinguished Oxford philosopher and theologian Dr. Richard Swinburne published an updated edition of his book, The Existence of God. In this book Swin- burne argues that although there are no good deductive arguments for 6 We also see if it has good fit with other observed phenomena and general rules. 7 Hereafter, let ‘IBE’ stand for ‘Inference to the Best Explanation’. 40 ￿￿￿￿ ’￿ ￿￿￿￿￿￿￿￿￿￿ ￿￿ . ￿￿￿￿￿￿￿￿￿ ’￿ ￿￿￿￿￿￿￿￿￿ the existence of God, a successful “P-inductive” argument can be built to show that the existence of God is more likely than the alternative, and therefore, is the best possible explanation for the universe (2004: 12-13).8 Acknowledging that the traditional starting points of the Argument from Design are problematic, in light of both Hume’s scepticism and the development of Evolutionary Theory, he instead sets out to simply argue that it is more probable that a designer god exists than not, due to the evident existence of ‘spatial order’ and ‘temporal order’. By spatial order, Swinburne is referring to the apparent arrangement of nature in orderly and purposeful ways, such as the structure of the human eye. The character Cleanthes in Hume’s Dialogues sums up the idea of spatial order nicely. Consider, anatomise the eye; survey its structure and con- trivance, and tell me, from your own feeling, if the idea of a contriver does not immediately flow in upon you with a force like that of sensation (1779: DCNR.3). Not only does Swinburne see spatial order in the arrangements of objects, but he considers nature to be like a ‘machine-making machine,’ in that it is constructed so as to give rise to the order which we see in the eye. He supposes that the theory of evolution can be fully incorporated into the design inference as a part of the intricate and ordered operations of the universe, and argues that it is highly improbable that humans would have evolved without an intelligent creator, god, to set up the universe-machine to produce them. A key supporting point for this argument is Swinburne’s supposition that the universe has been ‘fine-tuned’ to allow for the development of life. In brief, there are several fundamental forces in the universe which govern the interaction of matter, and if the relative strengths of those forces had been different by the smallest degree from what they are now, then cosmic evolution would have followed a different path, and human life would never have been able to evolve (Swinburne, 2004: 172-190). In light of the great number of alternative ways the universe could have been, Swinburne proposes that it is more likely than not that there is an 8 A ’P-inductive’ argument is one in which the premises add to the probability of the conclusion, and a correct P-inductive argument is one in which the premises make the conclusion more probable than its negation (Swinburne, 2004: 6). Swin- burne examines the probability of the conclusions of several arguments, for and against, the existence of god. He measures their relative strength in terms of the extent they confirm the hypothesis that God exists. Swinbure intends to show that if h is the hypothesis that God exists, and eb is the evidence from a particular argument for God’s existence, then P(h b1+b2+b3 [. . . ] bn) > P( h b1+b2+b3 [. . . ] bn). Therefore, it is more likely than not that God does exist. ￿￿￿￿ ’￿ ￿￿￿￿￿￿￿￿￿￿ ￿￿ . ￿￿￿￿￿￿￿￿￿ ’￿ ￿￿￿￿￿￿￿￿￿ 43 P(h e) > P( h e) In order to judge whether or not the design inference adds or detracts from his overarching case, Swinburne needs to decide if the probability of god’s existence, given the evidence from design, is greater or less than 0.5. This figure is then incorporated into the overall evidence for his cumulative case. As we have seen, Swinburne concludes that the probability of god existing, given the evidence we have that spatial and temporal order exist, is greater than the probability that temporal and spatial order could exist without god existing. Is he justified in reaching this conclusion? Swinburne’s approach may seem very subjective, but this is precisely what the Bayesian approach is meant to address. The approach allows one to make intuitive and general claims about the relative probabili- ties of various causes. Thus, we can make a claim about which is likely to be the actual ‘explanation’, out of a pool of competing explanations. So long as one’s conclusions are reasonable enough for most people to accept them, then general probabilities can be assigned to cases. Sub- sequently, these general probabilities can be added together to build a strong evidential case for whatever hypothesis is the most likely candi- date. For example, take Swinburne’s discussion of the fine-tuning of the universe. Naturally, there is no way to calculate how many possible cal- ibrations of the fundamental forces of the universe would have allowed cosmic evolution to give rise to life. However, it does seem reasonable to assume that there are a great many more ways in which the cosmic forces could have been arranged, such that human life never could have arisen. If the number of situations which would have not given rise to human life is judged to be greater than the number of situations which would have, then we should conclude that the possibility of human life not arising is statistically more probable. Given that we find ourselves in the less likely state of being alive (in an apparently ordered universe), then we are justified in thinking that it is probable that some explanatory hypothesis (which explains our current state) is true; in contrast to just assume that we have simply beat the odds. Think back to the example of the coin toss. If we find ourselves alive after the 50th toss, we might reasonably presume that the coin itself was not fair. Perhaps both sides of the coin were imprinted with a head. It seems more probable that there is some further explanation, which might account for the unlikely series of coin tosses, other than dumb luck. In thinking of the unlikely scenario in which we find ourselves in the universe, the god hypothesis explains why we find ourselves here in much the same way as the unfair coin hypothesis explains why the man with the gun did not shoot me dead. The point is simply that, if one was 44 ￿￿￿￿ ’￿ ￿￿￿￿￿￿￿￿￿￿ ￿￿ . ￿￿￿￿￿￿￿￿￿ ’￿ ￿￿￿￿￿￿￿￿￿ asked to make an educated guess based on the data available, one will either bet in favor, or against, the existence of god. For Swinburne, the unlikelihood of finding ourselves here means that we ought to allot the balance of probability in favor of the existence of god. If the probability is only a fraction more convincing in favor of the god hypothesis than not, then, we are justified in reasonably inferring that god really does exist. ￿ . ￿￿￿ ￿￿￿￿￿￿￿ ￿￿￿￿ ￿￿￿￿￿￿￿￿￿ ’￿ ￿￿- ￿￿￿￿￿￿￿￿￿ No scientist would object to occasional uses of Bayesian calculations, in fact such calculations are part of our day to day life, and can even be descriptive of the foundations of very successful science. Mackie admits that “we are justified in arguing inductively, in extrapolating observed regularities to unobserved cases” (1990: 147). A beautiful example of such successful reasoning was the discovery of the planet Neptune by German astronomer Johann Gottfried Galle. Galle discovered Neptune after several other astronomers had independantly predicted its exis- tence based on irregularities in the orbit of Uranus. Galle theorised that these irregularities were best accounted for by the gravitational influence of a postulated, yet undiscovered, eighth planet. The prediction that Neptune existed before it was discovered can be couched in Bayesian terms as: The probability that Neptune exists, given the evidence of Uranus’s orbit and our understanding of gravitation, is greater than the probability that Neptune does not exist. There is a key difference between this example of a very successful IBE and Swinburne’s argument for god’s existence. In the case of Nep- tune, there was a probable cause, which fell within the realm of normal science, which could potentially be investigated. If I am allowed to em- ploy a bit of my own probabilistic reasoning here, it seems that it is far more probable that an unobserved cause which operates within our universe can be discovered, than that we could discover a cause of the universe itself. I might propose, for example, that the cause of the current universe was a trans-dimensional, super-alien sneezing, and that the sneezing of such aliens always produces explosions of space-time that have ordered regularities as one of their inherent features. Now, it seems unlikely that we would ever be able to investigate if my sneezing-alien theory is cor- ￿￿￿￿ ’￿ ￿￿￿￿￿￿￿￿￿￿ ￿￿ . ￿￿￿￿￿￿￿￿￿ ’￿ ￿￿￿￿￿￿￿￿￿ 45 rect, or rather, if the god hypothesis is correct. Since the conditions of these possible causes are beyond the physical universe, then, the truth of these matters lies beyond our ability to investigate them as well. How- ever, this is simply an ad ignorantiam, and there is more we can say on this issue. When building an IBE argument, what makes an explanation ‘best’, “is always relative to the available competitors it faces, meaning that IBEs always involve comparative evaluations of evidential support among competing hypotheses” (Ratzsh). Swinburne argues that what makes god the best explanation is that the god hypothesis provides the sim- plest explanation for the observed phenomenon (the universe); while, it provides the greatest explanatory power when compared to other hy- potheses (2004: 82). We are justified in believing that this hypothesis is probably correct if “any gain of explanatory power would be outweighed by a corresponding loss of prior probability”, and if “any gain in prior probability would be outweighed by a corresponding loss of explanatory power” (Swinburne, 2004: 82). The god hypotheses, argues Swinburne, is the simplest hypotheses available, because all of the conditions neces- sary for the instantiation of the universe are available in one explanatory step; the actions of God. Positing any further beings beyond God gains us no explanatory power. Nonetheless, why this hypothesis has greater simplicity and explanatory power than competing explanations such as the sneezing super-alien theory, is very puzzling. In both theories, we have a practical end to investigation. We do not possess the proper investigatory tools to look beyond the proposed entity, in either case. It is not possible to discover any information about whether or not there is an explanation of god or the super-sneezing alien, and thus, our investigations must end with the postulation of some such being as an explanation for the universe. The sneeze of the alien has all the conditions necessary to create our universe, just like the proposed properties of God. Both theories postulate only one being with sufficient properties to create the current universe, and any other properties of such a being are things which we can only speculate at. Swinburne might argue that the infinite attributes of God, his om- nipotence and omniscience for example, would count as more simple properties than a super-alien with some definable set of finite properties. Nevertheless, I suggest that the burden of proof rests on Swinburne’s shoulders to show how some infinite set of properties that are sufficient to create an ordered universe are any more simplistic than some unde- fined set of finite properties, which are equally sufficient for creating an ordered universe. If both are fully sufficient for instantiating the uni- verse, how can we, from our limited perspective within the universe, pass judgment on which scenario is more likely? The properties of the