Are miracles impossible? Are miracles impossible?

The debate surrounding miracles has always been highly contentious, primarily because it has been one of the focal points for the argument between those who believe in a deity and those who do not. Thus, any attack on the authenticity of miracles is seen as an attack on religion. This is in fact true. Whilst it may not be the intention of someone seeking to deny the existence of miracles to also deny the validity of religion, most major religions have miracles explicitly written into their histories, if not their ceremonies. For example, if you deny miracles, you make a mockery of the Catholic idea of transubstantiation, a weekly article of faith for Catholics everywhere. Because of this, there is a lot of emotive and non-rational debate about miracles that has to be filtered out to view the solid, or more solid, arguments. The two main topics that arise from this question are what do we mean by a miracle, and can such an event come to pass. The definition of a miracle is essential in achieving a valid answer to the second question, as a loosely worded definition can have all sorts of unforeseen implications when examined closely. However, I do not intend to spend too long examining the semantics of the definition, rather the implications of the intended meanings. I intend to show that any definition of a miracle is unsatisfactory, leading to the conclusion that the whole notion of a miracle is incoherent, and is merely a piece of rhetoric or an exaggeration. The only slim chance I see for allowing the possibility of miracles raises many other problems, thus serving only to move the area of debate from miracles into more general metaphysics, and the nature of such a deity as might exist.

To begin with, we must define what we mean by a miracle. The classical definition comes down to us from St Thomas Aquinas, who defines a miracle as ä breach in the regular course of nature," in other words something that shouldn't happen normally. However, this is somewhat too vague for our exacting purposes, so I intend to use the definition given by David Hume in his essay Ön Miracles," which defines a miracle as ä violation of the laws of nature." This is certainly more explicit, and seems to amount to much the same meaning as Aquinas' definition, but it does still require some interpretation. We need to know what is meant by both "violation" and "law of nature." For Hume, we formulate a law of nature after having seen multiple instances of a particular occurrence in the same form, and he would want to say that we must also not have encountered a counter-instance. Alternatively, we can form a law from a single instance if we claim knowledge of causality. Either way, we must have exceptionless experience of a particular causal link, or as J. S. Mill calls it, a "complete induction." This concept of a natural law may be unsatisfactory, but we shall see whether we can accept it for the time being.

Having defined a miracle, we must now examine whether such an event can coherently occur. The first problem is that if a natural law is an exceptionless experience, and a miracle is contrary to a natural law, then surely the miracle must count as an exception. From here we must either accept that the supposed natural law is not in fact a law, or that for some reason the miracle does not count as a counter instance. It seems difficult to conceive of a consistent way of judging whether or not an event is a valid counter instance if we are going to start writing off some exceptions as unimportant. If we were to do that, then once we had formed any natural law, we could write off all counter-instances as unimportant, and claim that we never had to modify the law. This seems a patently ridiculous way of going about formulating laws, since if it had been adopted long ago, we would still be thinking in terms of alchemy or the like. It seems reasonable to say that any natural law has to take into account all events of the appropriate type, be they supporting instances or not. So now we have to deal with the problem that arises if we say that the natural law has to be reformulated to take the counter-instance into account. This leads to the alleged miracle being explained by a natural law, and thus not being in violation of it. If we choose this route, as it seems we must, then miracles cannot logically happen, since any event that cannot be explained by a current law of nature must necessarily generate a new law of nature that does explain it. Thus, no miracles are possible. However, it has been claimed that this line of argument leads to a circularity, since it bases its refutation of miracles on the uniformity of nature, and bases its assertion that nature is uniform on the fact that natural laws can explain all events, in other words, that there are no miracles. In order to avoid this accusation, several reformulations of the argument have been attempted, mostly based around redefining a miracle.

The first reformulation is attempted by J.L. Mackie, who seeks to alter the definition of a law of nature to a "basic law of working." This is a weaker version of Hume's definition, and put simply, allows for a limited number of counter-instances. It does this by allowing margin for external interference in the series of causes, of what type is not made clear. As Mackie puts it, basic laws of working "describe the ways in which the world ... works when left to itself, when not interfered with." This definition allows room for counter-instances when other factors have come into play, presumably something that would cause a miracle, such as a deity. It does however seem very vague, since surely one could argue that if we are free beings, then our actions 'interfere' with the way in which the world works, and thus exceptions would abound.

Another attempt at reformulation is undertaken by R. Swineburne, who seeks to allow that a law of nature can have a limited number of counter-instances, but also redefines a violation as än exception that would not be repeated under similar circumstances." This to me seems to encapsulate the whole notion of a miracle, that whilst it is a counter-instance to a law of nature, it does not destroy it, since it is a miracle, a one off event. Examples like transubstantiation could be postulated as problems, since they happen repeatedly, but one could then argue that they are no longer miraculous, since if this happens every week in Catholic churches across the world, it is hardly remarkable, and could potentially be formed into its own law of nature.

The final reformulation of the notion of a miracle seems to me to almost trivialise the notion, but still holds some appeal. D. Johnson argues that a miracle is in fact only ä violation of an apparent law of nature," in other words, it is only a miracle because we cannot explain it at the time it occurs, not because it actually violates a fundamental law of the universe. In other words, a radio would be a miracle to a caveman, as we could well believe. This reduces the notion of a miracle to a rhetorical concept, it is only miraculous because it is beyond our understanding. However, it does have some resonance with our ideas now, since we believe that we can explain some of the miracles in the Bible without recourse to divine intervention. Thus, they would indeed be classed as miracles for the people experiencing them at the time, but since we can explain them, we no longer regard them as miraculous.

There is one other formulation of a miracle that I have not yet discussed, that of a miracle as any event caused by divine intervention. I do not think that this is a very satisfactory way of classifying an event, since we cannot be sure if a deity did in fact cause the event if, as has been asserted, all events, alleged miracles included, can be explained by laws of nature. In addition there are several other problems faced by those who wish for a deity to cause their miracles. Firstly, one requires the existence of a deity. A discussion of this topic is beyond the scope of this essay, so I shall move swiftly on. A more relevant problem is that of interaction. If we require this deity to interact with our universe of extended matter, then it would seem that it would need to be at least partly composed of extended matter itself, and that leads to all kinds of problems. Extended matter is susceptible to entropy, in other words time and a finite life-span, is necessarily finite, since not everything is matter, and some would argue, is divisible, as all matter is. This final point is one made by Descartes in relation to the substance of mind. We do not want the human mind to be divisible, so it cannot be made out of matter, since matter is divisible. Alternatively, if we seek to avoid the problems of extended matter, our deity must be entirely non-physical, and we then have to ask how it can affect a substance that it has nothing in common with, whereupon we find ourselves in the realms of the dualist's interaction problem.

Additionally, if we are going to say that our deity has a part to play in causing events, we need to ask if that makes it part of the causal chain. If it does, then it would seem that our deity is subject to causal laws, and thus cannot be omnipotent, and is possibly even subject to natural laws, and thus loses many of the other powers traditionally associated with supreme beings. If it manages to affect causal links without becoming part of the causal chain, then this too needs to be explained, and we arrive at a similar problem to that of interaction.

It seems to me that if we are going to attempt to define a miracle, then using divine intervention to do it is a very weak position. However, any attempt to explain it in connection with laws of nature seems to be highly problematic. I have a great deal of sympathy with Swineburne's position, mostly on the grounds of his definition of a violation. I am not particularly happy with the idea that a law of nature is 'mostly right,' but if we are going to accept miracles, that would have to be the case. I would be happier with a formulation such that said that a law of nature was an exceptionless experience 'barring miracles.' However, this is obviously circular, and so must be rejected. I am left with the position that miracles are events that violate our current laws of nature, or to put it more succinctly, they are focused both temporally and culturally. Thus, miracles are, to later generations, or those in possession of some sort of 'divine' complete knowledge, not at all miraculous. Therefore I say that miracles are possible, but that rather than being something special in relation to the structure of the universe, they are rather a function of our own ignorance.


R.M Burns, The Great Debate on Miracles (London: Associated University Presses, 1981) D. Hume, An Enquiry Concerning Human Understanding (Bath: OUP, 1999) D. Johnson, Hume, Holism and Miracles (USA: Cornell University Press, 1999) J.S. Mill, A System of Logic (London: Longmans, Green and Co, 1965) Ed. E. Sosa and M. Tooley, Causation (Bath: OUP, 1993)

File translated from TEX by TTH, version 2.92.
On 24 Mar 2002, 23:49.