Philosopher Hugh McLachlan has written what in my opinion is a rather incoherent article on miracles for the New Scientist. He takes on David Hume and Richard Dawkins for taking the premises that miracles, by definition, violate natural law and that such events are used by the religious to support their faith.

McLachlan writes:

I would argue that, by definition, “laws of nature” are universal laws of the form “if A, then B”, or “all As are Bs”. Logically, they cannot be violated or transgressed, not even by God. If, even on one occasion, for whatever reason, there was an A without a B, then it would not be true to say “if A, then B”. What had been thought of as a natural law would in fact not be one.

He argues, therefore, that a “miracle” cannot be defined as an event that violates a law of nature, for that would be a violation of logic, which cannot occur. This is a circular and ultimately useless argument, however. Hume and Dawkins are defining a miracle as an occurrence that appears to violate the known laws of nature. McLachlan says this cannot happen because then the violated law would not be a law.

I dislike semantic arguments – ones that base a point on how a word is or should be defined. This is particularly problematic with words like “miracle” that do not have a technical or operational definition. One has to define what one means by “miracle” before discussing it in any meaningful way, and I believe that Hume and Dawkins did this.

I dislike McLachlan’s criticism of Hume’s definition because it misses the point. First of all, the unstated major premise of his argument is that we can know what the laws of nature are. We cannot – we can only imperfectly infer them. Therefore a definition of a miracle as something that actually violates the laws of nature is impossible to apply. A more useful definition would be something that seems to violate the known laws of nature – an apparent anomaly.

Identifying an anomaly is the best we humans can do – something that appears to violate our understanding of the laws of nature. McLachlan recognizes this himself when he writes:

“Laws” that appear firmly established are often overturned in science, yet we do not need to argue that a miracle must have occurred, assuming whoever reported the apparent overturning is telling the truth. Instead, the rational thing to do is to abandon the natural law or modify what we considered to be a true statement of it.”

What this really means is that there is no way to ever conclude that a miracle has occurred. At best we can identify a persistent anomaly – but we can never know if the anomaly is due to our limited knowledge of nature or due to a true violation of what is otherwise a consistent law. There are no miracles – only anomalies.

Then all that is left is to argue about the implications of persistent apparent anomalies. Within the realm of science this is clearly settled – you cannot use ignorance as a positive argument for an alleged phenomenon (no god-of-the-gaps arguments). And from a practical point of view it is best to assume that apparent anomalies are due to our limited knowledge and then set about to extend our knowledge through scientific investigation. This approach has worked very well so far.

From a philosophical point of view this is unresolvable. I believe that there are no anomalies so egregious and persistent that they challenge a purely materialistic view of reality. But others can believe what they wish.

But let’s get back to McLachlan’s logical argument – that a natural law is not a law if it is ever violated. This is, again, just a semantic argument and of little value. What Hume and Dawkins are really saying is that by such a definition a “miracle” occurs when a deific agent suspends the laws of nature as an act of will. Of course, then one could argue that the deific agent is merely part of a deeper set of natural laws – and again we are stuck in a circle of definitions. And once again we are back to – there are no miracles, only anomalies.

McLachlan really gets into trouble, in my opinion, when he tries to substitute his own definition for miracle:

Consider the Azande, an African tribe whose members believe all deaths and misfortunes are caused by either witchcraft or sorcery. Suppose a falling branch kills someone. On one level, the tribe accepts a scientific account of the incident in terms of, say, the effect of termites on wood. But on another level, they ask why did it come about that the particular person happened to be standing under the tree when the branch happened to fall?

McLachlan argues that a miracle may be defined by an occurrence that does not appear to violate natural laws but one in which the timing of events conspires to bring about an outcome that was desired by some unseen presumably deific agent.

McLachlan has not, however, solved any problems with the definition of miracle – he has just made them more subtle and insidious. There are two approaches we can take to this type of miracle. The first is to say that such alleged miracles are unfalsifiable. Nothing is happening that could not have happened just by chance alone. Occam’s Razor favors the explanation that there  is nothing more than “hyperactive agency detection” going on when ascribing the random fall of a branch to some agency.

However, depending on the specific claim, evidence may in fact be brought to bear. In the case of the Azande the unseen agent is believed to be witchcraft. That sounds like a testable claim. Are witchcraft effects reproducible? Do the “juxtaposition of events” that McLachlan says are an acceptable definition of “miracle” occur at a rate that is incompatible with random chance? If not then there is no reason to hypothesize that there is anything miraculous going on. And if there is, then we are back to the anomaly dilemma and McLachlan’s definition of miracle is no different than Hume’s or Dawkins’. Is McLachlan arguing that using witchcraft to ensure that a branch falls at the right moment to land on a particular person’s head is not a violation of cause and effect? McLachlan has solved nothing.

But McLachlan really goes off the rails with his conclusion:

So some people might think of “miracles” as particular juxtapositions of events, each of which has a correct and acceptable scientific explanation. This might be nonsensical, but it would be interesting to discover wherein the nonsense lies. We should be open not only to possible observations and experiences that might dislodge some of our accepted theories but to thoughts and ways of thinking that may challenge our notion of what acceptable theories and explanations can be like. We deceive ourselves if we imagine science has established that only scientific explanations are valid or that scientific explanations can take only one particular form.

I think we have discovered where the nonsense lies in McLachlan’s arguments. His last sentence is an utterly absurd self-contradiction. By definition – only scientific methods are valid within science. Science does not take “one particular form,” (a straw man argument) but all the forms of science must adhere to the rules of science. Unscientific methods and ways of knowing cannot be science.

McLachlan’s “miracles” are either unfalsifiable, which means they are simultaneously outside of the realm of science and are cut away by Occam’s razor, or they make testable claims and are part of science. They do not represent in any way a challenge to science or a separate yet valid method of gaining knowledge about the physical world.

Of course, to be clear, there is wisdom that lies outside the realm of science. Anything that involves a value judgment requires more than logic and empiricism – science may inform such things but does not encompass them. But this is not relevant to the question of miracles, which are (by any definition) about whether or not an event can be explained by current scientific knowledge.