This post serves a dual function. It is part of my series on Nicholas Everitt’s book The Non-Existence of God, and my series on Miracles.
Over the next few posts I will be covering chapter 6 of Everitt’s book, entitled “Arguments to and from Miracles”. This will serve as a nice, gentle and general introduction to the philosophical issues associated with miracles and miracle claims.
Everitt begins his chapter by noting that miracles are thought to be both indicators of God’s existence and (potentially) God’s character. He then observes that one of the problems associated with evaluating arguments to and from miracles is that they frequently rely on different senses of the term "miracle".
In this entry we will discuss the violation sense of the term "miracle".
1. Hume and Violation Miracles
We begin with the locus classicus: Hume’s discussion of violation miracles in his essay “Of Miracles”. This is a text to which we will return time and again throughout this series as it has shaped the contours of the modern philosophical discussion. In his essay, Hume defines miracles as follows:
- “Miracle” = a violation (or transgression) of a law of nature by a particular volition of the deity.
This violation sense of the term miracle is undoubtedly popular. The idea, presumably, is that the natural world is governed by a set of laws determining regular unfolding of events and that the only being who could violate such laws would have to be supernatural in origin.
The problem with this is that although there may well be a fixed set of natural laws, we are not necessarily any good at identifying them. Thus, it will be difficult to assess whether a putative violation is due the actual intervention of God, or due to an error in how we have been identifying the laws of nature. Hume, as we shall see, uses this as the basis for an argument against believing in miracles.
Before outlining Hume’s argument, Everitt stops to make a couple of points.
First, some believers argue that even if we could identify them, violations of the laws of nature would not necessarily be evidence for the existence of God. They contend that the violation must contribute to some divine purpose. Of course, this assumes that we have some alternative source of information about divine purposes.
Second, it is not necessary to assume -- even on this definition -- that miracles are rare. Indeed, some religious doctrines suggest that miracles might be commonplace. As examples of this consider some of the claims made concerning the success of petitionary prayer, and the Catholic doctrine of transubstantiation. That said, regular occurrence of miracles might create some problems from an evidential perspective, as we shall see.
2. Hume’s Argument
In his essay on miracles, Hume makes two simple arguments. The first is an in principle argument. It suggests that it is (in principle) impossible to rationally believe in the occurrence of violation miracles. The second is an in fact argument. It suggests that the existing testimony is weak.
The in principle argument relies on Hume’s mini-theory of rational belief. This has two elements to it: (i) beliefs must be proportionate to the available evidence and (ii) the prior improbability of an event increases the evidence needed to prove that the event occurred.
To see how Hume’s mini-theory works, consider a simple example: a friend of yours tells you that they saw a cow in a field. Assume your friend’s eyesight is normal and they have no reason to lie. Should you believe them? I would say yes: the presence of a cow in a field does not suffer from prior improbability and so relatively little evidence is needed to make belief in the presence of that cow rationally acceptable.
Now imagine a second version of this. This time around your friend claims that they saw a three-headed cow in the field. Should you believe them? No: the existence of a three-headed cow suffers from prior improbability and so you will need more than one eyewitness to convince you of its presence. You would probably need several (independent!) eyewitnesses, photographic evidence and may even like to confirm it for yourself.
Based on this example, Hume’s mini-theory of rationality seems unexceptionable. But now consider how it applies when you’re receiving testimony about the occurrence of a violation miracle. To make belief in this miracle rationally acceptable, two propositions need to be established:
- (1) That there is a law of nature L.
- (2) That event E (a violation of L) has occurred.
It is worth emphasising the point that both of these need to established, not just (2). But since our beliefs must be proportionate to the evidence, and since we could err in our identification of the laws of nature, Hume points out that:
- (3) The statement that E has occurred contradicts the statement that L is a law of nature.
In other words, any evidence that is sufficiently strong to prove the occurrence of E would also be evidence that counts against L’s status as a law of nature. It might be that L really is a law of nature and E really is a miracle, but it might also be that E is something that should encourage us to revise our commitment to L.
This is Hume’s in principle argument. Note that although Hume explicitly refers to belief in E that is based on testimony, the same argument holds even if you have observed the evidence for E yourself.
Hume follows this up with a discussion of the kinds of evidence that usually is presented to support belief in miracles and finds it to be not even of sufficient strength to warrant agnosticism. We won’t discuss that argument here.
3. Assessing Hume’s Argument
In assessing Hume’s argument, Everitt points out that Hume is not claiming that violation miracles are impossible. This is a common misinterpretation, one to which even the most sophisticated of Bayesian interpreters can fall prey. Rather, Hume is claiming that the following disjunctive conclusion must be reached:
- (4) Either it is rational to believe that E occurred but was not a violation miracle, or it is rational to believe that E did not occur.
Everitt thinks that many commentators fail to understand both the strengths and the limitations of this conclusion.
As an example, he cites Brian Davies who argues that although testimony may fail to provide rationally compelling evidence of miracles, the testimony could be corroborated by physical evidence. The idea being that historical events leave causal traces in the present that could be recovered.
Everitt responds by saying that this could very well be true, but that it misses the point. The point is not that testimony is particularly weak form of evidence; the point is that evidence for miracles and laws of nature complement one another. The more credible is the evidence for the miracle; the less credible is the evidence for the law of nature, and vice versa.
Everitt thinks that, when understood properly, Hume’s argument really does provide a decisive blow to arguments supporting the occurrence of violation miracles. But he qualifies this by saying that the decisive blow could be avoided by simply offering a different definition of miracles.
That is the extent of Everitt’s discussion of Hume’s argument. A few words are in order. Everitt studiously avoids the recent Bayesian interpretations and critiques of Hume’s argument. This is something for which he has been criticised. According to some of the Bayesian critiques, the disjunctive conclusion reached above is much too strong because there are conditions under which it may be permissible to believe in the occurrence of miracles.
We will consider some of these more sophisticated Bayesian critiques in future entries to this series. In the next post on Everitt’s chapter, we will consider two further arguments against the violation sense of miracles.