Richard Feynman once delivered a lecture in which he defined the scientific method by reference to attempts at falsifying ‘guesses’. A scientist following the method guesses a new physical model, computes the consequences of this guess and then performs an experiment in order to compare these consequences to what actually happens. The central premise of science is this: ‘if [the guess] disagrees with experiment, it’s wrong.’ And it really is more or less that simple.
Unfortunately, the converse of this statement is not true: if our guess and our experiment do agree, the guess could still be wrong. All we can do is consider it temporarily right. And although the specifics blur somewhat when we move away from objective laws and towards social phenomena, the fundamental principle remains: we create a hypothesis and then do our best to show that it is wrong. If we can’t, then maybe – just maybe – it is right.*
The important thing to notice here is that a scientific argument is inductive rather than deductive. Instead of considering premises that necessarily entail conclusions – as in mathematics and most formal logics – we allow the possibility that the conclusion is false while the premises are true. The idea is that the premises evidence the conclusion – i.e. they make it more likely to be true – but never prove it. A ‘good’ argument in this sense is one for which the probability of the conclusion being false while the premises are true (or the model being false despite the agreement with experiment) is low.
Now, the problem with using miracles to support the supernatural is essentially that such arguments are not ‘good’ ones according to this definition. Let’s use the example of a ‘miraculous healing’ evidencing Christianity to illustrate this. I contend that such an event does not appreciably increase the likelihood of Christian doctrine being true, despite what many will claim.
The reason for this is that there are other conclusions – ones involving well-studied phenomena such as confirmation bias – that are just as consistent with the healing as any religious claims. For example, is it more likely that there exists an all-powerful, telepathic master of the universe, or that your GP made a mistake when diagnosing you in the first instance? Or perhaps that your illness got better by improbable but perfectly natural means? I feel the answer to this should be obvious. Indeed, Tim Minchin’s comedy song ‘Thank You God’ does an excellent job of illustrating this.
Interestingly, when you question someone about precisely how their miraculous experience evidences their spiritual beliefs, they will often be stumped and have to resort to vague pseudo-arguments about the nature of ‘faith’. Even if their miracle did suggest the existence of some supernatural power, how did they infer that this was a god? How did they infer it was the God of the Bible? Or that Jesus was His son? The simple answer is that they couldn’t possibly have done so rationally. The ‘miracle’ is not scientific evidence for their supernatural claims.
*Note that this makes unfalsifiable hypotheses (i.e. those that cannot be falsified by observation; those that are untestable) scientifically useless: failing to falsify an unfalsifiable hypothesis does not lend support to the hypothesis.