Miracle arguments: Should we always prefer naturalistic explanations?

A common objection to arguments from miracles, such as the argument for the resurrection of Jesus, is that we should always prefer natural explanations (however improbable) to supernatural explanations. The conversation might go something like this:

Theist: “There is simply a ridiculous amount of evidence that Jesus rose from the dead. We have the empty tomb, the testimony of the appearances to disciples who believed they would die for their testimony, the conversion of Paul, and the early objections to Christianity grant these evidences. The best explantion for all these facts is the resurrection”

Atheist: “I can agree with all of those facts, however the resurrection is a supernatural explanation, not a natural explanation, and so it must always be rejected. Therefore a better explanation for these facts is that Jesus was an alien hologram who merely appeared to live and die and rise again, and appear to many people. When Jesus interacted with objects, that was just the aliens using advanced technology to make it seem like there was a man. But really, it was all aliens who were messing with people for fun. This is extremely unlikely, it is very implausible, but it is still a better explanation than any supernatural explanation”.

So then we have a principle to investigate: natural explanations are always preferable to supernatural explanations. How could we justify, or alternatively defeat, this principle?

Here is one such way:

  1. We know (independently) that natural objects exist
  2. We do not know (independently) that non-natural objects exist
  3. We should always prefer explanations that use objects of a class we know independently to exist
  4. We should never use explanations that involve non-natural objects

But this doesn’t seem to be very good. It would restrict scientists from ever positing new kinds of objects, and so we’d never come to believe in things like quarks. What explains our observations about protons and neutrons? Maybe quarks, or maybe it’s a mistake in our observations. We know mistakes exist, we don’t know quarks exist, so we can’t ever use quarks, and we must just be wrong. But that’s a bad conclusion

So maybe it’s just something special about supernatural explanations. And maybe no justification at all is given from the atheist. If no justification is given, then we can dismiss it without any argument. But let’s be generous and go further, to not only point out that it’s unjustified, but to demonstrate it false.

How can we do that? Here’s a simple thought experiment: suppose we live in the world of the book series *Mistborn*, where supernatural magic is reasonably common, and most people will know someone or knows someone who knows someone with some supernatural ability.

In this world, the atheist’s principle is clearly false. In this world, supernatural explanations are clearly justified, and are commonplace and accepted.

So the truth or falsehood of the atheist’s principle depends on what possible world we are in. The important difference between our world and the world of Mistborn seems to be the existence of the supernatural. So the atheist’s principle becomes something like “In worlds where the supernatural does not exist, we should not appeal to supernatural explanations”. But now the atheist begs the question: we are presenting evidence that the supernatural does exist in our world, and they can’t presuppose that the world doesn’t contain the supernatural in order to refute that evidence. That would be a circular argument.

So it doesn’t seem like we can justify the principle “we should always prefer natural explanations”.