I read the SEP entry on miracles a while ago and plan to take a shot at Hume's Of Miracles soon.

Before I get started - I cannot understand how miracles even make sense. Here's my thinking so far:

Premise: Our best predictions state some empirical quantity.

  1. the claim that a miracle can happen equals the claim that our best predictions are sometimes wrong and that our only and therefore best explanation of them is divine will.
  2. the claim that some miracle will happen equals the claim that our best predictions are not our best because of divine will.
  3. the claim that something was in fact a miracle equals the claim that our best explanation of it is divine will.

But it seems to be that 3 is weaker than 1 and 1 weaker than 2; yet, surely, if 3 then 1 and 2, so in effect any miracle means not believing our best empirical predictions, which is absurd.

Where's the error in this line of reasoning?

  • I reformatted your question and changed the title. Please feel free to revert my changes if I misconstrued your question.
    – DBK
    Commented Aug 18, 2014 at 1:22
  • Closely related (but not a duplicate) to philosophy.stackexchange.com/questions/914/… and philosophy.stackexchange.com/questions/14474/…
    – That Guy
    Commented Aug 18, 2014 at 4:41
  • Also related is Lewis's On Miracles, if you're looking for a different perspective. Commented Aug 18, 2014 at 18:21
  • 1
    The logical error is that "best prediction" means most-frequently-accurate-according-to-rigourous-statistical-analysis, not unfailingly-fully-correct. For 2: The claim that some non-specific miracle will happen is the claim that our usually best prediction will not always occur. If something happens 99.99% of the time, it's still our best prediction despite being wrong 1 in 10000 times. The absurdity arises because you confuse statistically-valid with true-without-exception, and because number 2 erroneously assumes that any exceptions stop something being most likely ("best").
    – AndrewC
    Commented Dec 31, 2014 at 21:40
  • ah ok, let me mull that over, ta
    – user6917
    Commented Jan 1, 2015 at 2:35

1 Answer 1


That's quite a simple one:

Let's use Hume's definition of "a violation of the laws of nature". The "laws of nature" are simply descriptions of the best experimental data we have.

Our best empirical predictions are under controlled circumstances and predict accurately under those conditions only.

For example, my friend Steffan may catch an apple before it collides with the earth in violation of our previous experimental data, but this does not contradict empiricism, even if in all experiments so far Steffan has not chosen to catch the apple.

Steffan's externality to experimental data means he is not included in the law of gravity. If Steffan ever caught an experimental apple, it would be experimental error.

(Of course this example is an analogy; I'm not asserting Steffan is divine.)

You may or may not believe I have a friend called Steffan, but you have to accept that if he exists, he may be able to disrupt science's prediction about what will happen to the apple with a to-him trivial intervention.

There's no need for anyone to reject science just because they believe Steffan exists and sometimes catches apples, especially if he's taken no interest in any apple experiments so far.

Your beliefs about science and about Steffan are neither mutually contradictory nor mutually supportive.

  • 1
    Mod moves 30 comments to chat. Please consider taking up extended discussion in our chat! Comments are really just for short -- well, comments :)
    – Joseph Weissman
    Commented Jan 2, 2015 at 17:41

You must log in to answer this question.