Testing the Supernatural

How Science Deals with the Unknown

In #articles

In an interesting paper entitled "Can Science Test Supernatural Worldviews", Yonatan Fishman proposes an unusual answer of "yes". I've spoken to this issue here and again here, where I refer to the term "supernatural" as a label-of-the-gaps, and trace its usage in ancient cosmology. I've stated that the term itself is either content-free or not defined specifically enough to be useful. However, I do make the following statement which I think points to Fishman's perspective:

I'd love to see something that really rises to the level of a miracle, like the 10 commandments etched in Pluto or something. Although technically a God-of-the-Gaps argument, it would be very interesting indeed!

We can ask the question, what would convince you of the existence of the supernatural, or of God? Fishman points out that if prayer actually worked, and only Catholic prayers were effective, that would in fact be evidence for the supernatural. What I find interesting about this line of thinking is that, even if you expect that there should be a natural explanation, if the person claiming supernatural agency continues to make specific predictions that turn out true - predictions that, at least on the face, seem to be unexpected from our current understanding - then you'd have to start thinking that they are on to something.

What I think is going on here is something a bit different than a God-of-the-gaps, at least in some contexts. One way of thinking about it is with an analogy to Dark Matter. In astronomy it was observed that galaxies rotated more rapidly than expected. Using the light emitted from the galaxy, you can estimate the amount of material needed to emit that light, and use that to determine rotation speeds. The amount of mass estimated from the light emitted is far less than the amount of mass estimated from the rotation speeds - somehow there is a large amount of mass unaccounted for. Astronomers named this extra mass "Dark Matter". Notice, that this is not an explanation for the extra mass, but a label for "something I don't know is causing this effect". One then has to make propositions of specific causes, like neutrinos, black holes, or even modifications of the laws of physics. From there one makes specific predictions, constructs methods for testing the propositions, and explores alternatives. The term, "Dark Matter", is still used as the umbrella term to label all of these alternatives.

In this way, the term "supernatural" may be a label for "something I don't know is causing this effect", where the effect may be a miracle claim, prayers answered, origin of the universe, origin of life, etc... Thus, the phrase "supernatural explanation" is a meaningless phrase - there can be no supernatural explanation, just as Dark Matter is not an explanation. However, as Fishman uses in his title, one could test "supernatural worldviews" - those constructions that use the label to suggest some unknown (and possibly unknowable) agency at work.

One might think that a scientific theory can't have anything like this in it, but that is not correct. For example, consider the wavefunction in quantum mechanics. Here we have an entity in a theory which is not directly observable - even in principle - yet the theory makes very specific predictions. It is possible to have such entities in a scientific theory, and we accept such entities in so far as the predictions which come from them are observed.

There is a certain appeal to this way of thinking, and I feel that an honest assessment of religious claims could proceed in this way. However, I think more people use "supernatural" as an excuse, a truly God-of-the-gaps, and stop all inquiry into alternatives. They see it as an explanation, not a label, but fail to make the next proper step of following the predictions - everything is a post-diction. You say that the universe is finely tuned for life because of God? Ok, what suggestion can you make for the next cosmological experiment to perform? What experiment would you propose to distinguish your theory (with supernatural entities) from a theory of physical necessity (i.e. the constants must be that way) or statistical necessity (i.e. the multiverse)? You say that the supernatural is needed to understand prayers being answered? Sure, but how would you test that? How would you explain the (many) failed experiments to demonstrate the efficacy of prayer, and propose an alternative? These experiments should be easy to propose, if one is not just applying a God-of-the-gaps. Of course, we don't see these experiments proposed.

So in summary, science cannot test supernatural explanations, because they are not defined. They can test specific predictions that incorporate unknown, and possibly unknowable, entities that have direct physical effects. In most cases of religious claims including supernatural effects there are no predictions made, and thus the term is empty and useless - an excuse for the intellectually lazy. Where specific predictions have been made (e.g. efficacy of prayer, existence of specific miracles, etc...) the predictions have not been supported. In this way, science can test God...and God has failed the test.