1. Are Supernatural Claims Testable?
Testability is often taken as an easy way to separate the natural from the supernatural.
For example, the Geological Society of America, in its position statement on evolution, asserts, "Creationism is not science because it invokes supernatural phenomena that cannot be tested." Yet, elsewhere in the same statement, it argues for the empirical confirmation of an ancient earth and the evolutionary origin of all life. This implies that creationism is testable, after all. A similar contradiction can be found in the statement of the United States National Academy of Sciences, Science and Creationism (1999), which asserts that creationism and other claims of supernatural intervention are scientifically untestable and, also, that these have in fact been falsified.
Even some naturalist philosophers, who might be expected to be more astute, make this error. M. Mahner and M. Bunge (1999, "The incompatibility of science and religion sustained: A reply to our critics", Science & Education 5(2), 189-199) assert that supernatural claims are in principle untestable but, in the same paper, argue also that many supernatural claims are incompatible with scientific findings.
Carrier grants that supernatural claims may well be testable. He believes, for example, that young Earth creationism has been conclusively falsified by empirical evidence.
One difficulty, according to Carrier, is that Christians try to make the Bible untestable by suitably re-interpreting the Bible so as to avoid any possible risk of scientific disproof. Thus we get hermeneutical inventions such as old Earth creationism, theistic evolution, or the notion that Gen.1-11 is meant to convey a purely theological message.
Carrier comments, "Supernaturalists now make every effort to place their beliefs beyond meaningful test precisely because their beliefs have failed countless tests—indeed, most have been decisively refuted. This is evidence of the desperation of their position but not of the untestability of more honest supernatural claims."
I think that Carrier is quite correct in this observation. The proper apologetic response is not to subvert the plain meaning of the biblical text--as many modern theologians are prone to do--but, rather, to challenge the naturalist interpretation of the empirical evidence.
Thus testability is not a satisfactory criterion to demarcate naturalism from supernaturalism..
2. Mind Versus Matter
So what does distinguish naturalism from supernaturalism? Carrier suggests that all formulations of naturalism reduce to this: naturalism is true if everything that exists is causally reducible to the nonmental.
In other words, if everything that exists--including thoughts, perceptions, and emotions--is causally reducible to different arrangements of matter-energy in space-time, then naturalism is true. According to Carrier, "This definition allows for the possibility of irreducibly mental things (like epiphenomena), so long as those things are fully and completely caused to exist by nonmental things (like brains)." [Considering thoughts to be mere epiphenomena means that they are caused by physical events but can themselves have no physical effect. This reduces our sense of freewill in, say, deciding to read a book, as mere illusion.]
This definition is consistent with the evolutionary assumption that everything that exists has naturally emerged from an intial state consisting only of matter-energy. It is closely related to the materialist view of physicist Lisa Randall, discussed in an earlier post How (Atheist) Scientists Lose Their Way.
In contrast, supernaturalism always involves some sort of mind over matter. Hence, as a distinguishing feature, Carrier proposes, "if naturalism is true, everything mental is caused by the nonmental, whereas if supernaturalism is true, at least one thing is not."
3. Refuting Naturalism
Let's examine some implications of Carrier's criterion.
First, note that the alleged disproof of supernaturalism, at least in its young Earth creationist form, assumes that naturalism is true. The difficulty is that we cannot directly observe the past. The best we can do scientifically is to reconstruct it via theoretical models, based on naturalist assumptions. However, if supernaturalism is in fact true then the past may have been quite different from naturalist projections. Perhaps, for example, the universe was created in mature form only a few thousand years ago.
Second, naturalism thus defined seems impossible to prove. How could one possibly demonstrate that all possible entities--at every time and place--are causally reducible to matter-energy? Supernaturalism, on the other hand, needs only a single counter-example. There have been numerous past claims of miracles, angels and the like. Naturalists must demonstrate all these claims to be false.
Third, concepts such as belief, truth, falsity, purpose, meaning, intention, and even naturalism are all mental entities that are about other things. As such, they can never be causally reduced to material things and still retain any genuine meaning. Hence, if naturalism were true, we could have no rational conception of it.
Indeed, if all our beliefs--as mental entities--have only physical causes then they can reflect only physical states rather than rational conclusions. As any philosopher knows, a physical "is" cannot generate a logical or moral "ought". It follows that belief in naturalism cannot be the result of a coherent train of thought. Therefore, any rational argument for naturalism is by its very nature self-refuting.
In sum, if we know anything, it must be that naturalism--as defined by Carrier--is false, for if naturalism were true, we would know nothing at all.