Part of the extended series Evidence for the Faith
Human rationality provides evidence that atheistic materialism is false. What follows here is a version of the Argument from Reason, which was originally made widely known by C. S. Lewis in Miracles, and was considerably expanded by Victor Reppert in C. S. Lewis’s Dangerous Idea: In Defense of the Argument from Reason.
Picture a box, a very large one, containing all the efficient physical causes in all of reality. On the atheistic-materialistic (A-M) view, there are no other causes. We say that causation is closed on the physical, or that there is causal closure there: that there cannot be any cause for any effect except for physical causes.
Now, on A-M, if rational inference exists at all (and it must), then it exists inside that box. It is a strictly physical process.
In contrast to that, the theistic view is that the universe of causation is larger, and that the very large box of physical causes is inside an even larger box that includes many other classes of causes. We would say that thoughts, in general, have causal roots in several different boxes. At this point I’m aware of the weakness of my analogy: boxing these things up is rather too cubical and physical of a way of thinking about them. I hope my point is clear, anyway, which is that causation is not closed on the physical, and that thoughts have causes from many different roots.
The A-M view disagrees. It says that a thought produced through rational inference is also a thought produced through entirely physical processes, and that logic in human thought is like the logic programmed into a computer. I don’t think that’s possibly true. There are many reasons for this.
First, there is the intentionality or “aboutness” problem. A-M proponents often ask us to think of the mind as being like a computer. It’s a useful exercise, because it helps us see how different rational thought is from mechanical signal processing. Logic gates in computers are never “about” the signal passing through them, nor are they “about” the information they carry. This is true of each individual switch, and it’s true of any large array of switches, for complexity cannot magically create “aboutness.” But rational inference is really a matter of being about the content being processed.
Modus ponens and modus tollens (Latin for, “two really common ways to run a logical deduction”) are algorithms. Plug any two true premises in, run the algorithm, and you get a sound conclusion—provided that the content actually is true, and provided also that the premises are related to each other in such a way that the algorithm actually applies. A computer’s logic is about the algorithm, not about the content; it will run the same algorithm equally well with any input. It doesn’t care whether the premises are true, and it doesn’t know whether they’re relevant to one another. it can only flip electronic switches according to a predetrmined program—for that’s what computers are: electronic switch-flippers.
Second, there is the related problem of the content’s being true or false. Physical systems cannot be true or false about anything. Nothing in my laptop is true about some other thing.
This is partly a consequence of the aboutness problem in physical objects, and it’s partly a matter of relationships. Consider the spreadsheet formula,
Microsoft Excel will return “TRUE” if you type those characters into a cell. But what is it in those characters that is true? Nothing. The expression is true only in the interpretation (and humans either have to ignore the first = or else interpret it differently than the second one). There is, after all, a reason we call it an expression: expressions always require intelligent interpretation; and the truth is not in the signs on the screen or the voltage states that produce the signs; the truth is in that which is being expressed.
Further: The computer is not reflecting on the values of 1 and 1, and considering whether their equality is something that obtains in wider reality. It’s not recognizing a correspondence between “1=1” and some general truth. The computer is throwing switches as it’s been programmed to throw them. It’s been set up to throw switches corresponding to the characters “TRUE” just in case the first number represented there is identical the second one. It knows nothing of truth. It knows nothing, actually, though it is a most impressive switch-throwing machine.
Third, there are clear differences between rational inferences and physical processes. Rational inferences don’t follow physical laws. There are no equal and opposite reactions in them, no inherent tendency towards entropy, no mass, no inertia, etc. They don’t have a size, a shape. They progress, but they don’t travel north, south, east, west, up or down.
As C.S. Lewis points out in the third chapter of Miracles, to the extent that we can ascribe inferences to physical causes, to that same extent we doubt their rationality. If Grandpa expresses an opinion, the response, “Grandpa says that because he’s tired,” is likely to be another way of saying, “Grandpa’s opinion probably can’t be trusted.”
Fourth, there is the problem of how safe it is to conclude that physical laws in physical brains could have any reliable effect of producing truth in response to their causal predecessors. There is no known mechanism whereby mere physical processes could reliably produce truth-related outputs.
I understand that evolution is commonly proposed as exactly that mechanism. There are two problems with that view, however, which Alvin Plantinga has put forth in his Evolutionary Argument Against Naturalism. To begin, evolution is not a truth-seeking activity of nature, it is a survival- and baby-producing activity. Beliefs need not be true to be effective in leading persons to survival and reproduction. Sure, there could be a relationship there, but if there is, it’s accidental, not essential. Further, it’s really quite a fantastic conjecture to suppose that in its producing humans who could survive in the bush and the caves, it produced a brain capable of non-Euclidean geometry, algebra with imaginary numbers (which turn out to be quite useful in electronics), and a whole huge host of other abstract ideas of which I have no clue, and yet which turn out to be classifiable in their contexts as true or false.
For those four reasons, or five if you split the last one in two, I seriously doubt that rational inference can be fully explained from within the merely physical box of causation. Furthermore, if what we name “rational inference” could be fully explained on physical causation alone, there would be little about it that was rational.
So if causation is closed on the physical, then rational inference is out the window. If one has concluded by means of logic that the physical world is all there is, then one has concluded that the world is the kind of place wherein logical, rational conclusions cannot be made. It is a conclusion that no one can conclude anything. That’s the contradiction inherent in atheistic materialism, and it’s one of the main reasons I’m convinced the world must be more than physical.
(Most of this content appeared in the form of a comment I wrote earlier this month.)
Update March 2: The Argument From Reason Redux