The Argument from Reason and Lewis's Post-Anscombe Revision

IV. The Argument of the First Edition
In the first edition of Miracles, Lewis presents the version of the argument from reason that Anscombe criticized. We can formalize it as follows:
1. No thought is valid if it can be fully explained as the result of irrational causes.
2. If naturalism is then all beliefs can be explained in terms of irrational causes.
3. Therefore, if naturalism is true, then no thought is valid.
4. If no thought is valid, then the thought, “materialism is true,” is not valid.
5. Therefore, if materialism is true, then the belief “materialism is true” is not valid.
6. A thesis whose truth entails the invalidity of the belief that it is true ought to be rejected, and its denial ought to be accepted.
7. Therefore, naturalism ought to be rejected, and its denial accepted.
This is the argument that drew the criticisms of Roman Catholic philosopher and
Wittgenstein student Elizabeth Anscombe. This critique is significant because of way in which it forced Lewis to develop and refine his arguments. Entirely too much attention has been paid to the putative psychological effects of this controversy, and conclusions concerning Lewis’s success as a Christian apologist have been drawn on the basis of his supposed emotional reaction to Anscombe’s challenge. Very often this is done without regard to the content of this exchange, and often this is done by people who show by their discussion to have no real understanding of the relevant philosophical issues. Such a procedure, as Bertrand Russell once said of a thesis in the philosophy of mathematics, has all the advantages of theft over honest toil.
V. Anscombe’s First Objection: Irrational vs. nonrational.
Is it correct for Lewis to talk about physically caused events as having irrational causes? Irrational beliefs, one would think, are beliefs that are formed in ways that conflict with reason: wishful thinking, for example, or through the use of fallacious arguments. On the other hand, when we speak of a thought having a non-rational cause, we need not be thinking that there is any conflict with reason.
While this is claim seems correct, it hardly puts an end to the argument from reason. The problem arises when we consider what a naturalist typically believes. Let’s take the theory of evolution as an example. Naturalists are always big on evolution, since invariably they must assign to the evolution process the task of producing the incredibly complex features, say, the human eye. Traditionally, the existence of the human eye was thought to be so efficient and complex that it had to be the handiwork of God. After all, the artificial replacement of vision by modern medicine is still the stuff of science fiction. Nevertheless, the naturalist is undaunted; she is persuaded that it is all the result of natural selection. However, a naturalist does not merely need to believe that we are the product of evolution through random variation and natural selection, she also has to believe that there was a process of scientific inference that led Charles Darwin to reach his conclusions about natural selection and how it works. Naturalism really does require the existence of rational inferences in order to be legitimate. Now a rational inference is a rational process. A valid or sound argument is an argument on paper in which the premises are true and the conclusion follows logically and inevitably from the premises and therefore must also be true. A rational inference, however, is not just a paper argument, it is an act of knowing on the part of a person that recognizes the content of the premises, accepts the premises as true, perceives the logical relationship between the premises, and concluded that the conclusion must be true as well. In short, the naturalist must believe not merely that their own beliefs were not produced by irrational causes, they must maintain that the conclusion that evolution is true was produced, in the mind of Charles Darwin and in their own mind, as the result of a rational process. Otherwise there would be no reason to prefer the deliverances of the natural sciences to a blind acceptance of the Book of Genesis as a way of forming one’s beliefs concerning the origin of species.
For that reason, it is possible to restate Lewis’s argument in such a way that it does not make reference to irrational causes, and indeed in Lewis’s revised chapter the phrase “irrational causes” does not appear.
1. No belief is rationally inferred if it can be fully explained in terms of nonrational causes.
2. If naturalism is true, then all beliefs can be fully explained in terms of nonrational causes.
3. Therefore, if naturalism is true, the no belief is rationally inferred.
4. If any thesis entails the conclusion that no belief is rationally inferred, then it should be accepted and its denial accepted.
5. Therefore, naturalism should be rejected and its denial accepted.
VI. Anscombe’s Second Objection: Skeptical Threats
Anscombe also objected to the idea that Lewis had argued that, if naturalism were true, then reasoning would not be valid. She asks, “What can you mean by valid beyond what would be indicated by the explanation you give for distinguishing between valid and invalid reasoning, and what in the naturalistic hypothesis prevents the explanation form being given or meaning what it does.” This is a paradigm case argument, and the point is this: We can ask whether this particular argument is a good one, but does it really make sense to argue that reasoning might itself be invalid? Anscombe maintains that since the argument that some particular piece of reasoning is invalid involves contrasting it with some other kinds of reasoning that are valid, the question “Could reasoning really be valid?” is really a nonsense question.
One way of using the argument from reason would be to use it as a skeptical threat argument. The idea is that if naturalism is true we will be unable to refute skeptical arguments against reasoning in general. The problem here is that it is far from clear that anyone, naturalist or not, can refute skepticism about reasoning, nor is it considered any great merit for any metaphysical theory that it would be possible to refute this kind of thoroughgoing skepticism. And, if we need to refute skepticism in order to accept some world-view, then it is not at all clear that theism will do that either. If we use our theistic beliefs to defend the basic principles of reasoning, then we would have to formulate that into an argument and then presuppose our ordinary canons of logical evaluation in the presentation of that very argument, thereby begging the question.
Rather, one can, it seems to me, present the argument from reason as a best explanation argument. One should assume, at least to begin with, that human beings do reach true conclusions by reasoning, and then try to show, given the fact that people do reach true conclusions by reasoning, that this is best explained in terms of a theistic metaphysics as opposed to a naturalistic metaphysics. Now if we present the argument in this way, and then an opponent comes along and says “I see that your argument presupposes that we have beliefs. I don’t think we do, so your argument fails, then we can reply to him by saying that if there are no beliefs then you don’t believe what you’re saying. Consequently the status of your own remarks as assertions is called into question by your own thesis that there are no beliefs, and that this is going to end up having a devastating effect on the very sciences on which you base your arguments. Presenting the argument in this way, it seems to me, gets around the problems based on the Paradigm Case argument.
VII. Anscombe’s Third Objection: The ambiguity of “Why,” “Because,” and “Explanation
The third and main Anscombe objection to Lewis’s argument is that he fails to distinguish between different senses of the terms “why,” “because” and “explanation.” There are, she suggests, four explanation-types which have to be distinguished.
1. Naturalistic causal explanations, typically subsuming the event in question under some physical law.
2. Logical explanation, showing the logical relationship between the premises and the conclusion.
3. Psychological explanations, explaining why a person believes as he/she does.
4. Personal history explanations, explaining how, as a matter of someone’s personal history, they came to hold a belief.
She suggests that arguments of different types can be compatible with one another. Thus a naturalistic causal explanation might be a complete answer to one type of question with respect to how someone’s belief came to be what it was, but that explanation might be compatible with an explanation of a different type.
Now what is interesting is that Lewis, in reformulating his own argument, not only draws the distinctions on which Anscombe had insisted; he actually makes these distinctions the centerpiece of his revised argument. He makes a distinction between Cause and Effect relations on the one hand, and Ground-and-Consequent relations on the other. Cause and effect relations say how a thought was produced, but ground-and-consequent relations indicate how thoughts are related to one another logically. However, in order to allow for rational inference, there must be a combination of ground-consequent and cause-effect relationships which, Lewis says, can’t exist if the world is as the naturalist says that it is.
Claiming that a thought has been rationally inferred is a claim about how that thought was caused. Any face-saving account of how we come to hold beliefs by rational inference must maintain that “One thought can cause another thought not by being, but by being, a ground for it.”
However, there are a number of features of thoughts as they occur in rational inference that set them apart from other beliefs.
Acts of thinking are no doubt events, but they are special sorts of events. They are “about” other things and can be true or false. Events in general are not “about” anything and cannot be true or false….Hence acts of inference can, and must be considered in two different lights. On the one hand they are subjective events in somebody’s psychological history. On the other hand, they are insights into, or knowings of, something other than themselves. So here we already have three features of acts of thinking as they occur in rational inference. First, these thoughts have to be about something else, and second, they can be true or false. Second, their propositional contents must cause other thoughts to take place. But there is more:
What from the first point of view is a psychological transition from thought A to thought B, at some particular moment in some particular mind is, from the thinker’s point of view a perception of an implication (if A, then B). When we are adopting the psychological point of view we may use the past tense, “B followed A in my thoughts.” But when we assert the implication we always use the present—“B follows from A.” If it ever “follows from” in the logical sense it does so always. And we cannot reject the second point of view as a subjective illusion without discrediting human knowledge. So now, in addition to the three features of thoughts as they occur in rational inference, we can add a fourth, that is, that the act of inference must be subsumed under a logical law. And the logical law according to which one thought follows another thought is true always. It is not local to any particular place or time; indeed laws of logical obtain in all possible worlds.
Lewis then argues that an act of knowing “is determined, in a sense, by what is known; we must know it to be thus because it is thus.” P’s being true somehow brings it about that we hold the belief that P is true. Ringing in my ears is a basis for knowing if it is caused by a ringing object; it is not knowledge if it is caused by a tinnitus. As Lewis puts it:
Anything that professes to explain our reasoning fully without introducing an act of knowing thus solely determined by what it knows, is really a theory that there is no reasoning. But this, as it seems, is what Naturalism is bound to do. It offers what professes to be a full account of our mental behaviour, but this account, on inspection, leaves no room for the acts of knowing or insight on which the whole value of our thinking, as a means to truth, depends.