C. S. Lewis on Rationality and Moral Reasoning

I was blessed this Christmas to finally receive a copy of C. S. Lewis's The Weight of Glory, and Other Addresses. While reading one of the essays within, "Why I am Not a Pacifist", I found a passage where Lewis takes the time to explain his own understanding of what rationality is, and how to reason properly. While the overall essay is about whether pacifism is morally correct, Lewis starts by explaining how we can tell whether we are correct about anything at all:

The most useful analogy here is that of Reason—by which I do not mean some separate faculty but, once more, the whole man judging, only judging this time not about good and evil, but about truth and falsehood. Now any concrete train of reasoning involves three elements: Firstly, there is the reception of facts to reason about. These facts are received either from our own senses, or from the report of other minds; that is, either experience or authority supplies us with our material. But each man's experience is so limited that the second source is the more usual; of every hundred facts upon which to reason, ninety-nine depend on authority. Secondly, there is the direct, simple act of the mind perceiving self-evident truth, as when we see that if A and B both equal C, then they equal each other. This act I call intuition. Thirdly, there is an art or skill of arranging the facts so as to yield a series of such intuitions which linked together produce a proof of the truth or falsehood of the proposition we are considering. Thus in a geometrical proof each step is seen by intuition, and to fail to see it is to be not a bad geometrician but an idiot. The skill comes in arranging the material into a series of intuitable "steps." Failure to do this does not mean idiocy, but only lack of ingenuity or invention. Failure to follow it need not mean idiocy, but either inattention or a defect of memory which forbids us to hold all the intuitions together.

Now all correction of errors in reasoning is really correction of the first or the third element. The second, the intuitional element, cannot be corrected if it is wrong, nor supplied if it is lacking. You can give the man new facts. You can invent a simpler proof, that is, a simpler concatenation of intuitable truths. But when you come to an absolute inability to see any one of the self-evident steps out of which the proof is built, then you can do nothing. No doubt this absolute inability is much rarer than we suppose. Every teacher knows that people are constantly protesting that they "can't see" some self-evident inference, but the supposed inability is usually a refusal to see, resulting either from some passion which wants not to see the truth in question or else from sloth which does not want to think at all. But when the inability is real, argument is at an end. You cannot produce rational intuition by argument, because argument depends upon rational intuition. Proof rests upon the unprovable which has to be just "seen." Hence faulty intuition is incorrigible. It does not follow that it cannot be trained by practice in attention and in the mortification of disturbing passions, or corrupted by the opposite habits. But it is not amenable to correction by argument.

Before leaving the subject of Reason, I must point out that authority not only combines with experience to produce the raw material, the "facts," but also has to be frequently used instead of reasoning itself as a method of getting conclusions. For example, few of us have followed the reasoning on which even ten percent of the truths we believe are based. We accept them on authority from the experts and are wise to do so, for though we are thereby sometimes deceived, yet we should have to live like savages if we did not.

Lewis goes on to explain how the same principles that allow reasoning in general can be applied to moral reasoning:

Now all three elements are found also in conscience. The facts, as before, come from experience and authority. I do not mean "moral facts" but those facts about actions without holding which we could not raise moral questions at all—for we should not even be discussing Pacifism if we did not know what war and killing meant, nor Chastity, if we had not yet learned what schoolmasters used to call "the facts of life." Secondly, there are the pure intuitions of utterly simple good and evil as such. Third, there is the process of argument by which you arrange the intuitions so as to convince a man that a particular act is wrong or right. And finally, there is authority as a substitute for argument, telling a man of some wrong or right which he would not otherwise have discovered, and rightly accepted if the man has good reason to believe the authority wiser and better than himself. The main difference between Reason and Conscience is an alarming one. It is thus: that while the unarguable intuitions on which all depend are liable to be corrupted by passion when we are considering truth and falsehood, they are much more liable, they are almost certain to be corrupted when we are considering good and evil. For then we are concerned with some action to be here and now done or left undone by ourselves. And we should not be considering that action all unless we had some wish either to do or not to do it, so that in this sphere we are bribed from the very beginning. Hence the value of authority in checking, or even superseding, our own activity is much greater in this sphere than in that of Reason. Hence, too, human beings must be trained in obedience to the moral intuitions almost before they have them, and years before they are rational enough to discuss them, or they will be corrupted before the time for discussion arrives.

These basic moral intuitions are the only element in Conscience which cannot be argued about; if there can be a difference of opinion which does not reveal one of the parties as a moral idiot, then it is not an intuition. They are the ultimate preferences of the will for love rather than hatred and happiness rather than misery. There are people so corrupted as to have lost even these, just as there are people who can't see the simplest proof, but in the main these can be said to be the voice of humanity as such. And they are unarguable. But here the trouble begins. People are constantly claiming this unarguable and unanswerable status for moral judgments which are not really intuitions at all but remote consequences or particular applications of them, eminently open to discussion since the consequences may be illogically drawn or the application falsely made.

...

This then, is our first canon for moral decisions. Conscience in the…sense (of) our judgment as to what is right, is a mixture of inarguable intuitions and highly arguable processes of reasoning or of submission to authority; and nothing is to be treated as an intuition unless it is such that no good man has ever dreamed of doubting. The man who "just feels" that total abstinence from drink or marriage is obligatory is to be treated like the man who "just feels sure" that Henry VIII is not by Shakespeare or that vaccination does no good. For a mere unargued conviction is in place only when we are dealing with the axiomatic; and these views are not axiomatic.

I believe most moral realists would agree with Lewis here. I imagine most moral non-realists would object that our moral intuitions have no truth content: just because we have "unarguable intuitions" about morality doesn't mean they have anything to do with reality. They're just a way people feel, perhaps generated by natural selection as a way to propagate the species. The trouble is if moral intuitions are invalid then how can we trust that our other intuitions, the ones that tell us if A and B equal C then A is equal to B, are valid? And if they are valid, then why not our moral intuitions? 

Some may (and do!) reply that we can trust natural selection to provide us with intuitions that correspond with reality (that are "true") because believing true things has survival value. They then argue that there is no moral reality for our moral intuitions to correspond to, so it is impossible for them to be "true" or "false": therefore only our non-moral rational intuitions are trustworthy. Besides the fact that this begs the question of whether moral realism is true or false, there is the question of whether we can believe that truth corresponds with survival value. Such a belief is based on inference, yet rational inference is the things in question: if our rational "intuitions" are just the result of natural selection trying to propagate the species, how can we trust the inference that truth corresponds with survival? If it was not the case, and our rational intuitions were as subjective and lacking in truth content as our moral intuitions, then we would never be able to discover that fact: rational inference would not result in truth, so we would never find out the truth.

Lewis talks about this more in his book Miracles:

It is clear that everything we know, beyond our own immediate sensations, is inferred from those sensations…Put in its most general form the inference would run, ‘Since I am presented with colours, sounds, shapes, pleasures and pains which I cannot perfectly predict or control, and since the more I investigate them the more regular their behaviour appears, therefore there must exist something other than myself and it must be systematic’. Inside this very general inference, all sorts of special trains of inference lead us to more detailed conclusions. We infer Evolution from fossils: we infer the existence of our own brains from what we find inside the skulls of other creatures like ourselves in the dissecting room.

All possible knowledge, then, depends on the validity of reasoning. If the feeling of certainty which we express by words like ‘must be’ and ‘therefore’ and ‘since’ is a real perception of how things outside our own minds really ‘must’ be well and good But if this certainty is merely a feeling in our own minds and not a genuine insight into realities beyond them—if it merely represents the way our minds happen to work—then we can have no knowledge. Unless human reasoning is valid no science can be true.

It follows that no account of the universe can be true unless that account leaves it possible for our thinking to be a real insight. A theory which explained everything else in the whole universe but which made it impossible to believe that our thinking was valid, would be utterly out of court. For the theory would Itself have been reached by thinking and if thinking is not valid that theory would, of course be itself demolished. It would have destroyed its own credentials. It would be an argument that proved no argument was sound—a proof that there are no such things as proofs—which is nonsense.

…

An attempt to get out of the difficulty might be made along the following lines. Even if thoughts are produced by irrational causes, still it might happen by mere accident that some of them were true —just as the black dog might, after all, have been really dangerous though the man’s reason for thinking it so was worthless. Now individuals whose thoughts happened, in this accidental way, to be truer than other people’s would have an advantage in the struggle for existence. And if habits of thought can be inherited, natural selection would gradually eliminate or weed out the people who have the less useful types of thought. It might therefore have come about by now that the present type of human mind — the sort of thought that has survived — was tolerably reliable.

But it won’t do. In the first place, this argument works only if there are such things as heredity, the struggle for existence, and elimination. But we know about these things — certainly about their existence in the past— only by inference. Unless, therefore, you start by assuming inference to be valid, you cannot know about them. You have to assume that inference is valid before you can even begin your argument for its validity. And a proof which sets out by assuming the thing you have to prove, is rubbish. But waive that point. Let heredity and the rest be granted. Even then you cannot show that our processes of thought yield truth unless you are allowed to argue ‘Because a thought is useful therefore it must be (at least partly) true’. But this is itself an inference. If you trust it, you are once more assuming that very validity which you set out to prove.

…

   Some Naturalists whim I have met attempt to escape by saying that there is no ground for believing our thoughts to be valid and that this does not worry them in the least. ‘We find that they work’, it is said, ‘and we admit that we cannot argue from this that they give us a true account of any external reality. But we don’t mind. We are not interested in truth. Our habits of thought seem to enable humanity to keep alive and that’s all we care about’…The real answer is that unless the Naturalists put forward Naturalism as a true theory, we have of course no dispute with them. You can argue with a man who says, ‘Rice is unwholesome ’ : but you neither can nor need argue with a man who says, ‘Rice is unwholesome, but I’m not saying this is true’. I feel also that this surrender of the claim to truth has all the air of an expedient adopted at the last moment. If the Naturalists do not claim to know any truths, ought they not to have warned us rather earlier of the fact? For really from all the books they have written, in which the behaviour of the remotest nebula the shyest photon and the most prehistoric man are described, one would have got the idea that they were claiming to give a true account of real things. The fact surely is that they nearly always are claiming to do so. The claim is surrendered only when the question discussed in this chapter is pressed; and when the crisis is over the claim is tacitly resumed.

When faced with this argument, in my experience one type of objection is typically raised: that the non-rational process of natural selection could theoretically create a brain that has axioms built into it that are rationally valid, because rationally valid thoughts are better for survival than rationally invalid thoughts.

My primary response is that we cannot assume that rationally valid thoughts have the kind of survival benefit that natural selection selects for. There is significant evidence against this; namely that of all animals humans are the only species known to be capable of reason. If reason is so evolutionarily beneficial, why do the vast majority of all living things lack it? And the lack of it doesn’t seem to be harming most of them; there are trillions of bacteria out there that are extremely successful at reproducing themselves without having any idea that “If A=B and A=C then B=C.” If our rational axioms are produced through the long process of natural selection it wouldn’t mean that those axioms are true, only that they are useful for survival. You might as well ask whether a tiger's claws are true! Natural selection, by it’s nature, preserves qualities that are useful for reproduction. As a non-rational process we can’t even pretend that it selects for brains that believe things that are true, only that it selects for brains that believe things that are useful for reproducing; and if our hard-coded rational inferences and axioms are only those that are useful for survival, then we can only trust them to be useful, not true. We may be able to trust our reasoning to help us catch our dinner and raise our children, but little more than that. We certainly can’t assume that what seems rationally valid to us will allow us to understand phenomena we did not encounter in our ancestral environment, such as the nature of galaxies billions of miles away, or allow us to understand philosophical questions such as whether Naturalism is true or false.

Some may object still, and claim that useful beliefs are the beliefs that are also most likely to be true. Yet we do not apply this principle anywhere else in life. It is useful for the Party Apparatchik to believe that the Party Leader is always correct: if he publicly disagrees then he could be killed. Yet when we can explain his belief in the Leader primarily by non-rational causes (ie, the fact that he will be gunned down otherwise) we discount the likelihood of the belief being true.

Perhaps the most pertinent example of this would be religion, or "supernaturalism". The vast majority of humanity, as far back as we have recorded history, has believed in the supernatural: in gods, spirits, ghosts, devils, angels, devas, curses, etc. Even today only 7% of the global population identify as atheists. What's more, we have reason to believe that these beliefs are at least partially hard coded into us through natural selection (or cultural evolution) because they are useful beliefs. We could argue whether the studies that show religious people are happier or healthier are true or not, but this is not necessary: it is a well established statistical fact that those who are religious are more likely to reproduce. Muslims, Orthodox Jews, Amish, Hindus, and Christians all have higher TFRs than atheists or agnostics. Reproductive success is the heart and center of natural selection: if any belief is evolutionary useful, then Islam and Christianity must be. There's far more evidence of that then there is that belief in the transitive property is useful for reproduction.

Yet if I argued that you should become a Muslim because they have the most reproductively useful beliefs (TFR of 2.9) I don't imagine anyone would be much convinced, because we all know that there is no necessary connection between something being reproductively useful, and it being true. Perhaps it is more likely to be true because it is reproductively successful, but given how all of us, Naturalist or Supernaturalist, believe that the vast majority of humans who have ever lived have been almost completely wrong about almost everything they believed that they couldn't verify with their own eyes (whether that's a belief in Thor or geocentricism or miasma) we clearly don't trust that useful beliefs are also true. There is no necessary connection.

So we believe our rational "intuitions" exist because they are useful (as we must believe, if we believe that they are the result of natural selection which only preserves what is useful) then we can only trust them to be useful (most of the time anyway: gills are a useful adaptation until you're thrown into a desert). They have as much connection to being true as Islam does, or wings, or placental birth.

Yet if our rational intuitions are only useful and not true then how can we believe that Naturalism is true? We don't directly observe Naturalism, it's a philosophical position derived through rational inference. So why should we believe that it's true? And if we don't believe it is true, then why should we believe that our rational intuitions are merely the product of irrational processes aimed at reproductive fitness?