Graphic Rule

2.508. The pursuit of probability

There was a time when the possibilities of analytic statements had been greatly developed, while the distinction between analytic and synthetic statements was still only faintly grasped. That time included the ancient Greeks, with their development of mathematics; and it also included the seventeenth century, with the further development of mathematics by Descartes and others. During this time there implicitly existed the hope that all statements whatever would turn out to be analytic, in other words that mere thinking without the aid of the senses would ultimately tell us the truth-value of every statement, in other words that the whole of science would become like mathematics, and specifically like what they regarded as the perfect form of mathematics, namely Euclid's Elements. This hope became clearest and strongest in the seventeenth century, when Spinoza composed an Ethics in Euclidean form, and Newton composed a kinetics in Euclidean form, and Steno composed a description of muscle in Euclidean form.

A consequence of this belief that all knowledge could become mathematical or analytical was that they expected and demanded to achieve in all spheres the wonderful certainty that we achieve in mathematics. All science whatever was to be absolutely certain and absolutely proved. Everything less than absolute certainty was useless and to be rejected. Descartes thought that he should reject, as absolutely false, everything about which he could imagine the slightest doubt. Accordingly, he had no use in his account of scientific method for the notion of hypothesis, or even for the notion of probability. It is a waste of time to be setting up hypotheses and estimating probabilities if we can obtain certainty. Every reasoning was either certain or no use at all. Pascal expressed the general attitude when he wrote: 'I am not content with the probable; I seek the sure.'

This hope was vain. Now that the distinction between analytic and synthetic statements has become much clearer, we see that it is false that all statements are analytic and ideally decidable by intellectual considerations alone. The question whether this lobster has recently shed his shell can never be reasonably decided without the aid of the senses; nor can the general question whether lobsters shed their shells from time to time. The laws of nature, if they exist, are not entailments, and cannot be discovered by mere logic and mathematics. The laws of nature hold between things; but entailments hold between propositions. Statements asserting laws of nature or facts of history are one and all synthetic, so that to decide on their truth-value without any appeal to experience would be folly. Even a great deal of what is commonly called mathematics is in fact synthetic and uncertain. For example, the geometrical theorems of Euclid are synthetic. What is analytic in Euclid's geometry is at most his proofs, that is, his statements that his theorems are entailed by his postulates. By far the greater number of the statements that seriously concern us are synthetic.

When people begin to realize that the Cartesian hope is vain, and that the absolute proof of mathematics never will be achieved for propositions describing the real world, they sometimes react with a despairing scepticism. They think that, if human thought cannot achieve mathematical certainty about existence, it is no good at all; and they resign themselves to the view that 'men cannot discover philosophical truths by the sole use of their natural faculties', or the more sweeping view that 'truth is impossible to attain'. 'I look on all sides', wrote Pascal (Pensées, xiv. 2, Havet), 'and everywhere I see nothing but darkness.' That was because to him darkness and certainty were the only alternatives. Hume was a sceptic for the same reason; he abandoned the Cartesian conviction that our knowledge could be mathematically certain, but retained the Cartesian conviction that it ought to be.

From this despairing scepticism it is an easy and common step to unreason. Pascal took it when he recommended us to decide whether there is a god by means of a wager (op. cit. x. 1, Havet). Many take it in the form of saying to themselves: 'Since certainty is unattainable, I can only believe whatever my intuition tells me.' Thus Newman (Grammar of Assent, p. 343), after deciding rightly that there cannot be a science of reasoning sufficient to compel certitude in concrete conclusions, infers wrongly that we should 'confess that there is no ultimate test of truth besides the testimony borne to truth by the mind itself'. Thus reason is brought into disrepute by being identified with the certainty and proof that are obtainable only concerning analytic statements, as it becomes clear that most statements are not analytic and cannot be settled by mere deduction.

The right reaction, to the discovery that mathematical certainty is impossible about most statements, is not to abandon reason, but to include in it the pursuit of probability as well as the pursuit of certainty. Then we can say, contrary to Hume, that it is by reason that we believe in the existence of bodies as well as in the truths of mathematics. The reasonable man does not say 'Give me certainty or I despair'. He looks for certainty about analytic statements only. About synthetic statements he is content with probability. On the considerations available to him at the time concerning a given statement, either the statement is more probable than its contradictory, or it is less probable, or it is equally probable. In the last case he suspends judgement; otherwise he adopts for the present the more probable of the two contradictories. In each pair of contradictories he habitually tries to estimate which is the more probable, and he habitually adopts that.

'All is uncertain', said Hume. But it is extremely unlikely that the probability of every proposition is exactly equal to the probability of its contradictory, so that we ought to suspend judgement about all propositions. It is still more unlikely that a pair of contradictories could both be improbable rather than probable. It is perfectly obvious, when you come to think of it, that some propositions are far more probable than their contradictories, and therefore ought to be adopted. If we were to follow Descartes's advice, and reject as false everything that is not mathematically proved, we should be rejecting both of two contradictories, which is absurd since one of them must be true. If we were to reject as 'invalid' every consideration that did not amount to a strict deductive proof, we should be perversely depriving ourselves of many reasonable aids to picking the true contradictory. The division of arguments into valid and invalid is a remnant of the exclusively deductive and mathematical way of looking at knowledge in general; it should be confined to mathematics. In questions of history and nature, we should give up bothering whether the proposition is certain or not. We should ask ourselves instead whether we will adopt the proposition or reject it, and in deciding we should remember the significant fact that to reject a proposition is to adopt its contradictory. It is not a 'burden of proof' that reason lays on us in existential and practical questions. It is a burden of judgement, of judging which is the more probable of the two contradictories in view of the available considerations.

Some people think that this is impossible because probability entails certainty, so that where nothing is certain nothing can be probable either. I myself argued this view at some length in my first book, The Province of Logic, London, 1931. I withdraw it now. I overlooked the fact that every statement has its contradictory, and that if the one is improbable the other must be probable.

To wild despairs such as 'truth is impossible to attain' the reasonable man opposes the following antiseptic reflection. Of any pair of contradictories, say 'there is a god' and 'there is no god', one is true and the other is false. Now each of these statements has been believed. Therefore somebody has believed a true statement. Therefore truth is possible to attain. If the sceptic replies that he meant that we cannot know for certain which is true, the reasonable man answers that we can often make a reasonable judgement as to which is more probably true, and then it is wise to be content with that. It is unwise to insist on all or nothing, for the result of doing that is to get nothing.

The probable must not be opposed to the true. Plato pointed out that it is wrong to say 'let us seek what is probable, not what is true' (Phaedrus 272 DE). That amounts to the immoral advice: 'seek to convince, and do not mind whether what you say is true.' The right maxim is: let us seek what is probably true rather than what is certainly true, since certainty is unobtainable outside mathematics. The probable is not opposed to the true or the false, but to the certain and the improbable.

2.509. Respect for evidence

Where synthetic statements are concerned, the pursuit of probability is the right middle way between two wrong extremes. One of these wrong extremes is the pursuit of certainty, which I have described. The other is the acquiescence in mere possibility. Some people speak as if they thought that, since certainty is unattainable, they might accept any proposition that is not known to be impossible. 'May it not be that ... ?', they ask; and the implication is that if it is possible it is actual. You can be sure that anyone who recommends a proposition by prefixing to it the words 'May it not be that ... ?' is less than a firstclass reasoner. For of course it may be. Every synthetic proposition is logically possible. But so is its contradictory. That too may be so. Hence the right question is not which of them may be so (for each of them may be so), but which of them has the better evidence. Similarly, the right question is not which of them has some evidence in its favour (for usually each of them has some evidence in its favour), but which of them has the weightier evidence in its favour.

I draw your attention to one specially bad and very common form of acquiescing in mere possibility. People sometimes adopt a proposition for true merely on the ground that they do not know it to be false. 'You cannot prove that it is false', they say; and they regard this as justifying them in holding it true. They speak as if our not knowing a certain statement to be false were good evidence that it is true. I call this the argument from ignorance, though I think it is not what everybody has meant by the phrase 'argument from ignorance'. It is fallacious, because ignorance is not a good ground for asserting anything except that we are ignorant. The question in matters of fact is not what we can or cannot prove, but which of the two contradictories has the better evidence. That we cannot prove that there is no god is irrelevant. The right question is which of the two contradictories, 'there is a god' and 'there is no god', has the better evidence. Even the enlightened Joseph Butler endorsed a case of this fallacy, when he wrote that 'due sense of the general ignorance of man would ... beget in us a disposition to take up and rest satisfied with any evidence whatever, which is real' (Sermon XV, § 10). No amount of ignorance can make it right to consider only one side of a question. However ignorant we may be, we should consider the evidence for both of the two contradictories, and decide which is the heavier.

The argument from ignorance is often concealed in the form of a question. When a man has no argument whatever in favour of his thesis that pigs have wings, he can still impose it on many of the unwary by putting it in the form of a question: Who can say whether after all pigs may not have wings? The implication is that, in view of the general ignorance of man, you would be a rash fool to assert that pigs have no wings.

The worst form of the argument from ignorance masquerading as a question is the 'how can' or 'how could' form. For example, 'the problem still remains how can anyone understand the difference between good and evil unless he has been given at birth a natural feeling for it'. The writer of this sentence implies that your ignorance how this can be done is good evidence that at cannot be done. But, of course, it is no such thing. If it were, we could prove that the moon is made of green cheese, thus: 'Unless the moon were made of green cheese, how could it have that patchy and crumbly appearance?'

The 'how can' form of argument is dishonest. By using this form the speaker conceals the fact that it is he who is making an assertion and thus incurring a responsibility. He insinuates falsely that the responsibility is all on you for not admitting the assertion. Instead of openly making his assertion and taking the responsibility for it, he insinuates that you ought to believe it unless you can answer some 'how' question. The honest thing would be to say: 'The moon has a patchy and crumbly appearance, and most things of that appearance are made of cheese, therefore the moon is probably made of cheese.' And then the weakness of the thought would be apparent. The advantage of the dishonest 'how could' form is that, while it must mean just this, it declines to confess it and take the responsibility for it.

Often the best way to meet one of these bogus questions is to reply: A question is not an argument, only statements can be arguments. Your question is an argument only if it is a way of stating that....' For thus you bring his assertion into the open and force him either to acknowledge it or to drop it. I recommend to you the idea that a question is not an argument. It invalidates a surprisingly large number of letters to The Times.

So much for the mistake of acquiescing in mere possibility. Concerning synthetic statements the reasonable man neither acquiesces in mere possibility nor demands mathematically certain proof, but estimates probabilities and adopts the more probable of the two contradictories.

The truth-value of a synthetic statement cannot be found by merely considering its meaning and entailments. It can be found only by both considering its meaning and doing something more. What more? What further criterion comes in? A great many synthetic statements profess to describe the world that is, or the world that has been or will be. This is done by all particular reports, such as 'the cat mewed when it saw her', and 'the sun will be eclipsed tomorrow', and 'Caesar crossed the Rubicon'. It is done by general reports, such as 'cats make a mewing sound but dogs do not', and 'eclipses of the sun are less frequent than those of the moon', and 'all bodies attract each other with a force reversely proportional to the square of their distance'. There is a class of synthetic statements which are all what we may call in a very general sense descriptions. This class includes all or most of the statements of history and natural science. It excludes statements of value, of goodness and badness.

The criterion of the truth-value of descriptive statements may be labelled 'correspondence'. These statements are false if they fail to correspond to the world, if what they assert to be so is not so. For example, if someone tells you that your waistcoat is on fire, you look down to see if it is. You look to see if what he asserts to be so is so, if his assertion corresponds to the facts.

How do we tell whether the statement corresponds to the world? It is impossible to answer this both generally and completely. It is a matter of gradually accumulating more and more knowledge, both about the meanings of statements and about the world they profess to describe. In the case of 'your waistcoat is on fire', it is a matter of knowing the rules for the use of 'your', the rules for the use of 'waistcoat', and the rest of the meaning, and then using your senses on the world to discover whether what you experience corresponds to what the statement asserts. The evidence for or against 'your waistcoat is on fire' is very simple. It is just what you can observe now in the region of your waistcoat. But evidence becomes a far more complicated and doubtful matter for statements describing absent parts of the world, and for general descriptions of the world, and for conditional descriptions of the world. Questions of judgement and choice soon enter in. Do you judge Tacitus to be a truthful and cautious historian? Was Fisher, or whoever it was, judicious in choosing the standard statistical deviation as he has? These matters will never be completely decided. The process of improving and adding to our decisions is the study of evidence.

We have seen two great domains of statement in which the criteria of truth are different. For analytic statements the criterion is selfcontradiction. For descriptive synthetic statements the criterion is the real world. There appears to be at least one more domain with its own criterion, and that is the domain of ethical and practical synthetic statements, such as 'we ought to tell the truth', 'Beethoven's music is better than Bach's', 'political equality is a good thing', 'disorder is undesirable'. In choosing between such a practical statement and its contradictory, we cannot use the criterion of selfcontradiction because they are synthetic statements. Nor can we use the criterion of correspondence with the real world; for these statements do not profess to describe the world, and therefore nothing the world can do will make them either correspond to it or fail to correspond.

In the first and second domains men are very largely agreed as to what the right criterion is and how to use it. At least, they tend very largely to use the same criteria in the same way and come out with the same results, though they disagree in their theoretical account of the nature of these criteria. But in the third domain, that of practical statements, men neither give the same theory of the criterion nor use the same criterion nor come out with the same results. In one sense there is no criterion for this domain; that is, there is no universally agreed and reliable criterion. But people have their own criteria. I have told you earlier some of my own principles of judgement.

Does this mean that we cannot properly speak of evidence for and against practical propositions, and that the notion of reason as including respect for evidence has no application here? No, it only means that in this sphere evidence is more doubtful and shifting than in the other two. There still is evidence of a kind. Plenty of arguments are produced on practical matters, arguments about the factual situation on which practice has to be based, about the consequences of proposed actions, about the criteria to be employed, about their relative strengths, about the way to apply them. Here as elsewhere it is possible to distinguish between the reasonable man who finds and respects the evidence, and the unreasonable man who does not. It is common and correct to distinguish between those who adopt policies after careful consideration and those who do not.

Reason, then, includes respect for evidence. The decision which of two contradictories is the more probable is to be made by examining the evidence for and against each. And this apparently applies to all classes of statements, though the evidence is of very different kinds in the different classes.

Let us observe some of the forms of disrespect for evidence.

One common form of disrespect for evidence is the habit of believing a proposition not because it has the better evidence but because its contradictory is painful. Somebody has said that 'I could not rest in a truth were I compelled to regard it as hateful'. Christians often recommend their doctrines on the ground that they are comforting, whereas their contradictories are depressing. You can find a striking example of this in Newman's Grammar of Assent, p. 305, which shows how powerful it can be.

Whether to believe the contradictory that has the better evidence, or to believe the one that gives more comfort, is one of our profoundest and most important decisions. Each alternative is often chosen. I am strongly in favour of choosing to believe the contradictory that has the better evidence, because that makes us more likely to believe truly. That one of a pair of contradictories frightens me is no evidence as to which of them is true. Believing truly is in the long run likely to comfort us more than believing falsely; and anyhow it is beneath our dignity as human beings not to seek a correct view of things.

Another form of disrespect for evidence is to reject it in favour of intuitions or hunches. Intuitions are necessary sometimes, namely when we have to make a decision but cannot in the time available find any evidence on which to base it. For instance, if you must decide this instant whether an attacker will shoot at your head or your heart. Furthermore, intuitions are sometimes good evidence in themselves. If there is a person whose intuitions in a certain field have turned out right more often than not in the past, then the fact that he now intuits a certain statement in the field to be true is good evidence that it is true. But an intuition is at best only one piece of evidence among other possible pieces. It is always capable of being overthrown by further considerations. It never justifies us in neglecting to look for other pieces of evidence, or in neglecting to put them also on the scales when they appear. And Bishop Gore's view that we are justified in believing an intuition if it gives us strength is very bad indeed.

Another form of disrespect for evidence, which may overlap with the foregoing, is the fanatic's deliberate hostility to evidence and rejection of it, what Professor Campbell called 'the blind uncritical devotion to an idea or cause which is so utterly sure of its own rectitude that "examination of the evidence" seems mere meaningless waste of labour' (Philosophy, 1950, p. 119). 'I believe it because it is absurd.' The fanatic, if he has to defend his view, does so by force or fury or intimidation or sarcasm or authority or mollification, all means which, while they often convince, never contribute to the determination of a truth-value.

The commonest form of disrespect for evidence is mere carelessness or thoughtlessness or failure to realize what is required. We assume that our meagre experience of some foreign nation is good enough evidence for generalizations about it, not out of any contempt for evidence, but out of mere ignorance of what sort of data such generalizations demand. We write home that 'the hotels here are' so-and-so, after having stayed in one and looked into two others. Respect for evidence involves knowing that evidence usually does not come without work, and hence involves searching for the good evidence.

A subtler form of disrespect for evidence is pretending that we have evidence when we have not. We may do this to others or to ourselves. 'Evidence is accumulating', we say. But how much has actually accumulated? For only that counts. To assume that in the future there will be more is to disrespect evidence.

Respect for evidence, or at least respect for the evidence of the senses, is the main element missing from the ideal of reason in Plato's dialogues. We may roughly say that he found reason merely in deductiveness, but we find it in deductiveness plus inductiveness. Mere deductiveness by itself constituted reasonable thinking according to Plato. But, since the distinction between analytic and synthetic propositions has become plain, we have to say that mere deductiveness is reasonable only in mathematics. Everywhere else it must be joined with inductiveness.

The theory of deduction and induction is logic, and logic is thus an important part of the development of the ideal of reason. It is a normative science; that is to say, it lays down norms of how we ought to think; it states an ideal. But do not take me as implying that every respected textbook of logic does this adequately, or even tries to. Among those that make a good shot at it are Pascal's Spirit of Geometry, the Port-Royal Logic, the third and fourth books of Locke's Essay on Human Understanding, John Stuart Mill's System of Logic, W. K. Clifford's Ethics of Belief, and Cohen and Nagel's Introduction to Logic and Scientific Method. Many other logic books are not normative at all, but parts of mathematics exploring the nature of entailment. In them you will find no practical help towards becoming reasonable. It is a pity that the same word 'logic' has to cover both a purely theoretical and a very practical inquiry.

2.510. Tentativeness

The reasonable man holds his views tentatively rather than dogmatically. He bears in mind the possibility that at some future time the weight of evidence may point the other way. He refrains from thinking that 'there cannot possibly ever be a good reason for changing this view, and so I will never listen to any argument against it'. His views are hypotheses rather than dogmas, laid down to be tested by their consequences and other connexions, and to be exchanged for their contradictories if the network thus revealed ever indicates so. Plato, to whom the tentativeness of reason was clear, made his 'Socrates' say that 'we must submit to the argument until somebody persuades us with a better one' (Rp. 388 E).

Tentativeness is not the same thing as hesitation or indecision. Holding one's opinions always tentatively neither is nor involves being always hesitating and indecisive. Descartes in a well known passage records his decision to act firmly on practical principles although they were provisional, and that was reasonable.

Tentativeness is not the same thing as the pursuit of probability; for we should be tentative about our mathematical opinions as well as about our physical and historical and practical ones. We may be wrong in thinking a certain proposition to be an analytic truth, as well as in thinking it to be synthetic truth. Euclid's postulate about parallels was for long mistakenly thought to be a mathematical certainty. The contrast between mathematics and chemistry is not that we are never mistaken in the one, whereas we are sometimes mistaken in the other. We are sometimes mistaken in each of them. The contrast is that the truth of true mathematical propositions depends only on their meaning, whereas the truth of true chemical propositions depends also on the nature of the world. Descartes should have extended to all his opinions the tentativeness which he rightly chose in practical matters.

Tentativeness does not involve always listening to every argument. If you were hurrying to a vital appointment, and a queer-looking stranger stopped you in the street and asked you to listen there and then to an argument that the earth is flat, reason would not require you to comply. Nor need you read all the volumes of the Society for Psychical Research before deciding that there are no ghosts. We could spend our whole lives listening to arguments and still not have heard all the persons who wish to persuade us. Hence we are obliged to choose what we will hear and when we will hear it; and we are not necessarily unreasonable because we have declined to listen to Mr A's argument or read Mr. B's book. The ideal of tentativeness does, however, involve never deciding that 'under no circumstances will I ever consider any further argument against this proposition, or reconsider an old argument, with a view to possibly changing my opinion'.

The tentativeness of reason sets many people against reason, because they cannot bear to be uncertain. 'Man cannot live by merely tentative beliefs', they say; 'he demands security of mind, an assured faith.' (This statement of the objection is taken from Campbell, Philosophy, 1950, p. 130.) This is a common and natural feeling; but it is childish and ought to be overcome, like the fear of being alone in a house at night. It amounts to taking the absurd position that 'I am going to have certainty at all costs, even if I have to stifle reason to get it'. The desire for security, like other desires, is to be gratified when it can be gratified without grave loss to other interests; but it is not to be gratified at all costs. We know the havoc that a nation causes when its unbridled desire for security drives it to be always acquiring a little more buffer territory at the expense of a neighbour. An individual's private desire for intellectual security can cause as great a proportion of havoc in the intellectual life of himself and his associates. We all can and should learn temperance in the indulgence of our desire for security as in the indulgence of all our desires.

And what is this metaphor, taken from the Bible, that 'man cannot live'? Of course it does not mean that those who are undogmatic in their beliefs die of being so. I suppose it means that they become less happy for being so. But they become more happy for knowing that they are doing the reasonable thing.

Those who think that human reason suffices, and those who think that only a god and faith in him suffice, have in common that they all think that something suffices. And in this they are all mistaken, for nothing suffices. We are always, in any case, going to have mistakes and sufferings, and finally we are going to cease existing. The question we must ask is not what will give us all the help we want, for nothing can do that; but what will give us the most help possible. And the answer is reason. Reason is more likely to avoid mistake and suffering than is faith or any other way.

Yeats once complained that 'the best lack all conviction'. He may have been right in the way he meant it; but -- a certain lack of conviction is the very thing that constitutes the best people.

2.511. The submission of reason

The essential part of bearing in mind the possibility that one is mistaken is not to begin each statement with the words 'I may be wrong but it seems to me that'. That would be very tiresome and quite useless. It is to allow free speech to others and weigh their ideas. Reasonableness includes listening to the other side, and giving the other side full liberty to argue. It includes submission to criticism.

The reasonable man behaves as a fallible being among fallible beings. The unreasonable man, on the contrary, sometimes talks as if you were fallible but he were a god. 'Don't trust man', he sometimes says, 'trust God.' This remark would be selfdefeating if he regarded himself as a man, for in that case he would be telling you not to trust himself and therefore not to trust this advice of his. So the implication is that when he speaks it is the voice of a god. Newman clearly makes this claim in the following sentence: 'Theological reasoning professes to be sustained by a more than human power, and to be guaranteed by a more than human authority' (Grammar of Assent, p. 377). Similarly, Rousseau at the beginning of his Discourse on Inequality, para. 7, quite clearly implies that whereas other men's books are merely human and contain lies, this book of Rousseau's is the voice of nature herself and therefore must be true, though he says he may have added something of his own without intending to. The idea that 'when you talk it is merely your subjective opinion, but when I talk it is the objective truth', is thoroughly bad-mannered and unreasonable. We are all in the same boat, the predicament of subjectivity. Whenever any of us talks it is his subjective opinion; but it may also be true.

Submission to criticism excludes keeping one's opinion secret. It involves making one's views known rather than concealing them. Reason is essentially public. The reasonable man submits to criticism, and he submits to the new evidence and new probabilities that criticism sometimes brings.

A very different kind of submission is sometimes demanded of reason, namely submission to authority. It is pointed out that no man can investigate all things for himself, and that we do in fact rely upon all sorts of authorities, and should be unreasonable not to do so. It is often suggested that a man has no right to have an opinion of his own on a subject about which he is ill informed. Even being in doubt about a proposition is disapproved by some people. Jeremy Bentham, and some others who shared his doubts about the thirty-nine articles of the Church of England, were induced, nevertheless, to sign them by a person who reproved their hesitation as 'presumption' (Robert M. Murray, English Social and Political Thinkers in the Nineteenth Century, i. 43).

It is true that no man can investigate for himself all the matters on which he needs to have an opinion, and that it would be foolish never to trust the authority of, say, a physician or an accountant. Yet the demand for submission to authority, as commonly made, is not a right conclusion from this, but on the contrary highly improper. For it is usually a demand that we should submit to a certain authority without having criticized his credentials, without having judged for ourselves whether he is justified in claiming to be an authority. It overlooks or suppresses the vitally important fact that each one of us has to decide who are the authorities.

The Medical Association is not the only set of persons claiming authority in medicine. There are the others, whom the Medical Association calls quacks. The patient has to decide between them; and sometimes he leaves one in disgust and goes to another. The Roman Church is not the only body claiming to be an authority about the gods. There are also the atheists, and the Moslems, whom the Roman Church calls infidels. There is no escape for any of us from choosing between these rival authorities. In adopting any one authority we are judging his rivals, in effect if not explicitly. Therefore we are all of us always inevitably judging the authorities, that is, criticizing them.

The only choices we have in this matter are whether to do this judging consciously or unconsciously, and whether to do it reasonably or unreasonably. Evidently it is better to do it consciously and reasonably. The person who submits to a priest because she found herself doing so when she woke up, and has never considered not doing so, is in a less reasonable and less safe position than the person who submits because she has criticized this authority and decided that it is better to submit. And it is better to review such decisions from time to time than to make them once for life. We should submit to no man in the sense of abandoning our own judgement for ever. Complete submission to an authority, far from being commendable, is a grave irresponsibility. We are responsible for all our opinions, however ignorant we may be in the field, because we are responsible for our choice of any authorities on whom we rely. All submission to an authority should be based on, and revocable by, our own judgement whether he is an authority; and this judgement should be revised from time to time in the light of the best considerations then available.

As to the doctrine that a man may have no right to an opinion on some matter, I suspect it of being held owing to ignorance of the fact that to adopt someone else's opinion is to have an opinion oneself. It is clearly wrong if it includes suspense of judgement in having an opinion. If a man is aware of a proposition he cannot help doing one of three, either adopting it or rejecting it or suspending judgement about it. If, however, we exclude the case of suspending judgement, as not properly falling under the expression 'having an opinion about', and say that having an opinion about a proposition is either accepting it or rejecting it, then it is true that sometimes a man may have no right to an opinion. But when? Precisely when he ought to suspend judgement about the opinion. And he ought to suspend judgement about it when he cannot find any good evidence for or against it; and this includes not being able to find an authority who has an opinion on the matter. If he can find an authority whom it is reasonable to follow on the matter, he ought to have an opinion because he ought to adopt the authority's opinion.

2.512. Practical reason

If a person confined his ideal of reason to what I have so far said, he would not call any action reasonable or unreasonable, except acts of thinking and of arriving at opinions and getting evidence for opinions. But we call actions of many sorts reasonable or unreasonable. Thereby we imply that there is such a thing as practical reason. We are right to do this. For we cannot help evaluating actions, and human actions are essentially thoughtful. Hence in evaluating actions we are evaluating something that involves thought. Hence our ideals for thought come into our ideals for action. Hence in evaluating actions we sometimes use the word that refers to our ideal for thought, and call them reasonable or unreasonable. Thinking includes choosing actions and principles of action, and this can be done well or badly.

Some great men have doubted whether there is such a thing as practical reason, or whether reason has anything to say about practice. That was because they confined reason to deductiveness and consistency. If we understand by 'reason' ideal thinking, as it is very common to do, and if we believe that thinking enters essentially into human action and largely helps to make it good or bad, then it is inconsistent for us to deny the possibility of practical reason. Practical reason is possible and to some extent actual. It is the same as wisdom.

Reason in action includes, first, all that reason in reflection includes. Reasonable action is based on opinion about the world reached by reasonable thinking.

As the reasonable thinker collects impartially all the considerations bearing on a question of fact, so the reasonable agent takes impartial account of all the interests affected and examines all the advantages and disadvantages of each course. There is, however, this difference of degree, that the urgency of time enters into action more than into theory, and makes it sometimes reasonable to act earlier without a full consideration instead of acting later with a full consideration.

Further, we all call it unreasonable to keep on doing two sorts of action such that the one sort defeats the aim of the other sort. Selfdefeatingness in practice is to be added to selfinconsistency in theory as part of unreasonableness. It is unreasonable, for example, to desire people's love and at the same time keep on hurting them. Such selfdefeatingness is often called inconsistency. But it is not a belief in two inconsistent propositions. The statement that 'I want people to love me' is consistent with the statement that 'I want to hurt people'. Selfdefeatingness is due to laws of nature, not laws of logic. It consists in pursuing two aims which nature has made incompatible; and only by knowing some laws of nature can we know what aims are incompatible.

We also call it reasonable to adapt one's means properly to one's ends, and unreasonable to try to achieve an end by a means which is probably ineffective, or which is effective but far too costly in relation to what the end is worth. To insist on achieving some end at all costs would be eminently unreasonable if we really meant 'at all costs'; but usually the context shows that we are referring only to a small part of all the possible costs.

We also call it reasonable to respect the aims and interests of all persons equally. Reason in this sense is referred to in some uses of the words 'impartiality', 'justice', and 'equality'.

So far I think I have mentioned only matters that most people include under reason in action. I myself include in it a further item which, I fear, is not so often included that I can claim that it is generally agreed. I mean the principle that the lessening of misery is the most important aim, and other aims should give way to it. If a man says he would rather all Englishmen were made miserable than that England should play second fiddle to the United States, I call him unreasonable precisely because he holds something else more important than the fight against misery. If a man says that a certain moral law is to be obeyed no matter how much misery it causes, because it is the command of a god or just because it is the moral law, I call him unreasonable for the same reason: he puts something else above the lessening of misery. In other words, pity is part of practical reason.

What is the relation of practical reason to the moral laws? I doubt whether Henry Sidgwick was right in believing that there is a 'common conviction that the fundamental precepts of morality are essentially reasonable' (Methods of Ethics, 5th ed., p. 383). I think that too many people dissociate morality from reason for this to be true. But the precepts of morality ought to be reasonable, and we ought to reject all moral laws that do not tend to diminish misery on the whole. The reasonable man desires to obey general principles that are impartial between men and men and could be obeyed by all to the common advantage. He submits himself to such rules of action as both are actually acknowledged and also impartially diminish misery.

2.513. Depreciations of reason

It is very common to depreciate reason or compare it unfavourably with something else. People say that reason tends to atrophy feeling, that it tends to make action feeble, that it is incompetent in certain spheres, that it cannot help to make us happy, that we should trust God not human reason, that to trust reason is pride, that reason cannot prescribe ends, that reason must be subordinated to faith or intuition, and that we should think not with reason but with the blood.

Yet it is absurd to talk against reason if the word means what I have suggested. Reason, I have suggested, is either the power to think or the good use of that power. As to reason as the power to think, nobody seriously holds the view that we had better not think, at least when it is thus expressed. And as to reason as the good use of the power to think, it is evident that to deny the goodness of that would be to say that something good was not good, a selfcontradiction.

How then do people come to talk against reason? They do so through fear of thought combined with the misconception that reason is a special faculty. Fear of thought is, of course, a common and natural occurrence, because thinking sometimes undermines a cherished belief or reveals an alarming situation. It is usually held in check by the fact that thinking is inevitable and sometimes supports a cherished belief or reveals a delightful situation.

The other element which leads people to talk against reason is the misconception that reason is a special faculty, not the general faculty of thinking but a special department of it. It is easy to regard reason as a faculty because in one correct sense the word does mean a faculty, namely the general faculty of thinking. It is also easy to regard reason as not being the general faculty of thinking but some specific or different faculty, owing to the other sense of the word, in which it means something different from the faculty of thinking. The sense in which reason is the power to think makes us regard reason as a power. But the sense in which reason is the ideal use of something makes us regard it as not the power of thinking. So we slide into regarding reason as some mental power other than the power of thinking. Reason now appears as one mental power among others, the others including probably intuition, faith, belief, memory, imagination, emotion, sight, hearing, and touch. The human mind, or the power of thinking, comes to seem like a toolbox; and these various faculties; reason, intuition, faith, and the rest -- are the tools in the box. Kant called reason a Werkzeug (Grundlegung, Pp. 395-6). Now the good use of a toolbox involves the good choice of which tool to use for each purpose. There are things that you can do well with a chisel but not with a hammer, and conversely. Hence we come to think that there are things you can do with faith but not with reason, and so on.

Once we have come to regard reason as a special faculty, and as only one among many tools available to the mind, we can safely depreciate it, which we should not dare to do if we regarded it as the power of thinking in general. This gives opportunity to that fear of thought which is latent in most of us. Whenever thought leads to results that distress us, we can now say it was because we used reason when we should have used some other faculty, like prentice carpenters trying to make a screwdriver do the work of a chisel. Now we can talk about 'the bounds of reason', and about 'areas where reason is incompetent', thus taking the liberty to reject the results of our thinking when they distress us. Now we can retort, against those who appeal to reason, that they are concentrating on one particular faculty and neglecting all the other ways of knowing. Thus we shut people off from using their reason on our favourite doctrines.

All this is a mistaken thing to do. We should not let the distressing results of some thinking seduce us into inventing the myth of reason as a special faculty alongside intuition and faith and the rest. There are no such special faculties. Reason, in the sense of a mental power other than the general power to think, is a fiction, a dummy set up to be knocked down by those who favour not thinking about certain matters. The human mind is not a box of tools from which you can select. It has only the one tool, thought. And our only choice is whether to use it badly or well, and whether to inquire and learn how to use it well. The English word 'reason' is sometimes the name of this tool, thought, and sometimes the name of the ideal use of this tool, which we dimly perceive and try to perceive more clearly. To be against reason is therefore either to be against thinking as well as possible, or to be against thinking at all.

The bounds of reason are as wide as the bounds of statement and belief. Anything whatever that can be stated or believed should be stated or believed as a result of thought, which is one sense of the word 'reason'; and the thought should be as good as possible, which is the other sense. There is no 'area where reason is incompetent', in the sense of a set of propositions which are to be adopted or rejected without thought. To say that reason is incompetent about a given proposition is to say that it is not good to search for the considerations for and against this proposition, or to weigh them against each other, or to adopt a view accordingly, or to revise this view from time to time, or to listen to criticism of it. On all choices between adopting a proposition and adopting its contradictory either reason is competent or nothing is. The Pope in September 1952 said to astronomers that, when the human intellect has done all it can, faith must carry on. The implied contrast is false, for the human intellect will not have done all it can until the human race is extinct.

This confusion leads also to the idea that those who praise reason are neglecting nonrational methods of knowing, or denying their value. If reason is thought of as one peculiar mental faculty, and we recognize other rival mental faculties alongside it, it seems that those who praise reason are neglecting these other instruments. But when we see that reason is not a special faculty, but the good or ideal use of the general faculty of forming statements and beliefs, this opposition falls to the ground. Every special way of producing beliefs that there may be, either intuition or sight or hearing or telepathy or what you will, is to be examined and given whatever weight seems reasonable after examination; but none of them is to be set up as autonomous and immune from criticism. None of them is even to be assigned a special sphere in which it is uncriticizable. For example, sight is not supreme in the sphere of the visible. We check and sometimes reject its deliverances by the rational use of other faculties, especially touch and memory. Reason is supreme because it is not a special faculty, but the best use of the whole faculty of forming beliefs for the sake of forming them truly. The only alternatives to thinking with reason are thinking unreasonably and not thinking.

The misconception of reason as a special faculty is also responsible for the idea that 'it is beyond the power of reason to prescribe ends' (Harrod in Mind for 1936). Reason is the good employment of thought, and thought both can and should adopt ends.

The idea that to trust reason is pride is, we now see, the idea that it is pride to try to use one of our powers as well as possible, namely the power of thinking. They might as well say that it is pride to try to run as fast as possible, or to preach a sermon as well as possible. Or perhaps they mean that we should use our power of thinking as well as possible to arrive at a conclusion, but then out of humility reject our result and believe instead what is told us by some authority. On the contrary, the doctrines of authorities should be considered before we reach our conclusion and weighed along with the other considerations before us. They are not opposed to reason but part of the evidence which reason weighs.

Any virtue may become an occasion of pride, for the peculiarity of the vice of pride is that it finds its opportunity precisely in the presence of a virtue. But the virtue called reason is not a necessary nor even a specially likely occasion of pride. On the contrary, the submission of our thought to criticism and argument and evidence is a great and good humility; for an important part of 'the worth of men consists in their liability to persuasion', as Whitehead has written (Adventures of Ideas, p. 105). This sort of humility is often conspicuously absent from those who depreciate reason.

It would be a very sad thing if the good employment of our power of thinking tended to atrophy feeling, but fortunately it does not. Brutality and schizophrenia tend to atrophy feeling; but reasonableness does not. It is true that the unreasonable man is often emotional in a way that the reasonable man is not; he tends to anger when contradicted. It is true also that reason tends to atrophy painful and harmful feelings; for good thinking includes reflecting upon the emotions, and deciding which are to be preferred, and training oneself accordingly. But the result of such reflection is not that we are to discourage all emotion. It is that we are to discourage anger and hate and gloom and envy and jealousy, but encourage love and pity and respect and joy and what may well be called the emotion of reasonableness, the sentiment of desiring to listen to both sides, and of enjoying taking one's decisions in the light of all available considerations. We tend wrongly to call it the work of reason only when we decide to discourage the emotion or repress the desire. It can also be the work of reason when we decide to encourage or gratify. Perhaps Plato is partly responsible for this misconception; his works have a tendency to suggest that the business of reason with the emotions is to suppress them all.

As to thinking 'with the blood', all thinking has to be done with the blood in any case. No one can think unless there is plenty of good blood in his brain. If this crass literalism exasperates some disciple of D. H. Lawrence into exclaiming that 'you know what I mean', the answer is that we know that he means that he demands the right to be dogmatic and listen to no argument, and he demands that we follow him blindly.

If you dislike my account of reason, do not because of that reject reason, but give a better account of it. I have recommended the pursuit of reasons and consistency, deductiveness and inductiveness, respect for evidence, tentativeness, and submission to criticism, because I think they are our most likely means of coming to hold true rather than false beliefs. And I have recommended selfcompatibility and impartiality and the rejection of misery because I think they are our most likely means of coming to good decisions. If you disagree, you should make your own account of what constitutes good thinking. You should not reject reason as such, because the word 'reason' is our name for the ideal of thinking, and our dignity as men demands that we conduct our power of thinking in the best possible way. Misology, the hatred of reason, is a great evil.

Graphic Rule