The Plan of these Lectures

Belief is a large and complicated subject. It has so many ramifications that one hardly knows where to begin the discussion, and when one has begun one hardly knows where to stop. But a series of lectures should have some sort of plan, as a play (even the most modern one) should have some sort of plot. After all, the lecturer has to communicate his thoughts to you in a one-dimensional order; though if it were only possible, a multi-dimensional mode of presentation, in which an argument or an exposition proceeded in several different directions at the same time, might suit the subject-matter better. It may be that somewhere in the universe there are intelligent beings who can communicate their thoughts in this multi-dimensional way. This might perhaps be one of the advantages of ‘the tongue of angels’, mentioned by St Paul in a celebrated passage. But it certainly cannot be done in any human tongue. I have to present my thoughts to you one after the other, and they will be less difficult to comprehend if I begin by telling you the plan upon which my one-dimensional series of remarks is to proceed.

The plan I have adopted is this (I admit it is a somewhat arbitrary one). There is to be a central topic, preceded by certain preliminary enquiries, and followed by certain corollaries. The central topic is one which suggests itself very naturally to anyone who reflects on the history of epistemology. He can hardly fail to notice that there are two very different ways of analysing belief. In the traditional way of treating the subject, it is assumed that believing is a special sort of mental occurrence (sometimes described as a ‘mental act’). This mental occurrence need not necessarily be introspected by the person in whom it occurs; but it always could be introspected by him if he took the trouble to attend to it. On this view, the main task of a philosophy of belief is to examine this introspectible mental occurrence, to analyse it in so far as it admits of analysis, and to distinguish it clearly from other mental occurrences with which it might possibly be confused, for example from the one which occurs when we merely consider a proposition neutrally, without either accepting it or rejecting it. I shall usually refer to this way of treating belief as ‘the Occurrence Analysis’ and sometimes as ‘the Traditional Analysis’, though it must be noticed that the tradition prevailed pretty generally until about a generation ago.

The modern way of treating belief is quite different. Believing something is now generally regarded not as an occurrence, introspectible or otherwise, but as a disposition. When we say of someone ‘he believes the proposition p’ it is held that we are making a dispositional statement about him, and that this is equivalent to a series of conditional statements describing what he would be likely to say or do or feel if such and such circumstances were to arise. For example, he would assert the proposition (aloud, or privately to himself) if he heard someone else denying it or expressing doubt of it. He would use it, when relevant, as a premiss in his inferences. If circumstances were to arise in which it made a practical difference whether p was true or false, he would act as if it were true. If p were falsified he would feel surprised, and would feel no surprise if it were verified.

On this view, then, believing a proposition is somewhat like being interested in cricket or having a distaste for gardening. Like these, a belief shows itself or manifests itself in various sorts of occurrences, some mental and some psycho-physical, but it is not itself an occurrence. Acquiring a belief, and losing it, are indeed occurrences, though we are not always able to assign precise dates to them. But the belief itself is not something which happens at a particular moment, but something which we have or possess throughout a period, long or short. And though it is liable to manifest itself by various sorts of occurrences, when and if suitable circumstances arise, none of these occurrences are themselves believings. The occurrent believings or ‘acts of believing’ which the traditional theorists discussed are on this view mythical entities, much like the ‘acts of knowing’ which the same traditional theorists also discussed. On this view, therefore, it makes no sense to say of someone at 4.35 p.m. ‘he is now believing that it is going to rain’, as it does make sense to say of him that he is now looking anxiously at the clouds or wondering whether to ask a passing motorist for a lift.

No doubt most exponents of the traditional analysis would be willing to admit that the sentence ‘Jones believes that p’ can be used in a dispositional way. If this were denied, they would be committed to the paradoxical conclusion that when Mr Jones turns his attention to something else, or falls asleep, and ceases for the time being to think of the proposition p, he ipso facto ceases to believe it. There is obviously a sense in which we can continue to believe a proposition for a long period, for many years perhaps, although we seldom think of that proposition at all. I have believed for a very long time that the earth is approximately spherical in shape, but the occasions on which I actually find myself giving my assent to that proposition are very few and far between.

But those who accept the traditional Occurrence Analysis would insist that this dispositional sense of the word ‘believe’ is secondary and derivative. We can say of someone, quite correctly, ‘he believes that Oxford will win the boat race this year’ and that he continues to believe it throughout the months of January and February. But on the traditional view, we mean by this that if at any time during that period he were to consider the proposition ‘Oxford is going to win the boat race’, an actual belief-occurrence would take place in him—a specific sort of experience which he could notice introspectively if he wished—and this proposition about the boat-race would be its object. Or, in Hume's version of the traditional analysis, if the idea ‘Oxford is going to win’ were to come into his mind at any time during that period, this idea would actually feel strong or forceful or vivacious to him. This feeling, in turn, might of course affect his actions. If he is an Oxford man he might make arrangements to go to London on Boat-race Day; Hume himself lays some stress on the effects of belief, and so does Cook Wilson, whose version of the traditional Occurrence Analysis is rather different. Nevertheless, it would be insisted that here again the occurrent believing—the actual belief-experience—is primary in ordine analysandi and the actions which the man does in circumstances to which the believed proposition is relevant are merely consequential.

With these explanations, I hope it will not be too misleading if the two contrasted ways of treating belief are called ‘the Occurrence Analysis’ on the one hand and ‘the Dispositional Analysis’ on the other. Sometimes too I shall call the first ‘the Traditional Analysis’ and the second ‘the Modern Analysis’, though this way of speaking is only roughly accurate. I have already pointed out that the Occurrence Analysis was still generally accepted until quite recent times; and it must now be added that the Dispositional Analysis (in a rather over-simplified form) was explicitly formulated by Alexander Bain nearly a century ago, and no doubt hints of it could be found in earlier writers. Indeed, there are some slight hints of it even in Hume himself, as we shall see later.1

The exposition and discussion of these two analyses of belief, the occurrence analysis and the dispositional analysis, will be the central topic of these lectures. But before we reach the central topic, we shall have to consider certain questions which arise whichever of the two analyses we accept. These questions concern the relation between belief and knowledge (Lecture 2 ‘The varieties of knowledge’, Lecture 3 ‘Belief and knowledge’); the relation between belief and evidence (Lecture 4 on the evidence of perception, memory and self-consciousness, Lecture 5 on the evidence of testimony); Locke's doctrine of degrees of assent and Newman's criticism of it (Lecture 6).

I shall then turn to the exposition of the traditional Occurrence Analysis (Lecture 7, ‘Hume's analysis of belief’; Lecture 8, ‘The entertaining of propositions’; Lecture 9 ‘Belief and “being under an impression that…”’). In the concluding lecture of Series I (Lecture 10) I shall consider an interesting difference of opinion within the traditional ‘occurrence’ school. We shall find that two of its most illustrious members, Descartes and Hume, take diametrically opposite views concerning the freedom of assent. This difference of opinion, like the one between Newman and Locke, has an important bearing on ‘the ethics of belief’. If belief is something wholly involuntary (and Hume maintains, at least in one well known passage, that it is2) there could not be such a thing as an ‘ethics’ of belief at all.

The other half of my central topic, the exposition and discussion of the Dispositional Analysis, the comparison of the two analyses, and the attempt to decide which (if either) is correct, will be reserved for Series II, and the first three lectures of that series will be devoted to it. Then, but not till then, we shall be in a position to consider what I called the corollaries. There are a number of important questions about belief which cannot be profitably discussed until we are in possession of the conceptual tools which our examination of the central topic will provide. The first and most obvious of these concerns half-belief (or perhaps ‘near-belief’ would be a better name for it), a subject which can hardly be discussed at all until we have learned and digested what the dispositional analysis has to teach us. We must also consider the claims that some beliefs have a self-verifying character: Vergil's remark possunt quia posse videntur—‘they can do it because they think they can’—will serve to illustrate what is meant. If there are self-verifying beliefs, this may be expected to throw some light on the nature of faith. We should all wish to know what kind of an attitude faith is: not religious faith only, though certainly this is the most important sort of faith, but other sorts of faith too, for example the faith which one human being may have in another. This topic may also be approached from another side, by asking what we mean by belief in someone or something and how believing ‘in’ is related to believing ‘that’. And here a lecturer under Lord Gifford's foundation will naturally be expected to pay some attention to the belief ‘in’ a world-outlook or world-view, because this type of belief ‘in’ is an essential part of the religious attitude. In Series II Lecture 9 I shall have a good deal to say about belief ‘in’ and about the relation between belief ‘in’ and belief ‘that’.3 It is a subject on which strange and unplausible views seem to be widely accepted.

It may seem to some of you that these ‘corollaries’ are far more interesting than the central topic itself, and more interesting too than most of the preliminary ones. In comparison with these great and fundamental questions, our enquiry into the occurrence analysis and the dispositional analysis may seem merely academic or even pedantic. What difference does it really make which of these two analyses of belief—if either—is the correct one? Who cares whether ‘Jones believes that p’ is or is not a purely dispositional statement about Mr Jones?

If anyone takes this view, of course I have some sympathy with him. My reply, however, is that in philosophy the longest way round is often the shortest way home. Let him regard the discussion of my central topic not as an end but as a means. If belief in a religious world-view is what interests us most, we shall be in a better position for considering this subject if we first pay some attention to the nature of belief in general; and one way of doing so is to ask what kind of a statement ‘Jones believes that p’ is, and what is the most adequate way of analysing it.

Nor must it be thought that belief is the exclusive preserve of theologians and philosophers of religion. This is not even true of belief ‘in’ (for example, one may believe in classical education or in taking daily cold baths.) Still less is it true of belief ‘that’. Belief is a state in which we are throughout our waking lives, and often too when we are dreaming. And this applies to the belief which does not amount to knowledge as well as to the belief which does. Belief on evidence which is less than conclusive is a very familiar condition. We are in it all the time about many subjects. We could not live without it. Nor would it be reasonable to try to. We must not suppose that there are only two alternatives, knowledge on the one side and helpless agnosticism on the other. We may still have good reasons for believing when we do not have conclusive ones. A reasonable being, situated as we are, cannot dispense with such beliefs. Very often conclusive evidence is not available to us, or not available as yet, at the time when practical decisions have to be taken, or theoretical inferences drawn, or difficult emotional adjustments made. Then we must make the best use we can of such evidence as we have. Sometimes, too, though less frequently, perhaps, than is supposed, we believe without any evidence at all, or on evidence so flimsy that it hardly deserves the name.

Belief, in fact, is one of the most commonplace and familiar things in the world. It is not something reserved for solemn occasions. And it is just these commonplace and familiar things which raise the most interesting philosophical problems. (Perception is another. What could be more commonplace than seeing or hearing?) As Bertrand Russell has said ‘Philosophy, if it cannot answer so many questions as we could wish, has at least the power of asking questions which increase the interest of the world and show the strangeness and wonder lying just below the surface even in the commonest things of daily life’.4

So much for the plan or plot of these lectures. I turn now to some general points which it may be well to mention briefly in this introductory discussion.

The Study of Belief Cuts across Traditional Boundaries

There is a traditional boundary between epistemology on the one hand, and moral or practical philosophy on the other. In these lectures we shall sometimes find ourselves crossing that boundary or standing on both sides of it at once. Obviously it will be necessary to say something about moral beliefs. Some difficult problems arise about them, if it be held (as it often is) that the function of moral sentences is to express the speaker's attitude of approval or disapproval.

It is also obvious that there is a connection between belief and will, or, to put it more generally, between belief and the practical side of our nature. We all know too well that there is such a thing as ‘wishful thinking’, the tendency to believe a proposition because we wish it to be true or fear the consequences which would follow if it were false. On the other hand, our beliefs in their turn affect our actions. They manifest themselves in what we do and in the practical decisions which we take. It must be remembered too that not all ‘doing’ is public and overt. Directing our thoughts to one subject rather than another is a kind of doing. It alters something, it makes a difference, even if only to what happens subsequently in ourselves. The difference which it makes may be an important one too. What we ‘think in our hearts’ may sometimes matter more than what we do publicly in the marketplace. It may have a powerful effect on our emotions and our wishes. A man's beliefs show themselves in the direction which his thoughts take, and sometimes in his voluntary decision to think about this subject rather than that. The most obvious examples of this are moral and religious beliefs. If a man believes that charity is the greatest of virtues, he will tend to think about the good points in his neighbour's character rather than the bad ones.

Moreover, something rather like the traditional free will problem arises about belief itself, as well as about the actions, overt or not, which we do ‘in the light of’ or ‘under the guidance of’ the beliefs which we hold. This is the issue we shall have to discuss when we consider the controversy between Descartes and Hume concerning the freedom of assent (Series I, Lecture 10). In some sense, and within limits, we are free to act as we choose and responsible for our choices. How far, if at all, are we free to assent as we choose? When a proposition comes into our minds, or is propounded to us by others in speech or in writing, are we free to give our assent or to withold it, as we choose? Anyone who speaks about the ‘Ethics of Belief’5 must surely think that in some sense we are responsible for holding the beliefs which we do hold, or at least for holding some of them. At first sight it may seem strange to speak of the ‘ethics’ of belief at all, and I am not sure that the word ‘ethics’ is the most appropriate. Nevertheless, we all of us do use such quasi-ethical expressions as ‘justifiable’, ‘unjustifiable’, ‘have a right to’, ‘have no right to’, when we are discussing the beliefs of others, and sometimes when we are discussing our own. What William believes may be true, but he has no justification for believing it on the evidence that he has, though others perhaps have, because they do have good evidence for it. Again, he has no justification for being absolutely convinced that there will be a frost to-night, on the evidence that he has, though he would be justified in holding a mild opinion that there will be. For we commonly think that there are degrees of belief; and when we raise these questions about a man's right to believe something or ask whether he is justified in believing it, we ask not only whether he is justified in believing such and such a proposition at all, on the evidence that he has, but also whether he has a right to believe it to the degree that he does—with the degree of assurance or confidence which he actually shows. We often find that others (and ourselves too) hold firm convictions on subjects about which we are only ‘entitled’ to have mild opinions. On the other hand, the term ‘reasonable’, the appropriate term of commendation for beliefs, is applied to actions as well, and here too it is a term of commendation.

Moreover, when we consider the relation between belief and action we come upon another topic which is traditionally supposed to fall within the moral philospher's domain. What are we to say of a man who fails to act upon some belief which he holds, in circumstances to which the proposition believed is obviously relevant? He proposes to catch the 4.45 train from Paddington. He believes, to all appearances sincerely, that the train will leave punctually, and his wish to catch it is also sincere. He has plenty of time to get to the station, there is no lack of transport to take him there, he has plenty of money to pay for a taxicab. Yet he arrives at the station five minutes late and misses the train. This is not unlike the problem of ακρασια (‘incontinence’ ‘weakness of will’) discussed by Aristotle in Book VII of the Nicomachean Ethics.6 In the example just given, the belief upon which the man fails to act is a belief about a particular event, the departure of a particular train on a particular day. But the same question arises about general beliefs, and this brings us even closer to the problem discussed by Aristotle. For example, no one doubts that all men are mortal. But there are many who die without having made a will.

In these two examples we are, of course, confronted with a special type of ‘unreasonable’ conduct; and here as elsewhere the term ‘unreasonable’ conveys some degree of disapproval or censure. The disapproval does not directly concern the student of belief, since the beliefs themselves, in these two examples, were perfectly reasonable ones; what was unreasonable was only the failure to act in accordance with them. But the student of belief does have to ask, as the moral philosopher also does, how that particular type of unreasonableness is possible. And this is an awkward question for anyone who holds, as some philosophers do, that believing a proposition p is just a disposition to act as if p were true in circumstances to which the proposition is obviously relevant. I shall discuss it in the first lecture of Series II.

There is another traditional boundary-line which we may have to cross, though with fear and trembling, in Series II of these lectures. I have already mentioned the curious class of beliefs which I call self-verifying ones,7 where believing a proposition appears to have a tendency at least to make the proposition true. Are there really cases where ‘thinking makes it so’? And if there are, shall we have to revise our ordinary notions about objectivity, and shall we also have to revise our ordinary assumptions about the unreasonableness or even blameworthiness of ‘wishful thinking’? These questions are interesting for their own sakes. But when we consider them, we find ourselves confronted with the traditional problem of the relation of mind and body. Some of the beliefs which do at least appear to have a self-verifying tendency are beliefs about the believer's own body. For example, if the patient believes that he is going to recover, this belief of his makes it more likely that he will recover. It may even be that he will not recover unless he believes that he will. Indeed, when Vergil said ‘they can do it because they think they can’, the ‘doing’ to which he referred was rowing fast enough to win a boat-race, and this is certainly a bodily activity.

Belief is supposed to be an epistemological subject, whereas the relation of mind and body is a metaphysical one; or at least it was formerly supposed to be, though on some modern views it is not a philosophical subject at all, but a scientific one, and the only philosophically important point about it is the muddle, or series of inter-connected muddles, which have led people to think that it could be settled by purely philosophical methods. But whichever view we take about the status of this ancient problem, it is certainly not an epistemological one. Yet if I am right, the philosophy of belief cannot altogether avoid considering it.

Finally, there is one other traditional border-line which we shall be obliged to cross sometimes, the borderline between epistemology and the philosophy or religion. But neither epistemologists nor philosophers of religion have ever taken that particular line of demarcation very seriously; and even if they had, it would still be the duty of a Gifford Lecturer to disregard it. You are more likely to complain of me for crossing it too rarely than too often.

The ‘Performatory’ Aspect of First-Person Belief-Sentences

I now turn to a point which may appear to be a purely linguistic one, the difference between first person and third person uses of the verb ‘to believe’. There are some philosophers who think that questions concerning language are the only ones which a philosopher should discuss. I do not subscribe to this view. But it seems to me that the difference between first-person and third-person belief-sentences is of considerable philosophical importance. We might easily be led into philosophical mistakes if we neglected it, for instance if all the examples we used in our investigation of belief were of the form ‘I believe that p’. First person belief sentences, or more accurately, first person belief sentences in the present tense, have certain peculiarities which other belief-sentences do not share. Nor is this merely a fact about the English language. It is true of some other languages as well, though I would not venture to assert that it is true of all.

The first and most obvious point to notice is this. When someone says in the first person and present tense ‘I believe that p’ (still more if he says ‘I believe in X’) he is not usually giving us a piece of autobiographical information. It is true that just occasionally he might be. He might be announcing to the world a discovery which he has recently made about himself, a surprising discovery which interests him and may be expected to interest others ‘Good heavens! I never used to believe it, but now I find that I do’ or even ‘I thought I did not believe it, but now I find that I do’.

But ordinarily when someone says ‘I believe that p’ he is not giving us a piece of information about himself. He is expressing an attitude, rather than telling us that he has it. And sometimes, in the act of expressing it, he is doing something more as well. Sometimes he is taking a stand in the face of a hostile or sceptical audience. ‘This is what I believe. Call me a fool or an idiot if you like.’ This presumably is the purpose of the books or essays which are sometimes written with the title What I believe. Here the man who says ‘I believe that p’ or ‘I believe in X’ or writes a book called ‘What I believe’ is making a kind of public declaration, or issuing something like a pronunciamento. It is a public act of self-commitment.

I cannot see, however, that the self-commitment is necessarily intended to be irreversible, as theologians who use the phrase often seem to assume. At any rate, a reasonable being should not be expected to commit himself irrevocably to the truth of a proposition, unless he has conclusive evidence that the proposition is indeed true, which he seldom has in cases of this kind. And even imperfectly reasonable beings like outselves do sometimes take a stand in this not-necessarily irrevocable way, without saying or implying or being taken by others to imply that nothing whatsoever could conceivably lead us to change our minds. ‘This is what I believe now, on the evidence that I have, however much the rest of you may disagree with me.’ When someone says this, he need not be taken to be asserting that his evidence is absolutely conclusive, or that he himself thinks it is, or that he will always continue to believe what he believes now. Indeed, if we interpret his utterance in that way, we may be showing a lack of respect for him and failing to treat him as a reasonable being: for it may be quite obvious that the evidence for what he believes cannot be conclusive. The same is true of a book with the title ‘What I believe’. We need not take the title to mean ‘What I irrevocably commit myself to’. Such an interpretation might sometimes be justified, but by no means always.

But more often the utterance ‘I believe that p’ has what J. L. Austin calls a performatory character. Or if this taking of a stand is itself a ‘performance’ in his sense, it is not a very common one. A man who says ‘I believe that p’ is not usually standing up on a soap-box, so to speak. There need be no emphasis on the first personal pronoun. (‘What I believe is this, whether you believe it or not.’) Such declarations are made sometimes, but rather rarely. Much more commonly, we are inviting our hearers to accept what we believe and are assuming that they will. And we are doing more than that. We are conveying to them, giving them to understand, that they will be justified in accepting it.

As Austin has shown, the first-person present tense utterance ‘I know that p’ has a guarantee-giving character. When someone says ‘I know that p’, he conveys to us, or gives us to understand, that we may justifiably rely upon the proposition without reserve. It is not of course that he says we may thus rely upon it. He has no need to, because this guarantee-giving function is a feature of the conventional usage of the first-person present tense expression ‘I know’. It follows that in saying ‘I know that p’ one is claiming to have conclusive evidence for the truth of the proposition p; otherwise one would not be in a position to issue this ‘hard’ or ‘cast iron’ guarantee.8

Something like this applies to the first-person present tense use of the verb ‘to believe’ and of similar expressions such as ‘think that’, ‘hold the opinion that’. The difference, of course a very important one, is that here the guarantee is weaker. Indeed, it might be said that when the utterance ‘I believe’ has the taking-a-stand character discussed just now, it also, as it were, offers a guarantee, though in this case the speaker has little or no expectation that the guarantee will be accepted, and does not much mind whether it is accepted or not.

‘Is Wilkinson in Oxford to-day?’ ‘Yes, I believe he is.’ In giving this answer, one conveys to one's hearer that he may safely rely to a certain degree upon the truth of the proposition ‘Wilkinson is in Oxford to-day’, but not that he may safely rely upon it without any reservations at all. For example, it will be worth while for him to ring up Wilkinson on the telephone, though not perhaps worth while making a special journey to his house in the suburbs, if time is short and other business is pressing. The guarantee would have been weaker still if our informant had said ‘I suspect’ (or ‘I rather think’) that Wilkinson is in Oxford to-day, but I am not at all sure’. Indeed, it might seem too much to say that this utterance gives us anything which deserved to be called a guarantee. All the same, it does gives us something. It is as if he had said ‘I do not venture to promise that you will find him if you go to see him, but there is at any rate a chance that you will’. He conveys to us that we may safely have a small but not entirely negligible degree of confidence with regard to the proposition ‘Wilkinson is in Oxford to-day’; and in so doing he claims that he himself has some evidence in favour of the proposition, though evidence which is nothing like conclusive.

Now of course any of these guarantees, whether hard ones or soft ones, may turn out to be worthless. If our informant said ‘I know that p’, and we find out afterwards that p is false, or even if we find out that it is true but that he had no conclusive evidence for it, we shall be entitled to reproach him for ‘letting us down’. It is as if he had offered us a cheque which the bank refuses to cash. If he only said ‘I believe that p’, and it turns out afterwards that p is false, we shall not be able to reproach him on that ground. This time the guarantee was not offered to us as a cast iron one, and the claim he tacitly made in offering it was not a claim to have conclusive evidence for the proposition. But even so we shall be entitled to reproach him if we find out afterwards that he had no evidence at all for p at the time when he said ‘I believe that p’. Conceivably we might discover this even if the proposition p turned out to be true. Wilkinson actually is in Oxford to-day; but our informant, it now turns out, had no evidence whatever for saying that he was, or evidence too weak to justify any degree of belief at all (e.g. the fact that anyone who is a Research Fellow of an Oxford college is in Oxford sometimes). He answered ‘Yes, I believe he is’ in an irresponsible manner, or just from a wish to please. So he was not entitled to give us any kind of guarantee on this matter, not even the relatively weak one which he did give.

So far we have been concerned with first-person present-tense belief utterances which are made publicly and addressed to others. There can be no doubt that ‘I believe that p’, considered as a piece of social intercourse, does have this performatory aspect. But what happens when one says ‘I believe that p’ privately—to oneself—whether aloud or under one's breath or by means of kinaesthetic or auditory images? If one said it aloud, the remark might of course be overheard by someone else, or recorded by the police by means of a concealed microphone; and similarly if it was written in one's private notebook or diary, someone else might read it afterwards. But this is irrelevant. The remark which was uttered or written down in the notebook was not addressed to anyone else. It was not a piece of social intercourse. Can it still have a performatory character?

It seems to me that even here one might conceivably be ‘taking a stand’; and a man might conceivably write a book or an essay with the title What I believe even though he had no intention whatever of publishing it or showing it to anyone else, or even though he tore it up immediately after he had written it. But if one takes a stand, there must be someone or something against which one takes it. What could this something or someone be? If I say to myself ‘I believe that p’ in the stand-taking way (still more perhaps if I say to myself ‘I believe in X’) I say it in the face of some resistance in myself. It may very well be that I do not want to believe that p; for example I do not want to believe that my friend Mr Postlelthwaite has got into trouble. I should much prefer this proposition to be false. I should be much happier if there were not the strong evidence for it which, I have to admit, there actually is. Belief may very well be reluctant. In this case, the reluctance is the result of emotional factors, my liking for Mr Postlethwaite, or the admiration I have hitherto had for him. Sometimes it is the result of habit. Just because one has believed q for many years, one is reluctant to give up this belief, and believe p instead, where p entails the falsity of q. There are examples of this even in the history of science. Sometimes, again, one's reluctance to believe has a purely intellectual origin. One is reluctant to believe a proposition because there is much evidence against it. At first it may have been very difficult for Europeans to believe that there is such a creature as the duck-billed platypus. How very unlikely it was that there should be a warmblooded furry animal with webbed feet, which lays eggs and has a beak like a duck! But they were obliged to believe it, however reluctantly, when the testimony of good witnesses became very strong. Hesitations and doubts were no longer justified, and if they persisted it was time to take a stand against them. At such point a sceptical zoologist might well say to himself in his study ‘Yes, I do believe there is such a creature after all’.

There is also the rather curious situation in which a man thinks that he believes something, but in a moment of unusual clarity and honesty is obliged to admit that he does not, and even that he ceased to believe it some time ago. He thinks he believes that the coalmines can still earn a profit. Has he not written many letters to the newspapers to say so? But this morning, in his bath, he suddenly realises that he no longer believes this, and has gradually been growing less sure about it ever since he wrote his last letter to The Times three weeks ago. Here he is ‘taking a stand’ when he says to himself ‘I no longer believe it’, because he would still like to think that he does. He is taking a stand against something in himself, against a lingering wish which he still has. If only he could still go on believing that the coalmines can be made to earn a profit, he could maintain and even enhance the public reputation he has already acquired as a coalmines-advocate, and might even be elected as Member of Parliament for Wigan.

No doubt there is something strange in these examples. How can a man ‘tell himself’ that he believes such and such a proposition, or again, that he no longer believes it? To some philosophers the whole idea of ‘telling oneself’ anything may seem absurd, and the idea of ‘taking a stand’ may seem to make no sense when there is no audience before whom or in the face of whom one takes it. But experience is against them. It is useless to object that something is logically impossible when it does actually happen. Instead, we must just admit that many of us are in some degree divided or dissociated personalities, and perhaps all of us are sometimes. Why indeed should we expect the contrary?

But when someone says ‘I believe that p’ to himself, we cannot easily find anything analogous to the guarantee-giving character which this utterance has in its ordinary social use; or if we ever can, it is only when the degree of dissociation has gone far beyond this, and has reached a stage where we have to speak of two or more distinct personalities connected with a single human organism; and then we have entered either the realm of psycho-pathology or the realm of psychical research or both at once. Perhaps in the Sally Beauchamp case the personality called Sally could be said to be giving a guarantee to Miss Beauchamp when she said ‘I believe that p’.

So much for the first-person present-tense utterance ‘I believe that so-and-so’. The performatory aspects of this are important and should not be neglected, as they were until quite recently; and they are sometimes present even when the sentence is said privately to oneself. We must notice, however, that they are confined to the first person use of the verb ‘believe’ (and other allied verbs such as ‘think that———’, ‘have the opinion that———’). ‘You believe that p’ or ‘he believes that p’ has no performatory force. And even within the first-person use, the performatory force vanishes when we shift from the present to the past tense. When a man says ‘I believed it once’ or ‘Yes, I thought so myself until yesterday’ he is neither taking a stand nor giving a guarantee.

What of the future tense? We can say of someone else ‘he will believe it by this time next week’ (or ‘when he is ten years older’) ‘though he does not believe it now’. But can one say it of oneself? Just possibly one can. A distinguished philosopher of the last generation is alleged to have said ‘I shall die a Behaviourist’. Was not this equivalent to saying ‘Before I die, I shall believe certain propositions which I do not yet believe’? An enthusiastic young socialist might say in a cynical moment ‘I shall be a Tory by the time I am sixty’; is not this equivalent to saying ‘By the time I am sixty I shall believe in the political principles of Edmund Burke, etc., which I now firmly reject’?

If that is the correct interpretation of these rather curious remarks, it would seem that they are on just the same footing as first person belief statements in the past tense, and have neither a stand-taking nor a guarantee-giving character. But we notice that both of them do express beliefs which the speakers at present hold—beliefs which they how hold (correctly or not) about their own future beliefs. And in this respect they do have a guarantee-giving character, and probably the philosopher's remark had a stand-taking character as well. We cannot tell just from reading the written words how strong the guarantees offered were. But if we had heard them spoken, we might have been able to notice how confidently they were uttered. If someone says ‘p’ in a tone which expresses no doubt, he can be taken as offering us the same ‘hard’ guarantee which he would have offered if he had said ‘I am completely convinced that p’, though in these particular examples the guarantee might well turn out to be valueless, and the speakers may have been very rash in offering them.

In any case, it should be noticed that first-person belief utterances, in whatever tense, are relatively uncommon; and so far as the present tense is concerned, they are especially uncommon when one is in a state of complete conviction. If one is completely sure that p, one normally says just ‘p’. Nor does one normally say it in a particularly emphatic way, but rather in the most matter of fact way possible. One does not say it emphatically, unless or until someone else has contradicted one's statement or expressed doubts of it, or unless one expects that someone will—and sometimes not even then.9 One just takes it to be a fact that p, for example that to-day is early closing day, and does not bother to express one's attitude about it. It is true that we speak of ‘saying something in a tone of conviction’. But as Professor Ryle has pointed out, there is no special tone of complete conviction. It is rather that a ‘doubting’ tone is absent.

If someone takes a proposition to be absolutely certain (whether on good grounds or not) he is no longer interested in his own mental attitude with regard to it. He is only interested in the proposition itself, for example the proposition that to-day is early closing day, and in the consequences which can be inferred from it either with certainty or with probability; for instance that it would be useless to call at the oculist's shop this afternoon to collect his new pair of spectacles, even though he did get a postcard this morning to say that they were ready. This is why he does not put his mental attitude into words, unless someone else expresses surprise or incredulity.

He is, however, interested in and concerned about his own mental attitude if it falls short of complete conviction; that is, if it is an attitude of opinion, more firm or less, and still more if it is only an attitude of surmising or suspecting. He is then more likely to say ‘I think that p’ (adding, perhaps, ‘but I am not quite sure’), or ‘I suspect that p’ (adding, perhaps, ‘but I am not at all sure’). But even then he need not say so, not even to himself. He does rely upon the proposition p up to a certain point but no further. But he need not necessarily tell others, or even himself, that he relies upon it to that degree. He will only do so if there is some special point in doing it, for example, if others ask him ‘what do you think about this?’ or if he himself thinks they ought to know what his attitude is about it, even though they do not ask.

After all, we have other and more important uses for our beliefs than telling our neighbours that we hold them. We do not go about all the time, or even much of the time, saying to others ‘I believe that p’, ‘I am nearly sure, though not quite sure, that q’ ‘I have a suspicion that r’. In this respect, beliefs are like emotional attitudes. We do not go about all the time saying ‘I rather like Mr X’ ‘I very much dislike Mr Y’.

Moreoever, beliefs resemble emotional attitudes in another and more disconcerting way. I may dislike Mr Y without knowing (‘realizing’) that I dislike him, or even when I think that I like him. Similarly, I may believe a proposition p without realizing that I believe it. It is possible to be mistaken, and in some sense sincerely mistaken, as to what one's own beliefs are, and also as to the degree of confidence with which one holds them. It is even possible to believe a proposition p when one thinks that one disbelieves it or that one has an ‘open mind’ about it. There are unconscious or repressed beliefs, and there are also subconscious ones (beliefs which we do not know that we have, though we could discover that we have them if we made a not impracticably-great effort of attention.) First-person present-tense belief utterances may therefore fail to convey to others what the speaker's beliefs actually are, because he himself does not always know what they are.

Finally, even if we did formulate all our belief-attitudes in words, and even if the attitudes thus formulated were always the ones we do actually hold, we still could not formulate all of them at once. I should still have my belief that the sun is larger than the earth while I was engaged in saying ‘I believe that the earth is larger than the moon’.

I conclude that we should be gravely misled if we supposed that the only, or even the main task of a philosophy of belief is the analysis of first-person present-tense belief statements, though we must not neglect the special and rather peculiar properties which these somewhat uncommon statements have.

Believing a Person and Believing a Proposition

There are two other points about the verb ‘to believe’ which we should notice. In ordinary English there are several different usages of this word. I have already mentioned the distinction between believing ‘in’ and believing ‘that’. This will be discussed in detail in Series II. But there are two other distinctions about which something may be said now. The first is the distinction between believing a proposition and believing a person.

Most commonly, the verb ‘to believe’ is followed by a that-clause. I believe that there will be a fog tonight. Here what is believed (the object of the belief) is a proposition, something which is either true or false. This is the usage I have myself adopted so far. Where we say ‘Smith believes p’, ‘Jones believed q until yesterday’, the letters p and q are what logicians call variables, and the values of these variables are propositions.

But we also speak sometimes of believing a person. He told me that there would be muffins for tea, and I believed him. Alas! there was only buttered toast. It may be that in the history of language this ‘personal’ usage is the earlier one. Nevertheless, it seems to be logically derivative. We believe a person (or fail to believe him) in so far as he asserts something,10 whether orally or in writing or in some other way, for example by means of signals or gestures. It is not sufficient that he should make some communication to us. The communication must take the special form of an assertion. It would not make sense to say ‘He asked me the way to the station and I believed him’ or ‘He invited me to come to the party, but I did not believe him’. Or if these sentences might conceivably make sense in some special cases, it would only be because we took his request or his invitation to be implying some assertion which he did not actually utter, the assertion that he did actually wish to know the way to the station, or the assertion that he had been authorised to issue invitations to the party. In short, we believe a person (or fail to believe him) in so far as he asserts a proposition. Moreover, we only believe him in so far as we believe the proposition which he asserts. Similarly, we only fail to believe him in so far as we fail to believe the proposition he asserts. He told me his story and I believed him. In order to believe him I must believe that the events described in the story did actually happen. Believing a proposition is primary, and believing a person is derivative, because it has to be defined in terms of believing a proposition.

Degrees of Belief

So much for the distinction between believing a proposition and believing a person. But there is another and more interesting distinction which we must consider. I hardly know whether to call it a difference between two usages of the word ‘believe’ or a difference of opinion as to what the correct usage is. At any rate, the word is in fact used in two different ways (whether it ought to be or not) and the difference between them is of very considerable philosophical importance.

In the one usage, belief admits of degrees. You may believe something very firmly, or fairly firmly, or mildly. A rough scale of degrees of belief may be constructed, ranging from conviction at the top end to suspecting at the bottom end, with various degrees of opinion somewhere in the middle. According to this usage of the word ‘believe’, it is possible to believe that it will rain tomorrow morning without being absolutely sure or completely convinced that it will. Moreover, if one's belief does fall short of the highest degree (i.e. does not amount to absolute conviction) one will admit that the proposition believed may be false after all. What one is here admitting is not merely the logical possibility of its being false; for since the proposition believed is an empirical or contingent one, its falsity would not in any case be logically impossible. One is admitting more than this, namely, that the rain may not as a matter of fact occur. One is, of course, claiming (rightly or wrongly) that one has evidence for the proposition believed. But one is not claiming that the evidence is conclusive. The man mentioned in an earlier example who was asked whether Wilkinson was in Oxford and replied ‘yes, I believe that he is’ was using the word ‘believe’ in this way; and it is a very familiar use of the word among reasonable men.

Nevertheless there are some (though I hope there are not many now listening to me) to whom this usage of the word ‘believe’ appears paradoxical or even outrageous. And unfortunately there is another usage of it, though not a very common one. We might call it, rather unkindly, the solemn sense of the word ‘believe’. According to this second usage of the word, belief does not admit of degrees, and we are not allowed to say of someone that he believes a proposition p at all unless he is absolutely convinced of it, completely sure about it. If he is not absolutely convinced that p, but still accepts the proposition with some degree of confidence, we have to say ‘he has the opinion that p’ or ‘he thinks that p’, but not that he believes it. Archibald has opinions, of course, like everyone else, but he does not believe anything. Belief, in this usage of the word, is a matter of all or nothing. Moreover, in this usage of the word ‘believe’, it would obviously be absurd to say of someone that he believes something firmly or strongly. It would be rather like saying that he slept slumbrously or sat in a sitting posture. For belief in this sense is a firm or strong attitude by definition (or shall I say, a stiff one?).

It cannot be denied that both these uses of the word ‘believe’ can claim some sanction from common speech, though the second, I think (the solemn or degree-less one), is mostly confined to special contexts, particularly religious and political ones. Some people perhaps would say that it is only applicable to belief in. Archibald has opinions, but the trouble with him is that the does not believe in anyone or anything. On this view belief in is a matter of all or nothing, even if belief that admits of degrees. I shall try to argue later11 that belief ‘in’ does admit of degrees, or at least that there is a usage of the phrase in which it does. For the present, however, it will be best to confine ourselves to belief ‘that’. There are problems enough about belief ‘that’, and we had better deal with them first, before tackling the still more complicated subject of belief ‘in’.

Confining ourselves, then, to belief ‘that’, let us ask whether belief does or does not admit of degrees. We can now see that this is at least partly a matter of terminological decision. Whatever terminology we adopt, we must agree that there is a more and a less about the way one ‘holds’ or ‘accepts’ a proposition. There are degrees of something here, whether we call them degrees of belief or not. Let us compare (1) suspecting that p, (2) holding the opinion that p, (3) being almost sure that p but not absolutely sure, and (4) being absolutely sure that p, completely and unreservedly convinced of it. All these four attitudes have something in common, and they have it in different degrees. I propose to say henceforward that all these four attitudes (and there may of course be others in between) are attitudes of believing. That is, I shall use the word ‘believe’ throughout these lectures in the sense in which belief does admit of degrees—the weak sense, if one pleases to call it so—and I shall refrain from using it in the ‘all or nothing’ sense in which it refers exclusively to complete and unreserved conviction. If it is complete conviction that I am talking about, I shall say so. I shall treat complete conviction as the highest degree of belief (‘highest’ is not equivalent to ‘best’) and shall avoid the usage in which nothing short of complete conviction is deemed worthy of being called belief at all12. If anyone dislikes this terminology and prefers to use the word ‘belief’ in the degreeless or all-or-nothing sense, he may substitute some phrase like ‘accepting a proposition’ whenever I speak of believing.

My terminology is not very different from Locke's. Locke speaks of degrees of assent and formulates his ‘Ethics of Belief’ in terms of degrees of assent.13 In a reasonable man, he thinks, the degree of his assent varies with the strength of the evidence for the proposition assented to, and one is not entitled to give a high degree of assent to a proposition for which the evidence is weak. But Locke's doctrine of degree of assent (and the ‘Ethics of Belief’ which goes with it), was vigorously challenged by one of the most celebrated authorities on our subject, Cardinal Newman, in the Grammar of Assent ch. VI (‘Assent considered as unconditional’). On Newman's view, we do not assent at all unless we assent with complete and unreserved conviction. Newman's criticisms of Locke will be discussed in Lecture 6 of this series.

• 1.

  Series I, Lecture 7, pp. 186–8.

• 2.

  Treatise (Clarendon Press, ed. Selby-Bigge), p. 624.

• 3.

  Aeneid V, line 231.

• 4.

  The Problems of Philosophy, pp. 24–5 (Home University Library).

• 5.

  The first writer to use this phase, so far as I know, was W. K. Clifford (see the essay with that title in his Lectures and Essays). The latest is Professor R. Chisholm in his Perceiving: a Philosophical Study, ch. 1.

• 6.

  Nicomachean Ethics, Book VII, ch. 3.

• 7.

  P. 23, above.

• 8.

  Similarly, when I say ‘I promise to do X’ I convey to you (though I do not say) that you may safely rely on my doing X. The performatory character of ‘I promise to—’ was noticed by Hume (Treatise, Book III, Part II, Section 5). But so far as I am aware, no one before Austin had noticed that the first person present tense use of cognitive verbs such as ‘know’ and ‘believe’ has a performatory character.

• 9.

  Newman remarks acutely that ‘those who are certain of a fact are indolent disputants’ (Grammar of Assent, p. 152, Longmans, 1947). This is still true when what they are convinced of is not a fact—provided that they are completely convinced of it.

• 10.

  Of course we may believe a person (or fail to believe him) when he denies something. He denied that he was in London last night, and the Dean of his College believed him. But for our present purpose we may assume that denying p is equivalent to asserting not-p.

• 11.

  In Series II, Lecture 9.

• 12.

  Cook Wilson (Statement and Inference, Part II, ch. 3) appears to regard belief as a state intermediate between opinion and conviction—stronger or firmer than opinion, weaker or less firm conviction. This usage too might perhaps claim some support from common speech.

• 13.

  Essay concerning Human understanding, Book IV, chs. 14–16, and ch. 19 (‘of Enthusiasm’).