You are here

Lecture 7: Hume's Analysis of Belief

Hume's analysis of Belief in the Treatise of Human Nature1 is familiar to all students of philosophy, and has been discussed so often that the reader may well think there is nothing more to be said about it. Yet he is, after all, the most celebrated exponent of the traditional Occurrence Analysis of belief and no one has stated it more forcibly than he did. There are other versions of the Occurrence Analysis (possibly better ones) but we must begin with his; and we shall see, I hope, that it is not quite so unplausible as some of his critics suppose.

On the other hand, we can also find in his pages suggestions, at least, of the modern Dispositional Analysis, especially in Sections 9 and 10 of Treatise, Book I, Part iii. (It is a mistake to read only Section 7.) The most obvious example is the passage about ‘not really believing’ in Book I, Part iii, Section 9.2 We may regard this as a felix culpa on his part. It had not struck him, as it strikes us, that these two ways of regarding belief are rather difficult to reconcile with each other, and he himself makes no attempt to reconcile them.

It should also be noticed that in the most ‘official’ formulation of his theory (Book I, Part iii, Section 7) he says nothing at all about degrees of belief. He does not say explicitly, as Newman did later, that assent admits of no degrees. But in this section he writes as if it were degreeless: strangely perhaps, for ‘force’ and ‘liveliness’ do admit of degrees. Nevertheless, in his discussion of the Probability of Causes (Book I, Part iii, Section 12) he does admit that there are degrees of belief, and tries to define probability in terms of them. And here, perhaps we have another felix culpa.

If we are to be fair to Hume's theory of belief and learn all he has to teach us, there are three preliminary points we must bear in mind. The first is the context (so to speak) of the theory, the part which his analysis of belief plays in the whole argument of Part iii of Book I of the Treatise. The conclusions he reaches about belief in Sections 7–10 are going to be used later, in his analysis of the Idea of Necessary Connection in Section 14. He finds the problem of Necessary Connection so difficult that he has to begin by ‘beating about all the neighbouring fields’ in the hope that something useful will turn up.3 His analysis of belief is the most important part of this preliminary exploration.

Thus we must not be surprised to find that his theory of belief is a somewhat narrow one, applicable to some sorts of belief but not to all. He is thinking most of the time of the sorts of belief which are relevant to his problem about the Idea of Necessary Connection: as when a barking sound, for example, induces the belief that there is a dog in the garden. It is, in fact, the sort of belief which we have when we take a perceived event or situation as a sign of another event or situation which is not at the moment perceived. But this sort of belief—the sort which arises from past experiences of ‘Constant Conjunctions’ (as when we see a flame and take this as a sign that we shall feel heat if we come nearer),—is not necessarily the only sort of belief we ever have, however important it is.

The second point we must bear in mind is this. In his discussion of belief, Hume claims to be asking a question which no philosopher has ever asked before. ‘This operation of the mind’ he says ‘seems hitherto to have been one of the greatest mysteries of philosophy: though no one has as much as suspected that there was any difficulty in explaining it.’4 Perhaps Hume exaggerates a little. The distinction between Belief and Knowledge has interested philosophers for a very long time, and was indeed first discussed by Plato. But what had chiefly concerned them was the fact that belief is inferior to knowledge, rather than the nature of belief itself.

What is largely new about Hume's discussion is the attention he pays to the phenomenology of belief itself, and not just to the relation between belief and knowledge. And since his problem was a new one, there was naturally no adequate ready-made terminology for discussing it. He points this out himself, and says he will therefore have to formulate his theory in a number of alternative ways, and that no formulation he can think of is altogether adequate. We shall be unfair to him (and fail to learn what he is trying to teach us) if we take one of his formulations by itself, for example the definition of belief in terms of liveliness,5 and neglect the other expressions he uses, which are intended to amplify this one and correct any misleading implications it may have. Some of his critics, I think, have made this mistake. He does of course say that a belief is ‘a lively idea related to or associated with a present impression’; but he uses a number of other words as well, for example ‘force’, ‘solidity’, ‘firmness’, ‘steadiness’, to describe this peculiar characteristic which he thinks believed ideas have, and other ideas have not.

The third point we must bear in mind is also a terminological one, and this is one which Hume himself did not notice. We must not allow ourselves to be misled by his use of the ‘idea language’—the terminology of ideas and impressions. It would be too much to say (as some philosophers do) that the idea language should be abolished altogether. But if we use it, we ought to use it much more carefully than Hume himself does. The term ‘idea’ in Hume can mean either a mental image or a concept (abstract idea) or a proposition (something true or false). It would make no sense to say that a mental image is believed, or disbelieved either. It may of course be used as a symbol (we may think of ‘in’ or ‘with’ images, as we may think ‘in’ or ‘with’ words) and what is symbolized by it may be believed. But Hume continually says that ideas are the objects of belief, although he also very often uses the term ‘idea’ in the image sense.

We must allow for this peculiarity in his language; in particular, when he says that believing consists in having a lively idea, we need not jump to the conclusion that by a ‘lively idea’ he must just mean ‘a vivid image’. I think he knew as well as anyone else that we may have vivid images without believing,6 and that we may believe without having vivid images. He does indeed hold (mistakenly, no doubt) that every concept we have must in the end be cashable by means of images and that if it cannot be so cashed it is not a genuine concept at all. This is the Imagist Theory of Concepts which he shares with Berkeley; and even this theory, I think, is not so utterly silly as it is made out to be. (Its defect is not that it is absurd, but that it is too narrow.) But whether true or false, it certainly does not imply that when we have an idea in mind we must be actually imaging, but only that we must be capable of producing appropriate images when needed.

So much for the preliminary points I wanted to mention. To understand Hume's theory of belief, we must remember the context in which he states it, and we must also allow for certain terminological difficulties: partly inevitable ones, arising from the novelty of his problem, and others which he might have avoided if he had used his own idea-language more carefully. Now we are ready to consider his theory, as stated in Treatise, Book I, Part iii, Section 7.

Hume's Problem

Hume's problems in this section may be stated as follows. There are three attitudes we may have towards a proposition. We may believe it; or we may disbelieve it; or we may barely entertain it, without either belief or disbelief. (Entertaining is what he calls ‘conceiving the ideas according to the proposition’.7) Hume seems to combine the two last—disbelieving and merely entertaining—under the one word ‘incredulity’. Plainly they are different, but the difference does not matter for his purpose. What is important for him is the fact that both alike differ from belief. What then is the difference between belief on the one hand, and disbelief or merely entertaining on the other: or, as he puts it himself, ‘Wherein consists the difference betwixt incredulity and belief?’8 That is the question which he is trying to answer.

In the first place, it is not a difference in the ideas which are before one's mind. The very same proposition may be disbelieved by someone at one time and believed at another. Again, the very same proposition may be disbelieved by you and believed by me—and merely entertained or considered by a third person without either belief or disbelief. Now suppose I pass from incredulity to belief. What happens? Obviously I do not add any new idea to those I was considering before. It is still the very same proposition that I have before my mind. What I believe now is the very same thing that I was incredulous about before.

Hume gives two further reasons for this contention.

(a) If belief did consist in adding some new idea to those already before our mind, it would be in our power to believe whatever we pleased. For ‘the mind’, he says, ‘has the command over all its ideas, and can separate, unite, mix, and vary them, as it pleases’.9 But Hume thinks it obvious that we cannot believe what we please. Belief, on his view, is something which arises in us independently of our choice. Hume gives no examples here. But we can easily think of some. When you look through the window and see rain falling heavily, you cannot help believing that the streets outside are wet. When I see a flame, I cannot help believing that I shall feel heat if I come closer to it. When I hear a barking sound I cannot help believing that there is a dog in the garden.

(b) If belief did consist in adding some other idea to those already before the mind, what could this additional idea be? The only possible candidate, Hume thinks, is the idea of existence; but this candidate must be rejected, because there is no idea of existence in the sense in which there is an idea of ‘red’ or of ‘cat’.10 To put it linguistically, the word ‘existence’ does not function in at all the same way as words like ‘redness’ or ‘roundness’, though grammatically it is an abstract noun as they are. As Hume puts it himself ‘The idea of existence is nothing different from the idea of any object’. To conceive of A, and to conceive of A as existent, are the same thing.11 ‘Any idea we please to form is the idea of a being, and the idea of a being is any idea we please to form.’ Hume gives a theological illustration. When I conceive of God, a certain complex idea is before my mind. And when I conceive of him as existent, no new idea has been added to the ideas of omnipotence, omniscience etc. which make up the meaning of the word ‘God’. Existence, in more modern language, is not an attribute or characteristic, as omnipotence etc. are. Thus ‘When I think of God, when I think of Him as existent, and when I believe Him to be existent, my idea of Him neither increases nor diminishes.’12

So far, Hume claims to have shown that believing does not consist in adding some extra idea to those which were before our minds already. When A believes a proposition p and B disbelieves it or just entertains it neutrally, the ideas before their minds are exactly the same, or to put it another way, the state of affairs they conceive of is exactly the same. If it were not, there would be no disagreement between A and B.

Then what is the difference between them—between the belief of the one person, and the ‘incredulity’ of the other? Hume concludes that the difference can only lie in the manner of conceiving, since what is conceived is the same in both cases.

‘Relations of Ideas’ and ‘Matters of Fact’

Before elaborating this answer, Hume points out (not very clearly) that his problem only arises with regard to propositions concerning matters of fact. It does not arise (he thinks) about a priori propositions such as those of Pure Mathematics or Formal Logic (what he calls ‘propositions about the relations of ideas’) I think his view here is something like this: when a mathematical statement, for instance, is put before us, such as ‘17 is the square root of 289’, the only question which arises is, do we fully understand it or not? (This would apply to mathematical proofs too, which are complex statements of the form ‘p entails q’.) If you really do understand the statement ‘17 is the square root of 289’, you ipso facto accept it. The contradictory of it (‘17 squared does not equal 289’) cannot be conceived, because it is self-contradictory. Hume's word for ‘self-contradictory’ is ‘absurd’; it does not make sense to say that 17 squared is not equal to 289. For in mathematics and in the other a priori sciences (those which are concerned with ‘the relations of ideas’ and not with matters of fact) ‘true’ is equivalent to ‘logically necessary’, and ‘false’ to ‘self-contradictory’.

With matter of fact propositions, however, it is very different. One of Hume's examples is the proposition ‘Caesar died in his bed’.13 If someone says this to me, I certainly do not assent to what he says. But I do ‘clearly understand his meaning’ ‘and form all the same ideas which he forms’. What he says is false, but it is not self-contradictory. In this case the imagination is free to conceive both sides of the question, but with an a priori proposition it is not. There is no conceivable alternative to the proposition ‘17 is the square root of 289’. I think this really amounts to saying that in the a priori sciences, such as mathematics, there is no room for belief, as distinct from knowledge. When once you fully understand what is being said, you know either that it is true (i.e. logically necessary) or else that it is absurd (i.e. logically impossible), and there is no room for mere belief at all.

But if this is Hume's view, it needs to be amended a little. In mathematics and other forms of a priori thinking there surely is room for believing or ‘merely believing’, as well as for knowledge, because, very often, one does not fully understand the proposition which one is considering; one does not always see all the implications of the relevant definitions, transformation-rules, axioms etc. (Mathematicians themselves speak sometimes of ‘conjectures’, and they would admit that they sometimes ‘have a hunch’, i.e. a mild belief, that such and such a proposition is true before they have found the proof.)

But it could be argued, I think, that when someone is said to ‘believe’ a mathematical or logical proposition (i.e. merely believes it, without knowing it to be true) his belief is not really about the mathematical or logical proposition itself, but about certain words or other symbols. What he is believing is that there is some logically-necessary proposition which these symbols symbolize, without grasping what exactly this proposition is. This is a special case of a very familiar situation, which is vulgarly called ‘believing what you do not understand’ (or do not fully understand)—as when people are said to believe some doctrine of Theology (or of Physics or Economics) ‘without understanding it’. In all these cases, their belief, I suggest, is really about a sentence or other set of symbols: and what they are believing is that there is some important and true propositions which is formulated by the sentence, but they do not know what exactly the proposition is. We all of us operate very often with uncashed or not fully cashed symbols; and we do this in a priori studies, such as arithmetic and logic, as well as elsewhere. And that is how belief (as distinct from knowledge) does after all have a place in a priori thinking.

Belief As ‘A Manner of Conceiving’

Hume's problem, however, as he sees it himself, is concerned only with empirical propositions. His problem is, ‘What is the difference between believing an empirical proposition, and being incredulous about it?’ But strictly speaking, even this formulation of the question is too wide. For we also have to exclude propositions about what is sensibly or introspectively evident: statements describing a present sense-impression or a present impression of reflection, in Hume's terminology; for example ‘This looks red to me now’ or ‘I am now feeling frightened’. Here again there is no room for belief, on Hume's view. We are just directly aware of these impressions at the moment when we experience them. His problem really concerns empirical propositions which go beyond what is actually present to sense (or introspection), as when I hear a barking sound and believe that there is a dog outside the door when I do not actually see the dog, or see a flame and believe that it is hot, though I do not at the moment feel the heat.

This is what Hume calls belief about matters of fact. ‘Matters of fact’ is a technical term of his, meaning empirical facts about something which is not at the moment perceived or introspected. The ‘believables’ which he is discussing are propositions about matters of fact in this sense of the phrase: and the question he is asking is, what is the difference between belief and incredulity with regard to these propositions?

So far, he has only said that the difference must lie in the manner of conceiving. For when I believe that there is a dog outside the door but you are incredulous about it, there is no difference between the ‘ideas’ we conceive of. To put it otherwise, what you are thinking of is the same as what I am thinking of. The difference is that you think of it in an incredulous manner and I in a believing manner. Similarly, if I come to believe something which I previously disbelieved or doubted, the change which occurs in me is just a change in my manner of conceiving—not in what I conceive of.

Perhaps we should all agree with Hume so far. Surely he is just saying (in rather peculiar language) that the difference between belief and incredulity is a difference in mental attitude as opposed to a difference in the object? And surely this is perfectly obvious? Perhaps it is, but one may wonder whether it seemed so before Hume wrote. However this may be, he is now going on to tell us what sort of a ‘manner of conceiving’ belief is; in what way a believed idea differs from one which is not believed. And this is the most important and the most controversial part of his theory.

When we pass from incredulity to belief, the idea which we are conceiving does not alter at all in respect of its content, as we have seen. But, he says, it does become more lively or vivacious. The difference in ‘the manner of conceiving’ amounts, then, to this: an idea which we believe is conceived in a lively or vivacious manner, or presents itself to our minds in a lively or vivacious way; whereas an idea which is not believed does not present itself to our minds in this lively way, but in a faint or feeble way. As I remarked earlier, it is important to notice that Hume also uses a number of other adjectives (besides ‘lively’ and ‘vivacious’) to describe this difference in the manner of conceiving. And he points out himself that none of them are altogether accurate. His hope is that if we consider all of them together, we shall be able to see for ourselves what he is talking about.

I will quote his own words. ‘An idea assented to feels different from a fictitious idea that the fancy alone presents to us: and this different feeling I endeavour to explain by calling it a superior force or vivacity or solidity or firmness or steadiness.’14 He adds that belief is ‘that act of mind which renders realities more present to us than fictions, causes them to weigh more in the thought, and gives them a superior influence on the passions and imagination’15 and further that belief ‘makes them [ideas] appear of greater importance; infixes them in the mind; and renders them the governing principles of all our actions.’16 (here we find a hint, at least, of the Dispositional Analysis of belief).

One may well think that some of Hume's other words—‘force’ ‘solidity’ ‘firmness’ ‘steadiness’—are much better than ‘liveliness’ for describing this difference of feel which there is between ideas which we believe and those we do not believe. It is unfortunate, perhaps, that ‘lively’ is the one Hume uses, when he wants to give a brief definition of belief (‘a lively idea related to or associated with a present impression’).17 This preference for the word ‘lively’ has naturally misled his readers, and especially his critics. The word ‘lively’ is too easily taken in a more or less literal sense to mean just ‘vivid’. And so Hume is supposed to be maintaining that believing something just consists in having a vivid mental image. Words like ‘solidity’ and ‘firmness’ and ‘steadiness’—just because they are obviously metaphorical—do not convey this misleading impression.

But he himself says quite plainly ‘I confess that it is impossible to explain perfectly this feeling or manner of conception. We may make use of words that express something near it. But its true and proper name is belief, which is a term that every one sufficiently understands in common life.’18 This is a pretty explicit warning that the words ‘lively’ and ‘vivacious’ are not to be taken literally (any more than the words ‘force’ ‘solidity’ ‘firmness’, etc.) and that a ‘lively idea’ must not be supposed to mean ‘a vivid mental image’. If Hume really had thought that believing just consists in having a vivid mental image associated with a present impression, he would not have had all these difficulties in conveying his meaning to the reader. It would not have been in the least ‘impossible to explain perfectly this feeling or manner of conception’, as he says it is. On the contrary, it would have been quite easy.

Hume illustrates the difference between ideas which are believed and ideas which are not believed by several examples. I shall now mention two of them.

(1) I remember a certain past incident which you have forgotten. I try to get you to remember it too. I describe it to you at length, and still you do not remember, though you understand perfectly what I say (i.e. have the same ideas as I have). But then suddenly some little detail which I mention ‘revives the whole, and gives (you) a perfect memory of everything’. And now the very same ideas ‘appear (to you) in a new light’ and ‘have in a manner a different feeling from what they had before’. ‘Without any other alteration, beside that of the feeling, they become immediately ideas of the memory and are assented to.’19

(2) Two people, A and B, read the same book. A reads it as a romance, and 5 as a ‘true history’. The same ideas are before the minds of both. But clearly these ideas feel very different to the two readers. B (who reads the book as a ‘true history’) ‘has a more lively conception of all the incidents’. He ‘enters deeper into the concerns’ of the persons described in the narrative, i.e. I suppose, has stronger emotions about them.20 Moreover, the direction of his thoughts is affected in a way that A's are not. He is led to consider other propositions which are implied or made probable by what he reads—propositions about the actions or characters of the persons described, about their friendships and enmities, and even about their personal appearance. This illustrates what Hume meant when he said that ideas which we believe ‘weigh more in our thoughts’ and have ‘a superior influence on the passions and imagination’ than those which we take to be fictitious.

In considering this rather curious illustration, we must remember that Hume himself was both a historian and a tough-minded lowland Scot. It is possible that a more flightly-minded Celt might react the opposite way. Yet if he did, would he not be at least half-believing the narrative—even though he did ‘read it as a romance’?

Relation to a Present Impression

So far then, Hume has argued that the difference between belief and incredulity is a difference in the manner of conceiving, not in the ideas conceived. An idea which we believe feels different from one which we do not believe. This ‘feel’ or felt quality which believed ideas have cannot be adequately described, though it is familiar to everyone. We may however approximately describe it by saying that an idea which we believe has a liveliness or force or strength or steadiness or solidity, which other ideas have not; adding, that it has an effect upon the subsequent course of our thoughts, and also upon our emotions and our actions, which other ideas have not.

Hume now goes on to complete his theory by explaining how this ‘strength’ or ‘solidity’ or ‘liveliness’ comes into existence: how an idea comes to have this peculiar felt quality.21 It acquires this liveliness or force, he says, by being related to or associated with a present impression. What sort of relation or association does he mean? It is the sort which arises from past experience of constant conjunctions. Suppose impressions of kind A and impressions of kind B (flame and heat for example) have been constantly conjoined in my past experience; then if an impression of kind A occurs again, it will bring to my mind the idea of B. When I now see a flame I shall think of heat.

But this is not all. The impression reminds me of its usual accompaniment. But it does something else as well. The impression, like all impressions, has an intrinsic forcefulness of vivacity. And it communicates some of this forcefulness or vivacity to the idea which is associated with it.22 (One is reminded of the ghosts in the Odyssey which were reanimated by drinking blood.) When I see the flame, I not only think of heat, I think of it in a lively or forceful manner. In other words, I believe that the heat is actually there. The intrinsic liveliness or forcefulness of the sense-impression A spreads, as it were, to the idea of B, and so the idea of B is not just conceived, but conceived in the lively or forceful ‘manner’ which constitutes believing.23 In fact, the association due to past constant conjunctions enables an idea to approach in some measure to an impression. When you see the flame, it is almost (though not quite) as if you felt the heat. When you hear the growling sound from behind you, it is almost as if you saw the dog, or felt it biting you.

Seeing is believing, it is said. The remark is obviously false, and Hume does not think it true. Nor does he think that believing is seeing. But he does think that believing approaches in some measure to seeing. When you believe, it is only an idea which is before your mind. But the idea has some of the forcefulness or solidity, the ‘feel of reality’, which actual sense-impressions have, and it gets this because there is an actual sense-impression present, with which this idea is associated by constant conjunctions.

One may illustrate this last point—about the way an impression communicates some of its forcefulness or vivacity to an idea associated with it by constant conjunction—by an example which Hume himself gives in the Appendix24 (though he uses it for a rather different purpose). If I am looking out of a basement window and see a pair of legs walking on the pavement outside, I shall believe that a complete human being is passing by. For in my past experience the sight of a pair of moving human legs has been constantly conjoined with the sight of a human head, arms and trunk. So when I see the moving legs, I vividly or forcefully conceive of the rest of a human body. And the vividness or force-fulness of this conception approaches the vividness of sensation itself: so much so, that if you asked me, I should almost certainly say that I saw a man walking past.

In strict and sober truth, all I actually saw was a pair of moving legs, and the rest of the man was only conceived of and believed in, not actually seen. But the part which I do see communicates some of its sensational force or vividness to the rest, because of the constant conjunction between moving legs and the rest of a human body which there has been throughout my past experience; so that it really is almost (though not quite) as if I were seeing the man as a whole. Consequently, if the belief turned out to be mistaken—if the thing was not a man at all, but a machine with wooden legs dressed up in trousers—I should be quite inclined to say that my senses had deceived me. This would not of course be true (there would be nothing erroneous about my visual perceptions themselves) but Hume's theory does explain why we are tempted to say it. For in such a case as this, the idea believed is so closely tied to the visual impression that it almost has the vividness or forcefulness of sense-perception itself.

We must now say something more about this associative link between impression and idea which plays such an important part in Hume's theory of belief. As we have seen, he usually holds that it is in the sort of linkage which is produced by experience of constant conjunctions, such as the conjunction between flame and heat. If my experience has been that whenever an event of kind A occurs, an event of kind B accompanies it, then perceiving an event of kind A will lead me to believe in the existence of its ‘usual attendant’, an event of kind B. Moreover, in some passages at any rate, he seems to think that this only happens when there is a causal relationship between the events A and B, the earlier of the two events being the cause, and the later the effect. It would not be relevant to discuss Hume's analysis of causation here. It is enough to remind the reader that the necessary connection which we suppose there is between cause and effect is analysed by him in subjective or psychological terms, and is indeed itself defined in terms of belief. We say there is a necessary connection between cause and effect, because when the one event is perceived, we have an irresistible felt inclination to believe that the other event will follow. The impression from which the idea of necessary connections is derived is an ‘impression of reflection’ (i.e. a datum of introspection). It is the felt passage of the mind from perceiving A to believing in the existence of its ‘usual attendant’ B.

In fact, Hume seems to think, usually, (1) that all believing results from experience of constant conjunctions. (2) that therefore all believing is of the sort which occurs in a causal inference: either (a) inferring an effect from a perceived cause, as when you see the trigger being pulled and believe that there is going to be a bang; or (b) inferring a cause from a perceived effect, as when you hear a bark and believe that there is a dog a few yards away. He says in the Appendix ‘We can never be induced to believe any matter of fact, except where its cause, or its effect, is present to us’.25

But at other times he is aware that this view is going to get him into some difficulties. If believing consists in having a lively or forceful idea associated with a present impression, why must the association be of the ‘constant conjunction’ sort? There are other relations, beside constant conjunction, which produce associative linkages; namely, the relations of resemblance and of contiguity. If A resembles B, or A has been contiguous with B in some past experience of mine, then on perceiving A I shall be led to think of B. But do associations of this sort produce belief? If not, why not? When A leads me to think of B because of resemblance or by contiguity, do I think of B in the believing way?

For example, there is a cloud in the sky which resembles a dragon. It has a dragon-like shape; so when I see it, I am led to think of a dragon. But do I believe that the cloud is a dragon, or that there are dragons at all? Again, some years ago I drove through Doncaster, and a circus, with a number of elephants, was passing along the main street. Here was a relation of contiguity between the houses, pavements etc. and the elephants. I drive through the town again this year, and on seeing the houses, streets etc. I am reminded of the elephants. But do I believe that the elephants are there or that they will shortly reappear? Obviously not. Here is an idea associated with a present impression. But it does not have the liveliness or forcefulness which believed ideas have. Why not?

Effects of Contiguity and Resemblance

Hume discusses this question in Sections 8, 9 and 10 of Book III. That is why he had to write these sections (which I fear are seldom read) as well as Section 7 in which his own theory of belief is stated. First, he argues that the presence of these other relations (resemblance or contiguity) does make belief easier and their absence makes belief more difficult, but they cannot by themselves produce it. Only constant conjunction can do so.

For example, philosophers find it easy to believe in the transmission of motion by impact, because in this case the cause and the effect resemble each other (both are movements); in fact, they find this belief so easy that they suppose, quite falsely, that in this case the effect could be inferred from the cause a priori without the aid of experience.26
Conversely, they find it very difficult to believe that events in the body can cause events in the mind (or vice versa) because in this case the cause and the effect are so utterly unlike each other: and this, despite the constant conjunction between the two sorts of events which provides such very strong empirical evidence that there is indeed a causal connection between them.27

(In fact, the astonishing maxim that causes must always resemble their effects has been regarded by some as a self-evident metaphysical principle, though there is not the slightest empirical ground for it.)

One may offer another, and more commonplace, illustration. When you survey the winter landscape, on a cold, dark day, in a snowstorm, don't you find it quite difficult to believe that summer will come again in six months' time, and even a little difficult to believe that it will ever come again? This is because of the contrast, the extreme lack of resemblance, between the present state of affairs which you actually see and feel, and the very different state of affairs you are thinking of. Again, when you come home from a journey abroad (especially if you come by air) you may find it quite difficult to believe that only yesterday you were in Rome—though you quite clearly remember that you were. The contrast between idea and perception is so striking that it almost prevents you from believing what you perfectly well remember. Conversely, you might find it quite easy to believe that you had never been away at all—that yesterday's experience was just like to-day's—though on reflection you can clearly remember that it was very different.

Hume illustrates the effects of contiguity by a curious example about pilgrimages. A pilgrimage to Palestine or to Mecca increases our belief in the events narrated in the Bible or the Koran, and ‘has the same influences on the vulgar as a new argument’.28 This is because the events narrated in the Sacred Book are said to have occurred in or near the places which the pilgrim actually sees. As the vulgar might say themselves, the events described in the Bible seem ‘more real’ or ‘are more real to them’ when they have visited Jerusalem, and have seen the River Jordan.

Hume admits, then, that relations other than constant conjunction can strengthen belief, and their absence can weaken it. An idea which is believed will tend to be more lively or forceful, if it is the idea of something contiguous with or resembling what is at present perceived. Indeed, he goes even farther. He thinks that in some exceptional cases an idea can be believed (have a felt forcefulness or liveliness) when it is not related to a present impression at all. He thinks that this happens in madmen29 (I am afraid he did not notice that there are many different kinds of madness). In madness, he thinks, belief may be produced by purely physiological causes. The ‘ferment of the blood and spirits’ is such that fact and fiction can no longer be distinguished. Every idea which comes into the mind is forceful or vivid, and is accordingly believed, no matter whether it is related to a present impression or not. (This is the state of mind which has since been called Primitive Credulity—a state in which every idea which comes into the mind is ipso facto believed. I shall have something more to say about it in a later lecture.30)

Are we to assume, then, that Hume is now giving up the theory of Belief which he holds elsewhere—the theory that believing just consists in having a lively or forceful idea associated with a present impression by a relation of constant conjunction? He has now said, apparently, that sometimes the forcefulness or liveliness is partly the result of another sort of associative linkage between impressions and idea, namely an association by contiguity or resemblance. As we have seen, he admits and even insists that the presence of either of these other relations (contiguity or resemblance) may strengthen or facilitate belief, and their absence may weaken belief or make it more difficult. And in the passage about the madmen, he has gone even further, and has admitted that an idea can be believed (be lively or forceful) when it is not related to a present impression at all. Yet Hume does not seem to have given up his own theory that the liveliness or forcefulness of believed ideas arises from past experiences of constant conjunctions. For he makes constant use of it later, especially in the section on the idea of necessary connection.31

Justifiable and Unjustifiable Beliefs

What are we to make of this apparent inconsistency? I think the solution is fairly clear. Hume pretty obviously holds that in so far as the forcefulness or vivacity of an idea depends upon mere resemblances or contiguity, or anything else which is other than constant conjunction, to that extent the belief is subnormal or silly or superstitious or pathological. For instance, the pilgrim may of course be justified in believing what he has read in the Bible; but at any rate he is not justified in believing it more firmly than before, just because he has visited Jerusalem. As a matter of psychological fact, a pilgrimage may ‘have the same effect on the vulgar as a new argument’. But in so far as it does, the vulgar are being silly or superstitious.

Or again, we may consider the two philosophical beliefs which Hume mentions: the belief that motion is transmitted by impact (where cause and effect resemble each other, both being movements) and the belief that bodily changes cause mental ones (where the cause does not resemble the effect at all). Philosophers may in fact believe the first proposition much more firmly than the second. But Hume clearly thinks that this difference between the two beliefs is not justified. The justification for both alike is simply the experience of constant conjunction (observed uniformity of sequence); and it is equally strong in both cases. A fortiori the madman's beliefs are silly or subnormal or unjustified, when he believes every idea which comes into his mind, no matter whether it is related to a present impression or not.

According to this interpretation, Hume does want to stick to his original definition of belief as a lively (‘strong’, ‘solid’ ‘forceful’) idea associated with a present impression by a relation of experienced constant conjunction. Only he now wants to say that it is a definition of reasonable or sensible or sane or intelligent belief: and if, or to the extent that, the forcefulness of the idea comes about in some other way, to that extent the belief is unreasonable or silly or unjustified or pathological.

Perhaps Hume would not like using the words ‘reasonable’ and ‘unreasonable’ himself, because he tends to associate ‘reason’ with deductive reasoning, and with the theories of Rationalist philosophers who hold that important truths about matters of fact can be established by purely deductive inference. But actually we all think there is such a thing as inductive reasonableness as well as deductive reasonableness (it is Hume himself who has taught us this) and neither is reducible to the other. And it is perfectly true that for beliefs about matters of fact, the sort of beliefs Hume is here talking about, it is the inductive sort of reasonableness which is relevant. A belief about matters of fact is reasonable or sensible or justifiable, if or to the extent that we are led to hold it by experience of observable regularities—what Hume calls constant conjunctions.

It is sometimes said that Hume's philosophy leaves no room at all for the distinction between reasonable and unreasonable beliefs. On the contrary, it could be argued, the whole point of the Treatise, on its constructive side, is to show what the distinction is. Hume is saying that in the sphere of matters of fact reasonable or sensible or justifiable beliefs are those—and only those—which are based on experienced regularities. On its destructive or sceptical side, on the other hand, the Treatise is a criticism of the Rationalistic way of drawing the distinction between reasonable and unreasonable beliefs, which tries to substitute deductive inference for learning from experience. The criticism was greatly needed at the time, and perhaps still is.

If I right so far, Hume's theory of Belief may be summed up as follows:

1. The difference between believing and not believing is a difference in the manner of conceiving an idea, and not in the content of the idea conceived.

2. We can roughly indicate what this manner of conceiving is, by saying that a believed idea is one which feels strong or forceful or lively or solid.

3. Except in the special case of madness, the forcefulness or liveliness of the idea arises from its relation to (its associative linkage with) a present impression; the idea gets its liveliness from its relation to something actually perceived (or introspected) at the moment.

4. In sensible or sober or sane belief, this associative link between idea and impression arises from past experiences of constant conjunctions. The impression A enlivens or strengthens the idea B, because A-like impressions and B-like impressions have been constantly conjoined in the past experience of the believer.

5. If, or to the extent that, the associative link between idea and impression is of another sort (if it is just association by resemblance or by contiguity), then or to that extent the belief is subnormal or silly or unjustifiable. And a fortiori the belief is subnormal or silly or unjustifiable if the liveliness or forcefulness of the idea does not arise from its relation to a present impression at all, but merely from purely physiological causes (or as Hume might have added, from the effects of hypnotic suggestion).

An Extension of Hume's Theory

Before criticizing this theory, I want to suggest a slight expansion or extension of it—one which Hume himself (I think) could quite well have accepted. For surely his interpretation of the crucial phrase ‘constant conjunction’ is too narrow, and consequently his conception of induction is too narrow as well. He seems to hold that the causal sort of constant conjunction is the only sort. And so he is led to say (I have quoted it already) ‘We can never be induced to believe any matter of fact, except where its cause, or its effect, is present to us’;32 and again, that causation is the only relation which ‘can be traced beyond our senses and informs us of existences and objects which we do not see or feel’33 (e.g. informs us that there is heat in the neighbourhood of the flame which we see, though we do not at present feel the heat). Consequently, he seems to think that when belief is sensible or sane or justifiable it is always some sort of causal inference, based on past experience of causal regularities.

But if this is Hume's view, it needs to be modified on his own showing, because it will not cover all the examples which he himself gives. Causal constant conjunctions (regularities of sequence) are not the only sort of constant conjunctions we experience. There are also constant conjunctions in which the conjuncta are coexistent, not successive—regularities of concomitance, as opposed to regularities of sequence.

Let us consider his own example about seeing the moving legs and believing there is a human body attached to them.34 No doubt there is a causal relation between the movements of the legs and events going on in the rest of the body (e.g. events in the brain and in the nerves running down the spinal cord). But the ordinary believer knows nothing about these. Yet he believes, and believes reasonably or ‘sensibly’, that the rest of a human body is there, when he sees no more than the moving legs. The relation on which his belief is based is a relation of constant concomitance or coexistence, not of constant succession. Whenever he has previously seen a pair of moving legs, they have been spatially (not temporally) conjoined with a head, arms and trunk.

It is true that Hume only uses this example once. But if we consider another, which he uses frequently, the one about flame and heat, it is not clear that this is a case of regular sequence either. On the face of it, it too is a case of constant concomitance or coexistence. And if we do start asking causal questions and consult the physicists or the chemists, they certainly will not tell us that the flame is the cause of the heat—at any rate not if you mean by ‘the flame’ the bright flickering thing which you actually see (which is what Hume himself means by it). They will say that both the flame and the heat are concomitant or coexistent effects of a common cause—namely the chemical process of combustion, which causes both light rays and heat rays to be radiated concomitantly.

It is easy to think of other examples. When I see a chimney pot, I believe that there is a fireplace somewhere underneath it. When I see a wall in the distance, I believe that it is hard; i.e. that these visible qualities are accompanied by the tangible quality of hardness. When I look at the sea I believe that there is land on the far side of it, beyond the horizon. Again, let us consider what happens when we recognize an object as being a cat, or as being a piece of ice. Such recognition is made possible by past experience of constant conjunctions. When I recognize something as a cat, all I see, perhaps, is a tabby-coloured thing with a round face and two ears; on seeing this, I believe that it has many other feline characteristics as well, e.g. soft fur, a liking for milk, a tendency to pursue mice, a capacity for purring. I believe this, because these characteristics have constantly accompanied that sort of shape and colouring in my past experience.

This is obviously an inductive belief, something learned from experience; it is just the sort of case which fits Hume's theory very well. It is very plausible to say here that an idea (that of cathood) is made lively or forceful by its relation to a present impression, and that this comes about through past experiences of a constant conjunction. But the constant conjunction in this sort of case is not a causal one. It is a constant concomitance, a regularity of coexistence, and not a regularity of sequence.

Moreover, in all these examples, the belief is sensible or justifiable, or reasonable: reasonable, that is, in the inductive way. It is a belief about a matter of fact, and a result of learning from experience. But the experience, in these cases, is an experience of constant concomitances, not of constant sequences. We are making an inductive inference, but it is not a causal inference. Hume does seem to think that all inductively-established beliefs are causal ones. But one can see nothing in his own empiricist principles which obliges him to hold such a narrow view of induction and of beliefs about matters of fact (such a narrow view, if you like, of learning from experience). And his own phrase constant conjunction can perfectly well be made to cover constancy of coexistence (constant concomitance) as well as constancy of succession. So in what I have just been saying, I have not been making an objection against his theory of belief, but merely suggesting a quite easy and obvious extension of it—an extension, too, which seems to be demanded by some of the examples he gives himself.

Before leaving this point, we may notice that it is relevant also to what he says about beliefs based on mere contiguity. What is ‘contiguity’ but a kind of concomitance or coexistence? Is there really anything wrong (silly, pathological) about beliefs based on mere contiguity? Nothing; unless the contiguity is irregular or inconstant, as for example when I believe that there is a bull behind this hedge now, merely because I once saw a bull there. But if I had walked down the lane many times in the past, and had always seen a bull behind the hedge, that would be quite a good ground for believing that there is a bull there now.

Similarly if very extraordinary events were always happening in Jerusalem or near the River Jordan, then a pilgrimage to Palestine would quite justifiably strengthen the pilgrims' belief in the Biblical narratives. But actually, of course, in most cases where belief is strengthened by the effects of contiguity, the contiguity is not constant or regular. Your belief in the reality of King Arthur may be strengthened by a visit to Caerleon, because he is supposed to have held his court there. But there is no regular or constant contiguity between persons or events of the Arthurian type and that particular part of Monmouthshire.

Thus, if I am right, the distinction Hume wants to make between sensible or sane or justifiable beliefs and silly or superstitious or pathological beliefs should be formulated as follows:

1. There are beliefs arising from experience of constant conjunctions, either conjunction of the causal sort, or conjunction of the concomitance sort. These beliefs, and these only, are sensible or sane or justifiable.

2. There are beliefs arising either (a) from mere resemblance, or (b) from conjunctions which are only occasional, not constant, (that is what ‘mere contiguity’ amounts to), or (c) from purely physiological causes, as in madness. These are silly or superstitious or subnormal or pathological.

We ourselves might prefer to formulate his distinction in more familiar or ‘ordinary’ language. We might wish to say that beliefs arising from experience of constant conjunctions (whether causal conjunctions or not) are reasonable; and if, or to the extent that, a belief arises in some other way, i.e. is independent of experience of constant conjunctions, then (or to that extent) it is unreasonable or non-reasonable. But in that case we must of course remember that the reasonableness, or lack of it, which is here in question is inductive reasonableness, not deductive. For the beliefs Hume is talking about are beliefs concerning matters of fact. A reasonable person, in this sense, is one who learns from experience and believes accordingly.

‘Common sense’ (at least in one usage of the term) is another name for this inductive sort of reasonableness. If we say that ‘Mr. A has no common sense’ we often mean that in forming his beliefs he is unaffected by or indifferent to his own experience of constant conjunctions. Similarly if we say to someone ‘Use your common sense’, this amounts to saying ‘Recall to your mind the now-relevant part of what you have learned from experience’.

The Weaknesses of Hume's Theory

We may now turn to the weaknesses of Hume's theory. The first and most obvious one is that it is a very narrow theory, as I hinted at the beginning. It fits some beliefs very well, and they are important ones too, but by no means all. And even if we extend it a little, in the way I have just suggested—by widening Hume's own notion of constant conjunctions—it is still a very narrow theory. Even within the sphere of ‘matters of fact’ there is a very important class of beliefs which Hume's definition of belief does not fit, namely, general empirical beliefs: for example, the belief that all men are mortal; that water expands when it freezes, that wood is combustible. In fact any inductive generalization will do as an example: and we ought to include among such generalizations not only propositions of the form ‘All A is B’. ‘No A is B’, but also propositions of the form ‘Most A's are B’. ‘Few A's are B’. ‘about 70 per cent of A's are B’, etc. (e.g. ‘Most Englishmen enjoy watching cricket’.).

It is very odd that Hume's theory of belief will not apply to inductive generalizations, because he was so particularly interested in induction. He obviously thinks that we do believe many inductive generalizations, and is more concerned than any other philosopher before him to enquire what the justification for them is. Nevertheless, his own theory of belief (so far as I can see) will not apply to inductive generalizations at all. It will not apply to general beliefs about matters of fact, but only to beliefs about particular matters of fact. This is because he insists that an idea which we believe must be related to or associated with a present impression.

Let us consider the generalization ‘Water always expands when it freezes’. If we are willing to use Hume's ‘idea’ language, we can agree that the (complex) idea conveyed by this sentence presents itself to our minds in a lively or strong or solid or forceful manner. This proposition ‘feels’ very different from a generalization we do not believe, e.g. ‘All swans are black’ ‘Glass is a good conductor of electricity’. So this part of Hume's theory—his contention that believing is ‘a manner of conceiving’ and his attempt to indicate what this manner of conceiving amounts to—applies well enough. But though the idea or set of ideas conveyed by an inductive generalization does have this character of strength or forcefulness or solidity, where is the relation to a present impression?

I can believe that water always expands when frozen, even though I am not now seeing or touching any water. No doubt some sense impression will be present when I consider this proposition. Sense impressions are always occurring so long as one is awake and conscious. When I consider this generalization about the freezing of water, I may be seeing a desk and hearing the sound of the traffic in the street outside my window. But the sense-impressions I happen to be having at the time are entirely irrelevant to this belief. If the relation to a present impression is essential (as Hume says it is) I could indeed believe that this freezing water which I see or touch is expanding or will shortly expand, but not that water in general expands when frozen. Similarly, to take his own favourite example, I could believe that this flame which I see across the room is hot, but not that all flames are hot.

General Beliefs and Habits of Believing

Now some followers of Hume (and possibly Hume himself) might try to get out of this difficulty by saying that there are no general beliefs, but only habits of singular belief. This is the line taken by F. P. Ramsey in his interesting and intriguing essay ‘General Propositions and Causality’.35 According to Ramsey, we do not really believe that all flames are hot: the truth is only that we have a habit of believing that each particular flame which we see is hot. The actual believings are all of them concerned with actually perceived particulars. Again, on this view one does not actually believe that all men are mortal. The truth is only that whenever one sees some particular man, one will believe him to be mortal. Thus in all our actual believings (as opposed to mere habits of belief) the idea believed would be related to a present impression.

We notice that there is a suggestion here of the Dispositional Analysis of belief, since a habit is one kind of (acquired) disposition. One could quite well reformulate Ramsey's theory by saying that he does not so much want to abolish general beliefs—which was the way I put it just now—but wants to give a dispositional analysis of them, as opposed to an occurrence analysis. Nevertheless, this theory of Ramsey's is not what is ordinarily called the Dispositional Analysis of Belief, but only a restricted and incomplete version of it.

It is not a Dispositional Analysis of Belief as such, but only of one sort of beliefs—namely general beliefs, where the propositions believed are inductive generalizations. Its restricted character is clearly shown when we consider what these habits or dispositions are which he is talking of. They are themselves habits of believing, dispositions to believe. There would be an obvious absurdity (a vicious circle) in suggesting that all belief is just a disposition to believe. Ramsey does not fall into this obvious error. He is not suggesting a dispositional analysis of singular beliefs, e.g. the belief that this man whom I see is mortal, or that this freezing water which I touch is expanding. He holds that singular beliefs, such as these, are actual introspectible occurrences.

So much to make clear what Ramsey's theory is. According to him, what is called a general belief is really a habit of forming singular beliefs; or, as he also says, a rule for forming singular beliefs (something like a recipe in a cookery-book). Quite possibly Hume himself would have accepted this. It is quite in line with his own doctrine that induction is a matter of custom or habit; and that what we learn from experience of constant conjunctions is certain habits of expectation—for example, we learn to expect that whenever a particular flame is seen, it will be hot. But attractive though this theory is (and very convenient for whitewashing Hume, as we should all like to do), there are serious difficulties in it.

If it were correct, it is not at all clear how we could use general beliefs as premisses in reasoning, or how there could be logical relations between them, e.g. compatibility or incompatibility. If you believe that whales are mammals, and I believe that they are fish, here are two general beliefs which are incompatible, inconsistent with each other. But it does not make sense to say that a habit of mine is inconsistent with a habit of yours—or consistent with it either. To put the same point rather differently, we all suppose that inductive generalizations are true or false. And a habit is not true or false; nor is a recipe in a cookery-book—though it may be expedient, or inexpedient, to act upon it.

Again, if we consider our mental attitude to inductive generalisations, it seems obvious that we assent to them, or dissent from them, in just the same way as we assent or dissent to singular empirical propositions. In fact, we believe them (or disbelieve them) in just the same way as we believe or disbelieve other empirical propositions.

It is indeed true that a dispositional analysis of all beliefs, including singular ones like ‘that flame over there is hot’, is very plausible, and we shall see presently that there are hints of it in Hume himself. But this half-way house, which combines a dispositional theory of general beliefs with an occurrence theory of singular beliefs, is not a satisfactory position to stop in, though it looks attractive at first sight. On the contrary, one has to go the whole hog, either one way or the other; and either say that ‘x believes p’ is always an occurrence-statement about him, no matter what sort of a proposition p is; or else say that it is always a dispositional statement about him, no matter what sort of a proposition p is.

If this is right, we shall just have to conclude that general beliefs about matters of fact fall outside the scope of Hume's theory altogether; for the reason I gave before, that though here too the proposition which we entertain is conceived in a lively or forceful manner, it is not related to a present impression.

Hume's theory of singular beliefs—e.g. the belief that ‘That flame over there is hot’—might still be correct, of course. Beliefs of this kind are certainly very important, and play an enormous part in our lives. And if Hume's theory does give the correct analysis of them, this is a very remarkable achievement. Only, he has not succeeded in giving us what I think he claims to give, namely an analysis of all belief about matters of fact, but only about one sort of belief about matters of fact—the sort in which one actually perceived event or situation is taken as a sign of some other event or situation which we do not perceive at the moment. This, then, is one ground for saying that Hume's theory of belief is a narrow or restricted one, though within its own limits it might still be substantially correct.

Beliefs to which Hume'S Theory does not Apply

The narrowness of Hume's theory may also be brought out in a different way as follows: his theory only applies to what one might call automatic beliefs: beliefs which arise in us immediately, as soon as the appropriate sense-impression occurs, without any previous process of questioning or deliberation or weighing of evidence for and against, and without any consideration of alternatives to the proposition believed.36

When I see a flame ten yards away and believe that it is hot, I do not weigh the evidence for and against the proposition ‘It is hot’. It never even occurs to me to ask whether it might be anything else. The logically-possible alternatives that it might be cold or cool instead, or that it might have no temperature-quality at all, just do not present themselves to my mind. It is true that in a sense I have evidence for the proposition which I accept in this automatic and unquestioning manner, and very good evidence too, the evidence of past constant conjunctions. I have seen a great many flames in the past, and every one of them felt hot when I was near enough to it to feel it at all.

But though in a sense I have good evidence for the propositions I believe, I do not consider this evidence at the time. I do not recall these past conjunctions of flame and heat, nor notice their relevance to the present situation. Quite obviously, I do not recall all of them. You might expect perhaps that I should recall one or two of them at any rate, and possibly also have a dim recollection that there have been a great many others. No! nothing of the kind happens. I do not actually recall any past examples of the conjunction of flame and heat, though no doubt I could recall many if I tried. I might recall one or two if the conjunction (though constant) was a relatively unfamiliar one: for instance if I had only recently begun to notice that ice floats in water (I might have lived most of my life in a tropical country, so that I never saw any ice before this winter). But when the conjunction is very familiar, like the conjunction of flame and heat, we certainly do not recall any previous examples of this constant conjunction. When we see a flame now, we just jump immediately to the conclusion that it is hot without recalling any past experiences at all.

Hume emphasises this point himself in Book I Part iii, Section 8. ‘The past experience, on which all our judgements concerning cause and effect depend, may operate on our mind in such an insensible manner as never to be taken notice of, and may even in some measure be unknown to us.’37 He illustrates this by the example of a traveller who ‘stops short in his journey upon meeting a river in his way’38 ‘Can we think’ (says Hume) ‘that on this occasion he reflects on any past experience, and calls to remembrance instances that he has seen or heard of, in order to discover the effects of water on animal bodies?’ Obviously he does nothing of the kind. Just seeing the water there in front of him is quite enough to make him believe that if he walks into it, he will sink. In such a case, and in innumerable others, ‘the mind makes the transition’ (from the perception to the consequent belief) ‘without the assistance of the memory’.39

But obviously there are beliefs which are arrived at in a very different way; cases where we do consider alternatives, where we do attend most carefully to the evidence pro and con., try very hard to recall any past experience which may be relevant; and then, but not till then, we make up our minds (as we say) to accept the proposition p, and reject the alternative propositions q and r. The most obvious example is a jury considering its verdict; or again a doctor making a difficult diagnosis, or a barrister arriving at an opinion on a very intricate legal question.

Or let us consider the belief of a classical scholar, that Homer was a historical person and really did write the Iliad and the Odyssey. This scholar may have spent half a lifetime studying the question, considering and reconsidering the evidence on both sides, continually hunting for fresh evidence. At least, after many years of thought and study, he ‘makes up his mind’ about the question. He passes from doubt and suspense of judgement to belief. This is very unlike the automatic and almost instantaneous process which Hume describes.

Moreover, a belief of this kind certainly does not consist in having a lively or forceful idea related to a present impression, any more than the general beliefs do which I mentioned before, e.g. the belief that water expands when frozen. It is a lively or forceful idea, if you like. No doubt the idea (proposition) ‘Homer was a historical person’ does now feel different to the classical scholar from the way it felt before, when he had not yet made up his mind. And we may try to describe this difference by saying that the idea is now lively or strong or forceful in a way it was not before. But here again no present impression enters into the situation. No doubt the scholar is experiencing sense impressions since he is awake and conscious. When he at last gives his assent to the proposition ‘Homer was a historical person’ he is seeing a wastepaper basket, and feeling rather cold, and hearing the coo of a pigeon outside the window. But these ‘present impressions’ have nothing whatever to do with his belief.

So much for the narrowness of Hume's theory. There are types of belief to which it will not apply at all; though it applies very well to one very important type of belief, namely the sort of belief which one may call sign-cognition—taking one perceived event or situation as a sign of another, because of past experience of constant conjunctions.

But it is a mistake to suppose that this is the only sort of belief we have.

Hume and the Dispositional Analysis

You will remember that the central topic of these lectures was to be the examination and comparison of two very different ways of treating belief: the traditional Occurrence Analysis in which belief is regarded as an introspectible happening, and the modern Dispositional Analysis.

It is pretty clear that Hume himself is, in this respect at least, a Traditionalist. He does think that believing is an introspectible experience. He insists that a proposition which we believe feels different to us from a proposition we do not believe, and he tries his best to indicate what this difference of ‘feel’ amounts to, by the use of words like ‘lively’ ‘strong’ ‘forceful’ ‘vivacious’ ‘solid’ ‘firm’ ‘steady’. These words of his are certainly intended to indicate an introspectible quality or characteristic, or an introspectibly-detectable way in which a believed idea presents itself to the mind.

Nevertheless, if we consider the literal meaning of these adjectives, we notice that all of them, except perhaps ‘lively’, are names of causal properties: that is, taken literally, they are dispositional words. If we attribute force to something, we are speaking of the amount of movement it is capable of producing in other things. If we call something ‘firm’ or’ steady’, that means literally that it has the capacity of resisting disturbing agencies of one sort or another. The word ‘solid’ too is usually a causal-property word, equivalent to something like ‘capable of resisting penetration (or deformation)’, though sometimes, as Locke noted, it is the name of a tangible quality.40 It is quite obvious that ‘strength’ is similarly a causal-property word: and even ‘lively’ (or ‘vivacious’) could be a causal-property word, though it might also mean perceptible or introspectible vividness. For instance, we might say that this kitten is a lively or vivacious creature, and this could be truly said of it even though it is fast asleep at the time.

So there is at any rate a hint or suggestion here that believed ideas differ from others in the effects they are liable to have on us—and not (or at any rate not only) in some introspectible quality they possess. And this is something much more like the Dispositional Analysis. Thus in Professor R. B. Braithwaite's version of the Dispositional Analysis (which I shall consider later41) believing that a flame is hot consists in being disposed to act as if this proposition were true. It might be suggested that this is what the ‘forcefulness’ or ‘strength’ of the believed idea would amount to—its tendency or liability to affect one's actions (where ‘tendency’ and ‘liability’ are dispositional words).

Moreover, though Hume himself obviously intends to use these causal-property words (‘strength’ ‘force’ ‘steadiness’ etc) in a metaphorical way to indicate an introspectible felt quality, he does also emphasise the effects which believed ideas have on us, and unbelieved ideas have not. Belief, he says, gives ideas ‘more force and influence’ ‘makes them appear of greater importance’ ‘infixes them in the mind’ and ‘renders them the governing principles of all our actions’. He also says that belief ‘gives them (ideas) a superior influence on the passions and imagination’—an influence which mere fictions do not have.42

Perhaps his own view is, that though we do have such dispositions when we believe some proposition (are disposed to ‘take the proposition seriously’, to act, and also to draw inferences as if it were true) these dispositions are merely consequences of the peculiar introspectible quality or ‘feel’ which the proposition has. Nevertheless, it is a very natural suggestion that believing something just consists in having these dispositions, that coming to believe something just consists in acquiring them (in acquiring a tendency to take a proposition seriously both in thought and in action) and ceasing to believe it consists in losing them. That is why I say that there are at least hints in Hume of the Dispositional Theory, though the theory he actually holds is the traditional Occurrence Theory, which treats belief as a special sort of introspectible happening. Hume also has some interesting things to say about a topic which cannot very well be discussed at all if we accept a pure and simple occurrence analysis: namely half-belief.43 Treatise Book I, Part iii, Section 10 (rather misleadingly entitled ‘Of the Influence of Belief’) contains a discussion of one important sort of half-belief, the sort of half-belief or near-belief which we sometimes have when we read the works of poets and dramatists.

  • 1.

    My page-references in this lecture are to Selby-Bigge's edition of the Treatise (Oxford University Press).

  • 2.

    S.B., pp. 113 ad fin.—115.

  • 3.

    S.B., p. 78.

  • 4.

    S.B., p. 628 (Appendix)

  • 5.

    S.B. p. 96.

  • 6.

    If he did not, why does he have all this difficulty—admitted, and indeed insisted on, by himself—in explaining to us what this ‘manner of conceiving’ is?

  • 7.

    S.B., p. 95.

  • 8.

    S.B., p. 95.

  • 9.

    S.B., pp. 623–4 (Appendix).

  • 10.

    S.B., p. 66.

  • 11.

    S.B., p. 94.

  • 12.

    S.B., p. 94.

  • 13.

    S.B., p. 95.

  • 14.

    S.B., p. 629 (Appendix) Hume's italics

  • 15.


  • 16.


  • 17.

    S.B., p. 96.

  • 18.

    S.B., p. 639 (Appendix).

  • 19.

    S.B., 627–8 (Appendix).

  • 20.

    S.B., pp. 97–98.

  • 21.

    Part III, Section 8, Of The Causes of Belief.

  • 22.

    See Hume's ‘General Maxim’ at beginning of Section 8 ‘Of the Causes of Belief’.

  • 23.

    S.B., p. 98.

  • 24.

    S.B., p. 626 (Appendix).

  • 25.

    S.B., p. 623.

  • 26.

    S.B., p. in. (The word ‘apparent’ is here equivalent to ‘evident’.)

  • 27.

    S.B., pp. 246–8 (Treatise Book I, Part IV, Section 5 ‘of the immateriality of the soul’).

  • 28.

    S.B., p. 110–11.

  • 29.

    Section 10 ‘Of the Influence of Belief’, S.B., p. 123.

  • 30.

    See Lecture 9, pp., 12–18.

  • 31.

    Book I, Part III, Section 14.

  • 32.

    S.B., p. 623 (Appendix)

  • 33.

    S.B., p. 74.

  • 34.

    S.B., p. 626 (Appendix).

  • 35.
    The Foundations of Mathematics, pp. 237–255.
  • 36.

    In a way Hume insists upon this narrowness himself in his section Of the Reason of Animals (Book I, Part III, Section 16). He is saying there that the human beliefs he has been analysing are the same sort of beliefs as animals have or may be supposed to have.

  • 37.

    S.B., p. 103.

  • 38.

    S.B., p. 103.

  • 39.

    S.B., p. 104.

  • 40.
    Essay, Book II, Ch. 4.
  • 41.

    In Vol. 2, Series II, Lecture 1 ‘Believing and Acting “as if”’.

  • 42.

    S.B., p. 629. (Appendix.)

  • 43.

    See Series II, Lecture 4.

From the book: