Assent and Being Under an Impression That

We have seen that in the traditional Occurrence Analysis of belief, attention is concentrated on a special sort of mental occurrence or mental act, which may be called assenting To assent to a proposition, we have to entertain that proposition; and in the last lecture I tried to say what entertaining a proposition amounts to. Assent, in the traditional theory, consists in taking up a further attitude to an entertained proposition. We have now to consider what happens when we assent to a proposition as well as entertaining it.

In the traditional theory it is usually held that assent admits of degrees. Though Newman, as we have seen, denies this, most other exponents of the Occurrence Analysis have accepted it, as Locke does. The usual view is that when we give our assent to a proposition, we give it with a greater or smaller degree of confidence. I have already said a good deal about this, and need not say much more about it now; I will just remind you of the traditional view that a rough scale of degrees of assent can be constructed, ranging from suspecting or surmising at the bottom end, to absolute sureness or complete conviction at the top end, with various degrees of opinion in the middle.

It is a curious question whether this series of degrees should be regarded as a continuous one: that is, whether between any two degrees of assent there must always be another. I do not see why it need be regarded as absolutely continuous. It is more plausible to suppose that confidence rises or falls in small finite jerks, as the intensity of sensations is supposed to do.

I will just add that there is another difference of degree which should be mentioned. We may assent very firmly to a proposition to-day, we may even be absolutely convinced of it; and yet to-morrow we may assent to it much less firmly or even withdraw our assent altogether, although our evidence for the proposition has not altered in the meantime.

In other words, assent may have different degrees of tenacity or stability over a period of time, as well as different degrees of intensity at any given moment. I shall have to consider this later, in Series II, when I discuss the relation between belief and action. The notion of the stability or tenacity of assent is relevant when we ask how a man can assent to a proposition p, and yet behave, in practice, as if p were false. For the moment we may put it this way: an act of assent (or assent occurrence) may be more confident or less; a habit of assent may have various degrees of stability or tenacity. For the present, it is the act of assent or assent-occurrence that we are concerned with.

The next point is that assent (whatever degree it may have) is the culmination, as it were, of a process: we might almost say, the resolution of a kind of conflict. Assent, at least when reasonable, is preceded by a state of wondering or questioning, in which several alternative propositions are before the mind, together with the evidence for each of them. There must be at least two, p and not-p, and there may be more: not-p may cover quite a large range of alternatives, q, r, etc. During the preliminary stage of wondering or questioning, we sit on the fence, as it were, between these alternatives. As we fix our mind on the evidence in favour of p we may have an inclination to assent to p; and then, turning our mind to the evidence for q, we may have an inclination to assent to q instead. But though we may have these inclinations to assent this way or that, we do not yet actually assent. We remain in a state of suspended judgement.

This state of suspended judgement may last only for a very short time. On my way to deliver a lecture I wonder whether my watch is slow. I look at the clock in the corridor and my state of wondering and suspense of judgement is very quickly terminated. But in other cases the state of suspended judgement may last for a very long time, because the evidence is indecisive, or because there is too little evidence either way, or again, because there is too much, and it is not clear whether on the whole it favours p, or q, or r. The question just has to be left unsettled for the time being, because there are other more urgent matters to be attended to; and when we start wondering about it again, tomorrow or next week, perhaps, we still cannot settle it. We may even remain in a state of suspended judgement about it for the rest of our lives, as the life-long Agnostic does about the question whether God exists.

But still, the state of wondering and suspended judgement, long or short, does often come to an end. We do eventually assent to one of the alternatives. We come off the fence on one side. We prefer or plump for one of the alternatives, accept it or commit ourselves to it, and reject the others; and if we are reasonable, we accept the one which is supported, on balance, by the evidence we have been considering and reject those which are not.

It is important to notice the preferential character of assent. One result of it is that we cannot believe without also disbelieving. If we accept this proposition, we are also rejecting that one. Assenting is something two-sided or two-faced. In assenting to p, one ipso facto dissents from its alternatives q and r; at any rate this is what we do if, or to the extent that, we are assenting reasonably. This is why we find it natural to describe assent in the language of choice. We speak not only of deciding to, (i.e. deciding to do something) but also of deciding that. After waiting for him for 1½ hours, I decided that John had missed the train. We also speak of ‘making up our mind that…’ as well as ‘making up our mind to do something’. After some doubt, I made up my mind that the bird was a lesser spotted woodpecker.

Assent, then, is in some respects analogous to decision: and the process of wondering and evidence-weighing, which precedes reasonable assent, is in some respects analogous to practical deliberation, where we consider several alternative actions and try to estimate the weight or force of the reasons for doing this one or that one.

Moreover, as I said before, assent can also be viewed as the resolution of a conflict.1 We describe it as ‘making up our mind’ because our mind has primarily been, as it were, divided; we had some inclination to assent to p, and also some inclination to assent to q, when p and q are mutually incompatible propositions.

We may summarize this account of assent as follows:

1. Assent is the taking up of an attitude towards an entertained proposition.
2. This attitude has two different features or components (a) Preference; (b) Confidence.

(a) Preference

Because of its preferential character, assenting always includes dissenting. In assenting to p, we ipso facto dissent from or reject its alternatives q and r. Perhaps it was this preferential character of assent which led Newman to say that assent is unconditional,2 and that ‘There is no medium between assenting and not assenting.’ If you assent at all, you do have to make up your mind between two or more alternatives. You cannot ‘have it both ways’; you must decide in favour of this alternative or that.

It is true that you may withdraw your decision later: you may even withdraw it half a minute after you made it. But then you are ceasing to assent; you are returning to the state of wondering or suspended judgement, ‘sitting on the fence’ again. But in so far as you do assent to p rather than q, and for as long as you assent to it, it is p that you decide for. You cannot partially decide for p, or half-choose it. This follows analytically from the concept of ‘choosing’. You are rather like a motorist who comes to a fork in the road. He can stop at the fork as long as he pleases. But if he goes on, he must choose between this road and that.

(b) Confidence

But though when we prefer or plump for p and reject q it is p that we prefer and no half-way house is possible about that, we may still have great confidence or little about the alternative preferred. We may rely upon it whole-heartedly, with no mental reservations at all. Then we have assented with complete conviction. Or we may rely upon it with considerable confidence but with some mental reservation, and then our assent is what we call ‘Having the opinion that’ or more usually ‘Thinking that’ (as opposed to ‘Being convinced that’). Or finally, the confidence we have about p may be small, although p-as opposed to q or r-still is the alternative which we prefer. This is what is traditionally called ‘suspecting that’, and is the lowest degree of assent.

This part of the traditional theory has been criticized, notably by Professor William Kneale.6 Perhaps confidence should not be called a feeling, still less an emotion. But provided we are willing to speak of acts of assenting at all, or assent-occurrences, it does seem to be true that we can assent more confidently or less; the question whether confidence should or should not be described as a feeling can be postponed until later. I shall have something more to say about it in Series II.7 It is one of the issues we have to consider when we try to decide between the Occurrence or ‘Act’ Analysis, on the one hand, and the Dispositional Analysis on the other.

As you may have noticed for yourselves, contemporary philosophers have a certain hostility towards the concept of ‘feeling’ as such; and perhaps this hostility has gone a little too far. For the moment let us leave the matter there, and just say that in the traditional Occurrence Analysis it is usually held that there are degrees of assent, and that we assent more confidently to some propositions and less confidently to others.

‘Being Under an Impression That…’

It will be seen that in the analysis of Assent which I have just been stating, reasonable assent is taken as the typical case. Here assent is preceded by a state of wondering or questioning, where we consider several alternative propositions, together with the evidence for each of them. Our assent, when given, is assent upon evidence, or reasoned assent.

But it happens sometimes, indeed quite frequently, that we accept propositions in a very different way from this. We accept them in an unquestioning manner; we just ‘take it for granted’ that p, without question or doubt, and without any previous process of weighing the evidence pro and con.

To complete this account of the traditional Occurrence Analysis, we must therefore consider a kind of appendix to it. This is concerned with the state of mind which Cook Wilson called ‘Being under an impression that…’.8 (His disciple Prichard called it ‘Thinking without question’.) Cook Wilson's own example, which I shall elaborate a little, is this: I see walking in front of me in the street a man who resembles my friend Tom in build and clothing. He is tall and rather stout, as Tom is; he has a dark blue overcoat and reddish hair, as Tom has; he walks with a slightly rolling gait as Tom does. I quicken my pace and go straight up to him, and then ‘without hesitation perform some act which it would be a liberty to take with anyone but an acquaintance’.9 But then I find that this person is not my friend Tom at all. He is a complete stranger. Covered with confusion, I apologize and say ‘I am so sorry, I thought you were my friend Tom Postlethwaite’.

Now in what sense did I think that the man was my friend Tom? As we have seen, the phrase ‘think that…’ is often used to express the intermediate degree of belief, the degree which is called ‘opinion’. But was this an opinion? Did I weigh the evidence for and against, and then assent with a mild degree of confidence? No, not at all. I just saw the man's back, and ‘jumped to the conclusion’ that he was my friend Tom, without any consideration of alternatives, without any weighing of evidence pro and con, and without any doubt. My visual percept simply recalled my friend Tom to my mind, and caused me to ‘be under the impression’ that the man was my friend Tom. One might also say ‘I took it for granted that the man was Tom’. One might add, perhaps, ‘it never even occurred to me that it might be someone else’. (This is the point of Prichard's term ‘thinking without question.’)

But surely we did have some evidence that the man in the street was our friend—not conclusive evidence by any means, but evidence all the same? Cook Wilson's reply to this, if I interpret him correctly, is that there was indeed evidence available to us (the observed facts about the colour of the man's overcoat, the colour of his hair, his height and girth, his gait) but we did not use it as evidence.10 For us, on this occasion, these percepts do not function as evidence, though for someone else, who is attentive and wide-awake, they might have. For us, in the unquestioning state of mind in which we are, they are merely associative cues. The blue overcoat, the bodily shape, the gait, the reddish hair remind us of our friend Tom, they cause the proposition ‘This is Tom’ to come into our mind, and we then accept this proposition without question or doubt. It never even occurs to us that the proposition might be false.

This state of mind, whatever name we choose for it, is very common. When one sees an approximately transparent liquid in a bath, one nearly always takes for granted that it is water. Conceivably it might be sulphuric acid, but one does not apply litmus paper to make sure. Because it looks like water one ‘takes’ it to be water in a completely unquestioning manner. If you give me something which looks like a pound note, I do not inspect it carefully with a magnifying glass to make sure it is not a forgery. Our ordinary attitude to testimony11 is much the same, for instance when someone tells me ‘It is still raining’ or ‘I met Jane at a tea-party yesterday’. It is only in rather special circumstances that one weighs the evidence about the speaker's veracity, asks oneself ‘Is he likely to be telling the truth?’ and then arrives at the rational opinion that he is telling it. Our usual attitude is to take for granted the veracity of other people's statements, and to accept their testimony ‘without question’ unless their testimony is very improbable on the face of it, or we already have reason to think them very unreliable people.

Moreover, Cook Wilson's example might lead us to suppose that when we are under an impression that p, the proposition p is invariably false (as it was in this example). But as he points out himself, at the end of the chapter, there are numerous instances in which this state of mind ‘does not attract attention, because, though unwarranted, it does not lead to a mistake’.12 When I am under an impression that p, p may very well be true; and very frequently, it is in fact true. In 9 cases out of 10, perhaps, the man I take (in this unquestioning manner) to be an acquaintance really is the man I take him to be. And in 999 cases out of a 1000 when I am ‘under an impression that’ the liquid in the bath is water, it actually is water.

It is however true, and important, that this state of mind may easily lead us into error. And when it does, the error we fall into is something quite different from a mistaken opinion. In opinion, we are aware already that the proposition assented to may be false, and that the evidence for it, though favourable, is not conclusive. So if it turns out later than the proposition is false, we are indeed surprised, but we do not experience the peculiar kind of ‘shock’ which we have when the man in the blue overcoat turns out to be a complete stranger. This is something we were not at all prepared for (as we would have been, if our state of mind had been one of reasonable opinion). It is almost as if the universe itself had ‘let us down’, since it never even occurred to us that the proposition we accepted might turn out to be false.

According to the traditional Occurrence Analysis, there are two factors in assent, preference and confidence. It will be noticed that both are absent in ‘being under an impression that…’. When I assent to a proposition p, I prefer it to its alternatives. Not only p but also other propositions q and r are before my mind, and I decide in favour of p, in preference to q and r. But in the situation we are now discussing there are no alternatives before the mind. In the example about the man in the street, the proposition ‘That is Tom’ is the only one which occurs to me. It ‘never even enters my head’ that the man might be someone else. There cannot be preferring when there are no alternatives (or rather, when we are aware of none), nor can there be consideration of grounds for preferring this alternative to that. As Cook Wilson says ‘there is a certain helplessness’13 about this state of mind.

More surprisingly, perhaps, the second factor in assent, confidence, is lacking as well. What is characteristic of this state of unquestioning acceptance is the absence of doubt or diffidence. And whereas confidence, according to the traditional account of assent, is something which varies in degree, there are no degrees at all in ‘being under an impression that’ as Cook Wilson describes it. The feeling of doubt or diffidence is totally absent.

On this view, reasonable assent, even reasonable but incorrect assent, is something of an achievement. It takes a certain amount of effort to consider alternatives, to weigh the evidence for and against each of them, before giving assent to one of them. And—equally important—it takes a good deal of effort to refrain from assenting when the evidence is conflicting, or when there is too little evidence to justify assent at all.

‘Primitive Credulity’

Can we throw any further light on this state of unquestioning acceptance, which Cook Wilson called ‘being under an impression that…’? We have seen that if the traditional Occurrence Analysis of assent is correct, we do not assent to a proposition when we accept it in this unquestioning way. Then how does this unquestioning acceptance differ from pure and simple entertaining? We spoke about jumping to a conclusion. The sight of the blue overcoat, reddish hair etc. suggests (by association) ‘That is my friend Tom’. Now this does explain why that particular proposition—rather than some other one—comes into my mind at that particular moment. But even so, why should I accept it? Why should I not just entertain it in a neutral way, without either accepting or rejecting it? The jump is explained in this associative way, but not the fact that it is a conclusion which is jumped to. Concluding, even of this not-very-rational kind, is something more than mere entertaining, though in order to conclude that p we do of course have to entertain p.

According to these philosophers—I shall use their own terminology for the moment—the natural tendency of the human mind is to believe any idea which comes before it, unless and until that idea is contradicted directly and obviously by sense-experience; it is a ‘natural’ tendency, in the sense that such a belief comes into existence unless and until there is some special factor to inhibit it. As Bagehot puts it himself ‘My theory is, that in the first instance a child believes everything’ (op. cit., p. 417). All sorts of queer beliefs flourish side by side in the mind of the primitive adult or the civilized child. Any idea which comes into their minds is at once believed, unless it very obviously clashes with perceived facts. Moreover, as Bagehot remarks (p. 419) even the mind of the civilized adult operates in this way in dreams; and also, we might add, when he is under hypnosis.

Thus, according to this doctrine, what needs explanation is not belief but rather the absence of belief—suspense of judgement, the questioning attitude when one weighs evidence and considers alternatives. Belief is the primitive and the ‘natural’ state for which we have a strong and spontaneous inclination. So the existence of this unquestioning acceptance (called by Cook Wilson ‘Being under an impression that…’) ought not to surprise us. It is just a manifestation of Primitive Credulity. What ought to surprise us is the fact that we sometimes manage to avoid it.

What are we to say of this doctrine? Clearly there is some truth in it. But before we try to decide how much, it would be well to alter the terminology which these philosophers use. As you will have noticed, it is much the same as Hume's. The ‘believing’ which is manifested in primitive credulity is not at all like the assent which is analysed in the traditional Occurrence Analysis. It is just the unquestioning acceptance of an idea: and by ‘idea’ they mean what we have been calling ‘a proposition’. (Strictly, perhaps, an ‘idea’ should be a proposition symbolized by mental images; but in their usage it also includes propositions formulated in words.)

Their doctrine, then, in my own terminology, comes to this: the human mind has a spontaneous (unacquired) tendency to accept without question any proposition which is presented to it; and this tendency operates as a matter of course, unless there is something else to hold it in check. The power of suspending judgement, of asking questions and weighing evidence, the power on which reasonable assent depends, is not something we possess from the beginning. It is an achievement, which has to be learned, often painfully. To put it in another way: the attitude of ‘being objective’ about a proposition which comes before one's mind and assenting to it only if the evidence which one has is on balance favourable to it, and then only with that degree of confidence which is warranted by the evidence, and of suspending judgement unless and until these conditions are fulfilled—this attitude is something which ‘goes against the grain’ of our natural tendencies.15 We have to acquire this attitude of being ‘objective’ and impartial, much as we have to acquire the power of controlling our instinctive desires.

No doubt every sane adult does acquire this power of suspending judgement in some degree (indeed, it is one of the essential constituents of sanity); but not necessarily about all subjects, nor at all times. Many a man can suspend judgement and weigh evidence very well about matters which fall within his own personal experience, or about matters on which he is an expert; and yet outside this field he may remain, as we say, ‘childishly credulous’. Moreover, this critical and questioning frame of mind is not only an achievement, it is a somewhat precarious one. Most of us can suspend judgement and weigh evidence when we are healthy and wide-awake, but not so easily when we are tired, or ill, or frightened, or angry. We quite often slip back into the state of primitive credulity under the stress of emotional excitement or fatigue; and as Bagehot says, we all slip back into it when we are dreaming.

I suggest then, that Cook Wilson's ‘Being under an impression that…’ may be regarded as a relapse into this state of primitive credulity. When we jump to a conclusion, a proposition is brought into our minds by association or by some other non-logical process. And if we are not wide-awake and on our guard (as we quite often are not) our spontaneous tendency to accept the proposition has free play; there is nothing to inhibit our primitive credulity, and so we accept the proposition in an unquestioning and undoubting way, without noticing what we are doing (or rather, without noticing what is happening to us).

As has been suggested already, the surprising thing, according to this account of the matter, is that quite often this unthinking acceptance does not occur and reasonable assent, in accordance with the evidence, does occur instead; or suspense of judgement, if the evidence is too weak or too conflicting to justify either assent or dissent.

We might explain this, as Bagehot does, by supposing that our primitive credulity has been to some extent chastened or disciplined by adverse experience. Propositions accepted without evidence are very often falsified, at any rate when they are propositions about easily observable matters of empirical fact, and this painful experience of falsification tends by degrees to inhibit our primitive credulity. If it be asked why falsification should be painful, or why it should matter to us at all whether a proposition we have accepted is false or true, the reply is that however credulous we are, we also have a desire for the truth both as a means and as an end. So as a result of this painful experience of falsification, we learn by degrees to restrict our spontaneous inclination-to-believe, and to adopt the more ‘canny’ policy of believing only those propositions which the available evidence favours. The point of this policy is that these are the propositions which are likely to be true. (That they are the ones which are likely to be true follows analytically from the definition of ‘evidence for’. To say that there is evidence E for a proposition p is to say that since E is the case, p has some likelihood of being true.)

Nevertheless, this painful and salutary lesson is never completely learned by any human creature. The policy of believing only in accordance with the evidence can be maintained only by constant effort. The price of maintaining it is ‘eternal vigilance’, as has been said of freedom. Even the most rational man cannot always keep it up. We all relapse at times into the state of primitive and uncritical credulity in which we accept propositions in an undoubting and unquestioning way: and that is why there is such a state as ‘being under an impression that…’. It must be pointed out that these speculations about primitive credulity are not to be found m Cook Wilson's own writings, nor in those of his disciples. So far as I am aware neither he nor his disciples mention ‘primitive credulity’ at all.

Impressions and Estimates

So far, my own aim has been just to expound this final part of the traditional Occurrence Analysis of belief, this Cook Wilsonian appendix, as it were, on ‘Being under an impression that’ or ‘Thinking without question’, and to state it as plausibly as I could. It cannot be denied that Cook Wilson has here put his finger on something important and interesting, something, too, the existence of which must be acknowledged (in one form or another) whatever analysis of belief we accept, and even if we have to conclude in the end that the whole traditional way of treating belief is mistaken, or inadequate.

Nevertheless, there are some criticisms which are bound to occur to us. We are inclined to say that Cook Wilson has exaggerated the difference between the two states of mind he distinguishes, reasonable belief on the one hand and ‘being under an impression that…’ on the other. And might there not be intermediate cases between the two? To be clear about these questions we need to consider different types of example. His own example, the mistake about the man in the street, is highly relevant; but there are others.

When we make a slip in a piece of deductive reasoning, an arithmetical calculation for instance, our state of mind is entirely different from the one in which we are when we have a reasonable but mistaken opinion, and our error is rightly described as ‘a failure to think’. Here we are just ‘jumping to a conclusion’ in an unreasonable, or rather, a non-reasonable way. (It is not that we are ‘reasoning badly’. The trouble is that at a certain stage in the process we are not reasoning at all.) But in Cook Wilson's own example, about the man whom we mistake for a friend, it is not nearly so clear that our state of mind is completely different from one of reasonable belief, though of course it is not completely reasonable either. There is really no denying that in this case we do in some sense ‘have evidence’ for the proposition we unquestioningly accept, weak though that evidence may be. Moreover, there is no denying that in some sense we accept the proposition ‘because of’ this evidence which we have. What is wrong with us is the unquestioning way in which we treat the evidence. We fail to consider that there might be adverse evidence as well: e.g. the fact that to-day is Tuesday and our friend is nearly always away from Oxford on Tuesdays, or just the fact that two different men often look very much alike from the back.

There are other states of mind which are quite clearly intermediate between Cook Wilson's two extremes. Indeed, we sometimes use Cook Wilson's own word ‘impression’ for describing states of mind which are at any rate partially reasonable.

We sometimes ask another person ‘What is your impression of Mr So and So?’ or ‘What impression have you formed of him?’ For instance, we might be enquiring of the Warden of St Benedict's College about one of his undergraduates, and the reply we receive is ‘My impression is that intellectually he is low Third Class, but in character as sound as a bell’. Sometimes we treat such answers with respect and attach considerable importance to them. The ‘impression’ need not be about a person. It might be about a place (‘What was your impression of Guatemala?’). Or it might be about a whole class of persons or a whole class of material objects. ‘My impression is that Oxford undergraduates work much harder than they did thirty years ago.’ ‘My impression is that most metals conduct electricity.’ On the other hand, when we say, as we sometimes do, ‘My general impression is…’ we might be referring to an individual person or place or object. This is because inductive generalizations can be made about an individual entity, for example, ‘John usually goes to Church on Sundays’ ‘It rains nearly every day in the Lake District’ and these ‘general impressions’ are akin to them or are, as it were, an inexplicit form of them.

Again, let us consider the process which is called ‘estimation’; for instance estimating how far away a certain hill is which you see (‘about 1½ miles’) or estimating on a foggy day that the range of visibility on this part of the road is between 80 and 100 yards. A more painful example is the mark which an examiner gives to a candidate's paper. All these estimates are no doubt fallible. But again, we think it proper to pay some attention to them, and faute de mieux we are willing to rely upon them in our actions and take quite important practical decisions on the strength of them.

We may also consider the unsupported dicta of experts. For instance a classical scholar says ‘This line cannot have been written by Virgil even though it does occur in quite a good manuscript of the Aeneid.’

In all these cases, the interesting point is that these people cannot produce evidence for their impression or their estimate. At any rate they cannot produce nearly enough evidence to justify it, and quite often they cannot produce any evidence at all. Why then do we rely on such impressions and estimates—our own or other people's—to the extent that we do, and why do we think it reasonable to rely on them? Because there is a sense in which such impressions and estimates are quite often based on abundant evidence, although the speaker cannot state the evidence, either to others or to himself, or at best can only state a very little of it. One may have evidence, though one cannot give it.

What has happened is roughly this. The speaker has in fact had a number of experiences (very many, perhaps) which are relevant to the question you ask him. These experiences have accumulated over a considerable period of time, and each of them has had an effect upon his mind, though he cannot recall most of them or even perhaps any of them. It is the cumulative effect of these past experiences which determines the ‘impression’ which he has concerning the person or thing or class of things or persons, about which you ask him, or the ‘estimate’ which he makes. In this respect, his state of mind does resemble reasonable belief. In one sense of the word ‘have’ he really does have a great deal of evidence for the impression or the estimate which he imparts to us.

But in another respect, his state of mind differs from reasonable belief. Although he has this evidence and his impression or estimate is in accordance with it, he cannot recall this evidence or state it (not even to himself) or, at the best, he can only recall a very little of it.

When we ask him ‘What is your impression of x?’ or indeed when he asks himself, he just finds himself inclining to a particular answer, perhaps very decidedly, but he cannot say why. He cannot ‘justify’ his answer or his estimate, or at best can only give a very incomplete justification of it. He cannot give the evidence which (in a way) he has. Sometimes, moreover, this state of mind comes even closer to reasonable belief. The impression or estimate need not be an absolutely unquestioning one. The man may consider several alternative answers, when we ask him for his impression of so and so or his estimate of such and such. ‘Let me think’, he says, or ‘Give me a little time to think’. What sort of thinking does he do? Rather an odd sort. He considers various alternative answers, dwells upon each of them for a time, and then selects the one which ‘feels right’. What he is really doing is to allow time for his relevant past experiences to work upon him. For he does ‘retain’ them, even though he cannot recall them individually, and they can still have effects upon his present consciousness.

Consequently his answer, when he gives it, does have the preferential character which belongs to reasonable assent. We cannot quite say that he decides for p rather than for q. ‘Decide’ is too active a word. It would be nearer the truth to say that p recommends itself to him more than q does. Furthermore, in this type of case (where the man says ‘Give me a little time to think’) the other feature which reasonable assent is supposed to have, namely a feeling of confidence which varies in degree, is often present also. The impression or estimate need not be absolutely undoubting, as it is in Cook Wilson's ‘being under an impression that’. The speaker may say ‘My strong impression is that…’. Or he may say ‘That is my impression, but of course it is no more than an impression’. In the same way, an estimate (e.g. about the range of visibility on a foggy day) may be more confident or less. It may be made with very little confidence indeed, thereby resembling the lowest degree of reasonable assent. ‘The best estimate I can make is that it is about 80 yards, but I may easily be wrong.’

I hope it has now been shown that there are intermediate cases between Cook Wilson's two extremes, reasonable belief or opinion on the one side, and ‘being under an impression that…’ on the other. Cook Wilson has made a hard and fast distinction, a distinction of black and white, where in fact there are many shades of grey in between.

But still, we should not have been able to see that there are these intermediate cases, or to appreciate their interest and importance, if he had not drawn our attention to the great difference there is between the two extremes. His doctrine of ‘being under an impression that’ was therefore a very notable addition to the traditional Occurrence Analysis of belief.

• 1.

  It is this aspect of the situation which Hume mentions, too briefly, in the Appendix to the Treatise, pp. 625–6 (Selby-Bigge's edition).

• 2.

  See Lecture 6 above.

• 3.
  Statement and Inference, Vol. 1, p. 102.
• 4.
  Literary Studies, Vol. 2, pp. 413–21, ‘The Emotion of Conviction’. See also below, pp. 212–216.
• 5.
  Treatise, Book I, Part III, Section 12.
• 6.
  Probability and Induction, pp. 13–18.
• 7.

  See Series II, Lecture 2—where Kneale's criticisms are discussed. (pp. 282–5).

• 8.
  Statement and Inference, Vol. I, Part II, Ch. 3, pp. 109–113.
• 9.
  Loc. cit., p. 109. According to an Oxford tradition, Cook Wilson slapped the man on the back.
• 10.
  Statement and Inference, Vol. 1, p. 113. ‘The reason in such a case is not a conscious reason for us, in the sense of a premiss from which we infer.’
• 11.

  See Lecture 5, above.

• 12.
  Statement and Inference, Vol. 1, p. 113 fin.
• 13.
  Statement and Inference, Vol. 1, p. 112.
• 14.

  It is most unfortunate that this brilliant little essay is omitted from the Everyman edition of Literary Studies.

• 15.

  Most of us rather dislike ‘sitting on the fence’ in a state of suspended judgement, or even disapprove of it. Lloyd George is supposed to have said of a political colleague ‘He has sat on the fence so long that the iron has entered even into his soul’.