CHAPTER 1
TWO KINDS OF POSSIBILITY
I first consider two pairs of propositions: their implications, and the kind of observation and argument by which their truth or their falsity would normally be established, if they were challenged.
The first pair
(a1) 'It will not happen now'
(a2) 'It cannot happen now'
The second pair
(b1) 'He will not do it now'
(b2) 'He cannot do it now'
To bring out the contrast between the pairs, one might choose more specific propositions as examples of the two types:
(A1) The gas will not escape now'
(A2) The gas cannot escape now'
(B1) 'Jones will not escape now'
(B2) 'Jones cannot escape now'
I am concerned with the difference between "will not' and the 'cannot' in the two cases: I shall argue that it is a different difference.
Consider the first pair first: to make the transition from 'It will not happen now' to 'It cannot happen now,' from the weaker to the stronger statement, is, at the very least, to claim that a certain type of evidence, or reason for believing, or source of knowledge, that the thing will not happen is available: not only will it not happen, but it is impossible that it should. Someone might agree that the thing won't happen, but he might at the same time dispute the stronger statement that it cannot happen, that it is impossible that it should. 'I agree with you,' he says, 'that in fact it will not happen now; but, all the same, it might; it is not altogether impossible. Your statement is too strong.'
We see that the contradictory of 'It cannot happen here and now' is not only 'It can, or could, happen here and now,' but also 'It may, or it might, happen here.' These are alternative ways of representing the possibility of its happening, when someone has asserted the impossibility. When the two disputants agree about the future course of events — that in fact the thing will not happen — they may still disagree about the possibility of its happening. If it is certain that it will not happen, there is no possibility that it will.
I do not myself know of any entirely adequate account of the logic of statements of the form 'It is certain that so-and-so will not happen.' But at least it is clear that he who asserts that it is certain that it will not happen here, or alternatively asserts that it cannot happen here, must show that there are very strong grounds for believing, or that there is some sure source of knowledge, that it will not happen here, when the truth of his statement is challenged. A man may say 'It will not happen here,' (e.g. This horse will not win the race'), and his statement may be true, and the utterance may be an entirely justifiable utterance in the circumstances, even though he was not in a position to support the statement with good grounds, or good evidence, or by citing a source of knowledge. But 'It cannot happen here,' which denies that it might happen, asserts, at the very least, that there are very good reasons for believing that it won't. Perhaps it is sometimes, or even generally, used to assert, or imply, much more than this: e.g. that there is some well-established natural law covering this happening; perhaps it asserts, or implies, that there is some cause, or sufficient condition, of its not happening. However this may be, at least it asserts that there are grounds for believing that it will not happen, which are strong enough to justify the assertion that it cannot happen. If no such grounds exist, the statement 'It can't happen here' is unsustainable and incorrect, even if the statement 'It won't happen here' is acceptable; for an objector could say 'Well, I admit that it didn't happen: but still it might have happened, and you were quite wrong when you said that it couldn't happen.'
We are familiar with the series: (1) 'It might happen here,' which might also be expressed as 'It could happen here;' (2) 'It is unlikely to happen here,' but 'It might' or 'It could,' and (3) 'It is certain that it won't happen here,' which might be expressed as 'It can't.'
Turn now to the second pair of the original propositions: either to the more general 'He won't do it now' and 'He can't do it now': or to the more specific 'He won't escape' and 'He can't escape.' The transition from 'He won't' to 'He can't' may here be altogether different; for when we say that he can't do it, we are not ordinarily saying that it can't be the case that he will do it. Given an appropriate verb of action, we would not ordinarily be denying that he might do it, that there is a possibility that he will. We are rather saying that he is not able to do it, that he lacks the means, or the authority, or the opportunity, to do it. The transition from 'He won't escape now' to 'He can't escape now' would not ordinarily be a transition from a weaker to a stronger statement about a future happening — about a possible eventuality, his escape. The conjunction of 'can' with a verb representing voluntary action would ordinarily preclude this interpretation; only in the past tense 'He can't have escaped' would ordinarily have as its contradictory 'He might have escaped,' where this is equivalent to 'It is not impossible that he did.' 'He could have escaped,' is ambiguous when taken out of context. It might be intended to be equivalent to 'It could be the case that he escaped', and 'It is not impossible that he did,' and 'He might have escaped;' or it might be intended to be equivalent to 'He had the ability, the means, and the opportunity, to escape.'
'He can't escape' may be informative as giving a reason or explanation why in fact he won't. He won't escape, because he can't. It is informative as a reason or explanation, because it may be the case that he would escape if he could, but he can't. Contrast this with the other pair: of the gas it is not similarly informative to explain the fact that it won't escape now by saying that it can't. It is not clear what it would mean to say of a gas that it would escape if it could. It is just conceivable that some non-anthropomorphic, literal sense might be given to this; for example, that it is a gas which has a tendency to escape from any container, but that now its escape is obstructed. But 'It would, if it could, but it can't' is not a literal and natural, and even less a scientific way, of speaking of the behaviour of a gas; more natural would be 'It would escape if it wasn't for the lead in the container, which makes it impossible,' Here 'It's impossible' does not actually give the reason why it won't escape; it indicates only that there is a specific cause of its not escaping, in spite of its tendency to escape. If I am asked to give you a reason for believing that it won't escape, I may explain to you why it can't. But I cannot quote its inability to escape as a reason for its not escaping. On the other hand, 'He would if he could, but he can't' is a very natural, and informative, way of speaking of the behaviour of a person. We now know in very general terms why he won't escape: he lacks the ability, or the means, or the opportunity to escape.
'He can't do it' (the governor is speaking about the prospect of his escape from prison) gives one of two possible reasons or explanations of why he won't. The other is that he does not want to do it, or is unwilling to do it; for, if he doesn't do it, it is generally either vouloir or pouvoir, the will or the power, that is lacking. If both were present, and present in full measure and without qualification, that is, if no other desire were more urgent, he generally would do it. I add the qualification 'generally,' because there is a third possible explanation, which may perhaps not be accounted for as a special case of having a conflicting desire: the explanation given in the words 'because he thinks he ought not to escape.' 'Because he doesn't want to,' or 'because he wants something else more,' explains his not doing it, in the same way that 'He can't do it' explains. It excludes one of two obvious possibilities, and therefore it tells one more about the situation. To the gas, and to inanimate things generally, the concept of wanting neither has, nor can be given, any application. Consequently, The gas cannot escape now,' taken by itself, is no kind of explanation of its not escaping. For it does not exclude one of these two possibilities.
The statement 'He cannot escape now' of course has as its contradiction 'He can.' 'He can escape' does not entail 'He might escape.' For it might be the case that, although he is able to escape, there is no possibility that he will want to. Nor does 'He might escape' entail 'He can escape.' 'He might escape' only entails 'He might be able to escape,' but not the unmodified statement that he can. Of course 'He cannot' does entail 'He will not,' but it does not entail 'It is certain that he won't,' or There is no possibility that he will,' both of which are contradicted by 'He might.' He cannot escape now, at this moment' is an unmodified, categorical statement about him, on a level with the categorical statement 'He is unwilling to.' He lacks the power now, at this moment. The question may be asked 'Is it certain that he cannot?' or 'Is it possible (might it be) that he can?' There is nothing odd in the question 'Can it be (i.e., is it possible) that he cannot escape?' That the two kinds of 'can' should occur in the same sentence may be disagreeable in style; but the sense is not unclear. It may be possible or impossible that he can escape, just as it may be possible or impossible that he actually will. It may be certain that he lacks die ability; and we may be in a position to point to some necessary condition of his having the ability which is lacking.
Just as we might in some cases seek for scientific explanation of a man's doing something, we look also for causal explanations of his power, or of his lacking the power, to do something. 'Why was I unable at that moment to do so-and[so?' is typically a causal question; my powers to do certain things come and go, are present at one moment and not at another, and I have every motive for discovering, if I can, the conditions on which their presence and their absence depend. Sometimes, of course — as, typically, in the case of escaping — the answer to the question 'Why can't he?' may be entirely obvious, and no experiment, or careful causal investigation, may be needed. It is often obvious that it is impossible, and out of the question, that a man should be able to do something, just as it is often obviously impossible, and out of the question, that he will want to do something. We may always look for sufficient conditions of the existence, or non-existence, of a power to do certain things. We need to acquire, and to hold on to, abilities, and we therefore need to know the conditions on which their existence depends.
There are some things which I can do at one moment and cannot do at the next; in this sense, powers can be very episodic, and impotence the matter of a moment. 'Why can't I do it now? I could do it a moment ago.' At one moment I can remember something that I want to remember: at the next moment I cannot. Evidently, there are also powers which, once acquired, last all one's life, and there are others that are intermediate between these two extremes. There are powers that are in their nature intermittent, and also powers that are not specific — e.g. the ability to speak French: just as there are desires and inclinations and interests that are in their nature intermittent, and also those that are not specific. I wish to concentrate on the power to do a specific thing on a particular occasion, because this is the fundamental kind of power, and of potentiality, which philosophers have tended to neglect, particularly in the context of an individual's freedom of decision.
You may now ask — What is a power, as this is attributed to human beings? How is it identified, and how is its existence or non-existence recognized? Is there a sharp line between powers to do something, attributed to human beings, and the causal properties of physical things, e.g. machines? I think there is: perhaps not a sharp line, but a line.
I have so far only said that the concept of a power to do some specific thing is complementary to the notion of a will, or of wanting to do something: associated with the concept of an action, the vouloir and pouvoir, 'want' and 'can', depend on each other for their sense. Whether someone lacks the power to do something is tested in actual performance, but only subject to the condition that the subject had the will to do the thing in question: whether someone lacks the will to do something may sometimes be revealed (in part, and only in part, and the lack is not conclusively established) in actual performance, but only subject to the condition that the power to do the thing in question exists. To speak of a man wanting to do a particular thing on a particular occasion (e.g. having the urge to speak, or to protest) is not to speak of a tendency, in that sense of 'tendency' in which a gas may have a tendency to escape, as some philosophers have suggested. A human desire or inclination to do something is not only, or primarily, revealed in actual performance, as is the tendency of a gas. I normally discover, decide, and know what I want to do, quite independently of observing my own patterns of behaviour; and I may authoritatively, though not infallibly, disclose my desires and interests to others, together with the conceptions and calculations that enter into them. It may be true that I did have an impulse to protest, even though no evidence of this is to be found in my behaviour, and even though no test would establish that I did, or did not, have at that moment that impulse; for to ask me about my impulse is not to apply a test, in the sense in which you may apply a test to establish that this particular gas has a tendency to escape at this particular time. For these reasons — or connected with these facts — a statement to the effect that someone — say, the speaker — wants, or wanted, to do something at a particular moment is not equivalent to any set of hypothetical statements about his observable behaviour. A desire to act may occur at a particular moment, and may be no less an episode in a man's biography, to be reported in a categorical statement, than a twinge of pain, or a blow on the head. A desire, or inclination, to do something at a particular moment does not need to be tested in performance — or to be tested at all — unlike a power to do something, which does in the last resort need to be tested in performance, if its existence is to be established beyond doubt.
Compare (a) 'At that moment I wanted to race him and overtake him but, of course, I didn't;' (b) 'At that moment I could have overtaken him, but, of course, I didn't.'
The first statement may in normal circumstances be known to be true, beyond doubt, and quite independently of the inductive test of parallel cases; in such a simple case the first person disclosure is authoritative, although not immune from various kinds of error, from self-deceit, or deceit. The question 'How do you know that you wanted to?', as opposed to 'Are you sure?', is normally senseless. In the second case, the inductive test of parallel cases is the relevant test, and the first person statement is not authoritative. The question 'How do you know that you could have' is far from senseless. Whether I have, or had, the power to overtake, as I believe, only experience and experiment can show. Whether I have, or had, the desire is not normally, and in the standard case, to be finally settled by experience and experiment. In the normal case the subject, although liable to correction, can claim to know directly what he wants to do. There can be no question of the source of his knowledge; it is direct knowledge, in the sense that it has no source.
The notion of a power, as expressed in the form 'I can do so-and-so,' 'He can do so-and-so,' is a very wide and general one. It includes the kind of power to do something which is an ability or skill, the power which is being in a position to do something, either in a literal or a metaphorical sense, a legal or customary power to do something (e.g. to marry people, if I am a priest), and many variations and combinations of these. Whenever, in respect of any action, the contrast between the attempt to do and the achievement is in place, the notion of the power to do it is also in place. Plainly the question 'Can you do it?' will sometimes be an absurd question; either the verb of action in itself, or the circumstances of the particular case, may show that the question does not arise. But, in general, it is intrinsic to the notion of an action that it is something that can be attempted, and attempted, successfully, that it is something that one may try to do, and that one may either succeed in doing or fail.
