In his recent ‘Pelican’ book entitled Ethics Mr. Nowell-Smith has subjected to courteous but very extensive criticisms the positive doctrine about free will which I briefly sketched at the close of a paper in Mind (October 1951) and which I set out more fully in a book written some twenty-five years ago (Scepticism and Construction). As in almost every instance these criticisms seem to me to arise directly or indirectly from a refusal to attach weight to the evidence of introspection—which to myself is not merely indispensable but central—I have not felt obliged by them to introduce into the version of the present work any modification of my general position. This by no means implies however that I regard Mr. Nowell-Smith's criticisms as unimportant. They seem to me an excellent example of how the problem of free will appears—I might almost say must appear—to an acute analyst whose ethical thinking is determined throughout by the linguistic approach. If as I much hope those who do me the honour of reading this book include some philosophers partial to the linguistic approach the detailed reply to Mr. Nowell-Smith which I now offer is likely to deal with a number of points to which they also feel objection.
It will be simplest if I take up Mr. Nowell-Smith's criticisms more or less in the order of their occurrence in his text.
1. Having previously argued that the expression ‘he could have acted otherwise’ must ordinarily in non-moral contexts be interpreted not categorically but hypothetically Nowell-Smith begins his section on Libertarianism by observing that ‘it would indeed be remarkable if modal forms which are normally used in a hypothetical way were used categorically in one type of case alone’ (p. 278).
I am afraid I do not see anything ‘remarkable’ about this. Where two or more different types of situation have important features in common we often use the same verbal expression to refer to them even if the difference also is in some contexts important provided that in the context of our discourse this difference is not one to which we have any particular interest in drawing attention. Now in the context of ordinary discourse about human behaviour we seldom have any particular interest in drawing attention to differences that are of importance to the moral philosopher. It would not be in the least surprising therefore if in ordinary discourse we used the identical expression ‘he could have acted otherwise’ to refer to situations which had not only an obvious identity but had also philosophically important differences. Whether in fact such differences exist is a matter that can be settled only by an ethically directed analysis and comparison of the situations to which the (identical) verbal expressions relate.
2. Nowell-Smith proceeds:
‘It is essential to notice that the categorical interpretation is supposed to be necessary only in a very small but very important part of the whole range of human choice. And this too is remarkable; for it implies that the words “free” and “choose” are logically different in moral and in non-moral cases.’ (pp. 278–9).
I find this even less ‘remarkable’. If there is one word of philosophical significance in the vocabulary of the English language that can beat ‘feeling’ in the diversity of its usages it is the word ‘free’. That is why it is of such paramount importance for any discussion of the problem of ‘free will’ to begin by carefully distinguishing the precise nature of the ‘freedom’ whose existence or non-existence is in dispute. To anyone who appreciates the distinctive character of our usage of the word ‘free’ when we are talking about freedom as a condition of moral responsibility it will certainly not come as a shock to learn that the ‘logic’ of the word ‘free’ is different in moral and in non-moral cases.
3. ‘Campbell insists that the question whether choice is “free” in a contra-causal sense must be settled by introspection’ (p. 280).
This does not quite represent my position. What I have insisted upon is that introspection provides highly important evidence for the settlement of the question not that it can per se settle it. I have always explicitly recognised it to be in principle possible that the subjective assurance of contra-causal freedom which in my view introspection reports may be illusory (e.g. on page 463 of the Mind article) and that various objections to accepting that assurance as veridical must be independently considered. Metaphysics science epistemology ethics and religion have all I think something relevant to contribute to a ‘settlement’.
4. The above misunderstanding has I think a connection with Nowell-Smith's doubts about the ‘propriety of the language’ I use in reporting the results of my own introspection. Contrary to what he apparently supposes to be my attitude I thoroughly agree with him when he says ‘That I know introspectively what it is like to choose may be true; but I cannot be said to know introspectively that my choice was contra-causal or unpredictable; and this is the point at issue’ (p. 281). I think he has failed to appreciate that what I report as discerned in introspection is (to take the crucial case of moral decision between rising to duty and yielding to strongest desire) not a contra-causal activity but a belief—seemingly ineradicable—that one is contra-causally active. That is to say it is a report of ‘what it is like’ to make a moral decision. My introspective report may itself be at fault (I must refer to the main text for the justification of its detail) but it is surely of the kind of thing which introspection is competent to discern?
Why do I regard the introspective report as offering ‘highly important evidence’? Briefly for this reason. We are trying to ascertain the true nature of the activity of ‘moral decision’ (in the sense of that expression just specified—there is of course a very different sense of it to which reference will be made later). And not until we have satisfied ourselves about that so far as I can see can we possibly hope to determine the kind of language which is appropriate in talking about moral decision. Now just because moral decision is an activity it can be directly apprehended only ‘from the inside’ by the agent actually engaged in it. But if that is so the introspective report which tells us how the agent as actually engaged in moral decision understands his activity is surely very good prima facie evidence about the nature of this activity and in consequence about the form of words best fitted to express it?
5. ‘The issue between determinists and libertarians is an issue about the way in which expressions such as “choose” “can” and “alternative possibilities” are to be construed; and this is surely an issue which is to be settled not by self-observation but by logical analysis’ (p. 281).
I take it that the question how these expressions ‘are to be construed’ is equivalent to the question ‘In what sense or senses of these expressions do they represent truly the situations to which they refer?’ Otherwise the issue between the determinists and Libertarians would seem to be being regarded as a purely verbal one; and I do not think Nowell-Smith wishes to go so far as that. He does apparently believe however that something called ‘logical analysis’ of these expressions can settle the issue.
If we are to consider this claim we must try to get clear at the outset just what the co-called ‘logical analysis’ is. What is meant by the ‘logical analysis’ of an expression? I suppose what must be meant is the examination of an expression with a view to eliciting its entailments compatibilities and incompatibilities vis ô vis other expressions. It seems patent however that the entailments etc. of expressions as distinct from those of concepts and propositions are not strictly of a ‘logical’ character. The words ‘can’ ‘choose’ ‘alternative possibilities’ merely as words have no logical relations whatsoever. If we are to speak at all of the ‘logic’ of these expressions (or of any other expressions) we can do so only on the assumption—which is certainly valid within limits—that there exist certain rules for the use of these expressions which rules however are not themselves logical rules. We can then legitimately enquire what ‘logically’ follows from these rules with regard to any of these expressions in the way of entailments compatibilities and incompatibilities vis ô vis other expressions. But it is crucial to remember—what is apt to be disguised by the simple use of the term ‘logical analysis’—that the rules for the use of these expressions have not themselves the binding force of ‘logic’.
Bearing this in mind let us then consider how such ‘logical analysis’ works in the kind of case that is now before us; and let us take as our example the expression ‘alternative possibilities’. Is there a ‘rule’ that this expression never entails and perhaps is always incompatible with expressions to the effect that when one of the possibilities has become fact the other could not have happened in the categorical sense of ‘could’? I am not aware of one. And if I am told that nevertheless it exists it is a fair question to ask ‘Who imposed it and what is his authority for doing so?’ It is obviously no use appealing to the authority of the dictionary. This does indeed give us certain general rules (which have a valid claim to our respect) for the use of the expression but it very properly says nothing one way or the other on the special point before us. Nor can it be argued that ‘custom’ prescribes the rule. Apart from the very dubious nature of the ‘authority’ of custom there is no such custom. The custom of libertarians is to use the expression in the context of moral decision as entailing the categorical statement that whichever of the ‘alternative possibilities’ is decided upon the decision could have been for the other. The custom of determinists is to use it in a sense which denies this. So by this route we reach an impasse. And so far as I can see there is only one way out. If we want to know how the expression ought to be used we must try to discover which use of it best conforms to the facts to which it refers—the special realm of fact about which the dispute here arises being that of ‘moral decision’. That as I have already urged entails among other things our studying the actual experience of making a moral decision which is the primary source of the libertarian's claim that his use of the expression is the proper one.
In short ‘logical analysis of expressions’ just will not ‘deliver the goods’. Sooner or later we must go to the facts; and among the relevant facts are those supplied by introspection.
6. ‘The libertarian regards explanation in terms of character as incompatible with genuine freedom and must therefore draw a contrast between “the self” and “the character.” But if “self-determined” is to mean “determined by the self” it is necessary to give some account of what the “self” is…’ (p. 283).
I agree; though I am not sure that anything much less than the extensive treatment I give to this problem in the present work would really have been useful. It is one of the more intractable difficulties confronting the writer on free will who aims at being constructive as well as critical that a diversity of basic problems will be relevant to his theory and he manifestly cannot defend and often cannot even expound his views on these without writing a book. I try to make some amends in the present work; and I would call attention especially to Lecture V in which I discuss the relation of self-identity to personal identity (which Nowell-Smith I think mistakenly treats as indistinguishable). In view of the complexity of that discussion I would perhaps be better advised to say no more here in answer to Nowell-Smith's very proper demand. But I shall venture nevertheless to add a few words.
I argue in the lecture referred to that a man's self-identity can be retained even where there is loss or suspension of personal identity—which latter is for me tantamount to identity of character of the self's dominant cognitive conative and emotive dispositions. And I accept in principle the consequence that a man may be properly liable to moral praise and blame for his past acts even if these have been committed at a time when through injury or disease he was not the same ‘person’ as he was before and perhaps became again after. But I must stress the words ‘in principle’. It may well be the case that the abrupt change induced in the man's personal identity or character was of such a kind that in the given situation no conflict of ‘duty’ with ‘strongest desire’ was present to his mind—as it would have been in his ‘normal’ state. In that event a necessary condition for the ascription of moral praise or blame in respect of the act will have been absent. Or again the change in personal identity might have been such that the conflict of desire with duty was a great deal more acute the desires opposed to duty having become much stronger with the result that a correspondingly greater effort is required of the man if he is to ‘rise to duty’. In that event though moral praise or blame will undoubtedly be in order appropriate allowance will fall to be made in passing moral judgment. It will be manifestly improper to condemn the man for succumbing to temptation as severely as one would if he had retained his normal character and been faced with the need of a less difficult moral effort.
It goes without saying however that in these cases of abrupt and drastic changes in personal identity wrought through injury or disease the onlooker is seldom in possession of the relevant information in a sufficiently accurate and detailed form to warrant a confident judgment on the man's moral merits and demerits in respect of particular actions. But there is nothing new about that. The practical obstacles in the way of passing confident moral judgment on others are formidable even in normal cases though not always adequately appreciated. Nor has this fact any tendency to show that the criterion being used for passing judgment is unsound. A criterion is not theoretically invalidated because it proves difficult or even impossible of practical application.
7. ‘If it is necessary to decide whether or not a man could have acted otherwise before ascribing responsibility it is necessary that we should have some criterion for deciding this; and on the libertarian theory such a criterion is quite impossible. For let us suppose that we know a great deal about his character and also that the temptation which he faced seems to be a fairly easy one for such a man to overcome. On the libertarian hypothesis this information will not be sufficient to enable us to conclude that he could have acted otherwise…’ (pp. 283–4).
It is possible that I have not fully understood the argument of this paragraph; for as I at any rate interpret the libertarian view nothing that Nowell-Smith says has any tendency to show that the question of a criterion presents any special difficulty for it. The libertarian view (based primarily upon the introspective report of the moral agent engaged in the act of moral decision) is that a man in the situation Nowell-Smith depicts could always have acted otherwise; though of course the effort required to rise to duty will be harder in proportion to the strength of the temptation and corresponding allowances will have to be made in passing moral judgment on the man—a practice which accords perfectly with our more carefully pondered moral assessments of men in ordinary life. There are indeed a few exceptional cases like that of the drug addict or the victim of some pathological obsession in which it is very difficult to say whether it was even possible for the temptation to be resisted. That difficult problem does not arise however in relation to the example Nowell-Smith has chosen and I do not propose to initiate a discussion of it here. I shall only say that with diffidence I incline to doubt whether even in pathological cases (always provided that the idea of a duty opposed to inclination is present to the man's mind at all) we ought to say more than that the effort required to vanquish desire is so extreme that it is virtually impossible for the man to rise to duty; or more accurately to rise to his full duty—the dipsomaniac for whom it is virtually impossible to desist altogether from alcohol within his reach may yet by a not too prodigious effort succeed in limiting his indulgence to a level far below that which would satisfy his inclinations. But whatever be the correct solution of this particular problem it does not seem to me to have any vital bearing upon the fundamental issue between libertarian and determinist.
8. ‘The libertarian theory involves putting a very special construction on the principle that “ought” implies “can” which it is very doubtful whether it can bear. If we take this principle in a common-sense way it is undoubtedly true. It is no longer my duty to keep a promise if I literally cannot do so. But when we say this we have in mind such possibilities as my being detained by the police or having a railway accident or the death of the promisee; and it is possible to discover empirically whether any of these exonerating conditions obtained. But if “cannot” be construed in such a way that it covers my being too dishonest a person or not making the necessary effort it is no longer obvious that “ought” implies “can.” These reasons for failure so far from exonerating are just what make a man culpable’ (p. 284).
To me it seems about as obvious as anything can be that ‘ought’ implies ‘can’ in a sense of ‘can’ which goes far beyond the absence of external obstacles to the fulfilment of the duty which is apparently all that Nowell-Smith (judging from his illustrations) is willing to allow. I do not even believe that any very adroit maieutic would be needed to deliver this truth from ‘common-sense’. For if on any grounds whatsoever—scientific psychological metaphysical or religious—one has come to believe that moral choices are all determined in the sense that there only appear to be and never really are genuinely open possibilities before a man so that he cannot act otherwise than he does then it seems every bit as evident nonsense to say that he morally ‘ought to act otherwise’ as it would be if there were external obstacles such as Nowell-Smith cites preventing his so acting. And of course the ‘cannot’ here does cover ‘being too dishonest a person’ since it is implied (in the determinist belief) that he is too dishonest a person because ‘that's the way he's made’; and it covers similarly his ‘not making the necessary effort’ since it is implied that he couldn't make the necessary effort. These ‘reasons for failure’ unlike those mentioned earlier are indeed ‘just what make a man culpable’; but only on the assumption that ‘reasons for failure’ here means ‘ways of failing’ and not ‘conditions necessitating failure’. Surely we must all agree that a man cannot be held morally culpable for that which he can't help doing? It matters not a jot ethically speaking whether what prevents him acting otherwise is the strong arm of the law or the metaphysical structure of the universe.
9. Nowell-Smith turns next to my mind a little belatedly to The Concept of ‘Trying’ in order to ask whether the libertarian can find in this concept a criterion of moral culpability and rather surprisingly in words which if I may say so I might have written myself he tells us that
‘Morally we blame people not for failing to live up to a certain standard but for not trying hard enough to do so; and this is because while we do not believe that they could always succeed we do believe that they could always try’ (p. 285).
So far so good. But when Nowell-Smith attempts to analyse the experience of trying of exerting moral effort he seems to me to say very odd things:
‘We all know what it feels like to make an effort. These feelings are phenomena or occurrences that we experience in the same way that we experience aches pains qualms and twinges. And if we take the introspective language of the libertarian seriously it would seem that the question “Did he try?” can be answered only by the man himself and that he answers it by observing whether or not one of these feelings occurred. The logical status of the question will be like that of “Did it hurt?” But on this view an effort is not something that a man makes; it is something that happens to (or inside) him; and it would be highly implausible to make the question of his responsibility turn on the occurrence or non-occurrence of such a feeling’ (p. 285).
But surely it will not do thus to assimilate the ‘feeling’ of moral effort to the ‘feeling’ of aches and pains and to conclude that since the latter are ‘passive’ so too must be the former? It is true that moral effort like any other kind of activity can be apprehended only in direct experience; and in that sense it is ‘felt’ as pains are ‘felt’. But there the resemblance ends. It clearly does not make sense to say that what we directly experience in directly experiencing activity is ‘something that happens to (or inside) us’. When a man asks ‘Did I try?’ he is asking about the occurrence or non-occurrence of certain ‘feelings’ only in the sense that he is asking whether or not he had the experience of exerting moral effort. There may indeed be certain ‘passive’ feelings that accompany the experience of exerting moral effort—cephalic tension and the like. But it is certainly not about the occurrence or non-occurrence of these that a man is asking when he asks in the context of the moral situation ‘Did I try?’ (It is not easy to make one's self as clear as one would wish on these matters in brief compass and perhaps I may be permitted to refer the reader for a fuller discussion not only to Lecture VIII of the present work but also to my paper—which may I think have escaped Nowell-Smith's notice—in the Proceedings of the Aristotelian Society for 1939–1940 entitled The Psychology of Effort of Will.)
Very rightly Nowell-Smith goes on to insist (though as we have seen he strangely thinks it to be inconsistent with the introspective language of libertarian psychology) that ‘if making an effort is to be relevant to responsibility it must be thought of as something which a man chooses to do or not to do’. That is to say in my language that in the situation of moral temptation the making of the effort (or the withholding of it) must be thought of as the outcome of the man's moral decision. But so understood he declares ‘unfortunately it is fatal to the categorical interpretation of “he could have acted otherwise” (p. 286).
Why is it fatal? If I understand him correctly the argument he offers is as follows. On the libertarian view a man who fails to act rightly (to ‘rise to duty’) is culpable if and only if he ‘could have tried harder than he did’. Now of ‘his failure to try harder’ we are told there are only three possible interpretations. Either his failure is inexplicable; or it is due to circumstances beyond his control; or it is due to ‘his not having tried to try as hard as he could have tried to try’. On the first two of these interpretations the man is not blameworthy but blameless so they won't do. The libertarian is thus forced back upon the third. But it can easily be shown that this won't do either since it leads to an infinite regress. It follows that we cannot intelligibly use ‘the failure to try harder’ as a criterion of culpability.
Now it seems to me that Nowell-Smith's criticism here rests entirely on an assumption which the libertarian would not tolerate for a moment. This is the assumption that we cannot intelligibly hold a man blameworthy for an act that is ‘inexplicable’. The assumption betokens I think a failure to appreciate that if the act of moral decision is a free act it is bound to be ‘inexplicable’ but that this kind of inexplicability by no means entails that the agent is not responsible for the act. That depends on the nature of the activity itself; and once again our primary clue to that must be as in all activity what it is like for the agent who directly experiences it. What in my view it is like for the agent actually engaged in moral decision I have perhaps too often already described. For him his decision to exert or withhold the moral effort required to rise to duty is certainly ‘inexplicable’ in the sense that the decision is a creative act which nothing determines save his doing of it. But because this ‘inexplicability’ is just the inexplicability inseparable from any creative activity it has no tendency to induce the man to disclaim responsibility for his decision and to deny his blameworthiness if the decision was to withhold the effort required and yield to the importunings of desire. To demand an ‘explanation’ of moral decision in terms of factors external to it is simply to assume that determinism is true. The libertarian who knows his business will reject any such demand as involving a plain petitio principit.
10. None of the criticisms of myself which I have so far noticed occasion me much surprise. By and large they seem to me the natural outcome of a standpoint which deliberately precludes consideration of the only direct evidence there can be of moral-decision activity viz. the evidence of the moral agent's own direct experience of it. Yet I must confess to being not altogether prepared for the final charge which Nowell-Smith brings against me. Complaining that I take as ‘the only case of moral choice to which appraisals are relevant that of a man who knows what he ought to do but is tempted to do something else’ he proceeds as follows:
‘Now this so far from being the only case is not even the commonest or most important. For in the great majority of cases of moral difficulty what is difficult is not to decide to do what one knows he ought to do but to decide what one ought to do.… Men who belong to a generation for whom the questioning of accepted principles has been no mere academic exercise and who have found themselves faced with momentous choices in situations not covered by their traditional rules will be less likely than their fathers were to suppose that the only sort of moral difficulty is that of resisting temptation’ (pp. 288–9).
Now this does rather stagger me on two distinct counts.
(a) I am indeed sorry if Nowell-Smith supposes me ignorant of the commonplace that there is often very great difficulty in deciding what one ought to do. But I must protest that there is nothing in my writings on ethics which could justify him in such a delusion. I presume there must have been a time when I enjoyed this state of innocence; but it was certainly long before I began to write—nay even before I began to read—philosophy. Moreover I do assure Nowell-Smith that my early awareness of the commonplace was no symptom of a singular precocity. Nowell-Smith in this respect far from alone among philosophers of his school seems to have curious ideas about the state of man prior to the Wittgensteinian revolution (or should I say ‘revelation’?). It is hard to think of any time in the history of civilisation at which reflective individuals have not ‘found themselves faced with momentous choices in situations not covered by their traditional rules’ and have not been led thereby to appreciate that there are other kinds of ‘moral difficulty’ besides the difficulty of resisting temptation. So far as concerns the comparatively recent generation which Nowell-Smith has specially in mind the suggestion that for them ‘the questioning of accepted principles’ was a ‘mere academic exercise’ must seem to the social historian to border on the ludicrous.
It is true of course that in discussing the problem of free will I have thought it proper to focus attention upon the type of moral difficulty which occurs in the situation of moral temptation. But the reason for that is very simple. Free will is commonly thought to be a pre-condition of legitimate moral praise and blame; and it is at least primarily in respect of a man's response to the situation of moral temptation that he is accounted by others (and most assuredly accounts himself) morally praiseworthy or blameworthy. A man's response to the theoretical difficulty of deciding what he ought to do is an utterly different sort of thing. It is a response of his intellect and knowledge not of his will—and we do not normally judge a man to be morally blameworthy for a failure of intellect or knowledge as we do for a failure of will.
(b) This brings me to my second point. The decision in cases of difficulty about what we ought to do which Nowell-Smith chides me for not seeing to be both a commoner and a more important case of moral choice than the decision whether or not to rise to duty in moments of temptation is neither a less nor a more common and neither a less nor a more important case of moral choice since it is not a case of moral choice at all. In order to be brief I must be somewhat dogmatic in discussing this; but refinements of statement hardly seem necessary where the principle (as it seems to me) is so perspicuously clear.
In the most common case of such difficulty where we are unsure whether in our given situation we ought to do X or Y the routine procedure (I presume it will be agreed) is to subject to a much closer examination than we have yet given them the implications and consequences of each of these alternative courses of action. In the light of the fresh considerations which then appear we are generally though not always in a position to decide which course best accords with the moral rules or principles we accept as binding. But in what conceivable sense can I be said to be thereby ‘choosing’—let alone morally choosing—that X or Y is what I ought to do? My decision is the conclusion of an argument determined by the relevant objective considerations as they appear to me and I no more ‘choose’ that X rather than Y is my duty than I ‘choose’ that 2 plus 2 equals 4 rather than 5.
In the less common case where it is seeming collisions or inadequacies within the accepted rules themselves that occasion the perplexity reflection is again the natural route to a decision about ‘what we ought to do’; though reflection now upon the credentials of the rules we have hitherto accepted more or less uncritically. This is the typical situation that gives rise to the kind of thinking we call ‘moral philosophy’. Its most usual result is to place the supposed rules in a new perspective in which their authority is seen to be not absolute but derivative; and in the light of the principle or principles from which we now see their authority to be derived we assess their relative claims and solve our problem. Again clearly our decision is determined by what appear to us the relevant objective considerations and ‘choice’ has no place whatsoever.
The one type of case (and it is comparatively rare) in which there seems to me any plausibility at all in assimilating a decision about what we ought to do to ‘choice’ is this. It may occur that after reflection has done its very best in the time available before action is required we can still find no ground for judging the duty claim of any one of the two or more competing courses (or rules) to be superior to the other or others. In that event there is a sense in which a choice though a wholly arbitrary one not only may but must be made in favour of one or another. But it is very necessary to ask What is the precise nature of this choice in favour of let us say X rather than Y or Z. Are we ‘choosing’ that X has the highest duty claim (as Nowell-Smith would appear to imply)? Surely not. It makes sense to speak of wishing that a certain proposition were true but no sense that I can see to speak of choosing that it be true. Are we then perhaps choosing to believe that X has the highest duty claim? Again surely not. How can we choose to believe what we in fact don't believe?—for ex hypothesi we are in possession of a belief reached by reflection that X has not a higher duty claim than Y or Z. What we are choosing I suggest is to act as if we believed X to have the highest duty claim even though we don't in fact believe it. If that be so our choice is not merely not a psychological monstrosity but is in the circumstances the one rational procedure—rational because it is the only way of escaping what we do believe to be wrong namely the futility of taking no action at all.
I must add one further word. Even if this wholly arbitrary choice should be whatever Nowell-Smith thinks that it is it would still be impossible to speak of it as a ‘moral choice’ in any sense relevant to the present discussion. It is in Nowell-Smith's phrase ‘a leap in the dark’; and he himself tells us that ‘just because it is a leap in the dark I doubt if we should be inclined to blame him [the leaper] if he leapt in what turned out to be the wrong direction’ (p. 288). Yet what Nowell-Smith by his own profession is supposed to be discussing are ‘cases of moral choice to which appraisals are relevant’.