The task before us is to provide a defence for the traditional conscience. What is needed is an understanding of morality which is capable of being defended at each of the three levels at which moral disputes take place and which can explain how it is that moral considerations have an integrity of their own while yet being open, as they plainly are, to the influence of divergent world-views. These requirements would be met if morality were thought of as essentially concerned with the fulfilment of men's needs as individuals and as members of society—with the necessary conditions of human well-being. There is disagreement, but not unlimited disagreement, as to what these needs are, for some human needs are so obvious and exigent that it is virtually impossible to overlook them—hence the widespread acceptance of moral platitudes—while others are more or less controversial, depending for their recognition on truths (or alleged truths) about human nature which are open to dispute among adherents of different world-views. Science, especially the psychological and social sciences, may well extend and deepen our understanding of human needs, so that the scientific humanist would not from this standpoint be wholly mistaken. His error would lie chiefly in his dogmatic insistence on regarding science as the sole accredited source of knowledge of human nature. And morality would have a point which all could appreciate.
If morality is conceived of in this way, it becomes easier to chart ‘the theological frontier of ethics’. For Christianity has its own characteristic conception of human needs; the needs of men are those of a creature whom God loves, has created, redeemed and destined for eternal life, and who, therefore, has an unconditional value. The moral law is not simply the arbitrary command of God, but is based upon the nature of man as God made him. And God does not make just rational beings whose entire duty is to exemplify a universal pattern, as rational humanists have always tended to suppose, but individual persons with their own particular vocations, and their own access to the means of grace. So that there is a place also for the characteristically romantic values of spontaneity and authenticity.
I want, in the remainder of this book, to develop this theme and to show how this conception of morality can provide a secure and intelligible basis for the demands of the traditional conscience. But this can only be done if the underlying philosophical thesis is defensible. What I have been expounding in this preliminary and tentative fashion could be described as a form of theological naturalism and, as such, it is open to all the criticisms that have been directed against naturalism in ethics. At this point my resolve to adhere to Lord Gifford's injunction that these lectures should be ‘popular’ is liable to come under some strain. For it is impossible to avoid a certain amount of technical moral philosophy. In discussing Hare's version of liberal humanism and in complaining that he fails to appreciate the extent to which moral attitudes and moral theories (his own included) tend to reflect divergent world-views I have so far dodged his very acute criticism of any attempt to derive moral judgements from statements of fact, whether of empirical or metaphysical fact.
As we have seen, Hare bases his critique upon an analysis of language. In the case of a word like ‘good’ it is necessary to distinguish between the meaning of the word ‘good’ which is given by its use as a term of commendation, i.e. its ‘evaluative meaning’; and the criteria for its application, which will consist in those characteristics in virtue of which the speaker resolves to commend the action, character or situation; its ‘descriptive meaning’. The criteria specified will vary with the moral code of the agent; the commendatory function of the word remains constant. A request for reasons for calling something good is to be met by indicating the characteristics in virtue of which we hold it to be good. Similarly with decisions. If called upon to justify a decision I must refer to the principle upon which it was based; and, if called upon to justify that, I must indicate the consequences of the universal application of the principle (for it is these which give content to the principle). And so, as Hare points out, we could eventually be driven back to a complete specification of the way of life of which the original principle is a part.
Thus Hare admits the relevance of world-views to morality, but insists that at each stage it is the free decision of the individual to pick out certain characteristics rather than others as objects of commendation which gives these the status of ‘reasons’. The agent could always choose otherwise and, if he did, he would, in so choosing, have adopted other reasons. The only constraint upon him is that of consistency. He must be prepared to accept the consequences of any choice he makes.
There is something distinctly odd about this notion of choosing what shall count as a reason,1 but Hare is prepared to accept this and other paradoxes in order to secure a single objective, which is to account for the ‘action-guiding’ function of moral language:
[All value-words] have it as their distinctive function either to commend or in some other way to guide choices or actions; and it is this essential feature which defies any analysis in purely factual terms. But to guide choices or actions, a moral judgement has to be such that, if a person assents to it, he must assent to some imperative sentence derivable from it.2
Hare maintains that in order to know how to use a ‘descriptive’ word, we should have to know ‘to what kinds of things it was properly applied’ and no more; whereas evaluative words have in addition ‘prescriptive meaning’. Thus if ‘good’ were to be thought of as being simply a descriptive word, there would have to be a meaning-rule which says that the word is applicable to a certain kind of man; if it is thought of as also having prescriptive meaning, anyone who learns it will be learning not merely to use a word in a certain way but to commend or prescribe for imitation a certain sort of man.
It is worth noticing in the first place that this very simple dichotomy leaves us uncertain how to classify a great deal of our vocabulary. Many words that we commonly use involve some sort of judgement or assessment on our part for their correct application, e.g., aesthetic judgement in the case of ‘subtle’ or ‘harmonious’, historical or political judgement in the case of ‘revolutionary’ or ‘influential’. The terms I have constantly employed in this book, ‘romantic’, ‘rational’, ‘liberal’ come naturally to mind. It would be quite unplausible to suggest that there are rules of language which determine uniquely how they should be applied. When one has achieved mastery of the English language, it is a separate task to learn to think historically or philosophically. Thus any educated Englishman knows the meaning of the word ‘revolutionary’; it may take a trained historian or student of politics to determine whether, in the light of certain evidence, a revolutionary situation existed. Yet it is equally clear that these judgements need not be prescriptive. The historian may or may not take an attitude to the facts or prescribe a certain type of behaviour for imitation. Moreover, if he does use such words predominantly to commend or condemn, as he very possibly may, he does not have complete freedom as to the sense he shall give them. Although there are not fixed rules of language which determine their application, the criteria for their use are very much more restricted than for the use of very general words like ‘good’ and ‘bad’. You cannot choose to call anything brilliant, subtle, or harmonious any more than you can call anyone just, merciful, or courageous.
Nevertheless Hare will insist on two points. The first is that ‘good’ at any rate, as the most general word of commendation, is independent of criteria, at least when used in a fully evaluative sense, and must be independent, if it is to exercise the function of guiding choices. The second is that, unless this point is appreciated, one cannot understand how our more specific moral vocabulary is open to criticism and capable of change and development. But, as critics of Hare have pointed out, there is a close relationship between the use of ‘good’ and of these other more specific expressions. Suppose I am asked my opinion of a philosophical work. I say that I find it obscure, dull, involved, lacking in logical rigour, pretentious, and incoherent. Does if make any sort of sense if I then pronounce it a good book, asserting my logical right to apply what criteria I like to the evaluation of philosophical works? To be sure it can still be a good book, if, in spite of this accumulation of vices, it has countervailing virtues, but if I close the list there it would seem that I am no longer free to praise it (or perhaps we ought to say, not that I cannot praise it, but that I cannot reasonably praise it). The only way in which Hare can take care of this sort of example, so far as I can see, is to regard ‘philosophical work’ as a functional expression; so that ‘good philosophical work’, makes clear the criteria to be employed, just as ‘good auger’ or ‘good hammer’ does; but philosophy does not have a function in any relevant sense. Competent judges may differ as to what are the requirements for good philosophy, and as to whether a particular work satisfies them, but implicit in their judgement is a readiness to produce reasons which are defensible.
I do not myself think that the moral case is fundamentally different. For example, I do not see how I can reasonably call a man good because he is a coward. Of course I can commend him if, in spite of or perhaps even because of being a coward, he has developed certain virtues. He may have been forced by constant humiliations to recognize his own weakness and so have attained an honesty and humility denied to many stronger men (though honesty and humility require courage of a sort). But I cannot (or cannot reasonably) commend him for his cowardice as such if it is to be moral commendation.
The prescriptivist answer to this is to remind us that ‘cowardice’ is itself a value word, or normally used as such, and that, in so far as I cannot praise a man for cowardice, it is because in doing so I should involve myself in a sort of evaluative contradiction, commending and condemning someone on account of the same characteristics in one and the same breath. The very language I use has valuations built into it, which I cannot remove at will. In Hare's phrase, it ‘incapsulates the standards of the society’.3 But, it will be said, I can explicitly dissociate myself from these standards; I can commend him on account of those neutral characteristics which are the ground for his being commonly condemned. I can praise him, if I choose, for running away in battle, telling lies rather than endure discomfort, agreeing with everyone for fear of giving offence, etc. I can list the characteristics commonly associated with cowardice and resolve to commend them.
So it is said. But if anyone actually took this line (in life, not in philosophical discussion) we should, surely, in this case also, want his reasons for commending the man. We should want to know the point or purpose of his cowardly behaviour, whom it purported to benefit and in what way; and if the speaker answered simply by developing in further detail the sort of behaviour he wanted us to approve—elaborating, so to say, the cowardly way of life—we should simply be non-plussed. We should not say, as the prescriptivist supposes we should, ‘Here is a man with a somewhat eccentric set of moral standards!’ We should say, I imagine, that he appeared to have no moral sense at all.
‘We should, no doubt, react in this way,’ the prescriptivist might reply, ‘but does this settle the argument? Is it not too short a way with prescriptivism to refer to what “we” should or should not say?’ The prescriptivist suspects that such a move is no more than a covert appeal to the traditional standards which have been ‘incapsulated’ in ‘our’ language. To allow this appeal is, he believes, effectively to rule out the possibility of criticizing these standards. And this is, without doubt, a justified complaint against some formulations of ‘naturalism’. Thus Hare says:
It is useful to have in our language both secondarily evaluative words like ‘industrious’ and primarily evaluative words like ‘good’; and we should therefore be suspicious, if any philosopher seeks to persuade us in the interest of concreteness to neglect the study of words like ‘good’ and concentrate on words like ‘industrious’ and ‘courageous’. The object of such a manoeuvre might be to convince us that all moral words have their descriptive meaning irremovably attached to them; but, fortunately for the usefulness of moral language in expressing changing standards, this is not so. To take this line would be to give an account of moral language which is, so far as it goes, true, but not sufficiently general (in the sense in which Newtonian mechanics is not sufficiently general). The account would suffice for the moral language of an irrevocably closed society, in which a change of moral standards was unthinkable; but it does not do justice to the moral language of a society like our own, in which some people sometimes think about ultimate moral questions, and in which, therefore, morality changes.4
That the meaning of moral words may change and that people's views about morality may change, and that existing standards are always, in principle, open to criticism, these are facts to which the prescriptivist constantly, and rightly, calls attention. This is surely a conclusive reason for rejecting the sort of naturalism (called by Hare ‘descriptivism’) which holds that the inference from a non-moral description of something to a moral conclusion about it is an inference whose validity is due solely to the meaning of the words in it.5 Against such a view Hare insists that ‘in saying that it is proper to call a certain kind of man “good” (for example a man who feeds his children, does not beat his wife, etc.) we are not just explaining the meaning of a word; it is not mere verbal instruction that we are giving, but something more: moral instruction’.6
In this he is clearly right. If someone wishes to state a case for the practice of wife-beating, which has been entirely respectable in some societies, (including our own in previous centuries), it is not enough to tell him that non-wife-beating is part of the definition of ‘good’ as applied to men. The question at issue is, however, whether it at all follows from this that ‘moral instruction’ is what Hare and other prescriptivists take it to be. Does it, in particular, follow that ‘moral instruction’ cannot be criticized or defended on objective grounds e.g. by reference, as Hampshire suggests, to the natural, or the supernatural order? In other words, is a philosopher who wishes to state a case for the objectivity of moral judgements committed to the view, which Hare identifies with naturalism, that ‘in saying that it is proper to call a certain kind of man “good”… we are just explaining the meaning of a word’?
It is not at all obvious that he is. In recent moral philosophy it has too easily been assumed that the choice lies between a ‘flat’ naturalism and a ‘sharp’ non-naturalism. The ‘flat’ naturalist maintains that inferences from ‘is’ to ‘ought’ are possible by way of meaning-rules alone. The ‘sharp’ non-naturalist denies this and asserts as a counter thesis the typical prescriptivist argument that, as a matter of logic, no imperative conclusion can be drawn from premisses which do not contain at least one imperative. All argument against either one of these positions is taken to be pro tanto an argument for the other, since no account is taken of any but strictly deductive inferences; although, in appraising intellectual as well as moral qualities, we are familiar with terms of assessment, like ‘reasonable’ and ‘implausible’, which require to be supported by reasons, but cannot be defined by reference to any list of determinate properties.
In talking earlier about cowardice, I was unsure whether to say that you could not call a man ‘good’ on account of cowardice or that you could not reasonably do so. I was tempted to say ‘you cannot commend cowardice’ on the ground that it does not make sense to say ‘cowardice is good’, so, that if someone went through the motions of commending cowardice, we should not understand what he meant. Part of the problem here is that the conditions in which we are prepared to say we understand what someone means are elastic. In the present case of someone who said ‘cowardice is good’ we should have no difficulty in surmising that he was in favour of cowardice and prepared to recommend cowardice to others. If, in reply to questioning, he insisted that by ‘good’ he meant ‘morally good’ it would be less easy to see what he meant, because it would be hard to envisage the form which a moral defence of cowardice might take. Still we might understand him as believing that such a defence was possible. If, however, we insisted against him that there is a rule of language in virtue of which ‘cowardice is morally good’ does not make sense, he might very reasonably feel that we had cut off the debate in a somewhat arbitrary fashion. If what he says is nonsense, it is not merely verbal nonsense, but moral nonsense, and it is because it is generally agreed to be moral nonsense that it is tempting to regard it as verbal nonsense too, so completely has this particular standard been incapsulated in our language. Indeed it may be the case, as I think it is, that the status of courage as a virtue is so secure that there is no language that does not incapsulate it, so that we could argue with some plausibility that to deny it involves a contradiction in terms. But even this does not mean that all criticism of the incapsulated standard is ruled out and the appeal to language alone settles the argument definitively. It means only that criticism in fact fails. No one has produced and perhaps no one can produce good moral reasons for rejecting this particular standard and for altering our language accordingly.
I have deliberately chosen an example in which the moral arguments appear to be conclusive and in which, therefore, the prescriptivist case has least plausibility. What are we to say about the more controversial cases in relation to which prescriptivism is intuitively far more convincing? Some of Hare's most illuminating passages describe the way in which a word's evaluative meaning may alter, while its descriptive meaning remains the same, or vice versa; how with some words the evaluative meaning is primary, in others the descriptive. An example of the way in which a word which used to be morally neutral may come, on the lips of certain people, to be a term of moral commendation, is the word ‘adult’. For most people it still refers simply to a stage in the biological development of the human individual and has little or no moral significance; they assume that, from a moral point of view, adults are as likely to be bad as good. It has purely ‘descriptive’ meaning. But, for some people, to be ‘adult’ is to display a certain range of approved attitudes, and we know well enough what these are likely to be. An ‘adult’ person is one who is morally autonomous, progressive in politics, and has liberal ideas about sex. One could scarcely call Mrs Mary Whitehouse ‘adult’. It would, surely, be absurd to criticize this usage on the ground that the word ‘adult’ simply does not mean this. That would be to place a purely linguistic ban on a putative advance in moral awareness. As Hare remarks, ‘to say that in all moral words the descriptive meaning is primary would be suitable to the moral language of a closed society.’ Nevertheless the new usage may reasonably be asked to show its credentials for reception into our moral vocabulary. The mere fact that a number of people employ the word ‘adult’ as a term of approval is not enough to constitute it a term of moral commendation though it may be enough to constitute it a term of approval (cf. ‘professional’). Before we can understand it as such we shall need to know what it is about being adult that is thought to be morally good and why. In point of fact we shall discover that this usage reflects an ideal, and this ideal is bound up more or less explicitly with a theory about the relationship between the psychological and the moral development of the individual, as observed in Western European society, and the theory determines the selection of qualities that are thought of as being characteristically adult. The ideal and the theory that goes with it is, in fact, related to a particular world-view; a variety of scientific humanism of the sort sponsored by Alex Comfort. Given the theory it is plausible to hold that in virtue of a meaning-rule anyone having these characteristics is adult and worthy of moral commendation; and, when we understand the theory, we can understand the usage, although we are not bound to accept it, unless we accept the theory. The theory is controversial; it can be criticized and it can be defended, but for neither criticism nor defence is it enough simply to opt for or against the qualities of character which the theory selects for approbation. We have to assess the merits of the opposing world-views. Because of the role of theory in the selection of qualities for approbation or disapprobation, it must be comparatively rare for precisely the same qualities to be approved by one party and, quite independently, disapproved by another. For instance, when Hume condemned ‘the whole train of monkish virtues’ the traits which he was attacking had been picked out for him by the moral theory implicit in monasticism and he was predisposed to condemn them by his total lack of sympathy with the ideals of monasticism. It is most improbable that, if Hume had never encountered the monastic tradition, he should have formed a con-attitude to just those traits of character, which were the objects of a pro-attitude on the part of St. Benedict.
Thus we can see that the language we use and the theories it presupposes may undergo change and be open to criticism of different sorts, including moral criticism, without being ‘evaluative’ as understood by the prescriptivist; that is, without the individual speaker being logically committed by his use of the language to behaving in certain ways, or making certain choices.
So also with cowardice. I can, of course, express myself as in favour of cowardice and resolve to live in a cowardly fashion. I can explicitly dissociate myself from the use of any language in which traditional approbation of courage is incapsulated. It does not, however, follow that in so doing I am taking a moral stand. Whether that is so depends on whether courage actually is a moral virtue, as traditionally believed, or whether, as I am now supposedly maintaining, cowardice is. Nothing in the argument has any tendency to show that there could not be reasons, even conclusive reasons of, a moral kind for the one position or the other (but not, of course, for both).
But, it will be objected, if this point were conceded, the word ‘good’ would have lost its function of guiding choices. There would no longer be any logical connection between my recognition that an action was good, and my choosing it. It would be perfectly reasonable for me to say ‘It's good; so what?’ Unless in calling an action cowardly, and so bad, I commit myself to not doing actions of this sort; unless, that is, I am using the words ‘cowardly’ and ‘bad’ in a fully evaluative sense as the prescriptivist understands it, there is an unbridgeable gap of a logical sort between my use of such moral language and my conduct. But clearly the whole point of moral language is to guide conduct.
This objection is at the heart of the prescriptivist case and it requires an answer. Quite obviously if a man says ‘X is wrong’, yet his conduct in respect of X is to all appearance precisely what it would have been had he not expressed this opinion, there is something amiss. Is he insincere or morally weak, or is it possible that he is morally defective and unable to understand the meaning of the word ‘wrong’? The prescriptivist says that the trouble is that he doesn't sincerely assent to the imperative: ‘Let me not do X’, and so isn't using ‘wrong’ in its evaluative sense; for, if he were using it in its evaluative sense, he would refrain from doing X. On this view the possible alternatives are merged in one—a sort of insincerity. And, as has often been pointed out, this has the paradoxical consequence that it is logically impossible not to do those things which we think we ought to do, or to do those things which we think we ought not to do (so long as ‘ought’ is used in a fully evaluative sense).
Hare7 displays great ingenuity and genuine insight in trying to mitigate this paradox, but even if it can be done, which I doubt, it remains very much open to question whether prescriptivism does, in any case, account itself for the action-guiding function of moral words. There is an ambiguity in the conception of guiding action. We could say that beliefs guide our action in so far as they provide reasons for acting in a certain way, or in so far as they supply motives for acting in a certain way. I do not myself see how the prescriptivist analysis of moral judgements shows them to be action-guiding in either sense. For, if we ask what on this view is the reason for doing what we ought to do, we are told that it is that in using the expression ‘I ought to do X’ we have expressed the intention of so doing; but this provides no reason for forming one intention rather than another. And if we ask what motive is provided for doing what we ought to do, we are told that the motive is that we have formed the intention of doing it, but no explanation is offered why we form this sort of intention (moral intention) at all.
But the prescriptivist is right to press his question: How do moral judgements guide action? And if we reject his answer we must look for another. We must consider how morality can provide reasons for action and also appropriate motives for it.
If we regard the concept of morality as logically bound up with people's needs and interests it becomes possible to do justice to these requirements, while at the same time avoiding both the ‘naturalistic fallacy’ and the ‘prescriptivist paradox’. Moral words will have (to use Stephen Toulmin's expression)8 a ‘gerundive’ force in virtue of which they cannot be defined in terms of expressions which lack this force. What Hare and others say about evaluative meaning can be applied mutatis mutandis to this gerundive force. To use ‘industrious’ or ‘adult’ evaluatively in this sense is to use it in such a way as to imply that the characteristics in virtue of which a person is said to be industrious or adult are such as to make him worthy or ‘meet’ to be imitated: which in turn implies that reasons of the appropriate kind could be given for the assessment. If the reasons are to be moral reasons, they must relate to some intelligible conception of human well-being. It is possible to call someone industrious in an evaluative sense in this way, i.e. in such a way as to imply that he is worth imitating, without actually imitating him or even intending to do so. In that case, no doubt, the speaker could be regarded as a hypocrite but not as failing to make a moral judgement at all.
On this view what a man commits himself to when he asserts that an action is wrong, or uses another form of words that implies it, is that there are sufficient reasons of a moral kind for not doing it, and these reasons can perfectly well be objective. To ask, ‘Why shouldn't I do what is wrong?’, which for the prescriptivist is a nonsense question, amounts to asking, ‘Why shouldn't I do what there are sufficient reasons of a moral kind for not doing?’ This is a nonsense question if what are looked for are moral reasons, but is perfectly intelligible if what is looked for is a justification of morality.
To adopt this idea of morality is to side finally with those who take morality to be constituted by the sort of considerations that count in favour of a moral judgement as against those who define a man's morality as consisting of the set of principles that he regards as overriding. So Heliogabalus must be described as choosing to live his life by aesthetic rather than moral principles because he thought interesting colour contrasts mattered more than men's lives. From a moral point of view moral principles ought to be overriding, but in fact they sometimes are not. This should not worry those who are accustomed to confess, ‘We have left undone those things which we ought to have done; and we have done those things which we ought not to have done’.
What I have sketched is simply a basic framework of morality. Any developed morality is a much more complex matter, but I am suggesting that the complications are systematically connected with the needs of men in their varied social and personal relationships. No first order morality is adequate which does not reflect these; and no philosophical account of morality is satisfactory which does not recognize them. As soon, however, as one has got beyond the barest necessities for any kind of human life, and the very basic morality that is demanded by them, one emerges into an area of possible controversy. Moralists differ, often as a matter of temperament, as to whether to emphasize the moral claims that derive from institutions as they are, ‘incapsulated’ in the language of custom and tradition, or whether to stress the moral criticism of the status quo. But although existing institutions are always in principle open to criticism and the current language similarly open to revision, as the prescriptivist rightly stresses, such criticism and such revision must, to be moral at all, relate to some conception of human nature and its possibilities.
In the light of these considerations we can also make sense of the confusion in morals and in moral philosophy which I attempted to illustrate in my opening chapter. I remarked about the writers of my imaginary letters to The Times, that ‘these writers are divided not only in their opinions but in their vocabulary. They inhabit different words of discourse’. That is to say, they express different moral viewpoints related to different conceptions of human well-being. These, to use Strawson's language, are based upon varied, and often conflicting, ‘pictures of man’ which in turn reflect different ‘profound general statements about man and the universe’. Christianity is one of these and, as such, has its own characteristic conception of the nature and scope of morality and of its place in human life.
If, then, we are looking for a system of normative ethics which we can wholeheartedly accept, it must be one whose rationale we can also understand and accept, and which makes sense both of our deepest moral convictions—what I have been calling our ‘intuitions’—and of what we have reason to believe about man and the universe. If it is among our deepest moral convictions that morality is objective and categorical, and that moral principles are overriding, it must do justice to this fact. If we are persuaded that there are moral principles which hold, irrespective of consequences, or which can be set aside on account of consequences only in exceptional circumstances, it must take account of this too. But it must, also, be able to justify this procedure—to explain why it is that we should trust these profound convictions of ours and not dismiss them as mere irrational prejudices, however deeply held. Indeed, it must do more than this; it must account for the fact that morality has an integrity and autonomy of its own, so that we may properly reject a total view of life because its ethical implications are unacceptable. This is what is involved in the vindication of what I have called ‘the traditional conscience’. Of course there is another alternative altogether open to modern man. He may abandon the traditional conscience and rejoice with Nietzsche in his consequent liberation. But my argument is addressed to those who are not prepared to accept that alternative. So, while leaving open the possibility, as I am bound to do, that an entirely secular world-view might provide the traditional conscience with the rationale it needs (although I do not know of one that does) I want now in what remains of this book to consider, in relation to Christian theism, the following questions:
- (i) Does it show reason why certain moral principles should be given more weight than they merit on purely utilitarian grounds?
- (ii) Does it offer a justification for morality?
- (iii) Does it do justice to the extent to which morality is autonomous?
- (iv) Does it provide a rationale for, and a critique of, moral intuitions?
Cf. G. J. Warnock, Contemporary Moral Philosophy, Macmillan (1967), pp. 46–7.
The Language of Morals, p. 171.
Freedom and Reason, p. 25.
Freedom and Reason, p. 25.
Freedom and Reason, p. 21, my italics.
Freedom and Reason, p. 23.
See especially Freedom and Reason, chapter V, ‘Backsliding’.
See The Place of Reason in Ethics, Cambridge University Press (1950), pp. 70–2.