Franklin is enthusiastic about analytical language philosophical techniques inspired by Ludwig Wittgenstein which dis-solve philosophical problems.

He is not embarrassed to say that such a concept puts human freedom outside the natural causal order, which he believes is causal and deterministic. His view is thus in the category of "metaphysical" freedoms such as agent causation, non-occurrent causation, and contra-causal freedom. He cites R. M. Hare, Wilfred Sellars, Richard Taylor, and others as holding similar views.

THE PROBLEM (pp.1-5)

The problem of freewill is always with us. Both libertarianism and determinism, and each in many forms, have been proved to the satisfaction of innumerable protagonists; while to the satisfaction of innumerable others the conflict has been dissolved, has been shown to rest on misunderstandings of the concepts involved. The thesis of this book is both more modest and more ambitious than any of these. It is more modest in that, though I have argued for one particular view, I have not claimed to present conclusive reasons for it; and in fact I have tried to show in what respects, and on what grounds, it could and would be rationally rejected by many men. It is more ambitious in that I believe I have placed the problem in a new perspective. I have suggested why there is this never-ending dispute; what underlies the contentions of each position; why at a given period one may be in the ascendent over others; why the protagonists so easily, in fact almost inevitably, overstate their cases; and finally, what might offer the best chance of future progress.

Let us begin with a simple statement of the problem, and of the positions which may be adopted. Men do not often doubt that in most circumstances they can freely decide between one course of action and another, in a way that certainly inanimate objects, and presumably also animals, cannot. Nor, in the realm of morals, do they often doubt that unless a man chose freely he cannot properly be praised or blamed for his action. These beliefs give rise to a problem, however, when they are considered together with other frequently held ones which may seem to leave no place for such freedom. There thus arises the philosophical problem of freewill. Historically the difficulty has often been that of reconciling freedom with beliefs in God's omniscience or foreordination of all things, which raise what may be called questions of predestination. In the present climate of opinion among British philosophers these issues have been largely handed over to theologians, and for reasons of space I shall largely have to leave them there. I am concerned with another set of considerations which give rise to what may be called the problem of determinism.

People not only make choices, they explain why choices are made and they predict what they will be. If we reflect on what is involved in explanation or prediction we may conclude that they always involve some more or less explicit and more or less precise generalisation to the effect that this sort of thing occurs in this sort of situation. The ideal limit of such a generalisation is a causal or scientific law to which there are no exceptions. Moreover, on further reflection we may decide we cannot doubt that such laws exist, even if undiscovered. For we can hardly doubt that in principle all that occurs is fully explicable, and only this ideal limit, it may seem, would give a really satisfactory explanation. In traditional language the Law of Causation, that every event has a cause, is equivalent to the Principle of Sufficient Reason, 'by virtue of which', as Leibniz formulates it, we consider that no fact can be real or existing and no proposition can be true unless there is a sufficient reason, why it should be thus and not otherwise, even though in most cases these reasons cannot be known to us.

And on this Principle (together with the Principle of Contradiction), he said, all our reasoning is based., Yet the existence of such causal laws may seem to imply that for a given agent in a given situation only one course of action is a real possibility, and may thus seem incompatible with his freedom.

To this general problem philosophers return many different answers, but they may conveniently be grouped into three types. One position is that our distinctive human power to choose implies 'contra-causal' or undetermined freedom. This is libertarianism. It divides into various forms, which differ as to whether all or only some, and if so which, choices imply indeterminism. For example, one form holds that we have undetermined freedom only when facing up to moral demands.

The second group consists of philosophers who accept the incompatibility which the libertarian alleges between causality and at least the usual notion of freedom; accept, as he does not, the principle of universal causation; and thus reject the belief in freedom. These are par excellence determinists. We may, however, distinguish two sub-groups. First, there are those who aim to break to men the news that they are in no significant sense free. Secondly, there are those who set out to reconcile determinism with some consciously transformed notion of freedom. They agree that the common or unphilosophical notion of freedom inclines to libertarianism, but argue that this is a vulgar error, and that true freedom is, say, the recognition of necessity. The less that such a philosopher emphasises or concedes that his view of freedom is not the common one, the more he merges into a third group. These deny that the suggested incompatibility between freedom and universal causation exists at all. They refuse to save the notion of freedom by transforming it; they believe it is not the common man but the philosopher who is confused. Since, they suggest, we all assume both that we are free and that all events can be explained, it is for the philosopher to show by careful analysis how these beliefs fit together, and not to sacrifice either to the other. If we thus examine carefully what is under our noses we will discover what we might have expected: that the apparent conflict is due to various conceptual errors, such as, in the realm of morals, a faulty analysis of the concept of responsibility, or of the assertion that the agent 'could have done something else'. We will then see that what we all mean by freedom, and not some truncated or transformed account of it, is compatible with, and perhaps requires, determinism. Traditionally there are only two classifications, 'libertarian' and 'determinist', to apply to all these positions; for the word 'indeterminist', which is sometimes found, seems to be nothing but a synonym for 'libertarian'. This leads to little confusion, except In the case of the third group of philosophers. 'Libertarian' suggests to them that we have an undetermined freedom, and 'determinist' that we are not as free as we thought; and it is the very point of their contention that both these suggestions are wrong. Since they claim that on a proper analysis the problem dissolves, I find it sometimes useful to call them `dissolutionists'. In general, there are some contexts, such as that of clarifying the precise libertarian position, where it is convenient to call all denials of libertarianism 'determinist'; and others, including many ethical ones, where the threefold classification is more helpful. But all such nomenclature is merely a matter of convenience, and I do not wish to pin on anyone a label to which he would object. ¹

¹ A determinist is commonly defined as one who asserts that all choices are caused. It might therefore seem that dissolutionism is necessarily a species of determinism, namely that which asserts the compatibility of determinism with freedom. But it is in fact a cross-classification. Some dissolutionists profess to be agnostic about universal causation, and therefore are not determinists. They maintain merely that whether or not choices are caused, they may still be free.

The central issue may be put in terms of a contrast between two notions, which may be called self-determinism and indeterminism. No one other than an absurd fatalist would suggest that a man is in normal circumstances at the mercy of his environment, as a rudderless boat is driven by the waves. It is we who decide how to act, and act as we decide. We respond to situations with more or less intelligence and integrity. Another person, or we in a different mood, might respond to the same situation quite differently. No account of a man's environment and external stimuli would usually enable us to know what he will do; the answer will depend on his own deliberation and choice. We might express this manifest ability of human beings to transcend a mere immediate reaction to stimuli by saying that they are self-determined. But this is quite compatible with saying that there are causal or scientific laws (for example, psychological ones about thought-processes, or physiological ones about brain-processes) which govern an agent's decision. The fact that a full causal account would have to be enormously complicated, that it would have to include statements about the agent as well as about his environment, and that it cannot in practice be achieved, does not show that it is in principle unobtainable. So self-determinism is compatible with determinism; indeed it is the only intelligent form of determinism. The determinist says that this self-determinism is the only sense (if any) in which a man is free, and the dissolutionist adds that it is in fact all we mean when we speak of human freedom. The libertarian, however, denies this. He insists that such laws, if they existed, would be incompatible with our most basic conception of freedom. His view may be put by saying that man's freedom is undetermined.¹

¹ He often objects to this word, for it may suggest that undetermined choices are purely random, and thus concede too much to the determinist. But I here use it only to refer to the feature of the libertarian position that it denies complete causal determinism, and this negative feature is appropriately expressed by a negative word.

In day-dreams, and in other relaxed or lazy moments, we may be conscious, in one sense of that vague word, without paying attention to anything. But as soon as we become involved in, or concerned with, any situation or train of thought we start to pay attention to it. The activities, such as choices, with which the freewill debate is concerned, are primarily ones which are performed attentively. It is this difficult and fundamental notion of attention which I wish to examine.

Though it is possible, with difficulty, to pay attention to more than one thing at once, our consciousness is normally focused on one. The focus, however, often changes, as when our attention is distracted by something else while we are engaged on some matter to which, in general, it is directed. It is not easy, however, to decide how often this occurs. We might be tempted to try to answer the question by paying attention to some topic, and at the same time noticing how often our attention is distracted. But if we do we are apt to find that what we had taken to be our reasonably sustained power of concentration dissolves into a kaleidoscope of shifting phenomena. Let someone read this paragraph while paying heed to the question how often his attention wanders ; and almost at once, I think, it will. The buzzing of a fly, the memory of a past holiday, wondering what the time is or whether the whole procedure is absurd—such things as these will immediately intervene. Moreover, if he succeeds in reaching the end of the paragraph he will find he can hardly remember much of it. This suggests that the whole approach is at fault, for it involves precisely that attempt to focus consciousness on more than one thing at once which we find so difficult. We are here so exercised in the second-order activity of trying to notice what we are conscious of, that we lose almost entirely our first-order capacity to concentrate on anything at all. If, on the other hand, forgetting this peculiar exercise, someone should, after reading a paragraph with attention, ask himself whether his attention had wandered, I think he will normally be unable to recollect that it had. This cannot be merely because these wanderings are immediately forgotten. For, apart from the inherent implausibility of that suggestion, there is the unquestionable difference in the speed and comprehension of his reading.

The capacity to pay attention doubtless varies greatly between individuals. Certain cases, such as the very highest reaches of disciplined religious meditation, are almost different in kind from achievements in the lower range. No doubt also in each individual the power varies in accordance with many factors. It is much greater when we are attracted to some subject matter than when we force ourselves to concentrate on it. It is also much greater when the subject matter on which we focus is itself changing; for example, most people can read a book, that is, keep their attention on a progression of words and sentences, for a much longer period than they can fixedly contemplate a single unchanging object. But for present purposes these matters need not concern us, and the following points which I have made are sufficient. One may be conscious without paying attention to anything. Activity of any sort, however, including coherent and connected thought, involves paying attention. Our power to do so can be exercised without break for much longer than we might conclude from the unfortunate attempt to introspect the process while it goes on. Yet nevertheless it is an essential feature of our paying attention, in all but the rarest and most highly disciplined cases, that the focus of our consciousness is continually changing.

It is these changes of focus which I wish to consider, and I wish to ask whether one can appropriately apply to them certain distinctions that have often been recognised in other contexts. The first distinction is that between an action and an occurrence, or between what a man does and what happens to him. This, as I shall later mention, seems an essential distinction in philosophy of mind. The second one is drawn, within the realm of action or activity, between an act and a choice. On some occasions a man may simply act, without any deliberate selection of his course of action, while on other occasions he chooses between two possible courses of conduct. The notion of a choice, as here used, implies that alternatives are considered, and all but one rejected. How, if at all, do these notions apply to the paying of attention?

A philosopher might be inclined to say that changes in our attention are mental acts, as opposed to occurrences such as twinges of pain. For if we put aside special cases, as when a subject attracts us with a virtually hypnotic power, or repels us so strongly as to produce a mental blockage, we may agree that we can always control the movement of our attention, even if only with difficulty. But to say that we can control it is not to say that we do. Our attention is continually being attracted by what we would like to think about, or distracted from what we are or ought to be thinking about, or, as it were, rebuffed by the dullness or unpleasantness of the subject matter. These influences account for the direction taken by a vast amount of our thinking, particularly in our less disciplined moments. They are part of our mental activity, in the sense that they are what goes on while we think. Yet in another sense they are what happens to us, rather than what we do. We seem therefore driven (though such moves are perilous in philosophy) to draw a further distinction, within the realm of mental activity, between the relatively passive and the more fundamentally active. In fact I think we must apply the distinctions between occurrence, action and choice in this more minute field, as we are accustomed to do in larger scale activities.

The cases where the movement of our attention is attracted, distracted or rebuffed involve nothing which, in the strictest sense, we do. A distinct, if rarer, phenomenon is the active switching of attention from one thing to another. A man who has to decide a matter may suddenly notice that he is day-dreaming about some topic suggested by his earlier train of thought, and the very realisation of this fact is enough to cause him to direct his attention immediately back to the issue he has to decide. This directing is quite different in kind from the unconscious distraction which led him to the day-dream, and as compared to the latter it is activity as opposed to what happened to him. Yet it is hardly a choice. He did not decide to stop day-dreaming, rather the realisation that he was day-dreaming led to his stopping it. Now within this directing of attention, which is itself a sub-class of the changes in our attention, there is a yet smaller sub-sub-class which seems to me to correspond strictly to the notion of a choice; and indeed it is a genuine choice, though of a minute kind. This arises when we consciously decide between pursuing, or dwelling on, this consideration or that. Whether our deliberation is concerned to reach some truth or some decision about how to act, we may be faced with a situation which can be described by saying that we are aware of two trains of thought which we might dwell on, and that we select one and we decide to pursue it rather than the other. I shall call this the selective directing of attention, and it is on this notion that I shall concentrate in the next section.

I have no definite suggestion as to how often the selective directing of attention may occur. I suspect that the frequency of its occurrence varies from individual to individual, and it surely varies according to the type of thought involved. Certainly it represents only a small proportion of the changes in our attention, but these changes as a whole are so frequent that even this sub-sub-class of them need not be a rare phenomenon. My own view would be that for any normal person, engaged in rigorous or careful thought, it is a common occurrence. In any case, the distinctions which I have attempted to draw do seem to me to apply to the very difficult phenomena in question, even if in certain cases the line between the various categories may be somewhat blurred. It follows that it is unconvincing to reduce the process of thought to any account which presents it merely as an interplay of attractive and distractive influences. Undoubtedly such influences occur, and govern a vast amount of our thinking. Further, we may have psychological evidence to suppose that there are others, of the existence or power of which we are frequently unaware. But if we attempt to reduce all mental activity to the passive interplay of various influences, the result is not a satisfactory piece of philosophical psychology.

Before going further I must relate my remarks to a discussion in Professor Stuart Hampshire's important book Thought and Action. ¹ The issue arises from a passage where he is concerned to emphasise the difference between thought and action. ² He is right, I think, basically to take the side of Spinoza against Descartes on this point, and to say that reaching a conclusion is not analogous to an act of will. We do not decide to believe, we decide or realise that we believe; and if the process has been honest, the conclusion is independent of our will. He is right also in the hints he gives of the necessary qualifications to this position, and his indications of the way in which our will to believe or disbelieve may influence the thinking process. Finally, he is right in saying that this may have deep moral implications. The difference arises when he argues that what goes on in us when we think systematically cannot be described as an action, because the will plays no part in it.

¹ London: Chatto & Windus, 1960. It is not my purpose to make a detailed comment on this work.
² Op. cit., pp. 153 et seq.

Thought, when it is most pure, is self-directing, as in the exercise of the intellect in deduction and in the following of an argument. When I use the active verbs of will and speak of directing my thoughts to a certain topic, or of concentrating my attention on it, I still contrast these acts of will, which start the process, from the process itself. Thought begins on its own path, governed by its universal rules, when the preliminary work of the will is done. No process of thought could be punctuated by acts of will, voluntary switchings of attention, and retain its status as a continuous process of thought. ¹

¹ Op. cit., pp. 153-4, my emphasis.

The point is impressive if we concentrate on the case of following an argument, or of expounding one when we have worked it out. Here the direction of thought is in an intelligible sense controlled by the argument, and the only element of will involved is that of renewing our attention if we notice that it has flagged.2 Perhaps this is all Hampshire meant. But in original or constructive thinking, the directing of attention seems ubiquitous and essential. Certainly from time to time it may lead to the recognition of a conceptual connection. 'Of course this follows from that,' we may say, or, 'But this won't go with what I concluded previously.' Still, the movement of our thought is directed not by a chain of argument but by our groping for one. It is controlled by the argument, not as the movement of a train is controlled by the rails, but only as that of an explorer is controlled by the country. In the last resort, if the italicised sentence of Hampshire's were true, then creative thinking would not be 'a continuous process of thought'.

4. INDETERMINISM AS THE SELECTIVE DIRECTING OF ATTENTION

I now argue that a libertarianism which is rationally constructed out of the common view of choice should locate its indeterminism in the selective directing of attention. In the first place, this suggestion is surely thoroughly congruous with the four cases where I suggested that the common view would insist on indeterminism. In a clash between what is believed to be our duty and temptation, what actually happens at the moment of decision? Is it not that we either direct our attention to the fact that this is our duty, and refuse to think about the attractiveness of the other course; or alternatively dwell on the attractiveness of the other course, and simply try to put our duty out of our mind as best we can? The second case of honesty in deliberation is even clearer, for it was defined by reference to the difficulty of keeping our attention fixed on matters which we ought to consider. The third case was that of moral perplexity. Here the position would seem to be that, having found no satisfactory rational solution of our dilemma and still being obliged to choose, we simply by an act of will direct our attention to one course rather than the other. Finally, in relation to the whole sphere of original thinking, does not the originality lie in the selection of one aspect after another for attention, in the hope of seeing familiar things in a new light, making new connections or questioning established ones? All else that there is even in genius is surely an achievement which either comes or does not; whence and how it does is a matter on which common men have no opinion and the learned no consensus. Creative freedom is exercised in grappling with the matter, not in achieving the result.

This may be reinforced by another consideration. In so far as the common view is based on the selective directing of attention, the most obvious cases of indeterminism will be those where the directing is most obvious. All such frequent activities are commonly best thrown into relief when we experience difficulty in performing them. The difficulties in selectively directing our attention are of two sorts: (a) The keeping of our attention fixed on some chosen consideration, in the face of some counter-consideration which attracts us or makes a demand on our attention; (b) when we are faced with difficulty in ordering a complicated set of material, and must therefore direct our attention systematically over the whole of it. We may add, as a cross-classification, that the directing of attention may be apparent either (i) as a decision or act of choice at the end of deliberation, or (ii) as an element in the deliberation itself. Now the four suggested paradigms of indeterminism represent fairly well the four possible combinations of these elements. Moral libertarianism is case (a) (i), where the choice, rather than the deliberation, is to be exercised in the face of the attraction to dwell on other matters. The second example is case (a) (ii), the third case is (b) (i) and the fourth is (b) (ii). The fact that these cases fit so neatly into this classification gives some support to the view that the classification expresses their underlying principle.

My suggestion seems equally compatible with cases where the actual choice is thought of as self-determined. First we may put aside cases where a man acts, as we often do in small matters, without really noticing what he is doing. The notion of directing attention, let alone of selectively directing it, need here hardly arise, but also we hardly speak of choice. Where there is choice between alternatives there is doubtless some selective directing of attention, but in some cases it is, as it were, essentially irrelevant to the final outcome. Sometimes the result of deliberation is to bring us to see that only one course really has anything (from our point of view) to commend it. We might say that the deliberation had its elements of indeterminism, in the sense that we might have explored either this or that aspect first, etc., but that the eventual decision did not. Again, a man may be of such firm character that he knows in advance that, however much he directs his attention to various aspects of his situation, it will not alter his decision. This, for example, might be Luther's case. Here one might say that though the precise course of the deliberation was undetermined the outcome was never in doubt. These seem to me the sort of cases where the common view would say that, given the man's character, only one course was a real possibility for him.

The picture at which we arrive is that in serious cases of deliberation and choice there is a frequent selective directing of attention, which I suggest should be seen as the basis of libertarianism. But its significance varies from case to case. Sometimes it may affect no more than the trivial details of a path which in general we could have predicted that the agent would traverse. There are innumerable intermediate cases in which it becomes relatively more important. At the other end of the range its existence in a given situation may be crucial. A man's whole action, and even his whole future, may depend on the minute movement of thought by which he chooses to attend to one consideration rather than to another. It is a matter of profound importance for our conception of man whether or not we see such a movement as only a special case of general laws, and thus as an example of what would always happen if just this situation should repeat itself.

I have spoken of this position as a rational reconstruction of the common view of choice. I do not suggest that the directing of attention is the only consideration which might commonly lead men to a belief in indeterminism. In fact I suspect that at least two other matters, related to each other, would play a part. The first is what I may call the lack of expectation of precise prediction. In understanding human conduct, and indeed generally outside the realm of the exact sciences, we must normally employ generalisations which are expected to have occasional exceptions; and this may incline men to the view that no precise regularities could in principle be obtained. The second related matter is the obvious fact of sheer human uniqueness. After all, if two candidates were to submit examination answers which were even largely the same in wording, the presumption that they were not independently produced would be irrebuttable. Yet both these considerations seem to me to lack cogency. My basic reason for this, which will emerge more fully when I discuss determinism, is frankly conceptual rather than based on the phenomenology of choice and action. It is that if the sorts of complexity and uniqueness which I have just mentioned cannot even in principle be reduced to regularities, in however complicated a way, then it would seem that they simply cannot be explained at all; whereas cases of the selective directing of attention allow an explanation of a radically different and non-determinist sort. The result is that I would not wish to defend any libertarianism other than that which I have outlined. I do not directly appeal to the common view in support of my position, as if to a majority verdict. But I believe my position is not only largely consistent with what most men would hold, but also is held for reasons which would appeal to them. Henceforth I shall call the view I have outlined 'libertarianism'. I shall now turn to consider some of the ways in which it has been alleged that the problem dissolves.

I must begin by stating my own position. I have elaborated a libertarian view based on the selective directing of attention, which I have presented as the most defensible form of the doctrine. In terms of the current debate this is the position I would hold. But it would be more accurate to say I am not a determinist. For if I had to bet on what the future holds I would put my money on the suggestion mentioned in Chapter II, Section 3, that the future course of discovery would produce what I called a scientific transformation of the problem. This suggestion is that an adequate neurophysiology will eventually erode the assumption usually made by determinist and libertarian alike; namely the assumption that we could in principle specify with unlimited pre vision the state of affairs which would turn out to be, or fail to be, a sufficient condition for its successor. The eroding of this assumption might occur in either of two ways. The first is that the relevant phenomena in the brain might be so minute that Heisenberg's Uncertainty Principle (assuming that it is not itself superseded) would suffice to raise a necessary imprecision in the specification of the total situation. Or, secondly, a new sort of complexity, hitherto completely unknown, might be brought to light. In other words, the 'emergent levels of relations between neurons in masses' postulated by Bullock, might in crucial cases be such that the concepts required for determinism were inapplicable. ¹ Any such scientific transformation would seem, however, for reasons I shall mention, to be much more akin to libertarianism than determinism.

¹ T. H. Bullock in Behavior and Evolution, eds. A. Roe amd G. G. Simpson, p. 166. Both these seem to me fundamentally different conceptions from the suggestion which has sometimes been made that the activity of a 'mental force' might be empirically discovered by its influence on a physical system, namely the brain, which otherwise would behave in a different manner in accordance with the relevant laws of physics.

My reasons for holding this view are various. In the first place, I am aware of more than those I can state here, though the others are, like the ones I shall mention, persuasive only. This seems to me appropriate. For if we are seeking to draw together separate puzzles into a coherent account which offers an ultimate explanation, then it would follow that the debate must be discursive. It would be impossible to indicate in advance what might be relevant, or to confine the discussion within set limits. A persuasive consideration drawn from one field may have to be balanced against an opposing one drawn from another. I shall here make one point, not with thoroughness but as an illustration of this discursiveness. It is that there are significant links between the whole debate about freewill and traditional religious beliefs. I do not mean to suggest that no theist can be a determinist. But there are, I think, relatively subterranean and easily missed elements of congruity which lead from theism to a denial of determinism, if not vice versa. I see in fact more than one such element, but one will do as illustration.

Traditional morality rests on desert and an ethic of intentions. As we might expect (whatever the problems of temporal and logical priority), this is paralleled in the traditional religious notion of a final moral accounting. Determinists who reject the traditional moral values for an ethic of consequences would have to reject or transform this religious notion too; what they could put in its place, if so minded, will not concern me here. I have argued, however, that a determinist may keep something like the traditional moral values, based on respect for personality, by substituting for the ontological responsibility of libertarianism a responsibility of self-acceptance. My present question is whether a parallel conceptual move could be made in the religious sphere.

It might seem a priori certain that it could be. For surely the theist's categories derive in the first place from our ideal of human personality at its best, before being transformed in an attempt to do justice to the notion of a being worthy of worship. If moral goodness is a pre-eminent constituent of our ideal, how could it fail to have analogical application to the divine? But the issue is not nearly so simple. For the relation of divine to human is fundamentally different from the relation of man to man, and it would seem to me that there are difficulties which might well be insuperable in any analogical extension at all.

First, I have suggested that an ethic of self-acceptance might have to justify the infliction of penalties on those who denied their responsibility by falling back on a social pressure view. In human relationships this would be entirely legitimate; society must Survive. But is there a divine analogue? God has no need to protect himself. Any notion of justified penalties, therefore, could get going only if there were self-acceptance. In traditional language God could justly condemn to hell only those who admitted they deserved it. Secondly, what good reason could they have for admitting it? I can understand what it would be, as between man and man, rationally to accept oneself and others as responsible agents, despite the belief that one could not (nomically) have done otherwise. For in whatever sense I may be responsible for my act, at least it is clear that those before whom I acknowledge it are not responsible for it. But in what sense would it be rational to accept responsibility before a being which had created me to be the very entity which, being just as it was, must commit the sins for which responsibility is being attributed? The closest human analogies might well seem to be, not with mature admission of one's faults, but with neurotic and compulsive guilt-feelings. Finally, if a man did in fact accept responsibility for his sins, his doing so would of course also be predetermined. Again, as between man and man, this need not destroy the seriousness and legitimacy of the attitude, since the others do not cause me to accept my responsibility. But God, in the last resort, would be the cause, and this does seem to destroy moral seriousness. A ventriloquist does not prove himself right by getting his dummy to agree with him.

My suggestion is that, even if a determinist wished to retain the notion of desert by basing it on self-acceptance, the move which seems possible in the purely moral sphere may be impossible in the religious. In arguing this I have of course partly transgressed my self-denying ordinance, and have taken sides on the theological question of whether predestination can be reconciled with the justice of God. For though I have not discussed how the crucial texts of a revelation are to be interpreted, I have in effect suggested that, even if they implied both predestination and divine justice (which in the case of Christianity I strongly doubt), there would remain a conceptual point, of the sort where philosophers have professional competence. It is that`justice' could (logically) be here employed only in a Pickwickian sense which was incompatible with its ordinary meaning. But my purpose has not been to pronounce even on the latter issue. It has been only to illustrate by means of an example the ramifications in which we may be involved in any final assessment of the freewill issue. Contemporary philosophy has found great profit in the method of isolating small and precise problems for individual treatment. It is therefore all the more important to insist that there may be some issues, including the present, where this method is inapplicable.

By this illustration I have tried to indicate the point of my proviso that I have more reasons for denying determinism than those now mentioned. For I would wish to elaborate a view within which the denial appeared as rationally integrated with, rather than incongruously attached to, other beliefs about man and the universe as a whole. I cannot attempt this now, but I can give my final assessment of the considerations discussed in this book.

To many philosophers it has seemed that reflection on practical and on theoretical reason tends to lead in different directions, and thus to provide the central tension in the freewill problem. Reflection on how we decide (whether in general or solely in moral situations) raises the demand for indeterminism; reflection on how we explain a fact to be the case leads naturally to a belief in regularity. I would agree that this is the natural division of the matters to be considered, but I do not think the dichotomy is as neat as this. I find that in both areas the arguments against determinism seem the stronger; though in the first case perhaps more than in the second, and in neither case for conclusive reasons. I shall begin with practical reason.

First, I repeat that the natural view of choice is not determinist, and that determinism must be an attitude superimposed on an earlier view. It is an attitude which I cannot achieve. I am, for example, quite unable to agree with Freud that we are inclined to believe in the undetermined freedom of our trivial but not of our fundamental choices. On the contrary, there stand out in my mind several important issues of past years where it seems to me, even now, that I well might have decided the other way, in which case my life would have been notably different to what it is. Further, at the more minute level of careful introspection, I think that the selective directing of attention is of fundamental importance. I am prepared to believe that all other phenomena which occur in deliberation, such as the attractions and distractions which play so large a part in it, are determined by causes of which I am not necessarily aware. For otherwise, since we do not in such cases select or choose our train of thought, there would be no reason why the outcome should be this rather than that. But the selective directing of attention alone seems to demand that we apply it to the notion of a deliberate choice between alternatives. Both its sheer phenomenological distinctiveness, and its importance for life and conduct, seem to me so great that I am not embarrassed to make it the exception to a general assumption of the determinism of mental phenomena.

Secondly, the moral issues also weigh deeply with me. The morality of an ethic of intentions, based on respect for personality, seems to me so important, and so well justified when understood, that I could not abandon it. I can understand, as I have tried to indicate, that a determinist could base a broadly similar morality on self-acceptance rather than indeterminism. But still it would be in an important sense a new morality. Here the onus of proof is very relevant. I do not feel at liberty to change my conception of man in such a fundamental respect unless the change is forced upon me.