These last few pages have been concerned rather to show that we can attach a meaning to an act that is self-determined but not issuing causally from the agent's character than to defend the existence of such acts. It may still be said that although direct experience gives us the notion of such acts, it is a mistaken notion; since, to mention one important reason, it conflicts with the observed facts of continuity between character and conduct. This criticism would certainly hold good, as we have already admitted, of a freedom in which 'anything may happen,' such as the freedom of mere chance or caprice that is rightly castigated. But the freedom alleged on the basis of self-activity is not a freedom in which 'anything may happen.' If we consider now what exactly is claimed for 'effort of will' we shall see, I think, that although absolutely hostile to complete continuity of character and conduct, it is not incompatible with the substantial degree of continuity which is all that the observed facts can justify us in inferring.

There are two main points to be observed in this connection. In the first place, the function claimed for effort of will is that by it we can reinforce a weaker but higher desire against a stronger but lower desire, bringing about thereby action in the line of the former. It is only in relation to an end already desired - although felt as desired with insufficient strength in relation to its value on the one hand, and to opposing desires on the other - that effort of will is held to operate. But this at once limits the field of the 'possibilities' in the way of action which belong to a freedom of the type defended. Action, although free, will be limited to those courses of action which the interests, guided by the intelligence, of the agent suggest to him as possible modes of self-satisfaction. Freedom exists only within this prescribed area. Thus it is just as inconsistent with our view of freedom as it is with the observed facts to suppose that an act can take place which is void of all continuity with the conative tendencies of the agent. The continuity admitted will, it is true, justify prediction of conduct only of a very general order. But it will justify that. Knowing within broad limits what kind of interests a person has, we are entitled to predict, within broad limits, what kind of response he will make to a practical stimulus. We may predict, for example, that the typical Philistine will not respond to a sudden access of wealth by endowing a Chair of Fine Art.

That is the first point. The second is no less important. Taking our guidance, as before, from the experience of effortful willing itself, we are aware that there is a gradation of situations from those in which only a slight and relatively easy effort of will is required to reinforce successfully the higher but weaker desire, through those in which greater and more difficult effort is required, up to those in which we feel that the effort required is so tremendous as to be well-nigh impossible. Now this gradation is determined by the ascending degrees of strength of the tendencies felt as opposing the weaker, higher desire. So that what we are recognising here, from the point of view of the freedom-claiming act itself, is that it is far more difficult for an agent to achieve the higher end, where it conflicts with tendencies that have become strongly entrenched in his nature. Thus where, for example, the powerful forces of a deeply engrained habit are opposed to the right course, we shall not expect the agent to follow that course. He may, we believe, but he will be a 'moral hero' if he does; and since we know that most men are not moral heroes, we realise that the probabilities are all against it. In this way approximate prediction is once more possible, even from the standpoint of freedom. The Libertarian is not compelled by his theory to ignore the enslaving chains of 'bad habits.' He will differ from the Determinist in his attitude to what look like notable discrepancies between formed character and conduct only in that while for him they are unlikely, for the Determinist they are impossible. He, like the Determinist, will be ready to explore more closely the inner recesses of the agent's personality with the expectation of finding that the discrepancy is at least less acute than appeared on the surface.

When these two considerations are duly weighed, the criticism that Libertarianism involves a rupture of continuity between character and conduct must, at least so far as it relates to our version of Libertariansim, be greatly modified. It must be so modified that it becomes no more than an allegation of what every unsophisticated person actually believes of himself, viz. that he is able to overcome the bias of his formed tendencies, but that this is more and more difficult for him the more powerful are the tendencies, and the more remote from their direction is the higher course in whose interest he feels it right to subordinate them. This is not a rupture of continuity which can legitimately be complained of on the score of incompatibility with the facts of observation. The facts of observation do not oblige us to assume a whit more continuity than this theory allows, although they do, I think, oblige us to assume as much. It is true that, so far as observation is concerned, we can adopt as an hypothesis the view that there is complete continuity, and that apparent 'breaks' would disappear for a fuller knowledge of the relevant facts. But the evidence of observation does not compel this hypothesis, and since the evidence of inner experience directly denies it, I sec no good reason for supposing it to be true.