Suppose you took the view that someone's choice, say Toby's choice to try to leave his secure job, was a real effect — but of something or other that definitely was not an effect itself. Maybe this thing was the forming of an intention, a forming of an intention to choose as he then did. It was a kind of pre-choice. Whether the pre-choice took place an instant before the fateful choice itself, or a long time before, this view wouldn't be determinism but more or less the opposite.

Shall we say that what Initiation Determinism comes to with respect to a particular choice is that it was an effect and everything that led up to it or was in the story behind it was an effect? Well, that would be something of the right sort, but when you think about it, it is more than a little unclear what led up to a particular choice — what it is for something to lead up to a choice.

We can do better. Let us avoid the difficulty by saying, in a way, that all choices and other conscious events are effects of heredity and environment. To make our theory more explicit we need the idea of a causal sequence or chain. The shortest complete one of these consists in an effect preceded by a causal circumstance and that causal circumstance preceded by a still earlier one for it — for all of it. Most sequences we think about are longer, chains with more links.

Everything within any causal sequence, which is to say everything but the parts of the initial causal circumstance and the final effect, whatever may be true of them, is both an effect and a cause. There aren't any gaps. Thus it is true of the sequence that the final effect was made necessary not only by middle or intermediate circumstances but also by what we are calling the initial circumstance. When you get the initial circumstance the end of any sequence is settled.

Although we are not likely to think of the possibility at first, the parts of a causal circumstance, even an initial causal circumstance, can occur at different times, even very different times. To go back to the lighting of the match, suppose we first describe a causal circumstance for it in terms of the match's being dry, in oxygen, struck, and the surface it is struck on being of the right kind. We take these four things as being at the same time, the time of the striking. But we get another perfectly good causal circumstance if we take the first three items together with something earlier that guaranteed that the surface was of the right kind at the time of striking.

The causal sequence for any choice or any other conscious event, we say, is such that the parts of the initial causal circumstance occur at different times. They are also of different sorts. Some of the earliest parts are neural facts and other bodily facts just prior to the very first mental event in the history of the person in question. The other parts of the initial causal circumstance are events in the person's environment then and thereafter, probably right up to the time of the choice or whatever.

To have a definite idea of these environmental events, we can restrict them to items that affect a person directly, not through intermediaries. So each of these items is the last item in some environmental story, presumably causal, and this item's immediate effect is a bodily or neural event of the person in question. Suppose Toby's saying something shirty to his boss makes his boss feel that he may have to take a difficult decision about him. Our initial causal circumstance for his feeling this at the moment doesn't include all of Toby's history leading up to his words, but only the very last bit.

What we arrive at, then, is the idea that each choice and other conscious event is the effect of a causal sequence whose initial circumstance has in it neural and other bodily events just before the first moment of consciousness of the person in question, and what can be called last environmental events then and thereafter.

Certainly, the sequence for almost any choice will be complicated beyond tracing. For a start, there will be a multitude of conscious events within the sequence. A lot of them will have to do with learning. There will also be a multitude of things so far unnoticed, mental dispositions. These are dispositions to think or feel this or that, and are best regarded as persisting neural structures. In my view they are what can properly be meant, by the way, by evocative talk of the subconscious or the unconscious mind.

However impossible it is to set out all the links of a single causal sequence, we have a clear idea of what this explanation of the mental event comes to. We don't have to know the details, any more than we have to know the details to understand the claim that what crushed the daisy in the valley was the avalanche that started up on the mountain.

As with Mind–Brain Determinism, it will help to have in mind a particular idea of consciousness and the brain, the Union Theory. What it comes to is that conscious events, events of subjectivity, are nomic correlates of simultaneous neural events. Here is a model or diagram of Initiation Determinism in terms of this theory. Or rather, a model or diagram of a very small part of such an Initiation Determinism.

FIGURE

M1 is Juliet's mental event, her feeling at a particular moment about having met a man called Toby. It goes with her neural event N1. As for E1, it is some last environmental event which comes a little later. The event El together with the pair made up of M1 and N1 are the two parts of a causal circumstance for what comes later, the pair made up of another mental event M2 and another neural event N2. Maybe M2 is a happy feeling of anticipation on Juliet's part. The model leaves out everything before the time when M1 happened, and so doesn't come near to showing the very early neural and bodily events in the long causal sequence.

This is a good moment to glance at something distracting that goes against all of this. It is a picture of the mind which admittedly comes from true things we say. We say some mental events are effects and some are causes. Seeing something is an effect and deciding something is a cause. We also say that some mental events are both effects and causes. These beliefs can in fact be explained in terms of the picture of the mind we have been developing, but some philosophers have done otherwise. They have come up with something called Interactionism (Eccles and Popper). Here is a little model of a version of it.¹

This Interactionism is not easy to take seriously, for several reasons. It is supposed to be the full story of this bit of Juliet's history. If it isn't, and we add to it, we are likely to get the sort of thing we ourselves have been developing. If it is the full story, M1 was a free-floating or ghostly event. What I mean is that it was something mental that was unconnected to any neural event at the same moment. That is about as hard to believe in as ghosts themselves, which if they existed would also be free-floating mentality.

Another problem is the explanation of N2. That was a neural event which came out of nothing neural. At the moment before N2 — at the moment of M1 — there was nothing happening neurally. That is not a promising move in theory-construction, and is hard to swallow. We don't think there are gaps in a brain's history, gaps in the history of a part of a physical body. If our present business is setting out and clarifying a theory, whether Initiation Determinism or Interactionism, we can't help but do so with an eye on probable truth.

Interactionism as sketched is a bad determinism, which must not delay us. What really needs attention is the kind of explanation of mental events that is radically different from all of what we have been considering so far. This different kind is in terms of Free Will and is indeterministic, not a matter of causation as we know it. It has often been assumed and talked about, and it has also been set out in full complexity by industrious philosophers. In fact, they say more than anyone is likely to have the fortitude to be able to consider.

A Free Will theory may have a source in religion, now American religion in particular (Ekstrom; Boyle et al.). It may have a source in politics, perhaps a conservative politics that is keen to credit people with a certain right to the considerable amount of private property they have, and to leave others without such a right to any more than the lesser amount they have. A Free Will theory may also have a source in something more widely shared, a desire to give us humans a certain dignity—a standing above the rest of what exists, including what the Free Will philosophers in question would not be inclined to describe as the other animals (Kane 1996).

But the main guiding aim of a Free Will theory is likely to be a related one. In fact, it enters into the three already mentioned. It is the aim of getting to a conclusion from which it will then follow that we can be taken as absolutely responsible for our choices and our ensuing actions. It is to get to something such that it will follow that we can be taken as responsible in a certain way. That is all-important. As noticed in passing at the start, there is more than one thing that can be called being held responsible for things or being credited with responsibility for things.

FIGURE

M3 is Juliet's mental event of seeing Toby on waking up one morning. M4 is her deciding a moment later to say to him that they should have a child together. This conscious event M4 will turn out to be what the model is really focused on, a free decision. The later item A is her still later action of actually saying to him that they should have a child. N3 is a neural event at the first time, and N4 a neural event at the second time, leading towards the action A. That leaves S, which is Juliet's self. It might be better to call it a Self. Certainly it is very different from the other items in the model. For a start, it is not an event, not something that occurs at a time. Rather, it is an enduring thing, moving through time from the first to the last moment shown.

The model leaves a lot out. It gives no indication of the kind of connection between the associated neural and mental bits. It gives no indication of the kind of connection between the neural bits themselves. It doesn't say anything either about the connection between the enduring thing S and, in particular, Juliet's piece of deciding, M4. The idea is that the first of these two things, her self, somehow gives rise to the second, the deciding. We will have to try to fill the model in.

Free Will sounds like something out of the past, as it is, but it can be brought up to date. Its supporters do this. So, as I have already implied, they do pay attention to neuroscience and neural events. They then face the question of the relation between mental and neural events, and in particular between M3 and N3 and also M4 and N4. They have two possibilities.

To speak in a loose way, they can deny necessity or nomic connection here and thereby allow for Free Will, or they can deny necessity or nomic connection at another point. Their better or least bad option, I think, is not to deny it here. This has been their common habit since they have paid attention to the brain and neuroscience, and it continues (Searle 2000; cf. Eccles and Popper).

Let us then suppose they say that each mental event is in nomic connection with the neural event at the same time. That is the one feature that this picture of the mind shares with our determinist one. But if M4 is tied in this way to N4, and the aim is to make Juliet responsible in a certain way for M4, we need to regard N4 as something other than an unavoidable effect. It can't be a real effect of N3 or anything else, or we will make its partner M4 unsatisfactorily inevitable

. This is the stage when we bring in an interpretation of physics, or rather an interpretation of the part of it that is Quantum Theory. It is in fact an interpretation of some mathematics, a way of saying what the mathematics could come to in terms of the actual world. We will be returning to the subject. What is important now is that the interpretation supposes that there are things that happen that are not effects but which are made probable by what happens before them. So we say that the neural partner N4 of the decision M4 is not an effect. It didn't have to happen. It was just something made probable by what went before, in particular the neural event N3, which went with Juliet's seeing Toby.

We now come on a first large problem. That the theory takes neural events to be made probable by antecedents is not just a case of its supporters granting something of what seems true. It is not just a necessary bow on their part in the direction of neuroscience. That is, it is not just a concession in the direction of real causal connection.

There is the reason that there has to be a pretty sure connection between N4 and subsequent neural events on the way to the action, A, or else the action A will be made uncertain. There will be too much chance that Juliet's words aren't at all to the effect that they should have a child, but are something else. If there is not a very high probability that items like N4 will be followed by the right other neural events, then actions we fully and absolutely intend will on too many occasions mysteriously not happen.

So the links after N4 have to be pretty tight. But then in factual consistency so do the neural links before N4. That is unfortunate, since the theory needs these earlier links to be pretty loose in order for Juliet to be held really responsible for what is tied to N4, her decision to speak up.

Can this large problem of seeming inconsistency really be dealt with? Maybe it can. Maybe we can tolerate the ad hoc idea that the earlier neural events are not so probable, in order to leave room for Free Will, but the later neural events are more probable, so that our behaviour doesn't involve many mysterious failures and surprises. Even so, we have not got very far in filling out the theory.

Soon after Quantum Mechanics was interpreted as the idea that there are truly unnecessitated events in the world, events not nomically connected with what precedes them, it was supposed that Free Will was thereby automatically saved. It was saved just by these random or chance events. That is, it was supposed that these events by themselves give us Free Will. But of course they don't. To put the matter briefly, I can't be any more responsible in the desired way for a mere chance event and what goes with it than for a necessitated event and what goes with it. A chance event comes out of nothing. I might be less responsible. In fact, until more was supposed by the hopeful philosophers, this was a case of out of the frying pan into the fire.

The theory of Free Will we are trying to put together tries to deal with this. The neural facts, as we have seen, leave it somehow uncertain that the decision M4 will happen. What is supposed to explain that it really does happen is Juliet's self and its activity (O'Connor 1995a; Searle 2000; Eccles and Popper; Kenny; Boyle et al.). So, in terms of the model, we need to know what S comes to and what its relation is to M4.

A couple of things are clear about S. If you want to explain a decision somehow, but not make it into a necessary event, it will be a good idea not to explain it by citing something that happens before it, a prior event that somehow gives rise to it. That will immediately raise the question of the prior event's being caused, and worse, being caused to cause the decision. An ongoing entity, a self or originator, seems better than an event.

Also, if the aim is to hold us responsible now in a way for a past choice or decision conceived as causally unconnected with our brains and characters and everything else, it is at least useful to have something definite on hand now to aim at, something that seems to be of the right sort. We need the right kind of object for our feelings. The same point applies to our having a certain moral credit.

But awful questions arise, first the question of what S comes to. What is the nature of a self? Necessarily, there is supposed to be a lot more to one of these selves than we ourselves had in mind at the beginning of chapter 3 in connection with the character of subjectivity of mental events. We speculated about an interdependent duality within mental events, involving an aspect for which another aspect exists. A self is far from being a mere aspect of mental events. It is some kind of real entity outside of them, with some kind of power with respect to them. What we can all hesitantly discern about a side of a mental event becomes something much larger. It might indeed be better to elevate it from a self to a Self.

To repeat, what is such a self supposed to be? We already have mental events and neural events on hand. What sort of thing is this originator? Is it a third sort? Does our existence involve the brain, and mental events or the flow of consciousness, and also something different from both? So it seems. Of what material is this self? We have an idea of the material of mental events or consciousness, so to speak, and an analysis of the material of neural events, but what about the self? Is it of no material at all?

If the question about its material is just out of place or wrong for some reason, what are we supposed to think about it? If it is somehow mental, in what sense is that? We need some information to be going on with, long before we get to such questions as whether it, unlike mental events, is not tied to but is entirely free of the brain—which, incidentally, would certainly go against what was called psychoneural intimacy.

There is a temptation into which some philosophers have fallen at this point in trying to provide it. It is the temptation to regard a self as a person within a person, a homunculus. There is a temptation to think of S as an inner person deciding M4. But this is terrible.

One familiar and smaller reason is that it seems we will then have to try to give an account of this inner Juliet along exactly the lines of the account we are now trying to give for Juliet, and so on.

Getting rhetorical or trying out deep thoughts is no help either. Nor is a deceptive kind of plain speaking. The latter happens when it is said, as if no more needed to be said, that this talk about a self is just about 'the mind' or 'a person'—in this case the real person Juliet. But plainly it's not all of the mind or all of Juliet, or even much of it, or even close to much of it, that is in question. We need something more definite. We can't just forget that the self is supposed to be able to overcome desires and the like which are surely elements of a person. If it chooses between inclinations, it doesn't include them.

In short, we have a second problem, still larger than the one about inconsistency in probabilities. It is a problem of clarity. All we can get hold of with respect to a self or originator is that it is in a kind of relation to decisions — we are told that it originates them. It is safe to say that no one has ever begun to answer the question of the nature of a self or originator. In fact no advocate has really faced up to the problem. All have slid by it. It seems to me likely that all future announcements of a self will be like the traditional and the recent ones (Reid; O'Connor 1995a; Clarke; Rowe).

So there are two large problems, one about inconsistency in matters of probability, and one about clarity in talk of the nature of the self. There is a third problem, related to the second, and such as to make the second worse. It is another problem of clarity. It is all very well, when asked about the relation between an originator and a decision, to say that it originates the decision. What does that mean? We do need to know what the connection is between S and the decision M4 in the model.

It is downright embarrassing to hear that the originator looks over the brain and selects neurons for activating in order to get what it wants, or rather the mental events it wants. This turns up in a large book on Free Will by a philosopher and a neurophysiologist (Eccles and Popper). It does of course involve the homunculus trouble, but that is not all. There is another much larger difficulty.

It is no good using ordinary mental verbs such as 'look over', select', and 'want' in order to try to describe what the relation is between an originator and a mental event. At any rate it is no good leaving the matter there. We want to know what these verbs come to in these uses, what is involved in the activities they describe. We need a general understanding of these activities. A general understanding of such activities is in fact exactly the concern of determinist and indeterminist philosophies of mind. Determinist philosophies understand such activities as a matter of standard effects. What is the opposed understanding — something in addition to the denial that the activities are such effects?

In the earlier discussion of causation in chapter 2, it was allowed that some of us ordinarily talk of our choices and decisions as being effects but would say they are not necessitated — not standard effects. And earlier we noticed various non-standard ideas of effects sometimes proposed by philosophers. There were ideas of events as just made probable by previous ones, or events vaguely owed to a vague power, or events having previous ones as required conditions, or events having a 'usual cause'. My own feeling is that most of us will really say, if we can be got to think about it without distractions, that our choices and decisions are best thought of in terms of standard effects. But forget about that.

In order to have a general understanding of the relation between an originator and a decision, let us try to think in terms of one or another of the non standard ideas of causation. It doesn't matter which one. We try out the idea, as various philosophers have, that the originator by itself in some non-standard sense causes mental events to happen. We noticed this possibility at the beginning of the chapter about causation. To say the least, problems arise.

But if the originator in the Juliet story was the same from start to finish, why did it somehow cause her decision M4 when it did, rather than at the earlier time of M3 or the later time of the action A? Why wasn't it always causing it throughout its entire career? Indeed, is there any point in saying that it 'caused' it in any sense when it occurred? Surely this causal language is so different from the standard one that it is baffling and as good as incomprehensible until more is said.

Moreover, if the originator is unchanging, how can it be said to cause endless numbers of different things, endless different decisions, as of course it is supposed to? What is the point of saying it caused Juliet to decide as she did when it might as well have resulted in her deciding the very opposite? That fact is fundamental to all non-standard causation. Remember that nothing made Juliet's decision happen, nothing necessitated it. The self didn't guarantee or ensure that decision. This last and overwhelming objection, by the way, is also one of several objections to something you may be tempted to think about, an originator that does change or develop over time.

One more remark here. Some of the philosophers who try to save Free Will by way of a self make use of a certain old idea taken from religion and theology. They even make use of the idea to try to define or explain the nature of a self. It is the idea of God as self-causing or cause-of-himself. What the philosophers of Free Will say about the self is that it, and certainly not an event in or of it, causes itself to make a decision — or something like that. It is hard for me to believe that time needs to be spent on trying to make sense of such stuff, wherever it turns up. Surely it does not make as much sense, even, as talk of causes as things that make other events probable, talk of causes as only required conditions, and so on.

My reason for this impatience is that causation as we know it, whatever it is, and all ideas of causation other than this one of self-causation, involve a relation between two things, a dyadic relation between two non-identical things. That is what you can call an axiom about causation. So if someone talks of exactly C causing exactly C, they are not talking at all of the sort of thing we know something about. They need to start explaining from the beginning, rather than assume they can start with a certain understanding on our part.

Can this bundle of problems be dealt with by making a certain move? Can we help out advocates of Free Will by offering them the use of our clearer ideas of causation? Of standard or ordinary causation?

In fact, have at least some of us not implicitly been thinking of the originator as something like a causal circumstance? Certainly this is natural. It avoids some difficulties, since it is satisfactorily impossible to think that a whole causal circumstance persists through time but has its effect only at one moment rather than another. It is also satisfactorily impossible to think that a causal circumstance can give rise to a multitude of different things, and at one moment to opposite things. But despite advantages, we simply cannot take the originator as a causal circumstance. It doesn't guarantee anything. Is there another possibility? What if we think of an originator as just a cause, just a part of a causal circumstance?

We now have an idea of Free Will, very different indeed from the model we have been considering. The originator becomes one constant or ongoing element of various causal circumstances for decisions. The other items in each particular circumstance might be different mental events, including desires, inclinations, and so on. We certainly face a difficulty. If such a causal circumstance really is a causal circumstance, then one particular decision and no other has to be the upshot. If it is to serve the ends of a Free Will theory, we are going to have to make the new story very different, and very mysterious.

It will have to be that when everything else is in place for a decision, it is up to the originator whether or not to pull its weight. Certainly an originator has traditionally been thought to be able to defeat desires, go against the person's whole nature, rise over the past, choose the path of duty, and so on. But then we seem to be back with the originator as decisive by itself, again something like a causal circumstance — which in fact it can't be. We get pretty much the same bafflements as before. How could what is unchanging give rise to something or its opposite? What is this obscure activity?

We do not get out of these various difficulties about the relation between an originator and decisions by turning to seemingly different conceptions of an originator, perhaps as a faculty of the mind called the Will, or an active power, or the Self-Conscious Mind. If is not clear what these entities are, but it does seem that the different ways of talking simply inherit the various difficulties. The Will, for example, is sometimes said to be a rational disposition (Kenny 1975). Such a thing has the power to produce something all by itself, but it may not. But then what are we to understand about its working when it does work? Why does it work at one time rather than another? How can it give rise to a different and opposite decision from the one it does give rise to? All this seems to be the same mystery again.

One summary of the mystery is that we are given no explanation of why or how decisions and the like are supposed to come about, and thus given no content to talk of their being in our control, or their being such that we are responsible for them. The sad fact is that these theories seem to fall back into being what we noticed earlier, and what they must of course try to add to, the idea owed to Quantum Theory that decisions and choices are a matter of mere chance events. This falling-back does indeed wreck responsibility and dignity and special rights and so on.

In my opinion the least embarrassing response to the request for an explanation of the relation between an originator and decisions is that an explanation cannot be given. We have to regard this relation as primitive or unanalysable. The situation is to be taken as like one with a language or a logical system. Not every term can be broken down into others — there has to be at least one term that is taken as clear without being explained in terms of others. We start with that and explain other things in terms of it.

What this response comes to is that there exists a relation between an originator and a decision such that the person in question can in a way be he responsible for the decision. More can be said about what it is to hold a person responsible, but nothing can be said about the relation itself. We just somehow understand it. This may seem suspicious, or a cop-out, and of course it does not help with the two earlier problems, about inconsistency in connection with probability and the very nature of an originator. But it is a possible position. Something like it it is held by a very acute philosopher who has been to the fore among Incompatibilists. Freedom is incompatible with determinism, he said, but this freedom remains a mystery (van Inwagen 2002).

After all this, it may come as no surprise, despite the temptations of the idea of an originator, that there are philosophers of Free Will who have given up entirely on the idea. They argue for origination without an originator or self. They have no truck with a non event or unchanging substance somehow giving rise to choices and decisions. They have nothing to say about self-causation. They take S right out of our model and try to make satisfactory sense of the decision in other ways. The general idea is that you can explain Free Will and save the desired kind of responsibility just by seeing that items like M3, N3, M4, N4, and A stand in certain relations. By such means you can explain why or how decisions come about and are in our control and why we are in a way responsible for them.

Of the various relations mentioned, perhaps the simplest is that a choice or decision, although it is not a standard effect, is owed to a reason. This reason may be identified somehow with a belief, a desire, a combination of those two, an intention, or something of the sort. To go back to the little model, but with S taken out of it, what explains M4, Juliet's deciding to say to Tony that they should have a child, is her reason, M3 — something like the look of him that happy morning. Indeterminism was true of the whole episode. It may be added, bravely, that M4 was not an effect in any sense at all, however special. M4 is explained just by the reason for it (cf. Ginet; Nozick; Searle 2000).

This certainly seems to fail. Speaking very generally, we have two ideas of reasons, the first being of what it is natural to call good reasons. Such a reason is a proposition that makes another proposition true. Take a first proposition that itself has two parts: if I am an uncle then someone is my nephew or niece, and I am an uncle. That is a conclusive reason for a second proposition: I do have a nephew or niece. Of course the first proposition does not cause the second — because, for a start, neither proposition is an event, something that happens. Also, the relation between the reason and the conclusion, as we say, is just a logical one.

Our second idea of a reason is of a belief or desire or whatever, a conscious event, that causes another such event, say another belief or a choice or decision. Such a reason may have as its content a reason in the first sense — a good reason. But the later event does not occur because of that logical relation, but rather because it is the effect of the earlier event. One proof is that the logical relation could exist, and the person in fact think of the good reason, without the second event's happening at all. We aren't always logical.

What follows from this is that if you try to explain an event by citing a reason, you are already in the business of giving a causal explanation of some sort. If you deny that you are in this business, you need to begin to try to tell a whole new story. You can't depend on reasons where they are no more than terms of logical relations. You need an explanation of events.

There are other attempts to make a relation of origination clear. Some use the ideas noticed in connection with an originator. A decision is said to be self-causing. Or it is said to be an effect only in the sense that it was preceded by something that made it probable, or was a required condition for it, or had an uncertain power to produce it, or was something like a 'usual cause'. If we no longer face the awful problems having to do with an unchanging originator sailing through time, we do have other problems with these ideas.

The main one is that these various items simply seem to fail to give us an explanation of the choice or decision. We get no reason to think that the choice or decision is in the control of the person. The simple fact is that we are to understand that the earlier event could have occurred entirely without the so-called effect. That must be true because all that is said is consistent with the assumed indeterminism — there being no causal circumstance for the so-called effect, no necessitation of it.

This is not necessarily the slightly controversial claim that all explanations of events are standard causal explanations. Rather, it can be the proposition that if you give up standard causation, you really do need to supply some other general idea of explanation. You can't leave a hole where there was something before. And you can't fill the hole by giving it the name of being some sort of funny cause. Funny causes, by definition, don't say why things actually happen.

As it seems to me, this is the situation of the philosopher (Kane 2002b) who has laboured most manfully to explain origination without recourse to an originator. To talk of choices and decisions as events in the sense of being probable, he adds a good deal. Choices and decisions are the results of efforts, the outcomes of struggles, the upshots of willings, the resolving of conflicts between duty and desire. They are, perhaps most importantly, self-forming actions.

All this is conveyed to us as if these descriptions themselves are supposed to give us an explanation of the coming-about of the decisions. But if the various verbs and locutions are deprived of a standard causal content, which they must be, and given only some content having to do with probabilities, the choices and decisions remain unexplained. For all that has been said, any one of them might never have happened.

That is not all that can be said at this point, particularly about the matter of probability and effects. Probability is a difficult and disputed subject, as was remarked earlier, even if it is clear that an event's having been made probable by something is not the same as the thing's having been a causal circumstance for it. But this is consistent with another clear and good idea — the possibility that our talk of the probability of an event's happening actually presupposes and depends on there really being a standard causal explanation of the event. This bad news for Free Will philosophy is roughly as follows.

What is it for an earlier event A to have made it 95 per cent probable that a later event B would occur? A good idea about this has to do with the fact, of which you have already heard, that typically we don't know exactly and fully what is in a causal circumstance for B. And that typically we have a pretty good idea. We know in what situation a causal circumstance tends to occur.

Suppose it has been our experience that in 95 per cent of the situations in which event A occurred, it was followed by B. We say B is 95 per cent probable with respect to A, and what this means is just that in 95 per cent of the situations in which A occurs, there is precisely a causal circumstance for B. That leaves B as just probable with respect to A, and not the effect of A as a causal circumstance. But the fact of probability simply presupposes that B is the standard effect of something. As I say, bad news if you want to put probability together with indeterminism.

Let us notice just one other attempt to make sense of origination. Some advocates of Free Will, including the philosopher lateiy mentioned, have said that decisions are explained teleologically, that is, in terms of their goal. They are explained by what they lead to. This ancient line of thought is owed to the fact that we can indeed say things like this: Birds have hollow bones because that enables them to fly better and We perspire because that reduces our bodily temperature. But this talk, as almost everyone agrees, ca of really give us the conclusion that effects by themselves explain their causes. That seems to be an astonishing idea.

Attempts have been made to make teleology less astonishing (G. A. Cohen). They have not succeeded, and nothing is going to get us to agree that the occurrence of a decision is explained just by what it results in. In connection with the birds and our perspiration, what will come to mind is an evolutionary story, which really is standardly causal. To say birds have hollow bones because that enables them to fly better is to say there is an evolutionary explanation of the hollow bones — some types of creature have survived because of the advantage to their predecessors of their hollow bones. If we turn the Free Will theory's teleological explanation into some standard causal one, however, we will defeat its main purpose (Honderich 1982).

Last but not least, it is likely that a Free Will theory really cannot get rid of the embarrassment of an originator. It has to have something that is going to be responsible. A past decision itself, whether it was probable or self-causing or teleological or anything else, isn't what we hold responsible for actions or give a kind of moral credit to for actions. If a philosopher says it is not a person in an ordinary sense who is responsible, something of certain traits, desires and so an, he will indeed need to offer us something more than a choice or decision in certain relations. We don't put past decisions in jail either.

Have I been too hard on the philosophy of Free Will, too judgementaI? Well, have a look for yourself at efforts to set out clear, consistent and complete accounts (Kane 2002a; O'Connor 1995b). As for our project of setting out a determinist philosophy, we need to finish it by looking at the relation of conscious events to subsequent actions. We will then come to a final judgement about the clarity, consistency, and completeness of the two philosophies, and then really look at the question which has already been pushing in, their truth.