by Ted Honderich
-- The Determinism and Freedom Philosophy Website --

How do our thoughts, feelings, choices and actions come about? In what follows here, the two kinds of traditional and still orthodox explanations are considered. The fundamental proposition of a defined and developed theory of determinism is laid out and compared with various ideas of free will or origination. This is Ch. 3 of Ted Honderich's large work A Theory of Determinism: The Mind, Neuroscience and Life-Hopes -- which is also Ch. 3 of the paperback Mind and Brain. In a nutshell, the determinist line of thought is that events of consciousness go together with brain events to make up psychoneural pairs, which pairs are effects of certain causal sequences. This is different from supposed explanations of mental events considered earlier, including the common view that there is something called interaction between mind and brain. The determinist line of thought is fundamentally different from those explanations of our existence considered after it -- indeterminism, free will, origination. The  determinist explanation derives from and depends on a preceding partial account of the relation of mind and brain, Mind Brain Connection. Also an earlier account of causal and lawlike connection in general, Causality or Causation, the Fundamental Fact Plainly Explained. For details of the books and other writings referred to, go to References. The sections of the inquiry below are as follows:






       3.6.1  Self-Causing Decisions
       3.6.2  Self-Justifying Decisions
       3.6.3  Teleology

       3.7.1  Homunculus
       3.7.2  'The Mind'
       3.7.3  Selected Elements of 'the Mind'
       3.7.4  The Person
       3.7.5  Originator

       3.8.1  Mental Verbs
       3.8.2  Self-Causing Cause
       3.8.3  The Originator Itself as Causal
       3.8.4  The Will
       3.8.5   Primitive Relation


The question we have been considering was expressed in this way: How is mind related to brain, or, what is the relation between mental and neural events? More particularly, we have been considering the question of exactly how a mental event is related to the particular neural event to which it is intimately related (p. 90) and thus with which it is simultaneous. The Identity Theory in its various forms attempts an answer to this question of the psychoneural relation. The Correlation Hypothesis gives a better answer, but one which will be supplemented. It will be supplemented, thereby giving us what will be called the Union Theory, in the course of our considering the question to which we now turn. It is this: What exactly is the explanation of mental events? In this diachronic rather than synchronic question our concern is with facts prior to a given mental event, those which provide an answer to the question of why it occurred.

Answers to the first question partly depend for their worth on, and can be tested by, what they include or entail by way of answers to the second question. Hence, in considering Identity Theories of the psychoneural relation, we took into account their entailments with respect to the explanation of mental events. We thus did in effect consider certain answers to the diachronic question of the explanation of mental events. The Correlation Hypothesis, however, was not considered in terms of its implications for the question of the explanation of the mental. The answer to be given to that question in this chapter, as will be anticipated, consorts with the Correlation Hypothesis—indeed supplements it—and raises no problems for it.

One criterion which must be satisfied by any arguable answer to the question of the explanation of the mental, as we have safely assumed, is that the answer must satisfy the conviction or axiom of mental indispensability: earlier mental events are essential parts of the explanation of typical mental events and of actions, (pp. 91 f.) A second criterion, as we also know, is psychoneural intimacy: mental events co-occur with neural events in particular neural locales as a matter of some sort of necessity, (p. 90) A third criterion, since not relevant to the question of the psychoneural relation, has so far not been considered.

Take any decision or choice, above all any decision or choice reached or made in a reflective or deliberative rather than a spontaneous way. If it would be intolerable to leave earlier mental events out of the explanation of a person's choice, it would be as intolerable to leave out something else—the person. Certainly I am convinced that I am part of the, explanation of my resolution to resist certain temptations, perhaps to put certain things out of mind. This conviction of personal indispensability, as we may call it, evidently is in connection with the conviction of mental indispensability, but is distinct. Consider a person's decision, and the earlier mental events taken as indispensable to it, say a certain perception and desire. We can feel no assurance that earlier mental events of roughly the same types, or of the very same types, if they occurred in the life of another person, would have been followed by the given decision. Evidently, and to say the least, the person makes a difference.

What is a person? The question most often addressed in this neighbourhood is that of what makes a person at any earlier time identical with a person at a later time. What are the logically necessary and sufficient conditions of the fact that the person who wrote earlier pages of this book is the same person writing later pages? Presumably the correct answer to this question, although it sometimes seems to be implied otherwise (Parfit, p. 202), will contain an answer to the question of what a person is. To know what is logically necessary and sufficient for something's being the same person as another is to know what is logically necessary and sufficient for being a person at all. I could perhaps be sure something is a person without being able to identify it with any previous person—having the concept of a person may not give one a criterion of personal identity. But having the correct criterion of personal identity, which is to say an account of what is necessary and sufficient to make x now the same person as y before, is surely to have the concept of a person.

The controversy about personal identity (Parfit, 1984; B. Williams, 1973; Perry, 1975) has perhaps not been finally resolved, and will not be resolved here. What is to be said here will be consistent with what can be taken as the arguable views. These do not include the view that a person is to be identified with an ongoing simple mental substance. The best-known version of this idea is of course that of Descartes, which makes the substance into the non-spatial pure ego. Any such idea has its origin, no doubt, in our perception of what was earlier called a subject, one part of the linked duality which is the nature of any mental event. (2.2) There is no adequate reason to elevate or reify a subject into an ongoing mental substance. The objections to any such idea are several (pp. 80 f.), as are the objections to any such idea's being adequate to our conception of a person. The principal objection of the latter kind is that the idea of a simple substance taken by itself entirely leaves out continuities of experience. It leaves out, that is, what is sometimes called psychological continuity and connectedness or mental relations. Such a simple substance, if one existed, might exist as or in an entity of no memories, character traits, or ongoing feelings. If such things are somehow taken as added in, we do of course have something quite different from the official view.

Whatever the correct account is of what makes the writer of the earlier pages identical with the writer of this page, the account must somehow include continuities of experience. It is essential to the fact of identity, as it is to the nature of a person at a time, that he is the individual distinguished by certain memories, dispositions of charac-ter, tendencies, persistent inclinations and desires, hopes, commit-ments, intentions, plans, and the like. It is no doubt true that my identity is not dependent on my settled political convictions, or my uncertain optimism, or my memory-image of a farm in a dell on the road to Kitchener. It is inconceivable that my identity could be independent of all or perhaps even most of my memories, tendencies, hopes and so on.

The arguable views of personal identity respect this fact to one degree or another. The better known views, very broadly speaking, fall into two categories. The first actually centres on continuities of experience. By far the most impressive account of personal identity available (Parfit, 1984), which may in fact be true, is at bottom an account in terms of continuities of experience.

"We are not separately existing entities, apart from our brains and bodies, and various interrelated physical and mental events. Our existence just involves the existence of our brains and bodies, and the doing of our deeds, and the thinking of our thoughts, and the occurrence of certain other physical and mental events. Our identity over time just involves (a) Relation R—psycho-logical connectedness and/or psychological continuity—either with the normal cause or with any cause, provided (b) that there is no different person who is R-related to us as we once were." (Parfit, 1984, p. 216)

The other category of views gives a fundamental place to bodily or brain continuity. I today am identical with the writer of a previous page of this book in virtue of a spatio-temporal line between an earlier living body or brain or brain-part and a present living body or brain or brain-part. It seems true, although perhaps in need of more argument than will be provided here, that this category of views in fact depends for its considerable recommendation on its at least implicit inclusion of the idea of continuities of experience, taken as rooted in the body or brain. If the stuff of my brain persisted despite the total absence of such continuities, I would surely not persist. If it is conceivable to suppose that the organization or structures of my brain might also persist in the total absence of such continuities, surely I would still not persist.

It may be that both of these categories of views, and also views which somehow amalgamate the categories (Nozick, 1981, Ch. 1), require an addition. Perhaps the addition will make for a further category of views. It has to do with an elusive continuity of consciousness in a sense so far unmentioned. By way of minimal and less than helpful description, a person is a continuing locus or tunnel or frame of consciousness or mental events. This seeming fact, which stands in some relation to the subject within a mental event, and which is not to be dramatized into a mental substance or ego, would account for what has often enough been considered by philosophers, a particular apprehension one would feel at the prospect of, say, future pain—after and despite an ending of one's continuities of experience. The apprehension has to do with the pain, not the loss of the given continuities. The seeming fact carries the conclusion that there is more to an ongoing person than continuities of experience with one or another cause.

The possible addition to views of personal identity, or third category of view, is not a matter to be pursued here. What is to be concluded, rather, is that answers to our question of the explanation of mental events must indeed satisfy a criterion in addition to the criteria of mental indispensability and of psychoneural intimacy. It is the criterion of the indispensability of the person, or of personal indispensability. It is that a person, conceived as an entity involving at least continuities of experience, is essential to the explanation of at least typical mental events and actions. My thoughts'and desires, and my actions, are in general owed to such things as my memories, character, hopes, and intentions. A decision or act of mine is owed to more than the obvious mental events which precede it, typically a certain desire and belief.

So much for one preliminary to consideration of the question of the explanation of mental events. The second, quite different, has to do with causation. The answer to be given to the question of explanation, and other answers to be considered, and also the answers looked at above (2.3)—although we did not pause over the matter—have to do with causal sequences. These are related to what are ordinarily called causal chains. The car stopped, we may carelessly suppose, because of a causal chain or chain or events which began with the driver's stepping on the brake pedal. The vague picture we may have is of some single line of mechanical events involving the pedal, mechanical connections, brake drums, and the like, linking the driver's action with the stopping of the car. However, as we shall of course grant, more was needed than such a single line of events. Had the car been coming down Highgate Hill, and it was icy, the car would not have stopped as it did, despite such a line of events.

A causal sequence is a sequence of events or conditions—individual properties or collections of them, strictly speaking—each of which, save those at the beginning or beginnings and at the end, is both an effect and also an element in a causal circumstance. Thus there are no gaps. Figure 9 is of a causal sequence for event 67, perhaps the stopping of the car. ei might be the driver's braking. The items in it are of course particular events or conditions rather than types, and are to be taken as having occurred at the given times. Wherever a line joins an earlier and a later event or condition, say e, and £.4, or &i and e-,, the earlier was cause of the later. That is, in part, if the earlier had not happened, the later would not have—the earlier was required for the later. The situation in question was then an ordinary one, not involving items which were only alternatively required for others, (p. 19)

To recall our earlier brief description of a causal circumstance (p. 47), we took it that one consists in no more than a set of events or conditions such that the set (1) necessitated the effect, (2) preceded it, and (3) was dependently-necessitated by it. In virtue of these three conditions it can be said that the set made the effect happen and explained it. Figure 9 is to be taken in its converging lines as specifying two different causal circumstances for e7, and one for the interim effect e4. The rule for interpreting such diagrams is partly and roughly as follows. We get a causal circumstance for the culminating effect by moving leftwards along all of the lines from that effect to the first set of events or conditions, which is one causal cirumstance for it. Thus e4, e5, and e6 was one causal circumstance for e7. We get a further causal circumstance for the final effect by omitting one of these events or conditions, and substituting all of the events or conditions in the causal circumstance for the omitted event or condition. Thus e1, e2, e3, e5, and e6 was a second causal circumstance for e7. As remarked in our earlier account of a causal circumstance (p. 20), it is not necessary that all its elements be simultaneous. The only other causal circumstance in Figure 9 is of course ei, e2,, and e3, for the interim effect e4.

Each item in the diagrammed sequence, save the last, is given as an element in one or more causal circumstances. The only two effects are e4 and e7. The diagram does not convey, about the other items, either that they are or that they are not effects. The question is left open. All events and conditions save those just mentioned, 64 and 67, can be labelled initial elements, which is to say that their causal back-grounds, if any, are not within the specified causal sequence. It may be, of course, that they are effects in terms of another causal sequence, perhaps one which includes the one we have and also has 67 as culmination. It is at least generally a matter of choice, governed by our own interests, what the initial items of a causal sequence are taken to be. There is also choice, evidently, in the matter of the final effect. We pay attention to a course of events up to a certain natural point and not beyond, leaving open the question of whether the final effect also has a causal role.

Whether or not we wish to pay attention to the fact, is every event or condition in the universe an effect? That is not our question here, and it will not be a fundamental question at any point in this inquiry into determinism. We shall not be concerned with universal or LaPlacean determinism, which asserts that every event or condition without exception—as we may say, every individual property—is an effect. Our fundamental concern is not with all events or conditions, but only with those, to speak quickly, which are of direct relevance to our own lives. That is, we are concerned with those which give rise to feeling, intention, decision, choice, action, and the like.

A causal sequence, then, putting aside its initial and final elements, consists in events or conditions each of which is both an effect and a part of a causal circumstance, or of several such circumstances. We might go considerably further in elaborating the account of a causal sequence, but there is no need to do so.

Our ultimate concern will have to do with one principal fact which certain causal sequences share with all others. It is that the final element, and all elements save initial ones, are necessitated events or conditions. All elements but the initial ones are necessitated by the initial elements, and typically also by circumstances internal to the sequence. Hence, and above all, the final effect of a causal sequence is necessitated by its initial elements.

A second important concern will have to do with the fact, about causal sequences however defined, that each earlier circumstance in an ordinary sequence is dependently-necessary to all that follows, (p. 29) This is of course entailed—to recall the summary mentioned above— just by specifying the effect of a causal circumstance as dependently necessitating it. Such facts about necessity, of course, are to be understood as they finally were in our inquiry into nomic connections. For a circumstance to necessitate an effect is for roughly this to be true: given the circumstance, the effect would still have occurred as it did, even given most conceivable changes in the universe.


It is a natural idea that both external or environmental events and also bodily events somehow cause mental events, and that mental events somehow cause bodily and external events. The salt dish I see causes retinal events and these cause the mental events in which some of my visual experience consists. My wants cause my actions and thus their effects, from the state of my shoe-laces to the state of my garden. The natural idea pertains to several subject-matters, including a large one to which we have not yet come, which is that of the relation between mental events and action. The natural idea evidently contributes to a number of answers or types of answers given to the question of the explanation of mental events.

The simplest of them can have the name of Causal Interactionism. It is, as usually expressed, that neural events somehow cause mental events and mental events somehow cause neural events—which latter causal fact gives us mental events, by way of intervening neural events, causing later mental events. It would be unkind to suppose that it makes all neural events causal with respect to mental events, since it is safe to say that many are not. Nor would it be kind to take it as making all neural events into effects of mental events. Many are not. Are all mental events to be regarded as causal with respect to neural events? And all mental events to be regarded as effects of neural events? We might answer yes in both cases, as is contemplated with another doctrine of mind and brain noticed earlier, Anomalous Monism. (2.3; Davidson, 1980e, p. 208] However, we need not settle the matter. Let us simply consider the view that some mental somehow cause some neural events, and some neural somehow cause some mental events.

It is to be noticed that this allows that the given events, of both sorts, may both be causal—causes or causal circumstances—and also be effects. Causal Interactionism has occasionally seemed to be regarded as the view, to speak only of mental events, that some of them are causal and only that, and some are effects and only that. None are both effects and also causal. Mental events that are causal may be characterized as the ones we regard as active. Those that are effects are the one we regard as passive. (Cf. pp. 174 f.) Decisions go in the first category, and sensory experience, or part of it, in the second. The general idea is decidedly odd, to say the least, although in part like things to which we shall come—certain indeterminist doctrines of the mind. Causal Interactionism is better taken as allowing mental events to have the ordinary dual role: both effects and also causal. The view as we have it, then, as already anticipated, also allows a further thing, that a given mental event may be the effect of a neural event and, by way of that neural event, also the effect of an earlier mental event.

Are we to suppose that a mental event is a cause as distinct from a causal circumstance for a neural event? And so with our causal neural events? Let us start with that assumption, which gives us a first instance of Causal Interactionism. What we have is the view modelled in Figure 10, where the arrow with a 'c' above it is to be read as 'is a cause of'. The view as it stands is either unacceptable or incomplete. As it stands, taken as complete, it leaves m (and m') free-floating. That is, m is not given in any relation with a simultaneous neural event. As it stands, to speak differently, the view makes m ghostly: it has the nature of a ghost, in the popular sense of something mental or spiritual not in the ordinary simultaneous connection with a body. More precisely, the view as it stands, taken as complete, offends against the axiom of psychoneural intimacy. It needs remarking, as well, that in m and n (and n and m') we supposedly have a standard case of cause of effect. We thus do also need a causal circumstance for n (and m'. There are no causes without causal circumstances, as we know. (Ch. 1)

We can deal with this last matter by making the other assumption— by taking the view that a mental event may itself be a causal circumstance for a neural event, and a neural event may itself be a causal circumstance for a mental event. This gives us a second instance of Causal Interactionism. (There is no conceptual barrier, of course, to what we call a single event functioning as a causal circumstance.) Or, indeed, we could try something else that would be in the spirit of the present enterprise, and thus get a third instance of Causal Interactionism. We could take it that a mental event may be a part of a wholly mental causal circumstance for a neural event, and a neural event may be part of a wholly neural causal circumstance for a mental event. However, if either of the second or the third variant is taken to be a complete view, it is as unacceptable as its predecessor, for the same reason. It offends against psychoneural intimacy. Mental events are again left free-floating or ghostly.

There is a second objection, not much less serious, to all these three versions of Causal Interactionism. To speak in terms of the first, and Figure 10, it is incredible that the causation of the neural event n includes no neural fact at ti. The objection is an empirical one, certainly, as its predecessor also is, but surely irresistible even in the absence of detailed evidence. It cannot be that n comes from nothing, neurally speaking, at t^. It cannot be sufficient that n can be said to have a neural causal antecedent which is cause of m, at some time prior to tj. While that gives us some neural causation for the neural event n, it leaves us with the neural gap at t-i.

It will be recalled, of course, that we have not excluded mental events from the class of physical events. They are within the physical realm as defined essentially in terms of spatiality. (pp. 87 ft.) Thus, n does have a physical causal antecedent at ti. This does not much affect the objection. Our characterization of mental events, as it needs to, distinguishes them sharply from other physical events, including neural events. With respect to almost all neural events, or at any rate typical neural events, it is impossible to deny them immediately prior neural causes. Just this is done by the views in question.

Causal Interactionism as we have it also faces a third objection, that it does not allow for personal indispensability, but it may be that what we have could be interpreted so as to allow for the role of memory, character, and the like. Let us not pursue this. We can instead take it that the intended views have been incompletely stated. There is more to them. They are to have added to them a proposition about mental events being in relation with simultaneous neural events. This acceptance of psychoneural intimacy, which will also deal with the neural gaps, will then require something more: elaboration of the causal story. The completion or enlargement of the views, and the causal elaboration, will turn them into quite different things. These will be noticed briefly at the end of this section.

If we set out to secure psychoneural intimacy in dealing with the question of the explanation of mental events, and we are also persuaded of a proposition noticed a moment ago and of which there is more to be said (pp. 164f.), the neural events are somehow or other owed to neural or other bodily causation involving no mysterious gaps, we are likely to come to a particular answer. It is one of several answers taken as suggested by the practice of neuroscience, and like Identity Theories. Certainly, as we shall see in more detail in Chapter 5, it is in general true that research into the brain and Central Nervous System proceeds on the assumption that neural events have explan-ations that are somehow neural or other otherwise bodily. The particular answer to the explanation question is also suggested by fairly common usages in neuroscience, and by occasional philos-ophical reflections on the part of those engaged in it—although the reflections are often enough accompanied by others than do not consort well with them. (Rosenblueth, 1970)

The answer, which is also an answer to the question of the psychoneural relation, consists in two things: the Correlation Hypoth-esis and the plain proposition that all neural correlates are elements in causal circumstances within causal sequences which are wholly neural or bodily—none of the elements of the contained causal circumstances is mental. What we have is ongoing non-mental activity, to be explained wholly in terms of non-mental causation, with mental events as nomic correlates of some of this activity. They are correlates in just the sense specified by the Correlation Hypothesis. That is, a neural event necessitates a simultaneous mental event, the mental being necessary to the neural. Also, the mental event dependently necessitates the neural event, the neural thus being dependently necessary to the mental, (pp. 47-8, 107-8)

This view, Neural Causation with Psychoneural Correlation, is as modelled in Figure 11. The vertical bar joining mental event m and neural event n (and m' and n'} represents them as in nomic correlation of the specified kind. Neural event n, and another neural event n*, form a causal circumstance for n', the neural correlate of m'. The model thus represents something of the natural assumption that the occurrence of such a neural event as n' is owed to more neural facts at ti than just the occurrence of the neural correlate of m. The model can be taken as modelling someone's noticing olives on the table [m], and his intending to have one (m'). The view, or something close to it, seemed to me for a considerable time to be the best answer to the question of the explanation of mental events. (Honderich, 198 la, b, 1984d) The fact of the matter is that the view, despite its having neuroscientific and other support, is for a particular reason untenable.

It proposes to make mental events indispensable to the explanation of later mental events and actions by making them nomic correlates. With respect to the indispensability of m to m', we have m necessary to and simultaneous with n, and, by way of the other two nomic connections, we have n dependently necessary to m'. Is the fact that event m is at least dependently necessary to m' enough to secure what we want, that m is an essential part of the explanation of m'? In ordinary cases of causal and other nomic connection, there is not overdetermination. (p. 21) We can take it, safely, that at least most neural and psychoneural connections are ordinary. Thus, with respect to Figure 11, we take it that there existed no causal circumstance for n' additional to the circumstance consisting in n and n*. Similarly, we take it that there existed no neural correlate for m' additional to the correlate n'. In this situation it is true that any full explanation of m', perhaps the mental event of intending to have an olive, includes n'. And any full explanation of n' includes n. And, finally, there is no full explanation of n that does not include m, perhaps the mental event of noticing the olives.

An attempt can be made to maintain, then, that the answer of Neural Causation with Psychoneural Correlation does respect the axiom of mental indispensability. Alas, some of the utilized facts enter into what makes the view no less than self-contradictory.

Event m, as remarked, is necessary to its simultaneous correlate n. So too is m' necessary to n'. But, given this latter proposition, n' in fact simply cannot be necessitated by the causal circumstance n, n*. It is precisely not true of n, n* that since it occurred, n' was guaranteed. More precisely, to recall more exactly the nature of a causal circumstance, the following is not true of n, n*: that since it occurred, whatever else logically consistent with it and with n' had occurred, n' would still have occurred. Event n' would not have occurred if its simultaneous correlate m' had been missing. An attempt was made by me to escape this inescapable upshot by way of an unacceptable doctrine about causation, of which the less said the better. (Honderich, 1981b, cf. 1986)

It is worth remarking, perhaps, that the self-contradictoriness of the view in question is made clear by the particular account of causation and of nomic connection to which we have come, but that it is not dependent on the distinctive features of that account. Almost all accounts of causation include what are often called causally sufficient conditions, akin to our causal circumstances. And any at all arguable account of nomic connection, applied to the psychoneural relation, is likely to make m' in a substantial way necessary to n'. The result is inconsistency.

The short story as we have it about the Theory of Neural Causation with Psychoneural Correlation is that if we have the mental correlates explanatorily indispensable, we cannot have the non-mental causation of the neural correlates. To add the non-mental causation of the neural correlates is to fall into contradiction. The same short story, put the other way on, is that if we have the non-mental causation of the neural correlates, we cannot have the mental correlates explanatorily indis-pensable. Of the general problems raised by the mind, perhaps none is more fundamental than the one of which this is a clear formulation. It is a formulation, of course, which derives from and presupposes a proper rejection of Identity Theories as tenable accounts of the psychoneural relation, and puts the Correlation Hypothesis in their place.

Several related theories have the great recommendation of proceed-ing from such a clear view of the problem. One, which owes much to the course of modern science, above all physics, is the theory of the Psychoneural Relation as a Macromicro Relation. (Kim, 1984) It can be put into terms with which we are familiar. Its basic proposition is that neural events are effects of wholly neural or other bodily causal sequences. The neural realm, however, consists in micro-events of which the macro-events are mental. That is, mental events stand to neural events in a macromicro relation—mental events are in a certain sense reducible to neural events. Mental events thus stand to neural events as, say, events of temperature stand to molecular events, or molecular events stand to atomic events. As a consequence of this nature of the psychoneural relation—a nature of which more needs to be said—and also the previous proposition about neural causation, a certain conclusion is drawn about mental events causing mental events. There is, in some sense, mental causation. The previous view we considered, Neural Causation with Psychoneural Correlation, sought to secure the indispensability of the mental by linking an earlier mental event to a later or to an action indirectly, by way of neural events. The present view, despite what may be a moment of uncertainty on the issue (Kim, 1984, p. 268), proposes a kind of direct connection.

To proceed again through the sequence of argument, we must accept as basic that neural events have only non-mental explanations, that they are effects of prior neural and other bodily events. Nothing, we are told, will pre-empt such an account of why a limb movement takes place, and nor will anything pre-empt such an account of why any neural event takes place. As for the second step, taking the psycho-neural relation as a macromicro relation, this is of course not to identify mental or macro-events with the related neural or micro-events. The macromicro relation is one such that the micro-events are constitutive, simpler, finer, or at a lower level. To say that, however, is not actually to specify the relation.

The relation is so understood as to reflect the general belief that a type of macro-event may go with more than one type of micro-event, perhaps many. To use a common term, a type of macro-event may be multiply realized—realized in or by more than one type of micro-event. For example, fractures or breaks in materials go with many types of events in underlying molecular structures. To come to the nub, any family b of events is macrocausal with respect to a family a of events under the following two conditions: (i) if there occurs an event of one type within family b, there necessarily occur events of one of some set of types within family a, and (ii) if there occur events of any one of the set of types in a, there necessarily occurs an event of the single type in b. The macromicro relation, so specified, is also referred to as a relation of supervenience or dependency. Events of family b supervene on or are dependent on events of family a.

So—to speak more particularly—mental events are macrocausal with respect to neural events, which is to say: (i) if there occurs an instance of one type of mental event, there necessarily occur neural events of some or other type, and (ii) if there occur such neural events, there necessarily occurs such a mental event. Suppose now, as in Figure 12, that neural events n together with neural event n* cause neural events n'. Events n are micro-events of which m is the macro-event, as represented by the vertical line with an 'M' on each side. So with n' and m'. In virtue of the microcosmic causal connection between n and n' we may conclude something about causal connection between the macrocosmic events m and m'. It is not certain exactly what conclusion is proposed to us.

It is said that the causal connection between m and m' is real, that m is not inert. Something along the same lines but more definite—very definite indeed—is also said. It is said that m is as much a cause of m', in virtue of the relations of these to n and n', as wholly familiar events are causes of their effects. Event m is as much cause of m' as fire indubitably is of smoke, which things are also macro-events whose micro-events are indubitably in causal connection. However, it is also said in various ways that the microcausal relations are fundamental. Microcausal relations underlie, ground, and explain the macrocausal ones. Macrocausal relations are in a certain sense epiphenomenal and less real. Ultimately the world is as it is because the microworld is as it is. Above all, to speak in terms of Figure 12, or rather, despite what appears to be conveyed by the figure, it is said there is only the neural causal path. There are not two causal paths to m'.

The first thing to be remarked about this view is that it is difficult to take it as determinate. This is a matter of what has just been mentioned, its conclusion. There are two possibilities. If the view is that there is in fact only one causal path or connection, the one holding between n and n', then the view is fundamentally identical to the one we have just considered and rejected, Neural Causation with Psycho-neural Correlation. They evidently share the proposition that neural events are the effects of wholly neural or at any rate non-mental causal sequences. It is not so immediately clear, but clear enough, that they they share the proposition of psychoneural nomic connection. The macromicro account of the psychoneural relation does certainly imply that mental and neural events are of certain different kinds, such that the neural are somehow constitutive of the mental, which is not to imply that they are identical. What can it mean, to say the more important thing, in definition of the macromicro relation, that if an a-family event occurs, a b-family event necessarily occurs, and if an b-family event occurs, some a-family event necessarily occurs? The matter is not attended to, but there is no room for choice. The necessity is evidently not logical, which rightly is not suggested, and therefore must be nomic. The macromicro relation is fundamentally nomic connection, and, as will have been noticed, its one-many character brings it wholly into line with the Correlation Hypothesis.

It is worth remarking in passing that the theory of the Psychoneural Relation as a Macromicro Relation, while it has the great distinction of starting in the right place, does exemplify a certain weakness of many theories of mind and brain. In the aim of making use of scientific advance and conceptions of it, various mind-brain theories pay insufficient attention to what must be their own more fundamental constituents, such as an idea of necessary connection. Giving these constituents attention is essential. One profit is seeing the weaknesses and strengths of elaborated theories more clearly as a result of seeing their likeness to starker theories and hypotheses.

To repeat, if the theory we are considering allows but one causal connection—between n and n'—it is at bottom identical with what we have discarded. However, as noted, the theory does speak of a causal connection between m and m' which is as indubitable and substantial as perfectly ordinary causal connections not involving the mind. Let us therefore make the other assumption, and remove any question mark about the causal connection between m and m'. Doing so does give us a distinctive view. It is also a view which has great problems.

(i) The least of these is that little reason is given to assert the mental causal connection. We must not suppose, of course, that it is guaranteed by the axiom of mental indispensability. That axiom is not the proposition that mental events are the effects of wholly mental causal circumstances, or of wholly mental causal sequences. It is rather the proposition that prior mental events enter essentially into any full explanations of later mental events and actions. The two are not identical, as will already be apparent. What then is the basis for the proposition that m is causal circumstance for m' ? (To take m as merely cause of m' would of course require us to provide a containing causal circumstance. To make it other than wholly mental would lead us off in the direction of a quite different theory.)

Obviously the intended basis for the proposition that m necessitates m' is the microcausal relation between n and n'. However, it is not clear that it is a basis. We are invited to see the psychoneural relation, and the relation between an earlier and a later mental event, as   . respectively like the relation between, say, temperature and molecular movement, and the relation between temperature and melting. But there are evident discontinuities. I do not conclude that the heat is melting the candle because of a proposition of science about a causal relation between micro-events. The latter proposition is not premiss for the former. Further, any argument from microfacts to macrofacts with respect, say, to heat and molecular motion, is unlikely to carry over to the neural and the mental. Or, to speak more carefully, the relevant likeness of the two domains needs first to be established.
Clearly it has not been. To be brief, mental events are not much like temperature or other macrofacts of scientific theory.

(ii) The view as we now have it, although quite distinct from Neural Causation with Psychoneural Correlation, has precisely the great disability of that view. It is contradictory in virtue of claiming that a neural event has a wholly neural causal explanation, and also claiming that something simultaneous and non-neural is necessary to that event. In Figure 12, we have a neural causal circumstance for n'. Also, the macromicro relation involving n' is such that n' necessitates m', and hence that m' is necessary to n'. If we hold on to the macromicro relations, we must give up the idea that there is a neural causal circumstance for n'. If we hold on to the causal proposition, we must give up the macromicro relations. By giving up the m-n macromicro relation we lose mental indispensability. (It may be supposed, carelessly, that the view preserves mental indispensability by way of the direct causal connection between m and m'. But the axiom of mental indispensability, as just remarked, is that earlier mental events are indispensable parts of any explanations of typical mental events and of actions. And we are supposing that there is a full explanation of a later mental event, m', which leaves out any earlier mental event. We have it that m' is necessitated by n', and n' is necessitated by a causal circumstance which does not include m.]

(iii) Lastly, the view is involved in overdetermination, as unpersuas-ive as overdetermination generally or very often is. We have a full explanation of m' by way of n' and the causal circumstance for it. No doubt there are difficulties in the way of formulating satisfactorily a general principle to exclude suppositions of overdetermination, as there are difficulties in satisfactorily formulating the wider principle of simplicity or parsimony. 'Plurality is not to be assumed without necessity', 'What can be done with fewer assumptions is done in vain with more'—neither of these has sufficient edge. (Moody, 1967) Still, there can be no doubt of the strength of simplicity. It is rightly taken, not merely as an end in itself, but as a means to truth. In part, this is owed to the pragmatic consideration that complexity has the possibility of concealing error. More importantly, it is owed to well-founded inductive beliefs about reality generally.

The theory of the Psychoneural Relation as Macromicro is therefore not a satisfactory answer to the question of the explanation of mental events. Nor are any other views of the same basic structure. To mention but one of these, it centrally involves the ideas that mental events stand in a basic kind of nomic correlation with neural events and hence that mental events somehow share any causal roles of the correlated neural events. (Mackie, 1979) Hence an earlier mental event m, the correlate of neural event n, is said like n to be causal with respect to n', the neural correlate of m'. Mental indispensability, it is supposed, is thus achieved.

This view, Psychoneural Causal Transfer, although its distinction between basic and less basic nomic correlation is consistent with our own view of nomic correlation, is open to objection with respect to the idea that one of two nomic correlates, of any kind, somehow transfers its causal power to the other. (Honderich, 1981b, p. 425) It is also open to the objection of which we know. There is contradiction in asserting, with whatever elaboration having to do with basic as against less basic sorts of nomic connection, that a given neural event has a mental correlate which is necessary to it, and that the neural event is the effect of a neural causal sequence.

Finally, and very briefly, the objection of self-contradiction is also to be made to certain views not of the same structure as those we have latterly been considering. Earlier in this section, in connection with Causal Interactionism, it was mentioned that the doctrine might be completed or supplemented in certain ways. Three possible com-pletions are modelled in Figure 13. What we have in Supplemented Causal Interactionism, despite the addition of the correlates missing from Causal Interactionism as previously defined (Figure 10), is recognizably of the same sort. All of the modelled views, however whatever else is to be said against them, are involved in the given se\t-contradiction. All of them, further, are open to the objection that although they avoid neural gaps and ghostly mental events (p. 152), they leave some neural events without a proper neural cause. This is true of the event n" in each of the three versions of Supplemented Causal Interactionism.

The third version (13c) introduces what we have not had until now, a mixed causal circumstance—made up of a mental and a neural element. There are other conceivable views involving such circum-stances but which do not involve the causal alternation of interaction-ism—sometimes a mental and sometimes a neural event being the effect. Two are modelled in Figure 14. Both suffer from the same self-contradiction. In 14a, event n' has a necessary condition m' outside the supposed causal sequence for it. In 14b there is not the same but a related truth about m'. The difference has to do with the fact that m' is necessary to n', but n' is dependently necessary to m'.

To return for a moment to interactionism, it is to be added that all of the versions of Supplemented Causal Interactionism, like all of the versions of plain Causal Interactionism (p. 151) are very unreassuring for a particular reason just touched upon. It is that in one way or another they involve a truly fundamental switching back and forth between brain and mind: sometimes it is neural events that are causally efficacious and mental events that are effects, and sometimes it is mental events that are causally efficacious and neural events that are effects. Proponents of interactionism, as it seems, are untroubled by this. One favours a view which in part is approximate to 13b, and writes of it as follows.

"It would . . . perspicuously represent the difference between spontaneous or stream-of-consciousness thinking and deliberate attentive thinking. It is plausible to suggest that when, for example, we walk in the country and allow our thoughts free rein, it is the neurology beneath them that, according to its own laws, directs their course. On the other hand when we force our thought on to some track, to perform a deduction, or to solve a problem, the stream of mental events is directed by some mental (logical) laws, and the mental descriptive level is hegemonic: here the mental descriptions drag the neural descriptions about according to the laws of sequence which belong to the mental...." (Thorp, 1980, pp. 91-2)

Whatever the support given by slight reflections about walks in the country and solving problems, and passivity and activity generally, of which more will be said (pp. 174f.), few things seem to me less plausible than the idea that mind and brain are involved in this sort of fundamental alternation or toing and froing, with the brain running the mind at one moment, the mind running the brain at the next. (Cf. P. F. Strawson, 1985, p. 67.) There is room for disagreement here, but it may be difficult, as it is for most philosophers of mind, to regard the idea as other than startling, an offence against the principle of simplicity or parsimony. It is not peculiar to the various interactionist pictures at which we have looked, all of which are of a determinist character. It is also a part of indeterminist interactionism. (3.4)


A satisfactory answer to the question of the explanation of mental events can only be arrived at by constructing it under the guidance of constraints with which we are now familiar. Indeed it is not far off the mark to say that a satisfactory answer—as satisfactory an answer as we can have—will not be greatly more than a sum of these. Our situation is like that of someone inquiring into the nature of a cell, organism, or machine to which he has no direct access, or only imperfect access. Still, he has a good deal of constraining knowledge of it. Something of this sort was once true of research into the gene, as is often noted. Our inquiry so far has made clear or suggested a number of constraints, which is not to say that we can have a completed list of such things. Some are the constraints on rational inquiry generally, of which it would be unwise to attempt an enumeration, despite the fact that we can make an easy beginning, very relevant to our recent reflections, with non-contradiction—the avoidance of one contradiction in par-ticular.

To bring together what we have, however, any answer to the question of the explanation of mental events must of course be in accord with mental realism. To conceive of the mental in some merely relational, formal, or abstract way is not to be in touch with the subject-matter. Secondly, any satisfactory answer must satisfy the principle of psychoneural intimacy, which will get further detailed support in our coming consideration of neuroscience. Thirdly, it must satisfy the conviction of mental indispensability, and fourthly, that of personal indispensability. Fifthly, a satisfactory answer to the question of the explanation of mental events should not involve any general proposition of overdetermination, or in other ways offend against simplicity or parsimony.

A sixth constraint has so far been no more than implied in passing. The Dualistic Identity Theory with Neural Causation (2.3) includes the proposition that neural events have wholly neural or bodily causes—which proposition was left undiscussed. Causal Interaction-ism was subsequently judged to be incredible in its suggestion that a neural event at a given time might be owed to no neural or other bodily event a moment before. (3.2) The proposition that typical neural events are owed to wholly neural causes was also part of the theory of Neural Causation with Psychoneural Correlation, and perhaps the theory of the Psychoneural Relation as Macromicro. (3.2) We thus tacitly accepted with respect to these views, or anyway did not dispute, that neural events are somehow or other owed to neural or other bodily causation—notably bodily causation in the case of neural events at the very periphery of the nervous system, close to the environment. Particularly given the neuroscientific reasons to which we shall come, it is impossible to deny the proposition that typical neural events are somehow owed to neural causation. It is an imprecise version of our sixth constraint, that of the neural or other bodily causation of neural events, which will get more attention shortly.

A seventh and final constraint on answers to the question of the explanation of mental events is as important as it is obvious. The existence of the constraint is a large part of what made necessary our initial inquiry into causation. (Ch. 1) There can be no hope for any answer to our question of explanation which does not rest on and indeed incorporate an adequate conception of causal connection. Any answer which does not make use of, and, it might be said, is not formed by such a conception, cannot properly have attention.

The answer which issues from these seven constraints is the Hypothesis on the Causation of Psychoneural Pairs. The hypothesis, which involves a subordinate proposition of importance, is roughly that mental events are explained by certain causal sequences. The very end of such a sequence, involving two mental events, is modelled in
Figure 15. The figure also models the Correlation Hypothesis. It is represented, as before, by no more than the vertical bars joining m and n, andm' and n'.

The side brackets [ ] in the figure represent the subordinate proposition, which is in fact essential to the present view of the mind. The Proposition of Psychoneural Pairs is that a mental and a neural event constitute what is to be called a a psychoneural pair.-For two things of any sort to be a pair in one sense of the word is for them to function together or as a unit. One dictionary definition of a pair in this sense is this: an article consisting of two corresponding parts not used separately, such as a pair of scissors, tongs, or trousers. [Concise
Oxford Dictionary] A psychoneural pair—if the comment is useful— stands in some relation to a pair in this ordinary sense rather than any other.

A psychoneural pair, more precisely, consists in a mental event and neural event, nomically correlated in the way of which we know, and (i) constituting only a single effect, which effect (ii) can also be regarded as a single cause. It is not the case that either the mental event by itself or the neural event by itself constitutes a single effect. If it is, so to speak, less true that a psychoneural pair is a single cause, it is, so to speak, sufficiently true.

In Figure 15, m, n constitutes such a pair, as does m', n'. To repeat, (i) the Proposition of Psychoneural Pairs does not assert or entail of any cause of such a pair, but rather denies, that it is a cause of a member of the pair taken separately, (ii) The proposition in a lesser sense also denies, of any effect of such a pair, that it is an effect of a member of the pair taken separately. Figure 15 thus (i) gives the circumstance ni, bi, ei as causal with respect to m, n but not with respect to either m or n taken separately. Also, m, n is given as a cause of m', n', but not of either m' or n' taken separately. Further, (ii) with respect to the latter connection, involving a pair as causal, it is only the pair that is shown as a cause, not either of its elements, m or n. The Proposition of Psychoneural Pairs, to make it clear by a contrast, thus is not modelled by Figure 16. There, the mental event m in itself is given as both effect and cause. So is the neural event n.

Figure 15, by way of another description, gives us m, say noticing the olives, as correlated with a neural event n, the two being members of a psychoneural pair, which is to say a single effect within a certain causal sequence, and also what can be taken as a single cause of a later pair, the pair of which one member is m', say intending to have an olive. All of this, as remarked, is the very end of a causal sequence. The items n1 and n2 are neural events or persisting neural structures, item b1 is a non-neural bodily event, and e1 is an environmental event. More will be said of each category.

The Proposition of Psychoneural Pairs may strike one as surprising, particularly in its first part. How can it be that something is an effect without its elements taken individually also being effects? How can something cause x without also causing each part of x? How can m, n be an effect of n1, b1, e1 without m being such an effect and n being such an effect? In fact this is not only possible, but a logical consequence of what we have already accepted. On reflection it is also intuitively correct.

Recall (1) that for a circumstance causally to necessitate something else is for it to be true, roughly, that given the circumstance, even if there were to occur anything other than logical excluders of circum-stance and effect, the effect would still occur. (1.3) Recall as well (2) that in virtue of the Correlation Hypothesis, a mental event is necessary to the occurrence of the simultaneous neural event. In Figure 15, m is necessary to n—n would not happen in the absence of m. (3) However, the non-occurrence of m would not logically exclude the occurrence of n. The absence of m, that is, would not logically entail the absence of n. It follows from these three propositions that there can be nothing that necessitates n by itself. There can be no thing of which n alone can be an effect.

There is roughly the same story about m—that nothing can necessitate it by itself. In virtue of the Correlation Hypothesis, n is dependently necessary to the occurrence of m, but n's being missing would not logically entail m's being missing. Hence there can be nothing which necessitates m by itself. There can be no thing of which m alone can be an effect.

Will it be thought that these conclusions about n and m are overstated? Will it be thought that all that the Correlation Hypothesis requires is, roughly, that if there occurs something which necessitates n, there must also be something which necessitates m? (And that there is a related truth beginning from the hypothesis and there being something that necessitates m?) The thought ignores what it is for a to necessitate b. It is for it to be true that if a occurs, so will b, despite any changes short of logical excluders of both. But with respect to the supposedly possible causal circumstance for n by itself, there logically can occur what will prevent the occurrence of the supposed effect n— to wit, the absence of m. That m, by a certain supposition, must be caused to happen, does not make its absence logically inconsistent withn.

Not that it is needed, there is a further reason why there can be no causal circumstance for either n or m taken separately. If causal circumstances necessitate their effects, they are also dependently necessary to them. Can there be something that is dependently necessary to m by itself? There is a related argument against the existence of such a thing. Given the Correlation Hypothesis, n necessitates m. Can it then be true that something is dependently necessary to m—that in the absence of the thing in question, and like things, m would not occur? Can it be true, more explicitly, that there is something such that if it and like things did not occur, m also would not occur, no matter what else were to occur other than logical excluders of both non-occurrences? The answer must be no, since the presence of n, which is not a logical excluder in the relevant sense, would guarantee the occurrence of m. There is a related argument against anything's being dependently necessary to n.

The idea that there can be effects whose elements are not effects, on further reflection, has application to things more ordinary than psychoneural pairs. Consider a car door—or anything else—shaped from a a sheet of steel by a press. It will be true that parts or properties of the door, say p1 and p2, are in nomic connection. If, say, each is necessary to the other, and neither logically excludes the other, then there can be nothing that necessitates either on its own. There can be no press, and no function of a press, which is such that given it, whatever were also to happen short of certain logical excluders, pi would be the effect. The same is true of p2..

So much for the matter of a mental and a neural event as a single effect. The other part of the Proposition of Psychoneural Pairs, which is of less importance, is that a pair is best regarded as a single cause. To return to Figure 15, it needs to be allowed that it is not false to assert of each of m and n that it is a cause of the later pair, m', n'. However, these are unusual causes, what might be called bound causes. Given the Correlation Hypothesis, one of them (n) necessitates the other, and the other dependently necessitates it. Equally, one of them (m) is necessary to the other, and the other is dependently necessary to it. It is peculiarly fitting to speak of two such things, if they have a common effect, as a single cause. If no new proposition is added to what we have by this way of speaking, it does at least give emphasis to what we have.

Before more is said of the recommendations of the Hypothesis on the Causation of Psychoneural Pairs, and its subordinate proposition, let us attend to some unfinished business. As has already been clear, and remarked upon, it is impossible wholly to separate the synchronic question of the psychoneural relation and the diachronic question of the explanation of mental events. We thus found ourselves, in connection with Identity Theories taken as answers to the first question, also considering them in connection with the second question. Now, in taking the Proposition of Psychoneural Pairs as part of the answer to the second question, we also have a further answer to the first question. We add the Proposition of Psychoneural Pairs to the Correlation Hypothesis and so come to a complete answer to the question of the nature of the psychoneural relation.

That is, what is true of a mental event and a neural event when they are intimately related is that they are nomic correlates, in the sense of the Correlation Hypothesis, and also that they constitute a single effect, and, as can be said, a single cause. This complete answer to the question of the psychoneural relation can do with a name. The answer resembles but is different from Identity Theories, so-called, and it is distant from traditional dualisms, (pp. 110 ff.) It does not make a mental and a neural event into a unity but it does bring them into a certain union. The answer, therefore, is tolerably named the Theory of Psychoneural Union or the Union Theory.

To be a bit more explicit about the comparisons, this picture of the psychoneural relation, the Union Theory of mind and brain, is of course akin to but none the less distinct from a number of particular views we have considered or passed by. It is akin to the particular Identity Theory, so-called, which we passed by, and which mistakenly takes the relation of identity to consist in shared causes and effects. (p. 93) So too is it akin to the Two-Level Identity Theory (2.3), and the theory of Psychoneural Causal Transfer (p. 161). It would be tedious to lay out all the specific differences between these several views and what we have. They are differences, as may easily be confirmed, which enable the present view to escape all the difficulties faced by its predecessors.

Let us rather return to our main business and consider the recommendations of the Hypothesis on the Causation of Psychoneural Pairs—whose full statement can be delayed for a while—taken in conjunction with the Correlation Hypothesis. What must come to mind as the first virtue of this picture is that it does not come near to falling into the contradiction we have been noticing, (p. 155 ff.) By way of the Correlation Thesis, which is essential for psychoneural inti-macy, we still have it that a mental event is necessary to the simultaneous neural event. However, that does not give us a necessary condition for the neural event which, so to speak, conflicts with a causal sequence for the neural event. The mental event, rather, is within the sequence. Event m is as much within the given sequence as is n. This is the result of our conclusion that m and n constitute a single effect. In escaping the given contradiction, further, we escape any difficulty about mental indispensability. As is clear, this picture of the mind secures mental indispensability in the most satisfactory way: by making mental events causes of later mental events and presum-ably actions.

As a consequence of having m as much within the causal sequence as n, to come to the second distinguishing feature of this picture of the mind, we do not embrace the proposition that a typical neural event is the effect of a wholly neural or other bodily causal sequence. It is the effect, rather, of a causal sequence which includes psychoneural pairs, very likely very many. It is of fundamental importance, of course, that this gives us no neural gaps in the history of any neural event. No neural event in the sequence comes out of nothing, neurally speaking. Every neural event is in a way owed to an immediately prior neural or at any rate a bodily event.

Is there any weakness or embarrassment in our exchanging (i) the proposition that a typical neural event, putting aside any environ-mental or pre-bodily history, is the effect of a wholly neural causal sequence for (ii) the proposition that a neural event is the effect of a causal sequence which is at every moment neural and also at some moments mental? There is none, but rather, as we shall come to see, the satisfaction of embracing what is surely a truth. A part of the argument—its empirical content—must wait until we come to our consideration of neuroscience. What can be said now on the point is partly that there is no tolerating the contradiction we have been noticing, that the proposition to which we have come enables us to escape it.

Also, the proposition is one of a general kind accepted in all the pictures of mind at which we have looked—save the least arguable, which are Eliminative Materialism and Local Idealism. That is, it is a proposition which asserts nomic connection—diachronic and in particular causal nomic connection—between the mental and the neural. Proponents of the Theory of Neural Causation with Psycho-physical Correlation (3.2), to speak only of them, are disinclined to have partly mental causation of the neural, diachronic psychoneural nomic correlation. In this they perhaps are like a number of neuroscientists. It is difficult to see what general ground they can have, since they already have nomic connection between simultaneous mental and neural events, synchronic psychoneural nomic correlation. There can be no more reason of a general kind against diachronic nomic connection. Both connections, and all arguable accounts of the mind, encounter the still intractable problem of the ultimate nature or mechanism of connection between the mental and the neural, (p. 112)

Further, it was remarked above that the sixth constraint on answers to the question of the explanation of mental events is that neural events are somehow or other owed to neural causation. A part of this constraint, to be somewhat more precise about it, and if we put aside first events at the periphery of the nervous system, mut be that there can be no neural event which has a full explanation which leaves out prior neural or other bodily events. This is a counterpart, for the neural, of the axiom of mental indispensability. Another stronger and more important part of the constraint is that no neural event, putting aside any environmental or pre-bodily history, has a causal history which contains neural or other bodily gaps—times at which there occur no neural or other bodily events. The view to which we have come absolutely satisfies the constraint of neural causation so specified, in both parts. It is not in accord with the excessive requirement that each neural event, post-environmentally, is the effect of a merely or wholly neural causal sequence.

To pass on to another constraint, the view we now have evidently does not conflict with psychoneural intimacy. As remarked several times, psychoneural intimacy is achieved or explained by the Corre-lation Hypothesis. It was argued in the last chapter that the unacceptability of Identity Theories, and the absence of logical or conceptual connection between the mental and the neural, made the Correlation Hypothesis the only possible means of satisfying the intimacy requirement. It is essential. We now have a proposition that was not on hand before, that a mental and a simultaneous neural event form a psychoneural pair. Might it be that the intimacy requirement is satisfied by the proposition of psychoneural pairs taken by itself.

The argument is impossible. The core of the idea of psychoneural intimacy is that there holds a direct connection of some necessary kind between a mental and a neural event. The Proposition of Psychoneural Pairs is that they derive from a single causal sequence and are in a sense one cause. But this falls short of being necessary connection directly between them. Indeed, the Proposition of Psycho-neural Pairs is consistent with the doctrine of parallelism (p. Ill) correctly named, or pre-established harmony, which denies direct connection between the mental and the neural. (The distance between the Proposition and the Correlation Hypothesis, incidentally, is also made clear by the fact that the hypothesis is logically consistent with mental and neural events not being effects at all.) Thus we shall certainly persist with the view that the Correlation Hypothesis is essential for psychoneural intimacy. It would be mistaken, even, to take it that the Proposition of Psychoneural Pairs even contributes to the achievement of psychoneural intimacy.

The Hypothesis on the Causation of Psychoneural Pairs -- we shall arrive at its full statement shortly—is of course consistent with the constraint of mental realism. How does it stand to the prohibition on general assumptions of overdetermination? We do not have overdeter-mination -- that is, an event's having two existing causal circumstances or a single circumstance containing an element plus a substitute for it. We do have something distantly related to it. In terms of Figure 15, it was allowed that it is not false to say n is a cause of the later causal pair m', n'. Also, by the Correlation Hypothesis, n necessitates m, which can be said to be a further cause of m', n'. Thus it can be said there are two routes from n to m', n'. Different but related remarks can be made about m, which dependently necessitates n. Greater detail might be provided, but is unnecessary. That the view we have has this feature distantly related to overdetermination is no significant objection to it.

The constraint of mental indispensability, as we know, is simply and effectively satisfied, without contradiction, by the placing of earlier mental events within the causal sequences for later mental events. Our hypothesis has the recommendation at which Identity Theories aim. The satisfaction of the constraint in so far as it has to do with actions will be considered in the next chapter. What of personal indispensability? The constraint having to do with persons is in essence that the explanations of typical mental events must include references to such things as memories, beliefs, intentions, resolutions, tendencies, commitments, plans, hopes, and a good deal more. A choice is certainly not to be explained just by the desire and belief that immediately precede it, or by these with their neural correlates. It is also to be explained by the continuities of experience which are at least partly constitutive of a person.

What all these continuities involve, as was in fact remarked earlier in connection with a wide conception of the mental (pp. 86 f.), are dispositions. Dispositions in general are persisting conditions or sets of conditions which may be parts of causal circumstances. (2.2) The particular dispositions which are or are involved in a man's memories, tendencies, and so on cannot be regarded as other than neural facts about him. Such dispositions, in brief, are persisting neural structures. The Hypothesis on the Causation of Neural Pairs evidently accommo-dates them. It does not explain mental events only in terms of preceding mental events and their neural correlates. It allows a role to personal dispositions. The fact is represented by n2 in Figure 15.

The hypothesis, to move now to a full statement of it, has to do with causal sequences of a certain length, and whose initial elements are of certain kinds. If we were to contemplate certain short causal sequences, or perhaps even long ones, with certain sorts of beginnings, we would fail to have a determinism. An indeterminist picture of the mind can of course include certain causal sequences for mental events and actions—sequences which begin with, or leave open the possibility of beginning with, acts of Free Will or the like, the sorts of thing often taken as needed to guarantee freedom and responsibility.

Still, the Hypothesis on the Causation of Psychoneural Pairs will not be a claim about the whole causal background or history of each mental event. If it were such, it would be a yet larger proposition than it is. It would concern not merely a person and his relevant background, so to speak, but also a history consisting in parents, ancestors, and each thing that touched on his life and in so doing contributed to the mental event in question. If Brown's thought was a result of Green's words, then such a large claim as to Brown's mental events would include some of Green's life, and of course antecedents of that. The Hypothesis on the Causation of Psychoneural Pairs, rather, specifies that mental events are owed to more limited causal sequences, whose initial elements fall into two kinds.

The first kind consists in neural and other bodily elements just prior to the first mental event in the existence of the person in question. We need not exercise ourselves about the question of exactly when, after fertilization, that event occurs. No initial elements of this kind are shown in Figure 15, of course—as remarked, what we have there is no more than the final upshot of a causal sequence.

The second kind of initial elements consists in environmental events then and thereafter (such as e1 in Figure 15) which directly affect the person. For an environmental event to affect a person directly, let us say, it must be the last event in an environmental causal sequence whose effect is a bodily or neural event. That is, the causal sequence as specified contains no environmental event later than the given event. All of these direct environmental elements will of course be non-mental. The thoughts and feelings of others certainly affect me, but by way of their bodies and actions and the non-mental effects of these. This is part of the large fact of sense-perception and sensation, the only channels between our environments and ourselves.

The Hypothesis on the Causation of Psychoneural Pairs, to come to a full statement, is thus as follows:

Each psychoneural pair, which is to say a mental event and a neural event which are a single effect and in a lesser sense a single cause— each such pair is in fact the effect of the initial elements of a certain causal sequence. The initial elements are (i) neural and other bodily elements just prior to the first mental event in the existence of the person in question, and (ii) direct environmental elements then and thereafter.

The hypothesis is evidently related to familiar beliefs, rightly en-trenched or anyway sturdy, as to a person or an aspect of a person being a product of heredity and environment. It is no grander than these beliefs, but more specific. The hypothesis, given our understanding of causal sequences generally, of course excludes the possibility that the mentioned sequences contain any gaps (3.1)—that is, each event within them, whether neural or mental, is both an effect and an element of a causal circumstance. Thus the mentioned sequences cannot contain acts of Free Will as traditionally conceived—these, as we shall see more fully, are or closely involve items which by definition are not effects. The hypothesis does not exclude the possibility, of course, that the causal sequence for a psychoneural pair includes other psychoneural pairs. It allows the possibility that typically just this will be the case, as in Figure 15. The hypothesis, as already remarked, and to speak of another kind of possible gap, is of course consistent with the proposition that there is no non-neural or non-bodily gap in the causal background of a psychoneural pair.

One final characterization of the hypothesis is of some importance. It is that it is consistent with what has several times been noticed, the distinction we make between active and passive mental events. This might, perhaps, have been added to our list of constraints. There is no doubt that such a distinction is in order. There may occur a sequence of active mental events when I engage in a little habit: I look from my college window first at the top of the Post Office Tower, then at the lowest part that is visible, then at a cornice of the college in the foreground, and then at a pillar. So too are there active mental events when I wonder if something read in a book is consistent with something read earlier, look back unsuccessfully at a previous chapter, try to recall the earlier item, succeed in thinking of it, turn to the right chapter to check. As against such sequences of active events, there are episodes when, as I say, feelings and inclinations come upon me, thoughts intrude, or will not go away. The passive parts of conscious-ness also include sensation and sense-perception, but not, as it seems, attending to a sensation or perception.

It is common in indeterminist accounts of the mind to regard at least some active mental events, in the above sense, as unnecessitated events of a certain kind. Whatever can be said in characterization of these latter events—we shall be considering the question fully—there is no possibility that the ordinary distinction between activity and passivity has to do fundamentally with such an idea. Nor need we despair of analysing activity, as some do, or say something so limited as that passivity is to be distinguished from activity by the fact that in the former a man is a 'bystander' or 'victim' with respect to his mental events. (Thalberg, 1978)

What is fundamental is that active experience, as it can be called, is purposive or goal-directed. It involves a reason. When I was looking out the window and ran my eye over the four items, I did so in the desire to complete a little ritual. The episode of my thought about the book and inconsistency was similarly goal-directed. In each case, to be clearer but brief, my experience was subject to a certain ongoing desire. There was, in a certain uncontentious sense, quite different from one to which we shall come (3.6), a ideological explanation of what occurred.

There is nothing in this familiar account of activity, which can be much elaborated by way of such further ideas as those of interim goals and a plan, that conflicts with the Hypothesis on the Causation of Psychoneural Pairs, or any other account of all mental events as effects. The hypothesis allows, evidently, that a sequence of mental events might be owed, importantly, to a dispositional desire. It also allows what may be necessary, that a sequence of mental events may involve an ongoing conscious desire. The latter fact may perhaps be doubted, as a consequence of having a certain image of the import of the given hypothesis together with its predecessor, the Correlation Hypothesis. The image is roughly, that the two hypotheses together make consciousness into a succession of somehow unrelated events. This may be taken to conflict with the idea of active experience as a matter in part of an ongoing conscious desire. However, the import of the two hypotheses is not caught by the image.

As remarked in connection with the Correlation Hypothesis (p. 117) it in fact does not 'atomize' consciousness or deny its continuity, despite the fact that the hypothesis has to do with discriminable mental events. Further, we are in no way committed to the absurdity that successive occurrents are necessarily different in character—that, by way of an example, aches do not persist. What we have is wholly consistent with regarding a sequence of mental events as incorporating a sequence which can be described as an ongoing conscious desire, indeed a governing desire.


We have already considered quite a number of objections that can be or have been attempted against the Hypothesis on the Causation of Psychoneural Pairs. This is so since various propositions raised against the Correlation Hypothesis can also be raised, sometimes with little or no alteration, against its companion. Some are those having to do with its contained ideas of consciousness and of a mental event. (2.5) Others are the Wittgensteinian objection to mechanism (2.6), and the objection having to do with the different domains of the mental and the neural. (2.7) We shall not return to these disputes.

The remainder of this chapter, rather, will be given over to a consideration of a mixed set of radical alternatives to the picture of the mind given by our two hypotheses. We shall consider the plethora of indeterminist theories and ideas of the mind. They attempt to give an account of choosing, deciding and like things which is wholly different from the account we have in our two hypotheses. It is the burden of our hypotheses, or rather one burden, that choosing and deciding are fundamentally on a level with other mental events—if choosing and deciding are of a unique character (p. 222) and if they are bound up with what has just been noticed, active as against passive experience (p. 174), they remain with other mental events a matter of nomic correlation and a certain causal background. Acceptance of one of the opposed plethora of indeterminist theories and ideas of the mind, which elevate choosing and deciding, is implied by very many of the objections mentioned above, which fact is quietly passed over, and by a good deal of inexplicit philosophical scepticism about determinism.

If the indeterminist theories and ideas have negative sources— sources in objection—they are also to be characterized in terms of positive sources. One is the conviction of mental indispensability, as well described as the conviction that epiphenomenalism is false. (Popper, 1982b, pp. 28, 41; Popper & Eccles, 1977, pp. 72-5, 86-8) But to suppose that the conviction of mental indispensability cannot be respected by a determinist view is of course mistaken. Certainly the determinism that has been expounded here, essentially because of its contained idea of psychoneural pairs, excludes the possibility, say, that Mozart's devising of his music, his creative imagination, could in principle be explained in a wholly neural way, without reference to Mozart's mental events.

Indeterminisms have a source too in religious commitments or aspirations. (Popper & Eccles, 1977, pp. 557-8) At least some of these, however, can coexist with determinism, as is made clear by episodes in the history of religion and theology. Indeterminist theories are also owed to a more vague spiritual inclination to have the mental somehow above or beyond the ordinary physical universe, or really different in kind. There is the inclination to have the mental emerge from all else, to have yet more than a unique character.

Indeterminist theories and ideas have their main sources in certain commitments and attitudes having to do with the nature of our ordinary unspeculative and unspiritual lives. The most discussed of these, whether or not it should be, is a certain commitment to moral responsibility, a commitment to moral responsibility conceived in a certain way. It is enough to remark, for the present, that indeterminist pictures derive from the attitude, among others, that we are responsible for at least some of our decisions and actions in the sense that we could have decided or acted otherwise than we did, given all things as they then were, and given all of the past as it was. To speak in a currently favoured way, but one which adds no further idea, it is true of my decision or action in this actual world that there is a possible world in which I do not so decide or act, and everything is the same in this possible world, save for my decision or action and its consequences. (Van Inwagen, 1974, 1983; Boyle, Grisez, Tollefsen, 1976) I can in this sense credit myself and others with responsibility, and hold myself and others responsible.

These true sources of indeterminist theories and thoughts are noted not idly but for a particular reason. The main sources determine the natures of the theories and the like, fix their definition—or what definition they can have. If the theories and the like are often enough implied by objections to determinism, and by inexplicit scepticism about it, attempts to set them out explicitly and completely are rare.

We need to try to clarify libertarianism in order to have a better view of the picture of mind given by our own two hypotheses. That picture, like any conception which competes with others, is partly to be judged by way of the strength of its competitors. This is so with respect to both conceptual adequacy, which is our present business, and of course also truth. There is—to look forward—also another equally large reason for trying to clarify and for considering indeterminist theories. Indeterminist assumptions and beliefs are not the property of philosophers, but of all of us. They are entrenched in our culture—by which I mean at least European or Western culture—and inform many of our attitudes, practices, and institutions. Indeterminist assumptions and beliefs thus enter into attitudes, practices, and institutions—including the practice of regarding people as responsible in the way noted— which will be our concern in considering the consequences of determinism. (Chs. 7-10) These attitudes and the like, and the assumptions and beliefs within them, are further understood by way of what can be called their best underlying theories, or so it must be hoped.

Before proceeding, we need to attend to one further matter having to do with those attitudes, assumptions, and so on, in particular the aspiration to regard ourselves as in a way responsible. It is possible to understate or to have too small a conception of the goal of indetermi-nist theories, and hence of their necessary nature, and hence of the difficulties they involve. It is also possible, however, to overstate our attitudes and hence the goal of theory. (G. Strawson, 1986, Ch. 2; cf. T. Nagel, 1986, Ch. 7) The result of this is the conclusion that indeterminist theories and thoughts are yet more conceptually doubtful than they have often been taken to be. A second and connected conclusion would be alarming if true, in that it would make a good deal of this book, like a multitude of others, in one way otiose. It is, however, a conclusion which is unlikely—sufficiently so as to put into doubt the distinctive and original argument for it.

The conclusion is that there is actually no need, in connection with the dispute about which of two kinds of freedom we have, one of them a matter of indeterminist theories, to consider the conceptual viability and the truth of determinism. All of the traditional debate about determinism, in connection with the two kinds of freedom, can be ignored, since the utter futility or even fatuity of indeterminist thinking can be fully established without considering the worth of determinism. That cannot be to say, incidentally, as is not made clear, that even on the given argument, determinist pictures of the mind can in general be ignored. Indeed, on the given argument, they must call for more concern since, so to speak, they have the subject-matter of our lives entirely to themselves. They must, on the given argument, be the only theories in which to locate the freedom we actually have. Still, it would be at least unsettling if time spent on determinist pictures just as means to the refutation of indeterminist pictures was time wasted.

The argument for that conclusion proceeds, as remarked, from a particular conception of the goal of indeterminist theories. It is said to be the goal of true responsibility, or responsibility in the strongest possible sense. (G. Strawson, 1986, p. 1) This responsibility is taken to consist in a certain self-determination, wholly different from self-determination or responsibility in a more familiar sense, where it has to do with one's actions flowing from one's own desires and the like. Self-determination of the required kind consists in determining one's own self, or determining how one is. (G. Strawson, p. 26) It is a creative power akin to that assigned to us by Existentialism in the dictum that existence precedes essence: that in my life I make my essential nature, that my life is at no point a product of my nature. (Sartre, 1957; M. Warnock, 1965, Ch. 5, 1973) No theory, it is said, can possibly give us this, as can be established by the following argument.

(1) One's rational action is necessarily a function of, or somehow determined by, how one is, mentally speaking. (2) If one is truly responsible for the action, one must therefore be truly responsible for how one is, mentally speaking. (3) To have the latter responsibility, one must have chosen the way one is, mentally speaking, at least in certain respects. (4) To have made that choice, one must have had principles of choice, reasons in the light of which one made the choice. (5) One must also have been truly responsible for these principles, and hence (6) have chosen them. (7) But for this choice, one must have had prior principles of choice. (8) One must also have been truly responsible for these principles, and hence .... We must thus conclude that true responsibility for one's actions, a determining of one's self, would require what is logically impossible, one's actually completing an infinite regress of choices of principles of choice. Indeterminist theories of the mind are therefore attempts to secure or elaborate the logically impossible. They are exercises in fatuity. (G. Strawson, 1986, p. 52; cf. T. Nagel, 1986, Ch. 7)

The argument, even when further filled in by its proponent, remains in several ways too rapid. Its fundamental idea, however, is not complex, and more or less independent of the matter of rationality. It is that since a certain conception of responsibility requires that I stand in a certain creative relation to an action, and since that action has a certain source or ground, I must also stand in the given creative relation to the source or ground, and so on back through an infinite regress. The essential weakness of the argument can be approached by way of an ad hominem route. A parallel argument can be attempted against conceptions of responsibility opposed to the indeterminist conception.

These, in fact noticed a moment ago and to be more fully considered in due course (Ch. 8), may briefly be said to take a man's responsibility for an action to consist in the fact that it flows from his own decisions and the like. It would not be enough, evidently, for the action just to be not contrary to desires of his. If only that were true, indeed, it could not be an action, or at any rate his action. But now, it can be argued, it must also be the case that the given desires must be desired by him. (Cf. Frankfurt, 1969, 19 71, 19 75) To speak differently, they must not be disconnected from his personality, nature, or identity. It may be remarked in this connection that there are drug addicts, say, who truly desire not to be subject to their cravings, and whose responsibility for their drug-taking is therefore at least in doubt. To allow that the given conception of responsibility has such a need to incorporate desires for desires evidently makes conceivable an argument about regress.

However, as can be anticipated, proponents of the conception are unlikely to be disturbed by the argument. They will reply, in short, that they do not require of all desires antecedent to an action that they themselves are desired. Enough is enough.

Similarly, proponents of indeterminisms can about as easily assert that the responsibility which they pursue faces no regress. There are several possibilities here, of which we need note only one. It can be asserted that the responsibility in question requires that I stand in a certain creative relation to an action, and hence in a certain creative relation to what it certainly has, a certain ground or source, but that either this ground or source or an earlier one is primitive. That is, it has no ground or source. An indeterminist theory of the mind may centre on this primitive episode. There will certainly be great difficulties about its description—as we shall see in what follows—but there need be no capitulation to a fiat, the fiat that any such episode must have a similar antecedent. To return to rationality for a moment, it can of course be said that there is no more difficulty about the primitive episode in this connection than with primitives in any logical system, morality or whatever, each of which must leave something unproved.


Over past centuries of their existence, indeterminist accounts of the mind have pretty well ignored the brain. This, of necessity, is no longer true. Philosophers who now espouse or are inclined to such accounts give some place to propositions about the psychoneural relation and claim to include a somehow neural explanation of the occurrence of neural events. It is of the essence of such theories, however, that they do not accept both of the Correlation Hypothesis and also the Hypothesis on the Causation of Psychoneural Pairs, or any related pair of propositions of nomic connection. The denial of one or both kinds of nomic connection -- which denial is the first of three elements of an indeterminist account of the mind -- is now based on physics, or rather, a certain customary interpretation of Quantum Theory.

That interpretation is of course to the effect that there are micro-events which are random or a matter of chance, which is to say not necessitated. Such an individual or token event is not to be explained as an effect of a causal circumstance or in terms of other kinds of nomic connection as we have defined it. It is not that it is merely inexplicable in the sense that it is impossible to know a causal circumstance or explanatory nomic correlate for it, which thing may exist. Rather, it is a truly random event, inexplicable-in the sense that there actually exists no causal circumstance or explanatory nomic correlate for it. The 'uncertainty' about it is not epistemological but ontological. In a traditional usage, if not one that is congenial to its advocates, it is a miraculous event. It is true that it is allowed that there is an antecedent probability of its occurring, perhaps a high probability, and this as a matter of necessity or law, but that it actually occurs is something which can have no standard explanation what-ever. Our present concern, however, is to consider indeterminist pictures of the mind rather than to consider their supposed basis in Quantum Theory. (5.6)

The first element of any indeterminist picture, to repeat, is a denial of one or both of our two hypotheses of nomic connection, or any like hypotheses. Macro-events which are the subject of these hypotheses are inferred to be random on the premiss that certain micro-events are random—which inference distinguishes the picture from near-deter-minisms. (pp. 8, 9) This is rightly taken as a logically necessary condition of the given sort of responsibility. A second element of any satisfactory indeterminist picture is what we shall call an originator, some persistent entity that originates certain mental events. We might, more traditionally, but less descriptively, or indeed mislead-ingly, name it a self, ego, mind, self-conscious mind, faculty of mind, active power, the will, the faculty of the will, rational capacity, agent, or person. As we shall see, some indeterminist pictures attempt to give relatively little place or role to such a thing. A third element, not entirely separable from the second, is an account of the relation of origination holding between the originator and the mental events.

All indeterminist accounts worth considering, to speak very summarily indeed, are thus to the effect that mental events stand in some relation to neural events, which neural events may have a certain non-nomic neural history and are in a certain connection with actions— and that certain of the mental events also stand in a certain relation to an originator. Only the terms of these relations are represented in Figure 17—the figure thus does not give all of an indeterminist theory. The originator o, whose nature is one of our problems, is some kind of persistent thing, represented by the arrow as moving through time. It stands in some relation to the mental event m' (as perhaps it previously did to m), which relation is another of our problems and is not represented. The mental event m' also stands in a certain relation to n' and hence to the non-nomically related neural event n and by way of it to the prior mental event, and also in some relation to a, which is an action. These unrepresented relations are also problems for us. As
before, m can be taken as someone's noticing olives and m' as his   . intending to have one. The action a can be taken as his reaching for one. The items n* and n** are included only to represent the fact that such a neural event as n' and any action will somehow be owed to more than a single preceding neural event—such as n oin'.

Typical indeterminists, if they now give some attention to the brain, do not go into relevant detail about what is necessarily the first element of their view, either a non-nomic neural history for neural events or a denial of the Correlation Hypothesis. One idea is a denial of the Correlation Hypothesis and any like proposition, and the retention of some proposition of neural and bodily causation. That is, roughly, the brain and Central Nervous System and body are left to run on nomically, as there is good reason to suppose they do, but nomic connection with mental events is denied. This would certainly free such mental events as decisions from constraint by the brain, and leave them as products of some kind of the originator. The idea is natural and common, but fatal to the enterprise in question.

One reason is that it would stand in the way of securing the desired kind of responsibility for a person's actions. In terms of Figure 17, if action a is left as the effect of a standard causal sequence of a wholly neural and bodily kind, we shall close off the possibility of the agent's having the desired kind of responsibility for it. Whatever is said of the level of mental events, and that of the originator, we shall not have it that the action could have been otherwise in the strong sense that is desired. It is neurally fixed. Let us take it, then, that an indeterminist theory of the mind includes the Correlation Hypothesis. Certainly any tolerable theory of this kind must include something like the hypothesis. It has sometimes been supposed, incidentally, that an indeterminist picture can make use of an Identity Theory or of Interactionism for the given purposes. (Popper &. Eccles, 1977; Thorp, 1980) What was said earlier against Identity Theories (2.3), and Interactionism (3.2), when we were not considering indeterminism, is as fatal here. It is the Correlation Hypothesis or something like it that must enter into an indeterminist theory.

What such a theory must deny, then, as its first element, is that neural events are effects of neural or otherwise bodily causal sequen-ces. It must deny the Hypothesis on the Causation of Psychoneural Pairs and any like hypothesis, and substitute some or other Proposition of Neural Indeterminacy. That is, roughly, neural events enter into sequences defined in terms of a conception noticed in our inquiry into nomic connection, and mistakenly assigned to all of us—what can be called probabilistic causation. (1.6) A probabilistic effect, roughly, can be stipulatively defined as an event whose probability of occurrence is raised by antecedent events and conditions, including the 'probabil-istic cause'. As our earlier inquiry showed, improving such a rough definition will involve difficulties in connection with probability itself and the theories of it. Improving on the definition of 'probabilistic effect', however, will also involve a particular act of stipulation of an unnerving kind—a decision as to how much the probability of a neural event must be increased by antecedents if it is to count as their probabilistic effect. Let us consider this second matter.

This decision will have to be taken not simply on general grounds, having to do with the nature of reality generally, but with reference to the formulation of an indeterminist account of mind and brain. The decision will have to be taken, more particularly, by attending to a number of constraints. One of these will be truth to neuroscience. However much science in general may be inclined to defer to physics, it is a plain fact, of which we shall see more, that neuroscience does not suggest, let alone propound, that the brain and Central Nervous System are prone to indeterminacy, let alone a great indeterminacy: that the most that can be said of neural events is that they have a low probability. On the contrary, it seems that the denial that neural events are necessitated can only be replaced by some proposition to the effect that their antecedents give them a very high probability.

A second constraint cuts in the opposite direction. If, so to speak, the originator is to be left with room for effective action, it seems that neural events cannot be near to being necessitated by their neural antecedents. If they are near to being necessitated in this way, and mental events are nomic correlates of them, then how are we to achieve a sufficient explanatory role for the originator and hence achieve the goal of a certain responsibility for mental events and actions? The two constraints therefore produce a dilemma for indeterminists. They are pulled towards too much and too little neural indeterminacy.

There is also a third constraint, which also fights with the second. If the originator is to act effectively, a secure link is needed between its mental event m' and the neural event n'. But there is the same requirement with the link between n' and the action a. The originator's plan for the olive runs the risk of failing at that stage. If we now have in mind the second constraint, that of leaving n' open to the originator's influence by reducing its probability relative to its neural antecedents, and also the third constraint, we will face a certain temptation. Or rather, the indeterminist will.

To speak in terms of Figure 17, an indeterminist may or indeed must be tempted to conclude (i) that the probability of the occurrence of n', given its neural antecedents, is p, (ii) that the probability of the action a, given its neural antecedents, is p', and (iii) that p' is greater than p, or indeed that a has a probability of 1 or even is an event necessitated by its neural antecedents. The first proposition would leave room for activity by the originator, and the second would secure its efficacy with respect to a. It is very clear, however, that the familiar and well-founded scepticism about ad hoc stipulations must at least work against the conclusion.

There are other large problems about what no one has ever specified—the Proposition of Neural Indeterminacy—but let us press on.


It was supposed by a few philosophers, and more scientists, when Quantum Theory was first seen as an ally to indeterminist views of the mind, that a denial of nomic connection and in particular of causation is all that is needed in order to secure the desired responsibility for decisions and actions. The Proposition of Neural Indeterminacy by itself may indeed seem to promise the conclusion that a vicious or virtuous decision at t could have been otherwise than it was, given everything as it was at t and before. However, the proposition provides at best only a logically necessary rather than a logically sufficient condition of what is desired. How could an unnecessitated or chance event be something for which the person in question could be responsible in the given way? How could anyone be disapproved of or censured in the given fundamental way for a mere chance event? Indeed, how could anyone be responsible in any way for a mere chance event? (Ayer, 1954c)

The matter is made a bit more complicated, but the question is not answered, by the postulation of high-probability chance events—as will be anticipated from what has been said already. In the end, however probable it is in advance, the fact that one thing happens rather than another—that a man on a given occasion decides to torture or kill rather than not—is something which has no cause whatever. Thus there is nothing that can provide a standard or fundamental explanation of the event, that particular outcome, (p. 59) Moreover, there is surely nothing that in any sense explains that particular outcome. It is, as it seems, easy to lose sight of this proposition. In fact, it is nothing other than the essence of the main contention on hand: that the event is no more than probable, however highly probable. It bears repeating that what we are to understand, at bottom, is that things might have been exactly the same up to t in two situations: one in which the event happens at t, and one in which it does not. Hence it seems there can be nothing that in any sense explains the event if it does occur.

Indeterminism by itself, then, cannot secure an indeterminist theory's goal of a certain responsibility. If indeterminist views of the mind are also aimed, as indeed they are, at securing or safeguarding other things, the postulation of undetermined neural events is again logically necessary but not sufficient.

All serious attempts to formulate an indeterminist theory of the mind thus have more in them than a denial that neural events are neurally necessitated. What is also included is some non-neural account of how it is that a mental event occurs. Inevitably this is again an account such that the mental event is not necessitated. It remains a chance event in the sense of being unnecessitated itself or having a suitable unnecessitated antecedent. However, the non-neural account does not leave a decision, say, as merely or only a chance event—as it would do, of course, if the decision were made a matter of only probabilistic causation, which is no part of the present story. The non-neural account of the decision attempts to explain it as arising in such a way that it is grounded or non-arbitrary. We shall take that to mean, just, that the decision or an antecedent of it is such that the agent can be responsible for it in the given way. In sum, then, an indeterminist theory will be to the effect that a decision's neural correlate is not necessitated, and the decision is not necessitated by non-neural facts, but somehow it is rendered grounded and not arbitary by them. As result of these non-neural facts, further, partly by way of the Correlation Hypothesis, a resulting action is somehow explained by the grounded decision. There are great problems with the last part, about the action, as we know, but let us press on.

The non-neural facts include some persisting entity, what we earlier labelled an originator, what is also named a self, mind, the will, active power, a person, and so on. However, such an entity is often and rightly taken as problematic, and, as a result, there are indeterminist theories which seek to have little to do with one. The originator is at least kept well to the rear. We shall first consider three of these, having to do with self-causing decisions, self-justifying decisions, and teleology.

3.6.1  Self-Causing Decisions

The first idea as to how an unnecessitated decision as so far understood may be non-arbitrary has to do, as we can say, with a way in which such a decision can in another understanding be necessitated. We have taken it, certainly, that the standard concept of a necessitated event is the idea of an event necessitated by other things than itself, and that the idea of an unnecessitated decision is the idea of a decision unnecessitated by things other than itself. Necessitation like all other causal relations, as we well know, involves two items. There can be no doubt whatever that this is fundamental to our standard conception of causation. A decision may be said to be unnecessitated in this sense but the idea added—an idea with a long history—that it is necessitated by a set of conditions and the like which includes itself. The other conditions are presumably desires, beliefs and so on. (Cf. Boyle, Gnsez, Tollefsen, 1976, p. II, p. 83, p. 90.)

A decision of this kind is therefore to be understood as having something like the nature that has sometimes been attributed to God. It is to a significant extent cause of itself, causa sui. Any such proposal, about decisions, God, or anything else, is deceptive. It is deceptive in that it is likely to be taken as making use of a clear idea with which we are familiar—the idea simply of a cause, or of necessitation, a causally sufficient condition, or the like. We are at least tempted to think that the proposal involves no more, so to speak, than a novel application of a satisfactory and settled idea. The temptation is to be resisted absolutely.

To repeat, whatever else is true of causation as we understand it, it involves two items, numerically different items. Causation as we understand it involves relations whose terms are non-identical. To speak of a self-caused thing is not to make use of a satisfactory and settled idea of a certain relation in a new way, but to introduce a wholly unfamiliar and different relation. To say of a thing that it is cause of itself, and to insist that the given idea of a cause is our standard idea, would be to fall not into mere obscurity but into self-contradiction. What must necessarily be intended by any consistent talk of a thing as cause of itself is a quite different relation which is consistent with its own terms being strictly identical.

Any of the standard causal relations noted in our own inquiry can be used to illustrate this. Suppose it is said that c caused e, and it is added that c was identical with e. Can we partly understand this by way of the idea that a cause is required for its effect—that if c had not happened, e would not have happened? Very evidently we cannot. That conditional, given the addition that c was identical with e, becomes the logical truth, to express it one way, that if c had not happened, c would not have happened. What is therefore needed is some quite different account of 'c caused e'. To my knowledge, no explicit account has ever been supplied, whether of decisions, God, or anything else. It is difficult to avoid the suspicion that no such account is possible.

That is, it is difficult to avoid the suspicion that it is impossible to give sense to talk of explanation of b by a, where the explanation is significantly similar to an explanation in terms of causation, and it is none the less true that b == a. It is easy enough to conceive, say, of a thing's being self-changing or self-governing, where that is the fact of one part of a thing being in causal relation to another part of the thing, or of one event in the history of a thing causing another such event, but this of course is nothing to the point. Self-guided missiles and the like are not near to being instances of the relevant sort of self-causation.

We must therefore resist the inclination to think that what is proposed, in order to make certain chance events other than arbitrary in the given sense, is the novel application of a settled idea. What we rather have is a no longer novel application of an unsettled idea. We are thus nowhere near having a clear idea of a grounded decision. In fact the conclusion that we have nothing clear is indubitably implied by what else must be said in or of the proposal. One thing already said is that a decision is to be understood as not standardly necessitated by either neural or non-neural facts. It follows that talk of its being necessitated, partly by itself, is to be understood quite differently. It is typically also said that the causation in question is to be distinguished from 'the causality which obtains between events' (Boyle, Grisez, and Tollefsen, 1976, p. 12) or 'event-causation' or 'Humean causation', where what is intended by these misleading descriptions is standard causation. Again, the explanation of something which involves self-reference is allowed to be a unique kind of explanation, an appeal to a unique kind of innate intelligibility. (Boyle, Grisez, and Tollefsen, 1976,pp. 83-5)

3.6.2 Self-Justifying Decisions

Let us look at a second attempt to clarify the idea or perhaps what can only be called the words 'chance or random but nonetheless grounded or non-arbitrary decision'. The attempt is one of a family, all the members of which try to explain something by way of the idea of the thing's 'turning back upon itself' or reflexivity. (Nozick, 1981)

A decision or choice, we may agree, is a matter of arriving at a view as to which of a number of reasons, for or against an action, have most weight. Or rather, we may agree, since it seems that reasons do not come with weights attached, the process is importantly one of assigning weights to reasons, of weighting them. This assigning or bestowing of weights, we are now told, is not necessitated by anything. However, to come to the principal point, it is grounded or non-arbitrary in that it refers to itself.

The decision may be self-subsuming; the weights it bestows may fix general principles that mandate not only the relevant act but also the bestowing of those (or similar) weights. The bestowal of weights yields both the action and (as a subsumption, not a repetition) that very bestowal. For example, consider the policy of choosing to track bestness: if the act weren't best you wouldn't do it, while if it were best you would. The decision to follow this policy may itself be an instance of it, subsumed under it. (Nozick, 1981, p. 300)

Is there the difficulty that some different bestowal of weights might have occurred, and hence that the given bestowal must count as arbitrary? To repeat,

can anything be said about why that one self-subsuming decision is made rather than another? No, the weights are bestowed in virtue of weights that come into effect in the very act of bestowal. This is the translation into this context of the notion of reflexivity: the phenomenon . . . has an 'inside' character when it holds in virtue of a feature bestowed by its holding or occurring. (Nozick, 1981, p. 304)

Further, such decisions bring into existence a self.

... the self can synthesize itself around this bestowing: 'I value things in this way'. If in that reflexive self-reference the I synthesized itself (in part) around the act of bestowing weight on reasons, then it will not be arbitrary. . .that that self bestowed those weights. (Nozick, 1981, p. 306)

The fundamental objection to this is that what we are given is a certain account of how decisions may be justified, which we do not need, and we are given no account of how they come about, which is what is needed and was promised. In any relevant sense of explanation, we are given no explanation of the occurrence of decisions. The greatly elaborated idea in what is actually said of decisions is a simple one: that a decision, like some other mental events, may apply in a justificatory way to itself. To take a clear example, a man in the grips of some idea of naturalness may make the decision d that only those decisions are to be trusted which are spontaneous ones, and he may have some tolerably clear conception of the spontaneity in question. If decision d was itself spontaneous in the given sense, then decision d is to be trusted. We can enter absolutely into his idea and outlook but have no answer at all to this question: Why did d occur? But precisely that is the question with which we are concerned.

Our question is indeed what was earlier called the diachronic question of the explanation of mental events, and to which we have been considering various answers, latterly the Hypothesis on the Causation of Psychoneural Pairs. The very heart of any Indeterminist Theory must also be an answer to this question. We surely get no answer to this question by having any account of how a decision is one to be trusted, a good or a rational one, one which 'tracks bestness' or fits a proper conception of one's own life, or any like thing. What we do get in what is offered is in analogy to explanations of the self-verifying nature or the incorrigibility or the like of, say, Descartes's proposition 'Cogito ergo sum', as distinct from an explanation of how it comes about that there occur mental events which can be individuated as being thoughts of that proposition.

Is there any reason to doubt the distinction between explanation in the relevant sense and justification, or, more generally, the rightly entrenched and pretty well universally accepted distinction between explanation and evaluation? Is there any reason to doubt the distinc-tion between answering the question of why or of why and how something occurred, and of what value it is? It is perhaps no surprise that none is supplied in the given account of decisions.

Obviously we typically explain the occurrence of decisions by citing evaluations by the agent, values of his. That is, we explain decisions by taking them as somehow owed to mental events and dispositions of an evaluative kind. It is part of the Hypothesis on the Causation of Psychoneural Pairs that such mental events and dispositions are precisely causes of decisions. The Indeterminist theory denies this account and seeks to give some alternative explanation of the occurrence of decisions and the like. The idea now under consider-ation, like the previous one in terms of what is called self-causation, sets out to make decisions somehow self-explanatory. However, it in fact offers an account of their self-justification, which is not on a level with and does not compete with either our typical account of the occurrence of decisions as owed to evaluative mental events and dispositions, or the causal hypothesis.

It may be that something called the self can be regarded as somehow brought into being, synthesized, delineated, developed, or whatever, by one's decisions. Certainly it is true that one's personal identity, being in part a matter of what were earlier called continuities of experience, is affected by at least certain of one's decisions. That a self can synthesize itself around a bestowing of weights, however, is nothing to the point. There is then at least guaranteed consistency, so to speak, between self and decision. What we are pursuing, however, is not a reason for thinking that self and decision do not conflict, or that the latter is not 'arbitrary' with respect to the former, but an explanation of how the decision comes about. If the justification or evaluation of a decision is nothing to the point, so too some sort of logical consistency or the like between decision and decision-created self is nothing to the point. Finally, with respect to what is said of decisions themselves, can the relevant question be thought to come into view in the second passage quoted (Nozick, 1981, p. 304) despite the persistent ambiguity? In fact the truth seems otherwise. In any case, no answer is given to the relevant question.

Both the previous account of decisions as self-causing and the account now on hand of decisions as self-justifying attempt to explain the supposed groundedness or non-arbitrariness of chance decisions by attending mainly or entirely to the decisions themselves—their own very nature. The central ideas about groundedness, it may appear, do not in fact much involve an explanatory self, ego, mind, self-conscious mind, faculty of mind, active power, rational capacity, agent, person, or the like—although much is said of the self, somehow conceived, in the present account. Both accounts, if they were to succeed in their aims, would have to give a real place to a somehow explanatory persisting entity and would have to clarify its nature effectively. The existence of such a thing could be no mere corollary but a central proposition. If the accounts are to achieve their goal of establishing a certain responsibility, they need not only mental events but such an entity. In the relevant sense, it is not decisions that we blame.

3.6.3 Teleology

Let us turn to a third attempt that may be made to give an account of the groundedness of decisions and choices, without facing the difficulties of a self. In answer to the question of why I decided to stop smoking, I say that I did so for health. Clearly this claim of mine is perfectly meaningful. It makes my decision intelligible, and evidently it may be true.

It is ordinary to think that such claims somehow are or reduce to ordinary causal explanations. They are so taken when they occur in biological science, where they are common and indeed fundamental. (Woodfield, 1976; Braithwaite, 1953; Sommerhoff, 1950) However, it may be denied, as in the present account of decisions, that they are or reduce to ordinary causal explanations. It is said, specifically, that such propositions as those about my stopping smoking are not to the effect that my desires and beliefs or the like caused my decision. Rather, it is said, such claims explain my decision by reference to a goal, end, or telos. These teleological claims, as we shall call them, are in fact as others describe them: teleological explanations. This view of them, dating back to Aristotle (Sorabji, 1980, Ch. 10), can be and often has been pressed into use to give an account of the groundedness of unnecessitated decisions.

Do teleological claims not reduce to ordinary causal explanations, and are they nevertheless explanatory? If we can persuade ourselves of this, we shall be enabled to see how an unnecessitated decision can none the less be grounded or non-arbitrary.

It will be as well first to look at the matter of teleological claims generally, independently of our present concern. Here are more examples. Birds have hollow bones because that enables them to fly better. We perspire in order to reduce bodily temperature. The gas is turned on because that will make the kettle boil. Shoe factories grow in size because such growth produces economies in production. In trying to think of them in the required way it is essential, to repeat, to exclude the ordinary idea that they are or reduce to ordinary causal explanations.

In each case we are given what supporters of the doctrine of teleological explanation regard as an explanandum and an explanans, an explanans that by itself is no less than complete. In the first case, the hollow bones of the birds are the explanandum, what is to be explained, and their better flight is a full explanans, that which explains. In each case, further, as of course must be allowed and sometimes is, whatever else is said, the explanandum does in fact have a causal role with respect to the explanans. Hollow bones are parts of causal circumstances for better flight. What we have, then, involves precisely a reversal of ordinary explanations, where causes and causal circumstances explain effects. In what is called teleological explan-ation we appear to have effects explaining their causes or causal circumstances.

The appearance cannot be the reality. Effects do not in any ordinary sense explain causal circumstances or causes. They do not explain them in the precise sense set out in our inquiry into nomic connection, having to do with temporal order and necessitation as against dependent-necessitation. (1.5) Nor are effects taken to explain causal circumstances, or any parts of them, by those who rely on an intuitive and unanalysed conception of explanation. In answer to the question of why something occurred, it is not possible to give as the sum total of explanation that it had a certain effect, whatever else may be said. This much is now explicitly allowed by defenders of teleological expla-nation, whatever the inclinations of their forebears.

Contemporary defenders of teleological explanation do maintain, as they must if they are to claim anything distinctive, that there is some related way of looking at the supposed explanandum and explanans, and at nothing in addition, and thereby getting an explanation of the supposed explanandum. If causes cannot be explained by effects, it is supposed, there do exist related possibilities. Let us look briefly at two ideas, suggested in what is the most distinguished recent attempt to defend true teleological explanation, or, as it is called, functional explanation. (G. A. Cohen, 1978, Ch. 9)

It seems sometimes to be suggested in this defence that a fact in the sense of a true proposition, as distinct from what makes the proposition true, does the explaining. The true proposition that the growth of shoe factories gives rise to economies in production, as distinct from anything else, explains the growth of the shoe factories. The true proposition that perspiration has the effect of reducing bodily temperature, as distinct from anything else, explains perspiration. An effect does not explain its cause, but the proposition that the cause has that effect explains the cause. This will not do. Certainly true propositions may be said to explain events, things, and the like—in general, spatio-temporal items. But what this must come to is that the events and the like are explained by what makes the true propositions true, which is to say events and the like.

Might something of that relevant kind really be the intention of saying that a fact does the explaining? Whether or not that is so, something of that kind is suggested in the second idea about the supposed explanans and explanandum. What explains the hollow bones of the birds? The answer given here is that it is their causal disposition to give rise to better flight. Why is the gas turned on? Because that state of affairs has a causal disposition to make the kettle boil. This will not do either, again for a clear reason.

To speak of a thing's causal disposition, as we know, is to speak of a property of that thing—a property such that when it is conjoined with certain other things, the resulting whole is a causal circumstance for the mentioned effect. But then to attempt to explain something by citing one of its causal dispositions is to attempt to explain part of a thing by just that very part. That claim, as we also know, is wholly obscure. Perhaps it does not require audacity to say that it is impossible. If it is impossible to explain the occurrence of a thing by citing its effect, or by citing merely a proposition, it is no less impossible to explain it by citing itself.

These two attempts to make sense of the idea that teleological claims are rightly called teleological explanations deserve more attention, and have been given some elsewhere. (Honderich, 1982a) However, without lingering, let us now briefly consider another view of teleological claims. There is in fact no mystery about them and they can be understood in such a way as to account for their appearance of explaining a cause by its effect.

To make a teleological claim may often be to intend, and to be understood as intending, an abbreviation of a certain undoubted explanation. Teleological claims, when they are not abbreviations, are in fact not explanations, let alone full or complete explanations, but explanation-claims: claims that explanations of a certain kind exist.
(Cf. Woodfield, 1976, p. 206; Mackie, 1974, p. 283, E. Nagel, 1979, p. 401) To say birds have hollow bones because that enables them to fly better is to say, here, that there does exist a standard causal explanation of the hollow bones an explanation which includes the fact that hollow bones have a causal role with respect to better flight. To say that shoe factories expand because that issues in economies is to say, here, that there exists a standard causal explanation of their expansion, which explanation includes the fact that expansion is part of a causal sequence for the economies.

What are the explanations? In the case of the hollow bones the explanation is of course an evolutionary one. This explanation, having to do with the persistence of some species and the disappearance of others, evidently includes the fact that hollow bones have the effect they do. In the case of the shoe factories, although there is the possibility of an explanation which is Darwinian in character, the best explanation is in terms of men's desires and beliefs. Men, wanting the benefit of economies in production, increased the size of shoe factories.
The fact that greater size does have the given effect is again part of the explanation. It is such Darwinian and purposive explanations, of course, which are typically abbreviated by teleological claims.

The principal point in this view of explanation-claims, which has been given without elaboration, is not a terminological one, a restriction on the use of the term 'explanation'. It can be allowed, however misleadingly, and despite what was said a moment ago, that we give an 'explanation' in saying that birds have hollow bones because that enables them to fly better. Certainly it can be said that teleological claims make events somehow intelligible. What cannot be allowed is that we have anything remotely like a full explanation of the hollow bones if we mention only the hollow bones, the better flight, and the causal role of the bones with respect to the flight. Partly as a result, what cannot be allowed is that what is in question is anything other than standard causal explanation. Both the Darwinian explanation and the given purposive explanation are of course of this form. In consonance with this, so-called teleological explanations are indeed better called explanation-claims.

The relevance of all this to indeterminist pictures of the mind will be clear. There is no possibility of explaining a decision d—no possibility of giving anything like a full explanation of it—simply by referring to its effect or by any related method. The ideas pertaining to causal facts and dispositions cannot help. However, given the proper understanding of teleological claims, there is more to be said. The teleological claim that I decided to stop smoking for my health, far from providing an explanation of my decision which does not involve necessitation, is to be taken as claiming the existence of precisely what indeterminist pictures of the mind deny, which is to say a standard causal explanation of the decision. Or, of course, the claim merely abbreviates the explanation. Should these analyses be denied, our situation is that we have no clarification whatever of the supposed groundedness of decisions.

It hardly needs remarking that there is a further disability of any view which seeks to explain groundedness in the given teleological way. It has exactly a disability shared with its predecessors, having to do with self-causing and self-justifying decisions. Were the teleological view to succeed in explaining a decision in the intended way, it would none the less not achieve the goal of safeguarding responsibility of the desired kind. It cannot be that we could hold a decision responsible. Some ongoing entity is required.


We turn now to theories which with good reason are less reluctant about an originator, or which in fact fully accept the necessity of one and are largely about it. To have that recommendation, of course, they must involve themselves in difficulty. Some involve themselves in more than difficulty.

3.7.1 Homunculus

What some seem to offer is an homunculus, a little person located within a person. This is suggested by the use of ordinary mental verbs. The originator perceives, chooses, and so on. It may even have 'a personality, something like an ethos or moral character', partly formed by its own actions. (Popper &. Eccles, 1977, pp. 472-3) Thus an originator is first located somehow at the centre of a person conceived in an ordinary way, and then, it may seem, the originator is itself regarded as such a person. The given picture of the mind, when made explicit, is absurd, and almost certainly involved in an intolerable regress. The originator will in turn require its own originator, and so on.

3.7.2 'The Mind'

A second idea is that what stands as the non-necessitating ground of a particular decision is no less than 'the mind' or 'the self-conscious mind' or some such, where this is one series of mental events, the one experienced by a particular person, or, better, such a series together with mental dispositions, (p. 86) The idea of the mind as mental events and dispositions, makes use of what is evidently the clearest and by far the most defensible conception of 'the mind'. It is in fact a conception at which we arrived without comment in considering consciousness and mental events, and which has informed our subsequent inquiry. Can the grounding originator which the Indeter-minist Theory needs be identified with the mind in this sense? It cannot. A part of one of several reasons is that not anything like all of the sequence of mental events which is my conscious history, or anything like all of my mental dispositions, can be thought to enter into the explanation of a given particular decision. Something more precise is needed, analagous to a causally relevant property or a set of such. Other objections to the view, as serious, will be noted in connection with the next one, where they also arise.

3.7.3 Selected Elements of 'the Mind'

Suppose we quickly and radically amend the view that what we need is all of a mind, and select certain items to constitute our originator. We thus come to a different idea about groundedness, which can also have brief treatment. There is the first difficulty that we already have these items within indeterminist theories as so far outlined. In terms of Figure 17 (p. 181), and thinking of mental event m', they are represented by the earlier mental event m and what can be called the neural structure n*. Are these now to be identified with the originator o, supposing that to be possible, and to be taken as entering into some new explanatory connection with m', over and above the connection we already have? The connection we already have involves the Proposition of Neural Indeterminacy, probabilistic causation, and the Correlation Hypothesis. We cannot discard that connection, if we are stay in touch at all with what is known empirically of the brain and mental events.

There is also a second great difficulty about the course contem-plated. If we take it, there will again be no chance of achieving the goal of an indeterminist theory, holding a man responsible in the given way. Here a certain distinction needs to b^ made clear. There is a conception of responsibility, to which we shall certainly be attending at a later stage, which has much to do with an agent's ongoing dispositions. The conception which gives definition to the Indeterminist Theory, however, is different. The theory sets out to secure that we could have decided otherwise than we did at t, given all things as they were at t and before. If one has this in mind, it becomes starkly clear that what grounds a decision cannot be the given dispositions or past mental events. Precisely these are central to what the responsible agent, so to speak, rises over or leaves behind. The whole enterprise of the theory is to achieve this escape or transcendence, importantly by way of the Proposition of Neural Indeterminacy.

3.7.4 The Person

These last two ideas as to the grounding of mental events have been noticed not only for themselves, but in introduction to something rather more popular which in a sense includes them. As indicated earlier, some indeterminist philosophers have the credit of being at least sceptical of an originator as traditionally conceived—a self, ego or whatever. One is thus scandalized by the supposition that an indeterminist view necessarily involves the 'dreadful and bizarre' idea of an Ego or Self. (Taylor, 1966, p. 134) A second hastens to say he will have no dealings with the 'extravagant' idea of a Self or Mind. (Thorp, 1980, p. 120) Both of these philosophers, and some others, take or contemplate what they regard as the far happier course of grounding mental events in, simply, a person, a living man or woman. When Tom decides to give up the unequal struggle, it is Tom himself who brings this decision about in some unnecessitated way. (van Inwagen, 1983, p. 4) In our way of speaking, the originator is no part or element or the like of Tom, but simply Tom.

It is very notable that none of these latter speculations includes an account of the nature of a person. That subject is fatally passed by. It was remarked earlier (p. 148) that a finally satisfactory account of the nature of a person will include continuities of experience and also, very likely, some idea of a locus, tunnel or frame for these. It can be speculated, further, that a person is also to be understood in terms of bodily or brain or brain-part continuity. Without attempting to press on to a specific answer to the question of personal identity and of the nature of a person, several things are clear.

Any tolerable account of a person, first, will be an account of greatly more than can with any reason at all be included in an entity which is to be explanatory of a given particular decision. To propose a selection of items faces the objection that we already have these relevant items in any tolerable indeterminist theory. The objection is a repetition of one made to the previous ideas as to groundedness, the attempted identification of the originator with 'the mind' or with selected elements of it. Another objection also transfers itself readily to the present fourth idea. To attempt to identify an originator and a person, since a person consists at least in part in a psychological past, a past of experience, is to tie an originated decision to precisely that of which it is to be independent, that which it is to transcend.

3.7.5 Originator

The outcome of our reflections as to the second element of an indeterminist theory is that it requires something which is indeed best suggested by the technical term or name of originator. How are we to attempt to regard this entity? It has traditionally been described as a substance or independent entity as distinct from an attribute, which is to say it enjoys a kind of independent existence. Also, it is a substance not in space. Despite Descartes, it seems unkind to proponents of indeterminist pictures of the mind to insist on either of these. An originator, we shall take it, may be an attribute of a person, rather than a substance, and may be in space. Certainly it must persist in time, be ongoing, for reasons we know. It would be yet more unkind to assign immortality to it, let alone an early choice of which side of the brain to concentrate on. (Popper & Eccles, 1977, p. 507)

It is natural and certainly traditional to take an originator as having a particularly fundamental unity and being simple rather than com-posite. This is presumably the same proposition as that the idea of an originator is primitive, not to be reduced to anything else—certainly it is no bundle or collection of things. Perhaps this is in some accord with what we are also to understand, that an originator is that to which, at least sometimes, we refer in using the first person pronoun, 'I'. It is said that it is something of which each of us necessarily has a pre-theoretical grasp. It is that which is misconceived by those who seek to give a unity to consciousness by speaking of merely a focus, tunnel or frame. It is that which is misconceived, also, by anyone inclined to the conception of consciousness and mental events set out earlier in this inquiry, and in particular to the minimal idea of a subject as distinct from content, (pp. 80 ff.)

To what extent can an originator be understood by way of that minimal idea of a subject as distinct from content? An originator, mainly because of its fundamental explanatory role with respect to mental events, cannot be identified with such a subject. An originator is not internal but external to mental events. The postulation of an originator requires, or at least makes it natural to take up, a quite different conception of mental events. It will be one which subtracts the subject from a mental event and makes it entirely a matter of content. Further, an originator as against a subject, even if it is attribute rather than substance, is not necessarily understood by way of an idea of interdependent existence—the mutual dependence of subject and content.

There is one other question of characterization which is uniquely difficult and very likely intractable, another question that proponents of indeterminist theories fail to get on to their agendas. What is an originator's ontological kind? Is an originator mental, or neural, or something else? Is it or is it not physical in the sense defined earlier, where physical events include both mental and neural? (2.2) To remember again our own conception of a mental event, if the subject is subtracted from it and elevated into an originator, and the event left as content, what is the nature of the originator?

It is of course customary, and perhaps essential to the goal of indeterminism, to locate it on the mental side of things, but in fact, it is also customarily distinguished from mental events—such as the decisions and choices it originates, or, on many views, other mental events which occur to it, or of which it is somehow the proprietor. Further, if it were mental, it presumably would have to be subject to something like the Correlation Hypothesis. It would then lack exactly the autonomy and creativity with which it is credited, which we shall be considering in a moment. The originator thus appears to be a different order of thing from neural things and also from mental things as usually conceived. What order of thing is it? For want of help from its proponents, and want of fortitude on my part, let us abandon this question. No answer of any useful content has ever been proposed.

In short, what has been and can be said of the sort of originator which the indeterminist requires is not satisfactory. If we put aside the suppositions that make it into a soul—immortality and so on—we have a conception of hardly any positive content. What we have is the thin idea of a unity of a wholly unspecified ontological kind. Needless to say, a principle of simplicity or parsimony must count against the supposition of this third category of being. Nor is there any effective argument to the effect that we have need of the complexity. It is nonsense to say we have a pre-theoretical grasp of something of this nature—neither mental nor otherwise physical. The supposition of an originator also raises further questions, of whether an indeterminist view satisfies the constraints of mental indispensability and personal indispensability. This has to do with the autonomy of an originator— its independence of desires, beliefs, and so on. Let us leave these matters unconsidered.
All of this is critical, but, from the point of view of precisely an indeterminist theory, as distinct from the philosophy of mind generally, it is not central. The central characterization of an originator must be in terms of its relationship to decisions and choices, its unique explanatory and responsibility-preserving role, its autonomous cre-ativity. We have not been paying much attention to that. Let us turn to it, the third element of an indeterminist theory.


3.8.1 Mental Verbs

In one indeterminist theory, the self-conscious mind is said to 'read out' from the brain, select from it at will according to its attention and interests, scan and unify it, probe and control it. (Popper & Eccles, 1977, pp. 282-3, 355, 363, 427) These are all relations between self-conscious mind and neural structures or activity. In other indetermin-ist speculations, a self or ego is presented as contemplating, evaluating, and judging mental events and dispositions, and of course somehow giving rise to them. One traditional theme is that such an entity considers and weighs up a conflicting desire and a feeling of moral obligation, and sides with one or the other.

One clear and fundamental objection to any such account is that the relation between originator and either neural or mental events is described in terms of ordinary mental verbs, principally verbs having to do with perceiving, controlling, choosing, and acting. Whatever else is to be said of these usages, there is the objection that they are neutral or unin formative. What we want to know is the nature of this contemplating, attending, choosing and so on. A determinist analysis of, say, my contemplating my desires, is given in terms of nomic connections, those specified in our first two hypotheses. It is no objection to this analysis merely to assert the analysandum. More importantly, we get no alternative analysis or account whatever of such an analysandum when what is offered to us is precisely that analysandum. What we are promised by the indeterminist is an account of contemplating, choosing and so on which presents them as unnecessitated but grounded pheneomena. We get no such thing by a recourse to the ordinary verbs themselves.

3.8.2 Self-Causing Cause

Talk of self-causation is used by indeterminists other than those noticed earlier, who speak of self-causing decisions. It may be said that an originator causes an event in or of itself—a change c in or of itself. This event enters into a wholly standard causal connection with, say, a decision d. Thus, it may be supposed, we have an answer to the question of the originator's relation to its decision.                      *

We need to ask, however, for more precision about the self-causing. Is it the case that what causes c is some other event in or of the originator? Suppose the answer is yes. If so, since we presumably here have another standard causal connection, how can it be that decision d is unnecessitated? To gain that end, the indeterminist may instead say of c that it is self-causing. What we have, then, is a self-causing event c of an originator o, which event enters into a standard causal connection with the decision d. That is, either c is itself a causal circumsance for d, or c is an element of a causal circumstance for d. On the latter assumption, the other elements of the causal circumstance for d will be mental items relevant to it, desires and the like. The latter assumption seems at least preferable, in that it includes essential elements within the story of the emergence of d.

The view might be claimed to lack one disability of taking decisions themselves as self-causing. That is, it does have an originator firmly on the scene, a persistent entity to the subject of judgements or feelings as to responsibility. Precisely the claimed superiority, however, rests on unexplained propositions. In what sense is change c 'in' or 'of the originator o? An answer is not easy to come by, or even to move towards, partly because an originator seems best conceived, as we have, as a simple unity.

Whatever is to be said of this, the view does have other disabilities. One is that to speak of self-causation of any kind, whether of decisions themselves or events which cause decisions, is, as we know, to speak of nothing clear. A second disability attaches to taking the self-causing originator-event as an element of a standard causal circumstance which includes other mental items, such as desires. These items already turn up in the attempted explanation of the decision by way of the Correlation Hypothesis and the Proposition of Neural Indeterminacy.

3.8.3 The Originator Itself as Causal

Another view goes further in the direction of making the originator a suitable subject of judgements and feelings as to responsibility. This it does by making the originator itself somehow causal, as distinct from any event of it or in it. Further, it is the originator on its own that is somehow causal, as distinct from a set of items including the originator. Of course some fitting account must be attempted of this causing, an account which makes the resulting decision other than necessitated but none the less grounded.

The substitution of the originator for a causal event, to begin with that, produces what is at least a certain surprise, given either of two assumptions. We may assume, first, an originator that is in fact unchanging throughout its existence. It is permanently eventless. The idea is in some accord with our description of it as a simple unity, and its being named a pure ego and the like. Or, we may assume an originator that is unchanging during a certain period of time. It is for this time eventless. In either case the surprising supposition is this: an originator o which is unchanging between t1 and t3 somehow causes a decision at t2.

What then is the so-called causal relation? Clearly it cannot include what is fundamental to standard causation, which is necessitation. It cannot be that originator o necessitates decision d at tg. Given that o at the earlier time was identical to o at t2, then, on the assumption of necessitation, d ought to have occurred at t1, or indeed d-type decisions ought to be being produced through the period from t1 to t2. Whatever is to be said of the mysterious relation, incidentally, it will carry a certain corollary satisfactory to the indeterminist. It will not        make sense to suppose that something else in any way caused o to produce d at ta, since there occurred no change in o.

To repeat, what is this so-called causal relation between o and d? We may be told, perforce, that what is in question when o causes d is not what we have called standard causation. That is no news. What is in question, it may be said, is a somehow more basic or older or less controversial conception of causation, or in fact the conception of causation. It is a conception of causation, so-called, encountered earlier in this inquiry, (pp. 57 ft.) A reminder is in order, and can be had by way of a typical exposition and defence of the conception, offered by a strong defender of an indeterminist theory.

Suppose someone throws a stone at a window and that the stone strikes the glass and the glass shatters in just the way we should expect. . . . Suppose further that God reveals to us that the glass did not have to shatter under these '         conditions, that there are possible worlds having exactly the same laws or nature as the actual world and having histories identical up to the instant at which the stone comes into contact with the glass, but in which the stone ,         rebounded from the intact glass. . . . Could this revelation really lead us to say that, despite appearances, the stone didn't cause the glass to break? That this is a logical consequence of the revelation? Wouldn't it be more reasonable to say this: that, while the stone did cause the window to break, it was not determined that it should; that it in fact caused the window to break, though, even if all conditions had been precisely the same, it might not have?

This case convinces me that whatever the facts of the matter may be, it is at any rate not part of the concept of causation that a cause—or even a cause plus the totality of its accompanying conditions—determines its effect, (van Inwagen, 1983, pp. 139-0)

Philosophers have a history of discovering scandals in the thought of their opponents. It seems to me a small scandal that this sort of persuasion, without even any colour given to it by talk of probability, should convince anyone to contemplate the idea that we have any such ordinary idea of causation. If we entirely put aside all of our firmly based causal preconceptions about stone and glass—the argument more or less depends on not doing so—what we have is the supposition that today there occurred an event, and it might not have occurred despite the fact that everything was as it was up to the moment of its occurrence. We have the related supposition that taking into account all relevant factors, we might get an exact and total repetition of them tomorrow, or daily for the rest of time, without the glass breaking. If so, it is to my mind absurd to speak of today's breaking of the glass as an effect, as something caused. For a clear reason, the event was no effect in any ordinary sense and we have no case of causation in any ordinary sense. The reason, expressed one way, is that there exists no explanation whatever of today's event. Anything and everything we can contemplate as a possible explanation, as we know, might have been followed by no event. Should we still be persuaded that the stone caused the given event since, although it did not necessitate it, it did 'produce' it? (van Inwagen, 1983, p. 4) Most certainly we should not. The stone 'produced' the breaking only in some wonderful sense consistent with the proposition that stone and all else might have been exactly as was, without a breaking. It is notable that the philosopher attracted to the given line of thought, earlier in his often admirable book, has the following to say: 'Causation is a morass into which I for one refuse to set foot. Or not until I am pushed.' (van Inwagen, 1983, p. 65) That reluctance, one might say, came home to roost.

To return to the originator o, and the occurrence of the decision d, we thus have no account of how the decision is more than a chance event. To say that the originator caused the decision, in the given sense, is in fact to add nothing whatever to the proposition we have, that the decision was unnecessitated. Needless to say, things do not become better, but still worse, when we contemplate what else is proposed, that what somehow causes something may in a further sense do so out of nothing. In the story of the stone we at least have a so-called causal event. In the story of the unchanging originator, no event at all occurs. If there is nothing that explains why the so-called effect occurs at all, there is also nothing that explains why the so-called effect occurs when it does.

3.8.4 The Will

It can be said of the originator, greatly more persuasively, that it is the will. (Kenny, 1975, 1978) It is one of the two great faculties of the mind, the other being the intellect, both of them being bound up with the creation and use of symbols. More specifically, the will is a capacity or power at least to intend, choose, or decide to do things. Further, it is a rational capacity or power, and thus to be distinguished from a natural or irrational capacity or power.

Aristotle . . . drew a sharp distinction between rational powers . . . and natural powers like the power of fire to burn. If all the necessary conditions for the exercise of a natural power were present, then, he maintained, the power was necessarily exercised: put the wood, appropriately dry, on the fire, and the fire will burn it; there are no two ways about it. Rational powers, however, are essentially, he argued, two-way powers, powers which can be exercised at will: a rational agent, presented with all the necessary external conditions for exercising a power, may choose not to do so. (Kenny, 1975, pp. 52-3)

It must be, I take it, that we are .to derive the idea of a rational power from a natural power, but this is not easy. It is not easy, in part, since we are not to understand that a natural power is a property of a thing such that the combination of it with other items forms a causal circumstance for some event. A natural power is nothing so clear or palpable. Some natural capacities are of our bodies, but they are not to be reduced to the structural parts and features in virtue of which we possess those capacities. They themselves are something other or something more than properties, or at any rate ordinary properties, of our bodies. This, it may seem, is other than fully clear.

We may, for want of something fully explicit, start from what we know, the clear idea of a natural power. We can then suppose that the rational power which is the will is in part this: a property of a person such that when it is conjoined with other things, the combination is such that it may or may not issue in a decision or whatever. There may be no difference whatever between a situation where the power is exercised, so to speak, and a situation where it is not. In fact, we are to understand, there is no situation whatever such that in it a rational power will have a standard effect, a necessitated event.

What must be said of this, in line with earlier remarks, is that we are offered no explanation of why the will gives rise to a decision when it does. In fact, it is at least arguable that no adequate content is given to speaking of the will as giving rise to a decision. We have no adequate idea of what relation is in question. It is not as if we have an articulated account of how a decision may be more than chance or random, which account we may dispute. Rather, it seems, we have no such articulated account.

What is implied by the description of the will as rational power? It may be conjectured that what enters into the exercise of the given power is a set of reasons pertaining to the decision or whatever. A reason may be specified by its prepositional content—'The train leaves at 5 p.m.'—and a decision may also be specified by its content—'I'll try to catch it'. There is thus what can be called, in a loose sense difficult to characterize, a logical connection between reasons and decision, a connection with which the will has something to do. Do we here have an account of the relation between the will and its decision?

It has been clear since Hume that causes do not give rise to their effects by entailing them. One thing does not cause another in virtue of the fact that a statement of the occurrence of the cause entails a statement of the occurrence of the effect. No such entailment is any part of the explanation by the cause of the effect. If, in what is said of reasons, there were intended a description of the occurrence of a decision in terms only of logical relations, the intention could not succeed. Certainly just this could not be carried forward into anything of the right sort. No amount of reflection on logical relations in themselves could bring us closer to having an explanation of why a certain event occurred. That is what we are after.

To look quickly at a further doctrine, it may well be true that in what is called practical reasoning, premisses which stand in some logically generative relation to a conclusion may cease to do so if a consistent addition is made to them. My conclusion that I will try to catch the train no longer stands when I learn that the London meeting to which I was going has been cancelled. Thus there is a certain want of analogy between a set of reasons and a causal circumstance. The reasons are defeasible. (Kenny, 1975, pp. 22-3) To see just that, however, is not to see how the logical relation between the unsup-plemented premisses and the conclusion gives anything like an explanation of an event, however defeasible. The point is made clearer by the fact that a determinist has no concern whatever to question or deny the given logical relations. What he will be sceptical of is quite different: that sense has been given to talk of a kind of explanation of events which is somehow akin to causal explanation and yet lacks its fundamental proposition, about necessitation.

The doctrine of the faculty of the will is impressively elaborated, more so than has been indicated, and it should have more attention. Some, not all well-judged, has been given elsewhere. (Honderich, 1980b) The doctrine is indeed ancient, but it is difficult to suppose that it owes its longevity to its own recommendations, as distinct from what it promises to us in terms of our aspirations.

3.8.5 Primitive Relation

It will not be surprising, given our survey, that some philosophers inclined to indeterminism take a certain step. They abandon the unequal struggle, and conclude that no analysis is possible of the relation between an originator and a decision. Their view is that there does exist a relation which satisfies the goal of indeterminist theories, but that no more can be said of it than that. What can be said, that is, is no more than this: a decision is not necessitated by an originator, but is related to it in such a way that it is a grounded or non-arbitrary event. The decision is such that the agent is responsible for it in the required sense—he could have decided otherwise given things just as they were.

It is typically the case that this unnerving recourse to an unanalys-able relation is somewhat concealed by certain talk. It may be said, for example, that the originator 'acts' with respect to the decision, or possesses 'active power', or of course is 'causal'. Taken together with the absence of any analysis of the relation, and the proposition that it is in fact unanalysable, it follows that we must divest the terms 'act', 'active power', and 'causal' of any ordinary sense. To say the originator acts, or has an active power, or causes a decision, can mean nothing whatever other than that it does not necessitate a decision but is so related to it as to make the agent in the given way responsible for it. There is exactly the same situation with various other terms and locutions that may be pressed into service.

There are also overt attempts to make more acceptable the claim of an unanalysable relation. Some are to the effect that there is nothing extraordinary about the given recourse. Certain of the things said here are themselves extraordinary. It is claimed that the assertion of an unanalysed and unanalysable 'causal' relation is in no way peculiar, since all talk of causation must involve such a primitive idea, all talk of causation is in the given way 'unclear'. The indeterminist, we are told, is in just the position of any philosopher concerned with standard or 'event' causation. 'He is not so much introducing mystery into the world as introducing more mystery into the world; the event causality with which we seem so comfortable is itself unfathomably mysterious, as any glance at a freshman metaphysics text will show.' (Thorp, 1980, p. 106; cf. Chisholm, 1966, pp. 21-2; Taylor, 1966, Chs. 2, 3)

The reply is not difficult. Despite the existence of problems with respect to our standard conception of causation, and dispute about it, the comparison with the given unanalysable 'causation' must surely be regarded as absurd. Our opening inquiry into our standard conception of causation, if it establishes nothing else, establishes that that conception is not unanalysable or in any remotely comparable sense 'mysterious' or 'unclear'.

Putting aside other like attempts to coat the pill of a primitive and unanalysable relation (Thorp, 1981, pp. 108-9), it needs to be said that those who wish to swallow the uncoated pill cannot be deterred. More plainly, there is no possibility of rejecting by philosophical or conceptual means the proposition that there exists a certain unanalys-able relation. Since, so to speak, nothing is said of it in itself, as distinct from what it secures, a certain responsibility, there is no account of it on offer to be examined for shortcomings.

At the same time, it is difficult to give attention to the supposition of a primitive relation between originator and decision. It is all very well to say, in a different sort of attempt to make the supposition tolerable, that all explanation must begin somewhere, with something taken as somehow primitive. (Boyle, Grisez, Tollefsen, 1976) However, there is all the difference in the world between beginning with a clear or contentful proposition taken as primitive, and beginning with something quite otherwise.

Figure 17 on p. 181 gave elements of indeterminist theories of the mind. Figure 18 can hardly be said to complete the picture, since, as we know, there are obstacles in the way of that enterprise. (Cf. G. Strawson, 1986, p. 32) It does add the Correlation Hypothesis, modelled by the vertical bars joining m and n, and m' and n', and indicates the obstacles. One is the relations of probabilistic causation, holding between n and n* and their 'upshot' n', and between n' and n** and their 'upshot', the action a. The further obstacle, over and above the mysteriousness of the originator o itself, is the relation between it and m', where m' is taken as a decision or like mental event.


The second hypothesis of the determinism being expounded, the Hypothesis on the Causation of Psychoneural Pairs, if opposed, must have something put in its place. The worth of the hypothesis, and the subordinate Proposition of Psychoneural Pairs, and the Union Theory, are partly fixed by what is available instead. If the earlier alternatives we considered—Neural Causation with Psychoneural Correlation and so on (3.2)—were open to objection, they were in a general way superior to the alternatives we have lately been considering. The indeterminist theories and ideas of the mind can hardly be regarded as conceptual challenges to the Hypothesis on the Causation of Psycho-neural Pairs. More precisely, these theories and ideas cannot be regarded as alternatives at the same level of contentfulness, clarity, coherence, and development.

This is a matter, first, of their first element, the denial of nomic connection and the substituted probabilistic account of the neural background of the correlates of mental events and actions. That account, or rather the promise of that account, is replete with difficulties. It is in internal tension, if indeed it can escape contradic-tion; it includes ad hoc stipulation, and so on. To remark on its ad hoc nature in particular, perhaps, is to move in the direction of what is not yet our official subject, the question of the empirical basis of our determinist theory and of indeterminist theories. However, there is not quite so sharp a line between empirical basis and conceptual adequacy as might be desired.

The second element of any indeterminist theory, an originator, is yet more unsatisfactory. There is no avoiding the need for an idea of an ongoing entity of a certain kind, and no avoiding the fact that the idea is rebarbative. It offends dramatically against the principle of simplicity. It has never escaped widespread philosophical and other suspicion.

Finally, in so far as the distinction between the second and the third elements can be maintained, the account that can be given of the originator's unique activity, the activity itself, is thin nearly to the point of non-existence. At bottom we are told no more of this activity, so-called, than that it is such as to secure that the person is in a certain sense responsible for decisions and actions.

Indeterminist theories and thoughts of the mind are thus conceptu-ally unsatisfactory. They are rightly described, in all their three elements, as consisting in 'obscure and panicky metaphysics'. (P. F. Strawson, 1962, p. 211) The conclusion would stand if we were to carry our inquiry further—if we were to consider the possibility of combining indeterminist ideas we have treated separately, and so on. There is very good reason to think the conclusion will be true of their successors in the future.

That is not all that is to be said of indeterminist theories and thoughts, however. It must be said for them, however reluctantly, that they are not without content. It comes to this:

In each of us there exists an ongoing entity or attribute which originates decisions, and hence actions; this origination is such as to give rise to a certain responsibility; our decisions are nomically correlated with neural events, but these neural events, like the subsequent actions, are unnecessitated.

We shall in due course look at what is said to be empirical evidence for the proposition. (5.5, 5.6) One other fact about this indeterminism, a larger fact to which we shall also come later, explains why its has a future. It is a response to what is itself a large and indubitable fact, a fact of our deep attitudes and beliefs, of which much will be said.

Editor's Postscript
This inquiry into the traditional and still orthodox accounts of the explanation of brain and consciousness is different indeed from Ted Honderich's subsequent thinking. The traditional and orthodox accounts consist in what can be called cranialism about consciousness. Compare the theory of Radical Externalism or Consciousness as Existence. Go to Radical Externalism: Honderich's Theory of Consciousness Discussed. Also the entries on consciousness in the index of this website and in particular the paper Consciousness as Existence, Devout Physicalism, Spiritualism.
If you wish to quote from the inquiry above, by the way, it would be a good idea to get hold of one of the books from which it comes. Books do not generally have mistakes introduced by scanning machines.

HOME to Det & Free Website front page
HOME to T.H. Website front page