where we recognize that the external reality is not open to perceptual scrutiny, and are seeking some new basis on which to acquire information about it, knowledge is blocked by the fact that no new basis is available. Indeed, there is no basis even for well-supported conjecture.
This is a bleak conclusion. And it surely further means that we have lost our only form of protection against the other two ways in which the denial of perceptual access was seen as threatening. Thus, once we take the external reality to be beyond the reach of both perception and knowledge, I do not see how we can continue to think of it as our world in anything approaching the ordinary sense. We can continue to think of it as the reality to which we corporeally belong—containing those substances (our bodies) with which our mentality has direct causal links. But we cannot come anywhere near thinking of it as the world in which we live and move and have our phenomenological being, in the way we ordinarily suppose. Nor, surely, can we continue to think of the reality as containing physical objects of the familiar types—objects such as trees, mountains, houses, and cars. Without perceptual access, it was always going to be difficult to think of these types as externally realized, since our very conception of them seems to involve thinking of the things which they characterize as having forms of sensible appearance. But, with the additional loss of epistemic access, we cannot even invoke the form of substitute we earlier envisaged, whereby the external items (though incapable of sensible appearance in the strict sense) reveal their presence by causing experiences which represent certain types of sensible situation. Putting everything together, it seems that, as we earlier feared, we face the wholesale collapse of our ordinary understanding of the nature of the external reality and our relationship to it.
In fact, in one respect, our situation has turned out to be worse than we anticipated. We might have thought, initially, that, whatever the problems over epistemic access to the external realm, we at least possess a rich knowledge of how things stand in the realm of our own mentality. In particular, we might have assumed that we have a rich knowledge of those experiential facts which ordinarily make it seem to us that we are perceptive inhabitants of a physical world, and so know that, even if our physical beliefs turn out to be unwarranted, they at least reflect genuine aspects of how our experience is organized. But, as we have seen, no such knowledge is in fact available. For our access to the mental realm itself depends, to a large degree, on our possession of physical knowledge. Both in our acquisition of beliefs about the minds of others and in most of what we take ourselves to know about our own experiential pasts, we crucially rely on our capacity
end p.241
 
to acquire and retain information about the physical world. So, with the loss of epistemic access to the physical realm, we lose it, in large measure, to the mental realm as well.
5 Taking Stock
It looks as if our whole investigation is now at an end. We set out to discover the nature of physical-item perception. Curiously, what we seem to have discovered is that there is no such phenomenon to be investigated. There is no account of what it is for a subject to perceive a physical item, because, with sense-qualia as the immediate objects of awareness, perceptual access to the external reality is excluded. To make matters worse, we also seem to have discovered that, in lacking perceptual access to the external reality, we lack epistemic access too. We lack this access both in our naive state, where we take for granted the perceptualist assumption, and in our state of philosophical enlightenment, where we have come to accept that this assumption is false. Finally, we seem to have established that, in our enlightened condition, we not only lack knowledge of the external reality, but also lack any rational basis for reaching significant conclusions about it; and this, of course, would mean that we lack any rational basis for retaining our ordinary physical beliefs. These are not the conclusions which we expected at the outset of our discussion; nor are they what we would have hoped for. But, given our acceptance of SQT, there seems to be no way of avoiding them.
This outcome is bad enough. But it also leaves us with a further problem. There is no denying that these negative conclusions about perception and knowledge are both surprising and disturbing: they undermine the whole basis on which we ordinarily form our beliefs and conduct our lives. But given that they are the conclusions—seemingly established by our philosophical arguments—the expectation must be that we shall now, however reluctantly, accept them. But the fact is that we do not. We acknowledge that the arguments seem to show that there is no perceptual or epistemic access to the external reality. But we find that our belief in such access remains intact. Nor is this just a matter of how we cognitively respond at the level of our ordinary thinking—the fact that we continue to take our perceptual and epistemic capacities for granted in the course of everyday life. For we also continue to endorse this common-sense view at the level of philosophical reflection, where we are fully aware of the case against it. This, at least, is what I find in my own case. Despite the arguments
end p.242
 
I have advanced, I remain convinced that I see, feel, and in other ways perceive physical items, and that these perceptual encounters provide me with a wealth of genuine information about their character and physical setting. And I can only assume that this will be the typical response of others who reflect on the situation.
Now this response, of course, does not count as an argument against the conclusions to which our discussion has brought us. For all we have shown, our failure to accept these conclusions might stem from some special feature of our psychological makeup. Perhaps we simply lack the psychological capacity, even in our reflective thinking, to accommodate such a radical and destructive departure from our natural ways of thought. And indeed, it would not be difficult to understand, whether in evolutionary or in creationist terms, why the common-sense outlook should be protected in this way. But what is also clear is that, given the nature of our cognitive response, we have no choice but to think that the issue of perception is not yet settled. If we cannot, even on reflection, eliminate the conviction that we are inhabitants of a physical world, whose ingredients are perceptually accessible to us, and whose character is revealed through perceptual encounter, we cannot avoid thinking that there is something wrong with the philosophical arguments which purport to show that that conviction is false. And, if we think that the arguments have somewhere gone astray, we are bound to feel a corresponding pressure towards trying to identify the place (if it exists) where this has happened.
end p.243
 
Abstract: The only way of avoiding the bleak conclusion of Part 5 would be to abandon physical realism for a form of idealism. The kind of idealism needed would be one that took the physical world to be something logically created by the organization of human sensory experience, or by some richer complex of non-physical facts in which that sensory organization centrally features. This would restore perceptual access to the physical world, since if the world is logically created in the way envisaged, awareness does not need to reach beyond the presented sense-qualia to make contact with the physical items. And, with the restoration of perceptual access, there would be the restoration of epistemic access too. This idealist position not only offers the only way of securing perceptual and epistemic access to the physical world but is also something whose truth can be established on independent grounds.
Keywords: epistemic access,idealism,perceptual knowledge,physical world,realism,sensory experience
Part Five The Idealist Solution
1 The New Option
I
So far, our investigation into the nature of perception has focused exclusively on two rival general approaches, and the more specific theories which fall under them. One approach is that of the strong, or full-blooded, version of direct realism (SDR). This accepts a realist view of the physical world—taking the world to be something whose existence is logically independent of the human mind, and something which is, in its basic character, metaphysically fundamental. And, in the framework of that realism, it asserts that, whenever a subject Φ-terminally perceives a physical item (i.e. perceives it, without there being any other physical item which he perceives more immediately), his perceptual relationship with it is something psychologically fundamental. The other approach is the broad, or flexible, version of the representative theory (BRT). This too accepts a realist view of the physical world, but, in contrast with SDR, insists that, whenever a subject perceives a physical item (and even when the perceiving is Φ-terminal), his perceptual relationship with it breaks down into (is constituted by the combination of) two components, one of which consists in the subject's being in some more fundamental psychological state—a state which is not in itself physically perceptive—and the other of which consists in certain additional facts, but ones which do not add anything to the subject's psychological condition at the relevant time.
If my arguments have been correct, neither of these approaches allows the development of a satisfactory account. The problem for SDR is over its handling of phenomenal content (the phenomenal manner in which the relevant physical item is perceived). If such content is taken to be just the presentational imprint of the external situation on the percipient's mind, cases of non-veridical perception (where sensible appearance is at variance
end p.244
 
with the external situation) are—incorrectly—excluded. But if such content is construed in the internalist way (so that the featuring of a quality in the content is ontologically separate from its external realization), the only way of making sense of its involvement in perceptual contact is in mediational terms, which is just what SDR rules out. The problem for BRT, in contrast, is concerned with the issue of perceptual contact. If the psychological states which are fundamentally involved in cases of supposed perception are not in themselves physically perceptive, it is hard to see how the subject's awareness can be thought of as reaching beyond the boundaries of his own mind; and it is quite impossible to see how there could be a reaching to anything external when BRT assumes the SQT (SDT) form which I have defended. The upshot is that, if these are the only positions available, and if my arguments have been correct, we do not have perceptual access to the physical world at all. And this further means that we do not have epistemic access either. What makes these conclusions so awkward is not just that they are unpalatable, but that we cannot bring our actual convictions into line with them—not even when we endeavour to consider the issue from the most objective, philosophically reflective, standpoint.
The situation is looking bleak. But there is still one possibility which we have not yet considered. So far, we have conducted the whole of our investigation within the framework of a realist view of the physical world; and, within that framework, SDR and BRT are the only available options. What we have still not explored is the possibility that this realist view is itself mistaken, and that its being so is what has led to our present difficulties. Adopting the framework of realism in our previous discussion was, of course, entirely reasonable. On the face of it, the realist view is the only position which does justice to our basic understanding of what a physical world is; and it is simply taken for granted in almost all current philosophical writing. It was only natural that we should begin by confining our attention to realist theories of perception, with the expectation that we would be able to reach a satisfactory outcome in these terms. But what is now inescapable is that the outcome in the realist framework has not been satisfactory. Our project of providing an account of the nature of perception has collapsed, and we are left with conclusions to which our philosophical reasoning commits us, but which we cannot bring ourselves to accept. In the light of this, it is clear that we can no longer avoid raising the question of whether the realist framework should be abandoned and an account of perception be sought along quite different lines.
end p.245
 
II
We must begin by considering in more detail what physical realism involves. As we have formulated it, the realist thesis is the conjunction of two claims: that the physical world is something whose existence is logically independent of the human mind (the mind-independence claim); and that the physical world is, in its basic character, metaphysically fundamental (the fundamentalist claim). Before we can approach the question of whether realism should be abandoned, we need to know how these two claims are to be understood.
The mind-independence claim is relatively straightforward. It means that facts about the human mind do not logically contribute to the existence of the physical world. But there are two points that need to be noted in this connection. The first is that when I speak here of 'facts about the human mind', I mean facts about the potential, as well as the actual, occurrence of human mentality, and I mean facts about the potential occurrence of mentality in potential, as well as actual, human subjects. So if someone thought that what logically creates the physical world, and the various facts it includes, is some set of general psychological laws—laws which concern some area of human mentality, but whose obtaining does not require the existence of human subjects—he would not be endorsing the independence-claim in the relevant sense, or have any chance of qualifying as a realist. The second point is that the claim of mind-independence should not be thought of as extending to every aspect of the physical world. Even a realist will recognize that certain physical facts are, for special reasons, logically dependent on the human mind—for example, facts about the secondary qualities of objects, construed in a Lockean fashion, and facts involving concepts like car and landmark, which imply that the objects in question are either of human design or serve certain human purposes. All that the realist is insisting is that the primary core of physical reality is mind-independent. This core will include, in particular, the existence of physical space, and the presence within it of the various forms of space-filling material object, characterized in primary-quality and non-anthropocentric terms. In case it is not obvious, I should stress that whenever, in the context of the present discussion, I speak of mind-dependence, or mind-independence, I always mean dependence on, or independence of, the human mind.
The fundamentalist claim is a little more complicated. A fact is metaphysically fundamental (or simply fundamental) if and only if it is constitutively
end p.246
 
basic, i.e. not constituted by other facts; and a fact F is constituted by a fact or set of facts F' if and only if it obtains in virtue of, and its obtaining is nothing over and above, the obtaining of F'. Now, irrespective of how the philosophical issues are to be resolved, it is undeniable that there is a wide range of physical facts which are not metaphysically fundamental in the relevant sense. And, indeed, this range will include many facts that belong to the primary core of physical reality which the realist takes to be mind-independent. For instance, the fact that one chunk of matter contains more atoms than another will belong to this primary core; but there is no denying that it is constituted by the combination of the relevant facts about the numbers of atoms in each chunk. It is because not all physical facts could be thought of as metaphysically fundamental that the realist claim is limited to the 'basic character' of the physical world. What is being asserted is that there is some non-empty set of physical facts which are metaphysically fundamental (constitutively basic), and that any other physical fact is ultimately constituted by some complex of facts drawn from this basic set, or by some such complex together with certain non-physical facts. Although this basic set will not include all the facts which the realist takes to be logically independent of the human mind, it will obviously be confined to them.
Even with its restriction to the basic character of the physical world, there is a theoretical risk that the fundamentalist claim, in its present form, will not serve the realist's purposes. This is because the realist might want to allow for the possibility that constitution within the physical realm is, in certain areas, infinitely regressive, so that there are some physical facts which are neither constitutively basic nor constituted by facts which are constitutively basic; and, indeed, he might want to allow for the possibility that this is so quite generally. In other words, while insisting that there is no reality more fundamental than the physical world—nothing non-physical which constitutively underlies it—he may want to allow for the possibility that, in certain areas, or perhaps quite generally, the physical world does not have a character which is basic in the relevant sense. To allow the realist this option, we would need to formulate the fundamentalist claim in a different way. We would need to express it as the claim that physical facts are never wholly constituted by non-physical facts. This would preserve the spirit of the original claim, since it would imply that there is nothing with a more fundamental status than the physical realm. But it would leave open the possibility that there is no level of physical reality that is not less fundamental than some other.
end p.247
 
For a realist who wants to leave room for a regress of physical constitution, formulating the fundamentalist claim in this way is what is strictly required. But, in my own discussion, I shall simply ignore this complication and retain the claim in its original form. In doing this, I am not limiting the significance of the conclusions I shall reach. The various points that I am going to make could all, with care, be re-expressed in a way that takes account of the possibility of regress. But the discussion will be altogether simpler and clearer if we set this possibility aside, and work on the assumption that the lines of physical constitution ultimately terminate. In any case, this assumption seems to me to be very plausible.
Because physical realism is the conjunction of these two claims, there are three ways in which we could think of rejecting it. Thus we could reject it by either (1) retaining the fundamentalist claim, but denying the claim of mind-independence, or (2) retaining the claim of mind-independence, but denying the fundamentalist claim, or (3) denying both claims.
The first option, it seems to me, can be quickly dismissed, since I cannot think of any remotely plausible way in which it could be developed. If we deny that there is even a primary core of physical facts which are logically independent of the human mind, we obviously have to offer some account of why this is so. How is it that facts about the human mind logically contribute to the existence of physical space and the presence of material objects within it? The obvious explanation would be that physical facts are, quite generally, constituted by non-physical facts, and that facts about human mentality form, or systematically feature in, the non-physical facts involved. But this explanation is explicitly excluded by the continued endorsement of the fundamentalist claim. The difficulty now is in seeing what viable alternative is left. The only possibility I can think of would be to say that the physical world, or some essential component of it, can be literally equated with something in the realm of human mentality. But there is nothing along these lines which even begins to make sense. The only suggestion which even comes to mind would be the crudest application of Berkeley's doctrine of esse est percipi—an application in which physical objects are identified with collections of human sensory ideas (in my terminology, collections of occurrences of sense-qualia). But such an identification would simply not leave the objects involved as physical objects in any recognizable sense. In particular, it would neither allow them to possess anything approaching the full complement of their qualities, nor allow them to form elements of a spatially unified whole. Even Berkeley, as I
end p.248
 
interpret him, was led to develop his doctrine in a more subtle way, to avoid the obvious objections.1
The second option, in which we reject the fundamentalist claim, while retaining that of mind-independence, can also be set aside, though here the reason is quite different. Our interest in the possibility of rejecting realism has arisen in a specific dialectical context. Thus we have been looking for a satisfactory theory of perception, and have found that, in the framework of realism, no such theory is forthcoming. Our interest in the possibility of a non-realist approach is with a view to solving this problem. Our hope is that, by abandoning realism in the right way, we shall be able to offer a radically new account of what physical-item perception involves, and thereby be able to show how perceptual access to the physical world is possible. Now what created the problem, in the framework of realism, was the combination of three factors: first, that, in cases of what we ordinarily take to be perception, the immediate objects of awareness are sense-qualia; second, that there is no way of understanding how perceptual awareness can reach beyond these qualia to items in the external, mind-independent reality; and third, that, from the standpoint of realism, perceptual awareness would have to reach to such items if it was to make contact with the physical world. But obviously, the abandoning of realism cannot help with this problem if it preserves the claim of mind-independence. For if the world has an existence which is logically independent of the human mind, we get exactly the same obstacle to perceptual access, irrespective of whether we keep the fundamentalist claim or not. Taking physical facts to be constituted by non-physical facts will not be of any assistance to the cause of perception unless it moves the physical world closer to the realm of human mentality.
It follows that, given our investigative concerns, we can confine our attention to the third option, in which physical realism is rejected in both its claims—in which we both deny that the existence of the physical world is logically independent of the human mind, and deny that it is, in its basic character, metaphysically fundamental. Even here, of course, there is a wide range of options available. But, from now on, I shall focus exclusively on a single non-realist position, which seems to me to offer the only real prospect of providing both a satisfactory account of the physical world and a satisfactory solution to the problem of perception. The position
1 See my essay 'Berkeley on the physical world' in J. Foster and H. Robinson (eds.), Essays on Berkeley.
end p.249
 
in question is a form of idealism—idealism of a broadly phenomenalistic kind. And since it is the only form of idealism that I shall be concerned with, I shall simply refer to it here as 'idealism' or 'the idealist position'. Let me start by offering a brief exposition of what this position involves.
III
As I have often stressed, human sensory experience is characterized by themes and regularities suggestive of the existence of an external three-dimensional spatial world and our own location as mobile percipients within it.2 This world-suggestive order can be thought of as stratified into a number of successively dependent levels, so that, at the base, we have themes and regularities which apply to sensory experience directly; at the next level up, we have themes and regularities which characterize the notional world whose existence is suggested by the base-level aspects of order; and so on. This stratification can be represented in different ways, according to how finely or crudely we distinguish the levels. But we can conveniently think of the whole series of levels as dividing into three broad phases. Thus, covering the base and near-to-base levels, there is a primary phase, in which the themes and regularities suggest the existence of a purely sensible world—a world consisting of a three-dimensional space, and, for any time, an arrangement of sensible qualities within it. Then, resting on that basis, there is a secondary phase, covering the middle ground, in which the themes and regularities of the sensible-quality arrangements thus suggested suggest, in turn, the existence of the physical world of our everyday beliefs—the world of ordinarily observable and sensibly characterized material objects. Finally, there is a third phase, in which the themes and regularities of this notional material world suggest the truth of certain scientific theories about its ultimate nature—a nature very distant from the perspective of our everyday beliefs. It may seem forced to speak of the aspects of order which emerge in the second and third phases as characterizing sensory experience, since it is not directly in terms of such experience that they are defined. But the point is that it is ultimately due to the way things work out at the level of sensory experience that these higher-level aspects obtain. The themes and regularities at the third phase are those which are directly fixed by the character of the notional material world
2 I am here ignoring, because it is irrelevant to my expository purposes, the sceptical challenge that was thought to arise for this claim as a result of the problem of perception. (See Part Four, Section 4, III.)
end p.250
 
which the second-phase themes and regularities suggest, and those at the second phase are directly fixed by the character of the notional world which the first-phase (directly experiential) themes and regularities suggest.
Now the fact that sensory experience is, in this phased way, characterized by such themes and regularities is presumably not accidental: it presumably reflects the fact that there is some unitary system of constraints which control the course of experience and oblige it to be orderly in this world-suggestive way. I shall speak of this system of constraints as the 'sensory organization'. In the framework of physical realism, the presence of this organization (this system of constraints) would be ultimately explained in terms of the existence of the physical world itself and the manner in which human subjects are functionally embodied within it: sensory experience would be organized (constrained) in the world-suggestive ways because the physical world existed as something logically independent of the human mind, and, by its causal control of our experiences, imposed a reflection of its character on them. What is distinctive about the idealist position is that it sees the lines of dependence as running in the other direction. Thus, rather than taking the physical world to be what is responsible for the organization, it takes the presence of the organization to be, or to be a central component of, what logically creates the physical world. In other words, it claims that the very existence of the physical world, and all the facts which make up its specific character, are constituted by the obtaining of the sensory organization, or by some richer complex of non-physical facts in which this organization centrally features. The basic idea is that, whether on its own, or in the context of some richer complex, the sensory organization creates a physical world by disposing things to appear systematically world-wise at the human empirical viewpoint; and it creates all the details of its specific character by disposing things to appear world-wise in the relevantly specific ways. Since the sensory organization ensures a series of successively dependent levels of world-suggestive order, there is provision for the physical world to have a corresponding series of levels of composition, beginning with a purely sensible composition, and terminating in what is revealed by the most penetrating forms of physical science. For the idealist, of course, there is no question of eliminating or downgrading the sensible aspects of the physical world in favour of the scientific. For he will see these sensible aspects as directly sustained by the non-physical factors which he takes to be metaphysically fundamental. Even the ES-aspects, which are vulnerable to elimination in the framework of realism, will take their place as basic ingredients of the physical world—idealistically underpinned by the relevant aspects of the sensory organization. So
end p.251
 
the idealist will have no difficulty in endorsing our common-sense beliefs that ripe tomatoes are genuinely characterized by the redness of their standard look, and that sugar genuinely possesses the sweetness of its standard taste. This is of crucial importance epistemologically, as we shall see presently.
It might seem that the assignment of ES-qualities to the physical world violates the realization-exclusion (RE) claim which I defended in Part Three—the claim that the ES-qualities are not capable of realization outside the content of sensory experience.3 For even though the physical realization of these qualities is not, from the idealist standpoint, something logically independent of facts about their sensory realization, even an idealist (of the kind I am envisaging) does not want to equate physical realization with a form of sensory realization. And indeed, he wants to allow that an ES-quality could be physically realized, in virtue of the relevant aspects of the sensory organization and whatever other facts may be needed, but there be no associated sensory realization at all. However, this conflict with the RE-claim is only superficial. For, in the context in which we advanced the claim—a context in which we were taking physical realism for granted—we were simply not interested in the possibility of a realization of the ES-qualities that was not fundamental, and so any question of an idealistically sustained mode of realization was irrelevant. In retrospect, and with the option of idealism now before us, we can simply reformulate the claim as asserting, more weakly, that the ES-qualities are not capable of fundamental realization outside the content of sensory experience. This claim still stands, and can be thought of as a more precise formulation of what the original argument was trying to establish. Likewise, and without disturbing the efficacy of the earlier argument, we can reformulate the realization-conferment (RC) claim to read, more strongly, that the appropriate featuring of an ES-quality in the content of sensory experience suffices for its fundamental realization.
Idealism takes the physical world to be logically created by the sensory organization, or by some richer complex of non-physical facts in which this organization centrally features; and this organization is the system of constraints on human sensory experience which oblige it to be orderly in the relevantly world-suggestive ways. In theory, it would be possible for the idealist to think of these constraints as autonomous natural necessities within the sensory realm. Thus, assuming that he has some way of thinking of different human subjects as forming an experiential community
3 See Part Three, Section 4.
end p.252
 
independently of their possession of a common physical world, then, in theory, he could construe the sensory organization as a law, or set of laws, requiring the sense-experiences of this community collectively to exhibit themes and regularities of the relevant sort. But, in practice, the suggestion of fundamental laws of this kind, with such a complex prescriptive content, and controlling events in what almost seems a purposive way, is hard to take seriously. And, for this reason, it is almost inevitable that the idealist will see the sensory organization as imposed on human subjects by something outside the sensory realm, and indeed outside the realm of human mentality altogether. In other words, he will suppose that there is some form of external reality—a reality which is separate from, and logically independent of, the realm of human mentality—which causally controls the course of human sensory experience, and whose modes of control form the sensory organization. In this respect, the idealist will find himself in agreement with the realist, who also accepts that there is an external reality which plays this causal role. But whereas the realist equates this reality with the physical world, the idealist takes it to be something which underlies the physical world—something which sustains the constraints on experience by which, wholly or in part, the physical world is created.
There is a variety of different forms which the idealist's external reality could take. But, given the special way in which sensory experience is constrained—the world-suggestiveness of the themes and regularities that the sensory organization ensures—it is plausible to suppose that this reality in some way 'encodes' the character of the relevant type of world, and that this encoding expresses itself through the ways in which the reality exerts its control over sensory experience. There are two contrasting ways in which such an idea could be developed. On the one hand, the idealist could take the external reality to be something which functions, both internally and in its causal control of human experience, in a purely mechanistic way—something which is governed, and linked with the human mind, by natural laws. Here the character and organization of the world suggested by sensory experience would in some way reflect the character and organization of the external reality itself—though the method of reflection may not be straightforward. On the other hand, the idealist could think that what directly controls our sensory experiences and ensures their world-suggestive character is a powerful rational agent (or conceivably team of agents), who (which) has a plan of the kind of world that he (it) wants the course of experience to suggest, and who (which) executes this plan by causing us to have experiences of the appropriate sorts. This, famously, was the approach of Bishop Berkeley, who took the external rational agent to
end p.253
 
be the Christian God. The first view of the external reality would, of course, bring the idealist into still closer agreement with the realist, who also takes the external reality to work in this mechanistic way. Indeed, curiously, it could lead to a situation in which there was no difference at all between the realist and idealist over the nature of the fundamental facts. Both might think that, at the level of what is fundamental, there is a certain kind of law-governed external reality, with a certain type of structure, and certain types of interactive link with human mentality, and they might agree in detail about the precise form that all this takes. The only difference would then be that, according to the realist, this external reality is what forms, in its basic character, the physical world, while, according to the idealist, it is what underlies it.
We have been thinking of the various ways in which the idealist might understand what is involved in the obtaining of the sensory organization; and we have noted that, in practice, he is almost certain to think of this organization as imposed by some form of external reality—a reality which is logically independent of the human mind, and whose causal control of human sensory experience forms the relevant system of constraints. What we should also notice is that there is a further way in which the recognition of an external reality may feature in the idealist's account. The opportunity for this further featuring arises from the open-ended way in which the idealist thesis has been formulated, as I shall now explain.
I did not define idealism as claiming that the physical world is logically created by the sensory organization, but as claiming that it is logically created either by this organization, or by a richer complex of facts in which it centrally features. So there is room for a range of more specific idealist accounts, according to whether the organization is taken to be creatively sufficient on its own, and, if it is not, what kind of richer complex is thought to be involved. One issue here will be concerned with whether, in addition to the organization, the idealist should think that there are certain further aspects of human psychology that contribute to the creation of the physical world—for example, the fact that we are disposed to respond to the sensory themes and regularities in certain interpretative and cognitive ways. But another issue will precisely be concerned with the role of the external reality. In one respect, of course, this reality is already directly involved in the idealistic creation, since it is the ways in which it causally controls the course of sensory experience that forms the sensory organization. But the idealist might also want to insist that the intrinsic character of this reality is directly involved as well. In other words, he might want to insist that part of what idealistically sustains the existence of the physical
end p.254
 
world is the fact that the reality which controls our sensory experiences, and ensures their world-suggestive character, is of a certain specific type. To take a simple example, a Berkeleian idealist is likely to insist that, in addition to the way in which God controls the course of sensory experience, the creation of the physical world is partly dependent on the fact that it is the Christian God, with his distinctive attributes and sublimity, who is playing this role. He is likely to insist that any other form of supernatural agent, or team of agents, exercising the same control, would at best create the systematic illusion of a physical world, or would do so unless it was acting under the authority of the Christian God.
One thing which clearly emerges from our recent discussion is that idealism, as I have defined it, is a highly generic position, which can be developed in a number of different ways, according to how certain key issues are resolved. Some of these issues relate to the nature of the sensory organization—to what it is that ultimately ensures the world-suggestive character of human sensory experience. Others concern the nature of the additional factors, if any, which combine with the organization to create the physical world. And, on both these fronts, the topic of the external reality prominently features. Recognizing that it leaves open these further issues is crucial to our understanding of what idealism, in the relevant sense, is. But, in what follows, it will be on the issue of idealism as such, rather than on its different versions, that we shall primarily need to focus.
IV
Idealism takes the physical world to be logically created by the sensory organization, or by some richer complex of non-physical facts in which this organization centrally features. As such, it rejects both parts of the realist thesis. It takes the physical world to be something whose existence is logically dependent on facts about the human mind, and it holds that physical facts are, in all cases, constituted by non-physical facts.
It is not difficult to see how, in rejecting realism in this way, idealism would dispose of the problem of perception. As we noted, this problem arose from the combination of three factors, namely: (1) that the immediate objects of awareness, in the relevant cases, are sense-qualia; (2) that there is no way of understanding how perceptual awareness can get beyond these qualia to items in the external reality; and (3) that, given the truth of physical realism, perceptual awareness would have to reach to external items to make contact with the physical world. The idealist position would immediately eliminate the third of these factors. It would mean that
end p.255
 
physical items were no longer constituents of a reality that was external to qualia in the relevant sense. Rather, they would derive their very existence from the constraints on the occurrence of qualia, or from some complex of facts in which such constraints centrally featured. It would remain the case that the subject's awareness did not reach beyond the sense-qualia occurring in his own mind. But, on the idealist account, it would not need to reach further to make contact with the factors by which the existence of physical items was constituted, and thereby make contact with the items themselves.
To illustrate, let us go back, once again, to the case of Pauline. We want it to turn out that, at any given time during the period when she photically encounters the apple, Pauline Φ-terminally sees a certain momentary stage of a portion of the apple's surface. In the framework of realism, we cannot understand how this perceptual contact is made. At any time, the only item immediately before Pauline's mind is a certain visual sense-quale, and, in the realist system, the only way in which the occurrence of this quale would be connected with the relevant portion-stage would be by a complex causal process, running through the subject's eyes and brain. It is impossible to see how this causal connection could turn the awareness of the quale into one which genuinely reaches to the external item. But, in the idealist's system, the situation is quite different. There is still a sense in which the contact with the portion-stage is mediated: the occurrence of the sense-quale does not in itself qualify as an awareness of the physical item. But, in the new situation, the mediating and mediated objects are ontologically linked. For the occurrence of the quale is itself an instance of the operation of the sensory constraints by which, on their own or in combination with other factors, the existence of the portion-stage is constituted. In such a case, it is no more difficult to see how the presentation of the quale succeeds in giving access to the portion-stage of the apple's surface than to see how the perceiving of this stage gives access to the persisting portion, or the perceiving of this portion gives access to the whole apple. Nor, indeed, is it difficult to see how, despite the mediation, the access to the physical item can—in accordance with the traditional conception of idealism—be thought of as direct. Moreover, exactly the same situation will obtain, under idealism, in any other case of what we would ordinarily take to be Φ-terminal perception.
By introducing this kind of ontological link, then, between the occurrences of sense-qualia and the relevant physical item, idealism avoids the problem of perception that arises in the framework of realism. It shows how the subject's awareness does not need to reach beyond what occurs in
end p.256
 
his own mind to make contact with things in the physical world. Noting this, of course, does not, on its own, provide us with a full account of how physical-item perception works in the idealist's system. Nor, indeed, will I attempt to provide such an account here; for that would require me to elaborate the idealist's system in far greater detail than I have space for in the present work. Nonetheless, there are three basic points about the idealist's treatment of perception that I want to underline.
In the first place, the idealist is claiming that, where there is Φ-terminal perception, the occurrence of the relevant sense-quale is an instance of the operation of the sensory constraints that are involved in the creation of the perceived item, and he rightly sees this ontological link as eliminating the problem encountered by the realist. But he is not claiming that any form of such a link would be suitable for the occurrence of such perception. For there is a wide variety of ways in which the constraints that are involved in the creation of a physical item can operate on sensory experience, and only a fraction of these would enable the subject to make Φ-terminal contact, or indeed any kind of perceptual contact, with the item in question. The majority of the ways would, at best, serve to indicate the presence of the item in some less direct fashion—for example, by revealing aspects of its causal influence, within the idealistically created world, on other physical things. Whether a mode of operation would serve to give perceptual access to the item becomes apparent once we consider it in the context of the whole idealistic creation.
Secondly, the perception of a physical item is something which occurs from a location within the physical world. So the idealist has to arrange things in such a way that, given any perception, the factors which idealistically create the physical world, and which, in the context of that creation, provide an appropriate ontological link between the relevant quale-occurrence and the relevant physical item, create the world as something which forms the subject's environment, and which puts the subject and the item into the appropriate environmental relationship. In broad outline, we can see how this would work out. The basic idea of the idealist approach is, as I have said, that, whether on its own or in the context of some richer complex, the sensory organization creates the physical world by disposing things to appear systematically world-wise at the human empirical viewpoint; and, in disposing things to appear systematically world-wise at this (our) empirical viewpoint, it disposes them to appear in ways which are not only suggestive of a certain kind of world, but of a world in which we are located and through which we move. So the idealist can take our location in the physical world to be sustained in the same way
end p.257
 
as the existence of the world itself. There is still, of course, the question of how anyone could think that the existence of the world was sustained in this way at all. On the face of it, disposing things to appear systematically world-wise at the human viewpoint would at best create a virtual world, rather than one that was real. But, at present, I am only trying to give an account of the idealist position, not to defend it.
Finally, even with the subject's location in the physical world in place, and assuming a suitable form on the relevant occasion, the idealist will not want to say that the Φ-terminal perceiving of a physical item depends on nothing more than there being a quale-occurrence which is ontologically linked with the item in a perceptually appropriate way. For, in order for there to be such a perceiving, the subject has to have a complete phenomenal experience, and this means that the occurrence of the quale has to combine with an appropriate level of interpretation—an interpretation which is suitably experiential and which represents the quale as something in the subject's environment. Moreover, relative to the way in which, within the created physical world, the subject, at the relevant time, is spatiotemporally related to the relevant item, and relative to any other aspects of the physical circumstances that may be relevant, the phenomenal experience has to be, to a sufficient degree, qualitatively appropriate to that item. As under realism, such appropriateness will not be entirely a matter of veridicality, but will also take account of what is normal, or normative, for the conditions in question.
By restoring our perceptual access to the physical world, idealism eliminates the problem of perception. And, in eliminating this problem, it eliminates the problem of knowledge which stems from it. It does not do this by supplying us with a new (a distinctively idealist) source of physical information—a way of gauging the character of the physical situation by inference from the character of the non-physical factors which constitutively underlie it. For reasons which we have already made clear, even the sensory aspects of these factors are not ones which we can effectively monitor without prior knowledge of the physical world.4 Rather, the way in which idealism helps the epistemological situation is by showing how our ordinary methods of acquiring putative physical information turn out to be well founded. Thus, by taking the physical world to be logically created in the way envisaged, it ensures our perceptual access to it; and, once we know that this access is in place, we can take physical facts to be open to perceptual scrutiny in the ordinary way. Admittedly, there is one factor in
4 See Part Four, Section 4, III.
end p.258
 
the situation which this account does not make explicit. If perception is to be a source of physical knowledge, the basic beliefs which it induces have to be true; and the basic beliefs in question are always, at least in part, about how things sensibly stand in the subject's environment. A crucial part of what enables idealism to represent the world as open to perceptual scrutiny, and thereby to vindicate our ordinary methods of acquiring putative information about it, is that—as I stressed earlier—it allows the world to possess a genuine sensible character, and does so even in respect of the secondary qualities which are vulnerable to a Lockean account in the framework of realism.
Idealism eliminates the problem of perception, and the problem of knowledge which stems from it. In default of any other solution, this makes the position very attractive: we would like there to be some way in which we can legitimately think of ourselves as having perceptual and epistemic access to the physical world. But it does not, of course, show the position to be true. Moreover, before we can begin to regard idealism as a serious option, there is a fundamental objection to it which needs to be overcome. For, on the face of it, it is essential to our very conception of a physical world that it be something whose existence is logically independent of the human mind. Thus when we consider the nature of the world in its own terms—a world of three-dimensional space and material objects, a world of fields and mountains, planets and stars—it seems to make no sense to suppose that all this derives its existence, wholly or in part, from facts about human mentality. It just seems obvious that we cannot do justice to the nature of the physical entities involved without according them an ontological autonomy in relation to our own mental lives and the ways in which these are organized. In short, no matter how awkward the consequences for our perceptual and epistemological situation, idealism seems to be a non-starter—automatically excluded by our basic understanding of what a physical world is.
This objection seems, initially, very powerful. But, on further reflection, I think we can see that the idealist has a satisfactory answer. The crucial point is that, from an idealist standpoint, the relationship between the physical world and human mentality is to be characterized in quite different ways, according to whether it is what is involved in the idealistic creation itself or what features within the realm of what is thus created. Obviously, as what is involved in the idealistic creation, the relationship is one of logical dependence. The physical world is, in all its aspects, logically created by the sensory organization, or by some complex of facts in which this organization centrally features, and the sensory organization is a
end p.259
 
system of constraints on human sensory experience. So, in terms of the idealistic creation, the existence of the physical world, and all the facts it contains, logically depend on facts about the human mind. On the other hand, the world which is thus created is not just created by facts which involve human mentality, but is created as something with which human mentality is intimately linked. It is created as a world which forms our perceptual environment—a world in which we are located and whose ingredients are perceptually accessible to us; and it is also created as a world in which we are functionally attached to human bodies, and in which, through such attachment, our mentality causally interacts with the states of these bodies and with events in the wider environment. In respect of this created situation, the relationship of the physical world to human mentality is not, even for an idealist, one of logical dependence. Rather—just as under realism—the two realms of phenomena are interactive parts of a larger psychophysical whole, and, as such, enjoy the same ontological status. In this context, the only way in which the physical is sometimes dependent on the mental—at least, the only way which has a bearing on what the realist would take to be the primary core of physical reality—is that human mental events sometimes have a causal influence on what happens physically. Indeed, in this context, so far from logically depending for its existence on the sensory organization, the physical world is itself responsible for imposing an organization on sensory experience. For a subject's sensory experiences fall under the causal control of the relevant centres in his brain (the brain of the body to which he is functionally attached), and these centres, in turn, are causally sensitive to the relevant forms of environmental input.
Now this distinction, between how human mentality relates to the physical world in the context of the idealistic creation and how it relates to the world in the context of what is created, gives the idealist a straightforward way of dealing with the objection. When we consider the physical world in its own terms, we consider it in a perspective in which our relationship to it is, in the way just specified, perceptual and interactive; and, in the context of this relationship, as I have just stressed, the existence of the physical world does not logically depend on facts about human mentality. So what the idealist can say is that it is just this fact which makes it seem that our basic understanding of the physical world and his own thesis are in conflict. And if this is what creates the appearance of conflict, then, of course, the appearance is illusory. For the fact that the physical world does not exhibit a logical dependence on human mentality in that context is entirely compatible with the claim that it is logically created in the way that the
end p.260
 
idealist envisages. There is no paradox in the suggestion that human mentality ultimately contributes to the logical creation of the physical world, but that, within the context of what is thus created, also features as something which does not thus contribute. We just need to be careful not to confuse the question of how things stand within the psychophysical reality which is idealistically created with the question of how things stand with respect to that creation.
This seems to me to be an effective way of meeting the objection. It does not, of course, show that there is no conflict of any kind between our basic understanding of the physical world and the idealist thesis; and it remains possible that further investigation will bring such a conflict to light. But what the objection was appealing to was an apparent conflict which is conspicuous when we first consider the issue—a conflict which just seems obvious when we compare our ordinary conception of the physical world with what the idealist is claiming. And what is now surely clear is that this initial appearance of conflict is something which the idealist can, from his own standpoint, adequately explain. He can point out that, even if idealism is true, it is bound to seem initially to be a non-starter simply because of the way in which the physical world and human mentality are related within the reality which is idealistically created—because, within this reality, the physical world exists as an interactive partner with human mentality, rather than as something logically dependent on it.
This disposes of the objection. What we still do not have, of course, is any positive argument in favour of idealism. So far, all we have found in its favour is that it eliminates the problem of perception, and the problem of knowledge which stems from it; and, as I have already conceded, this does not serve to establish its truth. Moreover, the onus of proof is still very heavily on the idealist. The basic idea of the idealist approach is that, whether on its own or in the context of some richer complex, the sensory organization creates a physical world by disposing things to appear systematically world-wise at the human empirical viewpoint; and it creates all the details of the world's specific character by disposing things to appear world-wise in the relevantly specific ways. But, on the face of it, as we have already noted, the most that this disposing would achieve would be the creation of a virtual world, rather than one that was real. Clearly it is up to the idealist to show why this initial view of the situation is wrong, and how the form of idealistic creation he envisages turns out to be effective.
Before we can be justified in invoking idealism as the answer to the problem of perception, we need, then, a positive argument in its favour. And it is this that I want now to try to provide. I want to try to show that, quite
end p.261
 
apart from the problems of perception and knowledge, there are considerations which oblige us to accept the idealist account. These considerations are ones which I have elaborated in much greater detail in my book The Case for Idealism.5 In the present context, where the discussion of idealism only forms the last phase of an investigation into the nature of perception, I am forced to deal with the issues more briefly, and with greater reliance on intuition. Even so, I hope to make the argument persuasive.
In developing this argument, I shall put the problem of perception on one side, as something not relevant to the issues which now concern us. I shall continue, however, to assume the correctness of the arguments in Parts Two and Three which created that problem. In particular, I shall continue to assume that we have succeeded in refuting SDR, and in establishing SQT as the correct account of phenomenal experience. Because of our recent discussion, there is also one respect in which I need to clarify my terminology. As I have stressed, the idealist accepts that, within the ordinary psychophysical reality—the reality which he takes to be idealistically created—the existence of the physical world is not logically dependent on the human mind. But, in what follows, whenever I speak of logical dependence on (independence of) the human mind, I shall be referring to how things are to be characterized when we stand back from this reality and consider the role of human mentality at the level of what is fundamental. In other words, I shall be using these expressions, as I have predominantly used them hitherto, in a way which allows me to say, without qualification, that the mind-independence of the physical world is something which the idealist denies, and hence something about which the realist and idealist have opposing views.
2 An Argument for Idealism
I
Let us speak of a reality as a physically relevant external (PRE) reality if and only if it is, in all its aspects, logically independent of the human mind, and, through its systematic causal control of human sensory experience, is what sustains the sensory organization. The realist is committed to accepting the existence of a PRE-reality. For he is committed to saying that such a reality is what forms the physical world in its primary (mind-independent) core—a core which includes the existence of physical space, and the presence
5 London: Routledge and Kegan Paul, 1982.
end p.262
 
within it of the various forms of material object. The idealist too will almost certainly accept the existence of such a reality as a way of accounting for the sensory organization. But, from his standpoint, this reality will not be identical with the physical world, but rather something which underlies it.
Now we can envisage the possibility that there is a PRE-reality, but that its structure and organization differ, in certain ways, from what all our empirical evidence suggests, and from what, in response to this evidence, we attribute to the physical world in our ordinary and scientific thinking. The issue on which I want to focus is that of how we ought to interpret things physically if we suppose that such a difference obtains. More precisely, I want to focus on this issue for a specific range of cases, where two further conditions are satisfied. First, the difference is not so great, overall, as to lend any intuitive plausibility to physical nihilism—to the view that there is no physical world at all. Thus it is nothing like the difference that would obtain if the PRE-reality consisted of a collection of interacting minds, or if it was of the kind envisaged by Berkeley. Secondly, the reason for the difference is that the PRE-reality is nomologically organized as if it had a different structure from the one which it actually has, and it is this organizationally simulated structure which then gets projected on to the human empirical viewpoint. We shall see in more detail presently just what this involves. I shall speak of cases of this kind as cases of relevant deviance—'deviance', because of the way in which how things empirically appear deviates from the character of the external situation, 'relevant', because of the specific conditions to which this deviance conforms.
As I said, the issue on which I want to focus is that of how we should interpret things physically if we suppose a case of such deviance to obtain. It will be best if, in the first instance, we pursue this issue in relation to a particular example.
II
Let us suppose that the PRE-reality, as it is in itself, is composed in the following way. There is an external three-dimensional space, S, and a certain stock of qualities capable of characterizing (of being instantiated by) elements (points, lines, regions) of S at times. The facts which compose the reality then either consist in, or are constituted by: (a) facts to the effect that certain elements of S are characterized by certain qualities at certain times, or over certain periods; (b) facts to the effect that certain causal relationships hold between different spatiotemporal instances of such
end p.263
 
quality-characterization; and (c) laws of nature (general facts of natural necessity), which govern both the spatiotemporal arrangement of qualities and the obtaining of causal relationships. By restricting what is fundamentally distributed over the elements of S to qualities, I am assuming that any persisting objects in S can be thought of as deriving their existence from (as things whose existence is constituted by) facts of quality-characterization, causal relationships, and nomological organization. (Thus, in the simplest case, we might think of the existence of such an object as constituted by the lawlike occurrence of a certain spatiotemporally and causally continuous series of region-moment instances of a certain quality.) But if anyone finds this approach too austere—for example, a realist who wants to include certain kinds of physical particle in his basic ontology—I am happy to expand the ontology of the PRE-reality accordingly. And, indeed, I would be happy to enrich the reality in any other ways that might be demanded. I have kept its basic ingredients to a minimum purely for simplicity of exposition, not because I am trying to set things up in a way that tilts things in favour of what I want to establish. I shall also assume, for ease of exposition, that S is an absolute and ontologically autonomous space, whose points and regions have identities which are both independent of a frame of reference and independent of time. And, so that this should not automatically exclude physical realism, which needs to equate S with physical space, I shall make the same assumption for physical space too. This last assumption, of course, is completely out of line with current scientific theory. But to provide a scientifically accurate account would make the ensuing discussion much more complicated, without affecting the substance of the philosophical issues involved.
In addition to how the PRE-reality is in itself, there is the matter of its relationship to us, and specifically, its role in sustaining the sensory organization. To take account of this, we have to suppose that there is a system of laws which make provision for, and regulate, forms of causal interaction between aspects of the PRE-reality and the realm of human mentality. A plausible supposition would be this: there is a certain class of complex persisting objects in S—the S-equivalents of human biological organisms—and a one-one correlation between these objects and human subjects, such that, for each such correlated pair, there is a set of laws prescribing ways in which the condition of the S-object is empowered to directly affect the mentality of the subject, and ways in which the mentality of the subject is empowered to directly affect the condition of the object. Of particular relevance here will be the laws which ensure that certain kinds of process in the S-objects cause the correlated subjects to have certain kinds of sensory
end p.264
 
experience. For the operation of these laws will form the final link in the chain whereby the makeup of the PRE-reality determines the nature of the sensory organization. The laws which connect the PRE-reality with human minds are not, strictly, ingredients of this reality, since the latter is required to be, through and through, mind-independent. But, for convenience, I shall often speak of them as laws of the PRE-reality, and as elements in the larger package of laws which collectively define its nomological organization. I should also make it clear that, in describing the relevant S-objects as the S-equivalents of human biological organisms, I was not excluding the option of equating them with such organisms. I was simply trying to stay neutral between that (the realist) option and taking the objects to be what underlie these organisms.
Now let us suppose that, with one crucial exception, the laws of the PRE-reality, both those that govern its internal workings and those that link it with human mentality, impose the same constraints on events across the whole of S and time. The exception is as follows. Within S there are two separate three-dimensional regions, R 1and R 2 , of the same shape and size, such that everything is nomologically organized, both internally and with respect to human mentality, exactly as if—by the standards of what would be required for organizational uniformity—R 1and R 2were interchanged. Thus suppose that there is some kind of mobile process in S whose instances would in general, under the laws, be made to follow a course of uniform motion in a straight line, unless affected by some further force. Then, if an instance of this kind of process comes (in the normal S-time continuous way) to some point on the boundary of R 1 , it instantaneously changes its location to the corresponding point on the boundary of R 2 , and continues in the corresponding straight line from there. And conversely, if an instance of the process comes (in the normal way) to the boundary of R 2 , it undergoes an exactly analogous shift to the boundary of R 1 . Quite generally, by the standards of how, in the rest of the space, things behave, interact, and interact with our minds, everything is organized, with respect to the boundaries of the two regions, as if each region had the other's location. We might express the point succinctly by saying that each region is functionally located where the other is actually located—the functional locations of the regions being the ones which we would need to assign to them to achieve organizational uniformity.
Everything is organized as if R 1and R 2were interchanged. And, as I have made clear, this relates not just to the constraints on behaviour in S, but also to the modes of interaction between what occurs in S and human mentality. In particular, then, it covers the ways in which situations in S
end p.265
 
affect human sensory experience. Crucially, this means that the organizational anomaly in the external reality is wholly concealed at the level of empirical appearance. At the human empirical viewpoint, everything seems to indicate that, within our own world, things behave in a completely uniform way across the whole of space. This, in turn, of course, shapes the character of our ordinary physical beliefs, which are directly responsive to what the empirical evidence suggests. And so the physical world of our empirical beliefs—both those at the level of our common-sense thinking and those which emerge through science—comes to mirror not the external reality as it is, but the reality which would obtain if the relevant regions were interchanged. In the world we empirically construct, the region corresponding to R 1gets located in the surroundings corresponding to the S-surroundings of R 2 , and the region corresponding to R 2gets located in the surroundings corresponding to the S-surroundings of R 1 . In other words, the topology which features in the content of our physical beliefs coincides with the functional topology of S rather than with its actual topology. For simplicity, let us assume that this is the only point of conflict between the structure and organization of the PRE-reality and those of the world that we empirically construct—the world which our empirical evidence suggests and which features in the physical beliefs that we form on that basis.
To bring the situation into sharper focus, let us make the example more concrete. Thus let us suppose that R 1and R 2respectively correspond to what, in the empirically constructed world, qualify as the regions of Oxford and Cambridge. Strictly speaking, of course, even with our assumption of the absolute and autonomous character of physical space, Oxford and Cambridge do not define fixed regions within it, since they are in constant motion. And this means that we can only correctly represent them as corresponding to R 1and R 2if we reconstrue the latter as spatiotemporally continuous sequences of regions-at-times. Such a reconstrual would be perfectly feasible, and would not affect the course of the argument. But rather than getting involved in such a complication, I shall simply pretend that the Earth is stationary, and that the two cities occupy the same regions of physical space at all times.
We are supposing that R 1and R 2respectively correspond to what, in the empirically constructed world, qualify as the regions of Oxford and Cambridge. And, of course, all the empirical evidence suggests to us that Oxford is in what qualifies as Oxfordshire, and Cambridge is in what qualifies as Cambridgeshire. But, because of the organizational anomaly in the external reality, the situation with respect to R 1and R 2is reversed, with the S-surroundings of R 1corresponding to the relevant portion of Cambridgeshire
end p.266
 
(i.e. to what remains of Cambridgeshire when the Cambridge-region is subtracted), and the S-surroundings of R 2corresponding to the relevant portion of Oxfordshire (i.e. to what remains of Oxfordshire when the Oxford-region is subtracted). So if someone were to drive from Oxford to Cambridge, the route of his journey in S would have a very different character from the route which was empirically apparent. The external process would begin in R 1 , the region corresponding to Oxford, but surrounded by what corresponds to Cambridgeshire. But as soon as the driver moved out of the Oxford-region, the process would instantaneously change its S-location to an area just outside R 2 , corresponding to an area in Oxfordshire normally thought to be just outside Oxford. The process would then continue in a way which coincided with the empirically apparent route until it reached the boundary of R 1 , when, as the empirical journey would find the driver entering Cambridge, it would once again instantaneously change its location and became a process moving through R 2 . (See diagram on next page.) So what would seem to the subject to be a spatially continuous journey in physical space would correspond to a process in S-space which became dramatically discontinuous whenever it reached (whether from the inside or the outside) the boundaries of R 1or R 2 . And of course it is not just that the discontinuities in the external process would not be apparent to the subject. Rather, the laws of nature would ensure that they remained completely undetected by anyone, at any time, by any empirical means. All the empirical evidence, past and future, actual and potential, from the most casual observation to the most searching experimental test, would seem to indicate that the external process was (in line with how things experientially appeared to the subject) spatially continuous.
I think that the nature of the case we are envisaging is by now clear. What still needs to be decided is what we should make of it. And one thing which must be conceded in advance is that, taken as a suggestion as to how things may actually be, what is envisaged is not at all plausible. It is not just that, ex hypothesi, it runs counter to all the empirical evidence. It is also that there would be something inherently puzzling about a reality which was organized in the way envisaged. If there is nothing which qualitatively distinguishes R 1and R 2from other regions in S—and there has been no suggestion of anything of that sort—it is surely very strange that the laws should treat them in such a distinctive fashion, and even stranger that they should do so in a way which contrives to prevent any empirical trace of what is happening. But none of this affects the role of the hypothetical case in our present discussion. All that presently matters is that what is being envisaged—however improbable—is coherently conceivable. And this is
end p.267
 
Process in S
surely so. However strange such a situation would be, there is nothing which logically excludes the suggestion that the PRE-reality is organized in this anomalous way.
Given that what is being envisaged is conceivable, the crucial question is how we should interpret it physically. How does the peculiar organization with respect to R 1and R 2affect the situation of Oxford and Cambridge in the physical world? It is essential, of course, that we should address this question from an appropriately detached standpoint. As we have already stressed, the physical beliefs that we form in the context of our ordinary empirical lives are automatically shaped by the empirical evidence. But the question which we now need to consider requires us to stand back from our empirically formed beliefs, and try to reach a verdict about their correctness
end p.268
 
when we take into account the external factors which are envisaged as lying behind them.
So, supposing ourselves to know that the PRE-reality is structured and organized in the relevant way, and that it is linked with the world of our empirical beliefs in the ways indicated, what conclusion should we reach about the physical situation? Formally, there are four options available. The first, and most drastic, would be the option of physical nihilism. Here, the discrepancy between how things externally are and how they empirically appear would be seen as excluding the existence of a physical world altogether. The second option, and a sort of modified version of the first, would be to say that, while there is a physical world, it does not contain the regions that we ordinarily think of as Oxford and Cambridge. There are, with respect to these notional regions, 'holes' in the fabric of physical space, and all that empirically suggests their presence, and the presence of objects and processes located in them, is just a form of systematic illusion. The third option would be to say that there is a physical world, and one which is replete with the Oxford and Cambridge regions, and that its structure and organization coincide with those of the external reality. This would involve saying that, contrary to what we ordinarily believe, and to what all the actual and potential empirical evidence suggests, Oxford is really in Cambridgeshire, and Cambridge is really in Oxfordshire. The fourth option would be to say that there is a physical world, and again one replete with the Oxford and Cambridge regions, but that its structure and organization are as they are empirically represented at the human viewpoint. This would involve saying that, despite the positions of R 1and R 2in the external reality, Oxford and Cambridge are, physically, where we ordinarily take them to be—Oxford in Oxfordshire, Cambridge in Cambridgeshire.
Now, of these four options, it seems to me that only the last has any plausibility. We can immediately dismiss the first option. It is just obvious that the local organizational quirk with respect to R 1and R 2 , and the resulting effects on empirical appearance, do not suffice to eliminate the physical world altogether. Equally, it only takes a moment's reflection to reject the second. For since, apart from the location of the relevant regions, how things empirically appear with respect to Oxford and Cambridge is accurate with respect to everything that pertains to R 1and R 2 , there is no temptation at all to think that the supposed Oxford and Cambridge portions of the physical world do not exist, and that all the apparent signs of their presence are illusory. But granted that we retain a belief in the physical world, and in those particular portions of it whose physical location is at issue, we surely want to gauge its topology by reference to how things stand
end p.269
 
empirically, rather than to how they undetectably stand in the external reality. We surely want to say that, when someone travels from Oxford to Cambridge, his journey is, as it empirically appears, genuinely continuous in physical space, and that the concealed discontinuities involved only pertain to the underlying processes in S. We surely want to avoid having to say that, each time the traveller reaches the critical boundary of one of the city-regions, he undergoes an instantaneous but imperceptible transference to another part of the country. We surely want to say that this transference is not with respect to the physical geography of England—the geography that we are trying to characterize in our maps and manuals—but only with respect to the external spatial arrangement which underlies it. In short, we surely want to say that, despite the S-locations of their external correlates, Oxford and Cambridge are physically located where we ordinarily believe and where all the empirical evidence suggests.
The basic point is that, to qualify as the physical world, something has to be our world in an epistemologically crucial sense; and to qualify as characterizing our world, in that sense, the topological arrangement of regions must surely conform to how things are disposed to appear, empirically, at our viewpoint. Of course, there can be aspects of this arrangement which we are not, in practice, equipped to detect: the regions to which they relate may be too remote, or the aspects may be too subtle to be captured by current investigative techniques. But what is surely ruled out is the suggestion that something qualifies as our world, in the relevant sense, but that things are so organized as to ensure that, even in an area where we seem fully equipped to monitor the situation, some aspect of its topology is systematically belied by all the actual and potential empirical evidence. If everything is organized so as to ensure that Oxford passes all the empirical tests for being in Oxfordshire, and Cambridge passes all the empirical tests for being in Cambridgeshire, that surely settles the issue of their true physical locations.
It might be objected that, while the physical world has to be our world in an epistemologically crucial sense, this only requires its topology to have a general conformity to how things are disposed to appear at our empirical viewpoint, and so allows for exceptional cases where some aspect of the physical topology is systematically at variance with the actual and potential empirical evidence. And this would mean that, in the particular case on which we are focusing, there is no obstacle to concluding that the physical locations of Oxford and Cambridge coincide with the S-locations of R 1and R 2 . I do not find this suggestion at all plausible. But, in response to anyone who does, I would simply develop the example one stage further. Thus,
end p.270
 
instead of envisaging just one case of a pair of S-regions which exemplify the relevant organizational phenomenon, I would suppose such cases to be a commonplace feature of S. And, if necessary, I would even suppose that, despite the anomaly involved in each case, there was a general pattern to which the various cases conformed—for example, that the pairs of regions were of a constant shape and size, that there was the same distance between the regions in each pair, and that the cases were distributed over S in a uniform fashion. Developing the example in this way would not, it seems to me, affect our judgement that there was still a physical world, and one without gaps. But, even from the standpoint of the objection, it would be impossible to suppose that, on such a scale and in such a regular way, things were organized so as to ensure that the topology of our world was systematically belied by the empirical evidence available at our viewpoint. It would be impossible to avoid thinking that the physical topology coincided with the organizationally simulated (functional) topology of S—in conformity with what all our empirical evidence (actual and potential) suggests.
Bearing in mind that I could always respond to the objection in this way, I shall, for convenience, keep to the example in its original (simple) form, and assume that it is to be physically interpreted in the way I suggested. As I have already said, this interpretation seems to me, in any case, to be the only one that is plausible.
III
We have been focusing on one example of a case of relevant deviance. In such cases, the structure and organization of the PRE-reality differ from what the empirical evidence suggests, and do so in such a way that (1) the difference is not so great, overall, as to lend intuitive plausibility to the conclusion that there is no physical world at all, and (2) the reason for the difference is that the reality is nomologically organized as if it possessed a different structure, and it is this organizationally simulated structure which gets projected on to the human empirical viewpoint. There are many other examples of relevant deviance that we could devise (indeed, the potential list is infinite), and any of them could be used to make the same general point. Some of these examples would again be concerned with the geometry of space. For instance, we could suppose that the external reality is organized, in respect of both its internal behaviour and its causal relations with the human mind, exactly as if a certain spherical region of space were rotated by a certain amount in a certain direction. Or again, we could
end p.271
 
suppose that the external reality is organized, in both respects, as if its space were more narrowly circumscribed, so that its functional boundaries enclose only a portion of what is enclosed by its actual boundaries. Or to take up a case elaborated earlier, we could suppose that the external item corresponding to physical space consists of two separate spaces organized as if they were joined.6 Other examples might concern different forms of structure, such as relations of qualitative sameness and difference. Thus we could suppose that, for a certain region of the external space, and two qualities, everything is organized exactly as if each quality as it occurs within this region was the same as the other as it occurs outside. Or more elaborately, we could suppose that there are two qualities which exchange their current distributions and functional roles every hour, so that everything is organized as if each of the qualities in hours 1, 3, 5, 7, . . . is the same as the other in hours 2, 4, 6, 8, . . . . In all these further examples, our intuitions would, it seems to me, follow the same course as in the case discussed: we would say that what is envisaged preserves the existence of the physical world—and a world without gaps—but that the relevant facts of physical structure conform to the organizationally simulated structure of the external reality, rather than to its actual structure. Thus if we supposed the external reality to comprise two spaces organized as if they were joined, our intuition would be that there was only one physical space, reflecting the distinctive (single-space simulating) character of the organization. Likewise, if we supposed that two external qualities exchanged their current distributions and functional roles every hour, we would trace the spatiotemporal paths of the relevant physical qualities in a way which restored organizational uniformity, rather than in a way which matched the external relations of qualitative sameness and difference. Moreover, I think it is clear that this would be our response in all cases of the relevant kind. And running through these responses would be the same underlying thought: that, as characterizing our world, in the epistemologically relevant sense, the physical structure must conform to the organizationally simulated structure of the external reality because it must conform to how things are disposed to appear at our empirical viewpoint.
All this, I fear, is rather quick. I would like to be able to spend more time elaborating some of these further examples and showing how the common pattern of interpretation applies. But, given the constraints of our discussion, we need, at this stage, to move on to our next investigative task. This
6 I introduced this case in Part Four, Section 4, III.
end p.272
 
is to consider how our findings with respect to relevant deviance bear on the issue of physical realism.
IV
If our intuitions have been sound, we know that, in any case of relevant deviance, there is some respect in which the structure of the physical world differs from the structure of the PRE-reality, and, consequently, there is some aspect of physical structure which cannot be equated with an aspect of analogous PRE-structure. Thus, in the example of the two regions, the location of the Oxford-region in Oxfordshire and the location of the Cambridge-region in Cambridgeshire cannot be equated with locational facts about S, since the corresponding S-regions occupy the reverse positions. This already seems to pose a problem for physical realism. Realism is committed to accepting the existence of a PRE-reality, and to equating it with the physical world in its primary core; and this involves taking the structural aspects of this reality to form the corresponding aspects of physical structure. So, for cases of relevant deviance, the correct account of how things stand physically and the realist account are in direct conflict.
This conflict looks awkward for the realist—and may in fact be so—but it does not suffice to show that his position is mistaken. The realist thesis, as we have formulated it, is only strictly concerned with the actual situation: it claims that the physical world is something whose existence is logically independent of the human mind, and is, in its basic character, metaphysically fundamental. But the cases of relevant deviance are, in the context of our discussion, only hypothetical: they represent ways in which we can coherently envisage how the situation might be. So the realist can still insist that his account of the physical world is right in fact. Nor, as things currently stand, would such a response be merely perverse. For, as we noted in the case of the regions, there seem to be good grounds for supposing that the kinds of deviance envisaged do not obtain. Thus, as well as running counter to what all the empirical evidence suggests, there would be something inherently puzzling about a situation in which the external organization was anomalous in this kind of way. So perhaps the realist can afford to concede that cases of relevant deviance are not amenable to a realist construal, but insist that the realist account is correct for how things actually are, or at least for how we are entitled to take them to be.
Whether the realist can really afford to limit the relevance of his claims in this way is, I think, debatable. But, for the sake of argument, let us suppose that he can. Even so, it seems to me that the underlying difficulty for
end p.273
 
his position remains. For, when we look into the issue more closely, I think we can see that the factors which make cases of relevant deviance resistant to a realist construal carry over, if in a less direct form, to cases to which the realist's approach seems better suited. In particular, they carry over to the case which represents the realist's best scenario, in which there is no conflict at all between how things stand in the PRE-reality and what the empirical evidence suggests. I shall speak of this as the case of zero deviance. What makes this case so propitious from the realist's standpoint is that it contains nothing which could even be thought to indicate that there is a qualitative difference between the external and physical structures, and so nothing which presents an immediate obstacle to identifying the two.
We need to begin by noting that, even when we suppose the actual situation to be one of zero deviance, we can envisage ways in which, without altering the external structure, cases of relevant deviance could have occurred. Let us confine out attention to deviance with respect to geometrical structure. Then the point is that, even when we assume that the geometry of the external space is wholly in line with the actual and potential empirical evidence, we can still envisage ways in which, with this geometry held constant, the laws governing the external reality could have been different, and different in a way which created an instance of the relevant kind of mismatch with how things empirically presented themselves at the human viewpoint. Thus, even when we suppose that, in actual fact, the external organization is entirely uniform with respect to the geometry of the external space, we can see how, without change to this geometry, things could have been organized, both in respect of what takes place within the space and in respect of causal relations with human mentality, exactly as if two regions were interchanged. Or again, with the same supposition about how things actually are, we can see how, keeping the geometry of the external space intact, things could have been organized, in both respects, exactly as if a certain spherical region were rotated by a certain amount in a certain direction. Or yet again, we can see how things could have been organized as if the external space were more narrowly circumscribed—as if its boundaries enclosed only a portion of its true extent. There is no limit to the number of examples that could be offered. We might also note that, as well as being able to envisage such possibilities in the form of what could have been, we can envisage them in the form of what could come to be. For, in each case, we can envisage the possibility of the laws which have held hitherto changing in the relevant way. But, although these possibilities of future deviance would serve our purposes just as well, I shall continue to
end p.274
 
focus on the cases of how things would have turned out, if the laws had always been different.
Even when we suppose, then, that the external geometry is entirely in line with the empirical evidence, we can envisage ways in which, with the same geometry, but suitably different laws, there could have been cases of relevant geometrical deviance. And just as we earlier raised the question of how the hypothetical cases of relevant deviance should be interpreted physically—cases where we envisaged the deviance as how, conceivably, things might actually be—so we can raise this same question for these merely possible cases. If things in the PRE-reality had been organized in a relevantly different way, creating the relevant kind of disparity between the external geometry and what was suggested by all the actual and potential empirical evidence, how would things have turned out for the geometry of physical space?
In the hypothetical cases, we reached our verdict about the relevant aspects of physical structure by appealing to a principle of empirical immanence. Thus, in the example on which we focused, after dismissing the suggestion that the envisaged situation was such as to preclude the existence of the physical world, or affect its repleteness, we were left having to decide between two views, one of which took the physical topology to coincide with the actual topology of the external reality, and the other of which took it to coincide with the organizationally simulated topology. And what led us to endorse the second of these views was the consideration that, to qualify as characterizing our world, the physical topology has to conform to how things are disposed to appear at our empirical viewpoint. And this is how we settled the physical interpretation of such cases of deviance quite generally. Now we cannot just assume that, because this consideration was decisive for the hypothetical cases of deviance, it will be decisive for the merely possible too. After all, there are many instances where something which is a conceptual requirement of a certain type of item in the actual world does not hold of that type in all possible worlds. For example, to qualify as water in the actual world, something has to be the sort of stuff that we standardly find in rivers and rain; but it is easy to imagine a possible world in which the stuff of rivers and rain is something quite different. Or again, to qualify as Moses in the actual world, someone has to satisfy a reasonable amount of what is recorded of 'Moses' in the Bible; but this does not mean that the person who does qualify as Moses could not have perished in the womb. Accordingly, it could still be suggested that considerations of empirical immanence only play a role in picking out what qualifies
end p.275
 
as physical structure in the actual world, and that it is then the nature of what is thus picked out, rather than any need to preserve conformity to the empirical evidence, which enables us to identify instances of physical structure in merely possible worlds—including those worlds which involve the envisaged forms of deviance.
However, I think it is clear that our intuitions about the hypothetical cases of relevant deviance do, in fact, carry over to the possible cases as well. Thus, continuing to focus on deviance with respect to geometrical structure, we surely want to say that if the external reality had been organized as if two of its regions were interchanged, then the locations of the corresponding physical regions would have coincided with the functional locations of these regions, in accordance with all the actual and potential empirical evidence. Likewise, we surely want to say that if the external reality had been organized as if its space were more narrowly circumscribed, then, in line with all the empirical evidence, the extent of physical space would have been correspondingly diminished. And this, it seems to me, would be our reaction in any case of this general type. In each such case, we would take the true physical geometry to be what shows up empirically at the human viewpoint, rather than what coincides with the concealed geometry of the external reality. In effect, our response would be to evaluate the physical significance of these merely possible cases as if they were actual. One way in which we can reinforce this point is by considering the same issue in reverse—taking the actual situation to be one of relevant deviance, and then raising the question of how things would have stood physically if this deviance had been eliminated. Thus suppose we take the actual situation to be one in which the external reality is organized as if two of its regions were interchanged, or as if the extent of its space were more narrowly circumscribed—obliging us to think of the physical geometry as coinciding with the organizationally simulated geometry—and then consider how things would have stood physically if, with the geometry of the external space held constant, the organization had been wholly in line with it. In this sort of case, there is nothing that could even tempt us to deny that the physical geometry would have been correspondingly different—coinciding with the external geometry, in accordance with the new (and deviance-eliminating) organization. And consistency demands an analogous response when the move from the actual situation to the possible one is, in terms of deviance, the other way round.
The merely possible cases of relevant deviance are to be interpreted, then, in the same way as the hypothetical, so that the physical geometry is taken to coincide with the geometry which the external organization simulates
end p.276
 
and projects on to the human empirical viewpoint. But this has crucial consequences for our evaluation of physical realism. To begin with, it means that, even in the case of zero deviance, where we can take the physical and external geometries to be qualitatively the same, the obtaining of the physical geometry logically depends on more than just the obtaining of the external geometry. It depends, in addition, on aspects of the external organization; for if this organization had been suitably different—though without change to the external geometry—the physical geometry would have been different too. In effect, the obtaining of the physical geometry depends not just on the external geometry, but on the fact that the latter is 'endorsed' by the external organization, and its character thereby translated into how things empirically appear. It follows that, even in the case of zero deviance, the facts of physical geometry cannot be equated with the facts of the external geometry. For if they could, their obtaining would not involve any additional factors. Moreover, the additional factors involved are not just to do with how the PRE-reality is organized internally. They also concern the ways in which it is organized in relation to human minds. In particular, they concern the ways in which the reality is nomologically disposed to affect human sensory experience; for it is these dispositions to affect sensory experience that determine the character of the geometry which presents itself at the human viewpoint. So, as well as being distinct from their external counterparts, the facts of physical geometry are logically dependent on facts about the human mind. And this means that they cannot even be assigned to the PRE-reality in a derivative form—as facts constituted by more basic PRE-facts—since this reality (as it is in itself) is wholly mind-independent. The same form of reasoning could be used to reach analogous conclusions about other kinds of structure, such as relationships of qualitative sameness and difference. But, for the purposes of our discussion, the results for the geometrical case will suffice.
The relevance of these results to the issue of realism is already clear. The realist, as we have said, is committed to accepting the existence of a PRE-reality, and to equating it with the physical world in its primary (mind-independent) core; and this involves taking the structural aspects of this reality to form the corresponding aspects of physical structure. As we have also noted, the scenario which is most favourable to the realist's cause is that of zero deviance, where we can at least take the physical and external structures to coincide qualitatively. It now turns out that, even when it is applied to this case, the realist account fails, and fails comprehensively. The physical and external geometries may coincide qualitatively. But they differ in their concrete existence, since the obtaining of the former depends
end p.277
 
on more than just the obtaining of the latter. Nor, indeed, can the facts of physical geometry be assigned to the PRE-reality in any way at all, since the additional factors on which their obtaining depends are partly concerned with human mentality. Finally, we can see that, with the exclusion of the physical geometry from the PRE-reality, all physical facts are excluded. For the exclusion of the physical geometry entails the exclusion of physical space; and all physical facts essentially involve the existence of physical space, being either facts about its character (or the character of certain of its elements), or facts about what exists or occurs within it. In short, even for the case of zero deviance, we are forced to conclude that the PRE-reality is wholly non-physical.
Because the case of zero deviance represents the realist's best scenario, the failure of realism for this case is enough to show its failure quite generally. But, for completeness, I should stress that the argument which we have employed for zero deviance could also be employed for any other type of case where an anti-realist argument is needed—any type of case where we envisage a PRE-reality, and where it is not obvious from the start that it is incompatible with the realist view.
V
From this refutation of physical realism, it is now only a short step to the endorsement of the idealist alternative. Indeed, this endorsement is already implicit in the points that have emerged.
The realist thesis comprises two claims: the mind-independence claim, to the effect that the existence of the physical world is logically independent of the human mind, and the fundamentalist claim, to the effect that the physical world is, in its basic character, metaphysically fundamental. The argument that we have brought against realism is, in itself, an argument against the first of these claims: it shows that there is not even a primary core of physical facts which can be assigned to an external, mind-independent reality. But, as we noted earlier, there is no prospect of being able to reject the mind-independence claim without rejecting the fundamentalist claim too.7 For the only way of accounting for the mind-dependence of the physical world is by supposing that the world is logically created by facts of a different kind, and that its mind-dependence is a consequence of its mode of creation. In other words, to account for the mind-dependence, we have to suppose that physical facts are, in all cases,
7 See Part Five, Section 1, II.
end p.278
 
constituted by non-physical facts, and that facts about human mentality form, or systematically feature in, the non-physical facts involved.
We already know, then, that the physical world is logically created by facts of a different kind, and ones which at least partly concern human mentality. But it is also clear that, whatever precisely these facts are, they have in some way to cover the obtaining of the sensory organization. This is clear from the considerations which showed the physical world to be mind-dependent. Thus we saw that, even in the case of zero deviance, the facts of physical geometry logically depend, for their obtaining, on aspects of the external organization, and that these aspects are at least partly to do with the dispositions of the external reality to affect human sensory experience. In effect, we saw that what determines the character of the physical geometry is not the external geometry on its own, but the fact that this geometry is 'endorsed' by the external organization, and its character thereby projected on to the human empirical viewpoint. And, in this respect, zero deviance is just the limiting case of relevant deviance. For, with relevant deviance, the character of the physical geometry is determined by the fact that the external organization simulates a certain kind of geometry, and thereby projects it on to the human viewpoint. Moreover, although our focus was on the geometrical case, these results, as I made clear, carry over to other forms of structure as well. Now, in the present context, we cannot make any assumption about the character of the PRE-reality. But it is clear that the only way of doing justice to the results which emerged in that earlier discussion—the only way of accounting for the way in which mind-dependence was seen to characterize physical structure in the situations envisaged—is if we suppose that a crucial part of what equips the facts that create the physical world to play that creative role is that they fix the way things are disposed to appear, empirically, at the human viewpoint. But what directly fixes the way things are thus disposed to appear is, of course, the sensory organization. So, whatever else they involve, the facts which create the physical world must at least cover the obtaining of this organization. They may cover it directly, by simply including facts which are explicitly about the constraints on sensory experience; or they may cover it indirectly, by including a range of facts—for example about the character of the PRE-reality and its nomological links with the human mind—in which the obtaining of these constraints is implicit. But, either way, the covering of this organization will play a central role in both securing the existence of the physical world and determining its specific character.
The upshot is that, in rejecting realism in the way that we have, we are committed to accepting the truth of idealism. We are committed to saying
end p.279
 
that the physical world is, in all its aspects, logically created (that its existence and all the facts it contains are constituted) by the sensory organization, or by some richer complex of non-physical facts in which this organization centrally features.
3 The Unfinished Story
I have tried to establish the truth of idealism. If I have been successful, then, in the context of our larger discussion, this is a welcome result. For it eliminates the problem of perception. It remains the case that the immediate objects of awareness involved in perception are sense-qualia; and there is no way of understanding how perceptual awareness can reach beyond these qualia to items in the mind-independent reality. But, given our idealist account of the physical world, awareness does not need to reach further than the qualia to make contact with the factors by which the existence of physical items is constituted, and thereby to make contact with these items themselves. In fact, in the context of our larger discussion, the establishing of idealism is a doubly welcome result. For, in eliminating the problem of perception, it also eliminates the associated problem of knowledge. It does this, not by opening up a new route to physical knowledge, but by showing that, with perceptual access to the physical world in place, information about the world can be acquired in the normal, perception-based way. Physical knowledge is now available because the physical world is open to perceptual scrutiny.
I have tried to establish the truth of idealism, but I have not provided, nor come near to providing, a fully elaborated idealist theory. To provide such a theory, there are a number of areas where I would need to develop the idealist account in a more detailed way. One area where I would need to pursue matters further concerns the nature of time. Time is an ingredient of both the physical world and the underlying reality by which the world is created. But does this mean that there is a single time-dimension, occurring at both levels? And if not, how are the separate dimensions related? And, in any case, how are we to understand the nature of time in its pre-physical form? I dealt with these questions at length in The Case for Idealism, but am not sure to what extent I would continue to endorse the answers I there gave. Another area where there would be a need for further elaboration concerns the nature of human subjects. Such subjects have to be, in themselves, non-physical, since they feature in the idealistic creation. But there is still the question of the kinds of non-physical things they are.
end p.280
 
And there is even the question of whether, in the last analysis, they should be thought of as things (substances) at all, or merely as organized bundles of mental items. On all this, I would certainly endorse, point by point, the Cartesian account of the subject that I defended in The Immaterial Self. I would also need to say something about how human subjects form a unified group for the purposes of the idealistic creation. The idealism for which I have argued is not solipsistic—a separate created world for each subject. In line with common sense, I have assumed that there is a single physical world which is communally ours, and a single (intersubjective) sensory organization, which contributes to its creation. And this involves thinking of human subjects as forming some kind of communal group independently of what results from the creation of their shared world. At the very least, it involves supposing that there is, pre-physically, provision for certain forms of causal interaction between different subjects, either directly, or via their functional links with a common external reality.
But perhaps the most important area for further investigation is that which arises from the form of the idealist thesis itself. This thesis claims that the physical world is logically created either by the sensory organization or by some richer complex of non-physical facts in which this organization centrally features. As we noted earlier, there is room for a range of more specific idealist accounts according to how the choices implicit in this formulation are resolved. Should we think of the physical world as created by the sensory organization on its own, or by a richer complex of facts? And if the latter, what should we take this complex to be?
The most crucial issues here are concerned with the role of the PRE-reality. To account for the presence of the sensory organization, we can reasonably assume that there is some form of PRE-reality, whose causal control of our sensory experiences is what forms the relevant sensory constraints; and, given that this reality is what imposes these constraints, it is directly involved in the creation of the physical world. But, as we noted, there is also the option of saying that the intrinsic character of this reality is directly involved as well—that part of what allows there to be a physical world is the fact that the reality which constrains our experience is of a certain specific type. It is not hard to see why we might find such an idea attractive. It is true that, in reaching our idealist conclusion, the crucial consideration was that the physical world has to be something whose structure conforms to how things are disposed to appear at our empirical viewpoint; and if this were the only requirement, we could happily think of the world as created by the sensory organization alone. But, in order to be able to think of the created world as a real world, we also need to be able to credit
end p.281
 
it with a certain kind of objectivity in relation to us. And the trouble with taking the sensory organization to be the only constitutively relevant factor is that this objectivity would then seem to be lacking: there just does not seem to be a sufficiently significant difference between the presence of a world which is created by the sensory organization alone and there merely being a systematic impression of such a world from the human viewpoint. But if the idealistically created world is to have the requisite objectivity—to amount to significantly more than the way things systematically seem to us—then this presumably has to come from some kind of externalistic underpinning, some way in which the empirical story is underwritten by the external factors which lie behind it. In The Case for Idealism, I argued that no such underpinning was necessary—that the sensory organization on its own was sufficient. But I am now inclined to think that this argument was flawed and its conclusion mistaken.
If we do think that the nature of the PRE-reality is needed as a direct contributor to the creation of the physical world, this immediately raises the question of what types of reality would be equipped to play this role. Two things are already clear, and in effect implicit in our earlier discussion. First, the kind of PRE-reality which occurs in the case of zero deviance is suitably equipped. Here, the structure and organization of the reality will coincide with how things empirically appear at the human viewpoint. If the empirical story needs to be appropriately underwritten by external factors in order to yield a matching physical world, it is hard to see how this need could be met in a more clear-cut and comprehensive way. Secondly, the kind of PRE-reality which occurs in cases of relevant deviance is also suitably equipped. In these cases, the structure and organization of the PRE-reality do not coincide with how things empirically appear. But the failure of coincidence is limited in its extent: thus, as we stipulated, it is not so great as to lend intuitive plausibility to physical nihilism. And, even with respect to the area of deviance, the way things appear does, in a less direct way, reflect the nature of the external reality; for the reality is organized so as to simulate the possession of a certain structure, and it is this simulated structure which gets projected, in all its details, on to the human empirical viewpoint. This too seems to meet any need for an externalistic underpinning in a clear-cut and comprehensive way. The delicate questions arise when we consider ways in which the conditions required for relevant deviance could be relaxed. Is it really necessary, for example, to impose some limit on the degree to which the structure and organization of the external reality can deviate from what is suggested by the empirical evidence? Or will any degree of deviance allow for the creation of the physical
end p.282
 
world—even a deviance which, prior to our acceptance of idealism, would have seemed incompatible with the existence of such a world—so long as the empirical situation is sustained by the external organization in the relevant way? And, in any case, does there have to be this kind of sustainment at all? How, for example, should we respond to the scenario postulated by Berkeley, where human sensory experience is directly controlled by God, and where the sensory organization is formed by the consistent policies of this control? The empirical story would still be, in its way, externalistically underwritten: it would be the story authored, and, in a sense, authenticated, by God. But is this enough for the creation of a real world?
These are intriguing questions, and, along with the issues mentioned earlier, they are ones which any comprehensive study of idealism would need to address. But neither they nor those other issues are matters that I have the space to pursue here. Nor is there any need for me to do so. The relevance of idealism in the present context is that it enables us to bring our study of the nature of perception to a satisfactory conclusion, and this is something which it does irrespective of how these further matters are settled. Just by knowing the truth of the idealist thesis, we know, in its essentials, how the perception of physical items is to be ultimately understood; and, crucially, we thereby know how, despite our account of phenomenal experience, such perception is possible.
end p.283
 
end p.284