Beginning with PPR (1914) and ending (for the present) with "Berkeley's Denial of Material Substance" (Philosophical Review, 1954), I have treated the subject of sense-perception on nine main occasions. The various essays were written in different contexts, and each usually without reference to its predecessors. It will be obvious to anyone who may take the trouble to collate them that the definitions of such terms as "sensibile," "sense-datum," "sensum," etc., which are stated or implied in some of them, are not consistent with those which are stated or implied in others. This seems to me to be a matter of very little interest except to the minute historians of minor philosophers, and I shall not waste time in discussing it. I now think that the least unsatisfactory treatment is to be found in the two latest, viz., "Elementary Reflexions on Sense-perception" (Philosophy, 1952) and the already mentioned "Berkeley's Denial of Material Substance." I would, however, warn possible readers that the word "sensum" is used in the former, and the words "sensibile" and "sense-datum" are not, whilst the opposite is true of the latter article.

      The contributions of Professors Price, Marc-Wogau, and Yolton are wholly concerned with certain aspects of this question, and I shall now consider them in turn.

(A) PROFESSOR PRICE'S PAPER.

      Professor Price is concerned mainly with the following two questions: --

1. Does the "act-object" analysis apply to all sensations, or do some at least of them require instead the "internal accusative" analysis?
2. Can the so-called "Sensum Theory" deal satisfactorily with certain experiences which a person would commonly express by saying that an object O "looks" or "sounds" so-and-so to him?

(1) Range of Applicability of the "Act-Object" Analysis.

      I begin by accepting practically all that Professor Price says in Sections I and II of his paper. All that I need say is that in writings later than The Mind and its Place in Nature I have recognised and tried to deal with some of the points which I had hitherto overlooked.

      In "Berkeley's Denial of Material Substance," e.g., I admit that the "internal accusative" analysis is quite plausible in regard to vague, peripheral, and unusual visual sensations, whilst the "act-object" analysis seems obviously appropriate to the visual sensations which are an essential factor in ostensibly seeing a body of definite outline in the middle of the field of view. Again, although I nowhere deal with sensations of smell, I have considered rather fully (in "Normal Cognition, Clairvoyance, and Telepathy" and in later writings) the phenomenological likenesses and unlikenesses between

1. intra-somatic and extrasomatic ostensible perceptions; and
2. among the latter, between ostensible seeing, hearing, and touching.

      All this, however, serves only to underline my substantial agreement with Professor Price's conclusion in Section II. The distinction between sensations to which the "act-object" analysis seems appropriate, and those to which the "internal accusative" analysis seems appropriate, has to be made within sensations of the same sense, and not just between the sensations of certain senses and those of others. And there are always marginal cases, where it seems arbitrary to say that the one type of analysis is more or less applicable than the other.

      Professor Price's doctrine, as I understand it, comes to this. When a sensation of any kind is occurring, there is always a sensation of a total sense-field of the corresponding kind. Let us call such an experience an "integral sensation," of the visual kind, of the auditory kind, and so on. The "act-object" analysis applies to integral sensations of every kind. To have an integral sensation of any kind always consists in sensing a total sense-field of that kind, or (to put it otherwise) in having such a total sense-field sensibly presented to one.

      But the total sense-field presented in an integral sensation may take three alternative forms.

1. It may not be appreciably differentiated at all. An example would be the total visual field of a person who is gazing up into a cloudless sky.
2. It may be differentiated into several sharply localised and bounded items, standing out from a relatively undifferentiated background. An example would be if the blue sky, in the former example, had a number of small fairly definite flecks of white cloud scattered about it.
3. It may be differentiated, but not in that particular way. An example might be the visual field of a person looking into an iridescent mist.

      In cases (a) and (c) we should not be inclined to say that the integral sensation is differentiated into a number of constituent sensations, each with its own object. But in case (b) we are inclined to say this; and then we apply the "act-object" analysis, not only to the integral sensation, but also to each of these constituent sensations. In cases (a) and (c) we may say that there is no sensum at all, but only a sense-field, undifferentiated in the one case and differentiated in the other. Or, if we like, we can say that in each of these cases there is a single sensum, undifferentiated in the one case and differentiated in the other.

      It must be admitted, however, that there are sensory experiences which it is natural to express in language which seems to imply the "internal-accusative" analysis, and which it would be extremely strained and unnatural to express in terms which imply the "act-object" analysis. The most obvious examples are certain organic sensory experiences, such as one naturally expresses by saying: "I feel tired," "I feel sick," and so on.

      Professor Price deals with these as follows. In such cases a person is having an integral organic sensation of an organic sense-field, which is very little differentiated, and is certainly not differentiated into outstanding localised organic sensa, such as an ache here and a tickle there. He perceptually accepts this organic sense-field as a state of his own body. Now on many occasions a person regards his own body as part of himself. On other occasions he regards his body, not as a part of himself, but as one thing in the material world, to which he stands in certain uniquely intimate relations. A person tends to use phrases like "I feel tired" when

1. the integral organic sensation, by which he is perceiving his own body, is not differentiated into a number of constituent sensations with localised organic sensa as their objects; and
2. he is taking his body, so perceived, as a part of himself.

      Professor Price suggests that there are occasions (rare for most of us, but not uncommon in the lives of certain poets and nature-mystics) when something analogous to condition (b) is fulfilled even in the case of visual sensation. The percipient takes, not only his body, as intrasomatically perceived, but also the total object of his visual perception, to be included in himself. He is most likely to do this when his visual field is not differentiated into a number of outstanding strongly localised and bounded sensa. In such cases he tends to describe his visual experience in terms suggestive of the "internal accusative" analysis. (The reader will find a good example in Byron's Childe Harold, Canto III Stanza 72.)

      I will now make a few comments on Professor Price's doctrine: --

      (i) I wonder why Professor Price is so sure that the "act-object" analysis applies to all integral sensations. Is he really sure of this, or would he be content to say that it is not obviously inapplicable to any such sensation? For my part, I would not be prepared to go further than this. The cases where the "act-object" analysis seems most obviously applicable are those partial sensations which occur as essential factors in ostensible perceptions of bodies with sharp outlines, or of physical events (such as a flash of forked lightning) which are presented as definitely localised and shaped. It is difficult to believe that there is not something significant in the high positive correlation between these two features in a sensation. Yet the latter feature is conspicuously absent in most cases of an integral sensation of a total sense-field.

      (ii) We must, no doubt, distinguish between

1. the question of what kind of analysis is applicable to a sensation, and
2. the question whether the sensibile which it presents (when the "act-object" analysis seems plainly applicable) could conceivably exist except as a sense-datum to one particular person on one particular occasion.

      Now it is illuminating, in this connexion, to compare and contrast in this respect

1. the visual sense-datum involved in ostensibly seeing a near-by familiar body of sharp outline in the middle of the field of view, and
2. a vivid and detailed "visual image" (recognised by the experient to be such) of such a body.

      (iii) In view of this, might there not be something to be said for the following opinion, which I take to have been held, e.g., by Stout and by Prichard? The "internal-accusative" analysis applies alike to all sensations, integral or partial. The so-called "act-object analysis" is not really an analysis. It is a statement, not about the internal constitution of any sensation, but about the part which certain sensations play in certain ostensible perceptions. No doubt there are intrinsic dissimilarities between those sensations which are, and those which are not, fitted to play this part. But these intrinsic dissimilarities concern the "content" or "quale" of the two kinds of sensation, and not the kinds of analysis applicable to the one and to the other.

(2) Appearing so-and-so" and the "Sensum Theory."

      The question which Professor Price discusses under this head may be stated as follows. Does every proposition of the form: That body appears φ to S entail a proposition of the form: S is sensing a certain sensibile as characterised by φ? Here "appearing φ" is used as a general name to cover "looking φ," "feeling φ," "sounding φ," etc., though most of the cases discussed by Professor Price are in fact instances of looking so-and-so. It is to be clearly understood that "appearing φ" is used here in such a way that to appear φ neither entails nor excludes being in fact φ.

      I think that Professor Price tacitly makes the following two assumptions.

1. That any quality which a sensibile is sensed as having must in fact belong to it, and the sensibile must have that quality in the very same determinate form as that in which it is sensed as having it.
2. That in no case is the sensibile which a person senses identical with any part of the surface of the body which he ostensibly perceives through sensing it. Nor a fortiori is it identical with any part of any other body.

      But I do not think that either of these assumptions is essential to Professor Price's argument here. What he is concerned to do is to call our attention to a series of cases of the following kind. In all of them one would say: "That body appears φ to S." But, as one goes along the series, one would be increasingly disinclined to admit that S is (or even in some cases that he could be) sensing a sensibile as φ.

      I agree with Professor Price that no serious difficulty for the general principle under discussion arises until we come to cases where the mode of appearing is appropriate to one sense (e.g., sight), whilst the characteristic which the perceived body is said to appear to have is appropriate to another sense (e.g., touch). An obvious example is when a person, looking at a block of ice or at a picture of a field covered with snow, says: "It looks cold."

      I should be inclined to treat such cases as follows. When a person, who has often both seen and felt such things as snow and ice, merely looks at such an object, the sensibile which he visually senses is no doubt subtly modified in a characteristic way through the excitement of traces left by his past associated tactual experiences. But I see no reason to think that this modification consists in the sensibile literally having the quality of sensible coldness, in addition to such qualities as hue, brightness, distribution of light and shade, etc. I should suppose that it is these latter qualities which are modified in a subtle and characteristic way. But we have no simple adjective available to name such characteristic modifications of hue, brightness, distribution of light and shade, etc., and none for the visual gestalt-quality in which they are integrated. So we express its presence by using the verbally paradoxical adjective "cold-looking."

      The following analogy may be helpful. When a person says that something looks cream-coloured, no one supposes that the sensibile which he senses is made of cream. We know that he uses this phrase simply because there is no simple adjective, like "red" or "purple," available to name the hue in question. Why should we not treat such phrases as "cold-looking," "hard-looking," etc., on similar lines mutatis mutandis?

      The next critical point in Professor Price's series is where the characteristic, which a perceived object is said to appear to have, is or involves something of such a nature that it cannot be literally present to any sense. Examples are: "He looks angry," "He looks ill," etc. To be angry, e.g., is to be feeling certain emotions, to be disposed to speak and act in certain ways, and so on. These are plainly not sensible qualities of any kind.

      The analysis which Professor Price offers of O appears φ to S. in such cases, seems to come to the following.

1. The sensation, in and through which S is perceiving O, is a sensing by him of a sensibile which he senses as having a sensible quality ψ, of a certain characteristic kind.
2. In S's past experience objects which have presented that kind of appearance have generally or always been found to have the non-sensible characteristic φ.

      I think that this is a plausible account of the circumstances under which such experiences happen, but that it is hardly an adequate account of the experiences themselves. I would suggest that when a creature (man or animal of the higher kind), whose past experiences have been of the kind described under (b), has a sensation of the kind described under (a), he feels certain characteristic emotions (e.g., apprehension) towards O, and that dispositions in him to react in certain ways are stirred and certain feelings arise in connexion with this. These emotions and feelings blend with the purely cognitive factor in the experience. I suggest that the experience which a person might express by saying "O looks angry to me," and which a dog might have but could not express in words, is this blended state of cognition, emotion, and feeling. And I would suggest a similar account of other such experiences.

(B) PROFESSOR MARC-WOGAU'S PAPER.

      Professor Marc-Wogau confines his discussion to visual perception of bodies, and I shall here follow his lead. I think that the question which he is primarily concerned to discuss may be stated as follows: --

      Let us grant, for the sake of argument, that whenever a person is ostensibly seeing a certain body, e.g., a cricket-ball, he is directly apprehending a certain particular existent, which sensibly presents itself to him as having certain qualities, e.g., brownness, roundness, convexity, etc. Let us also grant, for the sake of argument, that such a directly apprehended particular must have any characteristic which it sensibly presents itself as having, and that it must have it in the precise determinate form in which it presents itself as having it. It has been alleged that, even when the ostensible perception is veridical and non-hallucinatory (e.g., when there really is a cricket-ball in front of the percipient, and when his ostensible perception really is evoked by the stimulus of light coming to his eyes from it) the particular which he directly apprehends is never identical with the body which he sees or with any part of it, and a fortiori is never identical with any other body or any part of one. I will call this doctrine "the non-corporeality of visual sensibilia." It is this doctrine which Professor Marc-Wogau is concerned to discuss. His thesis is that, whether it be true or false, the reasons which have been alleged for it are inconclusive or positively fallacious.

      I shall not waste time in discussing whether the non-corporeality of visual sensibilia becomes analytic, if we substitute the phrase "visual sensum" for the phrase "particular which a person directly apprehends when he ostensibly sees a body." That depends on how certain philosophers may have defined a certain technical term. But a person must already have persuaded himself that such particulars are never bodies or parts of bodies, before he would make that property a part of his definition of the technical term "visual sensum," by which he proposes to call them.

      Professor Marc-Wogau is obviously right when he says that we need to be clear as to what we mean by "body" and by "part of a body" before we can appraise the doctrine in question. Suppose, e.g., you allow that the name "body" may be given to any complex of sensibilia inter-related in a certain characteristic intimate way, and that the phrase "being a part of a body" may be applied to the relationship of being a constituent of such a complex. Then some of the arguments which have been used in support of the non-corporeality of visual sensibilia cease to be relevant.

      In this connexion Professor Marc-Wogau makes some highly pertinent comments on statements which I have made in various places about "the common-sense view of bodies." The essential point is that I seem to suggest that certain beliefs about the nature of a body are implicit in the language and behaviour of plain men; that these can be formulated, and constitute an essential part of the connotation of the word "body" and the phrase "part of a body," as plain men understand them; and that it can be seen on reflexion that the sensibile which a person senses when he sees a body never answers to the conditions required for being a "body" or a "part of a body," in the popularly accepted sense of those words. Professor Marc-Wogau says that I seem to ascribe a philosophic theory to plain men; and that it is very doubtful whether they have one, and whether (if they have) I have formulated it correctly. He remarks that sometimes I seem to make it an objection to a philosophic theory about bodies and sense-perception that it would "shock common-sense," and at other times advise common-sense to "go out and hang itself" like Judas Iscariot.

      What I would now say about this is the following.

      (1) I still think that ordinary language and practice about bodies and parts of bodies do at least strongly suggest that the words "body" and "part of a body" connote certain characteristic properties, for most men at most times, in the entities to which they are applied. I would hardly describe these beliefs as constituting a "philosophic theory." But that phrase is no doubt highly elastic, and I do not wish to dispute about words.

      (2) At any rate I would admit (and indeed assert) this. When a professional philosopher, who has reflected on these topics and has been led thereby to draw distinctions which are not recognised and not needed in our ordinary practical dealings with bodies, tries to formulate what he takes to be suggested by ordinary language and practice, the result of his efforts will always be open to the following objections.

1. The beliefs which he ascribes to the plain man are likely to be severally much clearer and collectively much more coherent than those which plain men hold at most times.
2. It will be of very little use to enquire of supposedly representative plain men whether they do or do not hold the beliefs ascribed to them. The witness will not understand your questions, or see the point of your putting them to him, until you have got him to see distinctions which he would not have noticed if left to himself. At that stage he has ceased to be a "representative plain man," and his conditioning at your hands has probably biased him in favour of an affirmative or of a negative answer to your questions.

      (3) In view of this, it is perhaps presumptuous for a philosopher to describe his formulation of what he takes to be connoted by "body" and by "part of a body" as "the common-sense belief." The important thing is that he should formulate it clearly, and that it should not obviously conflict with common-sense notions, vague, incoherent, and half-baked as these may be. I think that the account which I have offered fulfils these modest conditions.

      (4) Lastly, I do not think that the fact that a philosophic theory (e.g the Leibnitzian theory that what we ostensibly perceive as a body is in fact a certain kind of collection of very low-grade minds) would "shock common-sense," is by itself any good reason for doubting it. As one who is, in his non-professional hours, the plainest of plain men, I would like to do what I can for "poor dear common-sense." But I refuse to regard that thing of rags and tatters as an oracle. And, as a professional philosopher, I should not hang my head or feel wistful, if I should find myself obliged (as a result of recognising distinctions which the plain man ignores, and of viewing synoptically facts which he views only severally, and of some of which he is entirely ignorant) to "shock common-sense" quite severely.

      I will now pass from these generalities to Professor Marc-Wogau's criticism of certain specific arguments, which have been put forward in support of the non-corporeality of visual sensibilia. Before doing so I will make two explanatory remarks.

      (1) The word "see" is used in ordinary speech in a looser and in a stricter sense. In the looser sense one would talk of seeing a body (e.g., a certain cricket-ball). In the stricter sense one would say that, from any one position at any one moment, one sees only a certain part of the external surface of the body which one is seeing in the looser sense. Now the only question worth discussing here is whether or not the visual sensibile which a person senses, when he "sees" (in the looser sense) a certain body, is or is not ever identical with that part of the external surface of that body which he is then "seeing" in the strict sense

      (2) Suppose I were looking at a body whose surface is variegated in colour, e.g., a geographical globe with the representation of the U.S.A. facing me, and with the various States represented by adjoined patches of various colours. Then what I understand by the statement that the sensibile which I am visually sensing is identical with the part of the surface of the globe which I am strictly seeing, is this. The total coloured particular which I am visually sensing is a part of a certain complete coloured surface of spherical shape (the rest of which I am not visually sensing), in precisely the way in which the representation of the State of Nebraska (which I am visually sensing) is a part of the representation of the U.S.A. (which I also am visually sensing). I assume that what Professor Marc-Wogau questions is the cogency of various arguments which have been adduced to show that such propositions as this are never true.

      Of the three specific arguments which Professor Marc-Wogau considers, I need not dilate upon what he calls the "Argument from Delusive Perceptual Situations." As he says, I have discussed that argument pretty fully in Mind, Vol. LVI, Pp. 104-107, and have admitted it to be inconclusive. I have not altered my opinion. But I continue to think that the argument has some weight. The other two arguments are

1. an argument from continuity, and
2. an argument based on the finite velocity of light.

(1) Argument from Continuity.

      This argument is based on the alleged continuity in the series of sensibilia visually sensed by a person when the same part of the same body (e.g., the whole of the top of a penny) is seen from a series of positions at various distances from it and at various angles to the normal through its centre. Professor Marc-Wogau complained, quite justifiably, that those who use this argument have generally ignored the fact of "phenomenal constancy," which has been established by experimental psychologists. I alleged that this merely shifts the point of application of the argument to the dividing line between the sub-class of this series of sensibilia within which phenomenal constancy holds and the sub-class for which it breaks down.

      To this he answers that there is no reason why all the sensibilia on the one side of this line should not be identical with the top of the penny, whilst none of those on the other side of it are so. (It may be remarked that, strictly speaking, there would be, on the view in question, only one sensibile on the former side of the dividing line, though it would answer to a number of different descriptions, each of the form "the sensibile sensed by X from position P.") Professor Marc-Wogau suggests that the difference between the two sub-groups of sensibilia might be compared with the difference between a set of trees which could be called a "wood" and a set which could be called a "grove."

      This analogy seems to me to be faulty.

1. Since one of the sub-groups would, as I have pointed out, contain only one member, it would be analogous to a single tree and neither to a wood nor a grove.
2. When we say that a "wood" and a "grove" melt imperceptibly into each other, all that we mean is that there is a rather indefinite range of size and density within which either name is equally applicable to a collection of trees.

1. being identical with a certain part of the surface of a certain body, and
2. not being a part of the surface of any body.

      What seems to me so paradoxical is to suppose that such a series of visual sensations, evoked under conditions which vary continuously should fall into two sub-classes, having objects of such radically different ontological status.

(2) Argument from the finite velocity of Light.

      I have dealt very fully with this argument in Scientific Thought, and again (with special reference to Professor Marc-Wogau) in Mind, Vol. LVI, Pp. 120-124. All that I wish to add here is the following.

      Professor Marc-Wogau says that we need not consider the argument except in cases where the body seen is so near to the percipient's body that the time-interval between the emission or reflection of light from the former and its arrival at the latter is extremely small compared with the duration of the percipient's specious present. The reason which he gives is that, in the case of a very distant body, e.g., the sun, it is obvious, from a mere comparison of the characteristics which the sensibile is sensed as having with those which the body is known to have, that the former cannot be identical with a part of the surface of the latter. The argument in question is therefore unnecessary in such cases

      That is no doubt true. But the fact that the argument is not needed in the case of very distant bodies does not affect its validity in such cases. Nor do I think that Professor Marc-Wogau would claim that it does. But, if that be admitted, the question of continuity comes in. The case where the body seen is near to the percipient's body cannot fairly be considered in isolation from cases where it is very remote. We see bodies which are at all sorts of distances, from close at hand to many millions of miles away. The external causal conditions of the visual perceptions are, so far as we know, precisely similar in kind in all these cases. Is it really credible that there is a certain range of distance, on one side of which the immediate objects of visual sensations are parts of the surfaces of bodies emitting light to the percipient's eye, and on the other side of which they are of a wholly different nature?

      To sum up. I think that all the arguments for the non-corporeality of visual sensibilia rest on considerations of continuity. In view of the continuity in the external conditions of our visual sensations, I find it very hard to believe that some of the visual sensibilia which we sense are parts of the surfaces of the bodies which we see, and that others are not parts of the surface of any body, if to be a "body" and to be "a part of the surface of a body" be understood in the simple literal way which I have tried to state and illustrate. Now I also find it very hard to believe that all the visual sensibilia which we sense are parts of the surfaces of the bodies which we see, if "body" and "part of the surface of a body" are understood in that way. Therefore I am strongly inclined to think that none of them are. I admit that neither severally nor collectively are the arguments conclusive. What I may call Professor Marc-Wogau's "half-and-half" theory is logically possible; but it is the kind of theory of which I can only say: "If it should be true, I'll eat my hat!"

      I will conclude, however, by adding this remark. I have been mainly concerned to work out, for good or ill, theories which presuppose the non-corporeality of visual and other sensibilia. But I have always recognised that there are other alternative views of sense-perception which an intelligent and instructed philosopher might take. Moreover, as I have grown older I have realised more and more that the plausibility of that presupposition rests on certain assumptions, e.g., the analysis of ostensible perception into sensation plus perceptual acceptance founded upon it, the "act-object" analysis of sensation, and the assumption of inerrancy. And I realise that I was formerly inclined to take all these too much for granted.

(C) PROFESSOR YOLTON'S PAPER.

      I shall discuss Professor Yolton's paper under the following three headings, viz.,

1. the three "alternative ontologies" which he mentions,
2. the notion of "ontological construction," and
3. the alleged "phenomenalist" and "dualist" strands in my writings.

(1) "Phenomenalism," "Phenomenalist Realism," and "Dualist Realism."

      Professor Yolton defines the first two of these theories in terms of what he calls "sense-qualities." I think he must mean what I should call "sense-qualified occurrents," and I shall so interpret his statements. On that interpretation one might say, e.g., that Berkeley held a form of "Phenomenalism," and that Lord Russell has propounded in some of his works a form of "Phenomenal Realism."

      I note that Professor Yolton includes among "sensible qualities" the property of being made of oak. Surely that involves having certain causal or dispositional properties, which are certainly not sense-given, in the way in which, e.g., redness and coldness are. However that may be, it seems to me certain that causal and dispositional properties are an essential element in the notion of a body. But "phenomenalists" and "phenomenal realists" might wish to give an account of such properties as inertia, elasticity, gravitational mass, etc., which "dualist realists" would not accept, and vice versa.

      I take "Dualist Realism" to involve at least the following propositions.

1. That there certainly or probably are "bodies," in the sense of more or less persistent substances within closed surfaces, e.g., spheres, cubes, etc., which have
   1. non-dispositional "extensible qualities" diffused over their surfaces or throughout their volumes,
   2. dispositional properties, e.g., inertial mass, elasticity, etc., and
   3. positions in a spatial order of at least three dimensions.
2. That when a person ostensibly sees or touches a body, he is immediately aware of certain sense-qualified occurrents, but that these are in no case bodies or parts of the surfaces of bodies, in the literal way in which one such sense-qualified occurrent often is a spatial part of another. They are particulars of a radically different kind. And it may be doubted whether any of them ever exists except as the objective factor in a particular sensation had by a particular individual on a particular occasion.

      Now, if this be what is implied by "dualist realism," it seems plain that there is prima facie another alternative, which might be called "direct realism." This would accept (1) and reject (2) in the above summary. I think that the phrase "to reject (2)" covers two very different attitudes, which might be called "pre-critical" and "post-critical." At the pre-critical stage no clear distinction is drawn between the notion of what I have called "sense-qualified occurrents" and the notion of bodies or parts of the surfaces of bodies. Speaking in terms of a conceptual distinction which is not clearly recognised by the persons concerned, we, who have recognised that distinction, can say that the percipient simply takes for granted (at any rate in perceptual situations which are not regarded by him as decidedly abnormal) that the particular which he is visually or tactually sensing is literally a part of the surface of the body which he is ostensibly seeing or touching. At the post-critical stage we have philosophers, who have recognised the distinction in question, and are well aware of the arguments against identifying the particulars sensed with parts of the surfaces of bodies seen or touched, but nevertheless hold that the former are identical with the latter, at any rate in normal perceptual situations.

      I suppose that Reid would be an adherent of what I should call "post-critical direct realism." I agree with Professor Yolton and Professor Marc-Wogau that it is dangerous, and perhaps almost a contradiction in terms, to ascribe any philosophical theory to plain men. Yet I am inclined to think that the language and behaviour of plain men and of philosophers in their non-professional hours, implies or suggests what I should call "pre-critical direct realism."

      When all the relevant facts, viz., those of physics, physiology, anatomy, and psychology, are taken into account, I think that direct realism is very difficult to maintain. I do not doubt that, with enough ingenuity and a good deal of special pleading, it could be saved from downright refutation. But I confess that I do not think that it is worth such intellectual acrobatics. On the other hand, I do not wish to depart further from pre-critical direct realism than I am forced to do by a fair synoptic consideration of all the relevant facts. That is not because I regard any proposition as sacrosanct, merely because it seems to me to be implied or strongly suggested by common language and everyday behaviour. At most I should say that there is a prima facie case for treating such propositions seriously, and seeing how far and in what sense they can be maintained in face of relevant facts which were unknown when ordinary language was formed, and which are seldom or never viewed synoptically even by those to whom they are known.

      I think that a form of dualist realism can be stated, which fulfils these various conditions on the whole better than any alternative theory known to me. I have in the main tried to work out such a theory. I have never doubted that other types of theory, better in some ways but perhaps not so good in others, can be coherently formulated and plausibly defended. I think, however, that most of them demand more boldness and speculative originality than I have ever possessed, and that it was better for me to stick to my last.

(2) "Ontological Construction."

      Professor Yolton defines this as an attempt to derive physical objects from instantaneous punctiform event-particles. He rightly contrasts it with what he calls "linguistic construction," which he defines as an attempt to derive, from the phenomenal world, meaningful concepts to apply to the physical world.

      I am sure that Professor Yolton exaggerates the significance of what I have said about ontological construction, in this sense, in relation to my account of the physical world. I suggest that this is because he ignores the very special context in which it occurs, and the very special purpose which I had in mind.

      So far as I am aware, all that I have said on this topic will be found on Pp. 587-603 of The Mind and its Place in Nature. It occurs in a chapter devoted to the nature of the unity of a mind. The context is a discussion of the question whether it is possible to take the notion of mental event as primary, and to regard a mind as a certain kind of complex composed of appropriately inter-related mental events. In order to elucidate this I threw out for discussion the question whether it is possible to take the notion of physical event as primary, and to regard a body as a certain kind of complex composed of such events.

      I did not want to devote much space to the development of what was a side-issue of a side-issue. I therefore stated the case in terms of literally punctiform instantaneous qualified event-particles, and talked as if I thought that these might be actually existent particulars, and as if bodies might be literally composed of these. But I never seriously believed this to be a possible view. Professor Yolton has taken all this too seriously and with too little reference to its peculiar context. I do not think that it would be profitable to pursue the question further here.

(3) The ''Phenomenalist'' and the "dualist" strands.

      The above misunderstanding (as I think it) does not, however, affect the validity of what I take to be Professor Yolton's main criticism. This, if I am not mistaken, comes to the following. I make great use of the notion of logical construction, and in particular of Whitehead's Method of Extensive Abstraction. But there is a standing ambiguity as to what I supposed to be accomplished by this. Sometimes I write as if I held that statements in which physical-object words and phrases occur can be replaced, without loss or gain of meaning, by more complicated statements which are about nothing but sense-data, sensible qualities and sensible relations. At other times I write as if I held that physical objects are particular existents of a certain kind, and sense-data (or the sensations in which they are the objective factor) are particular existents of a radically different kind; that there are intimate causal relations between certain of the former and certain of the latter; and that the logical constructions serve only to define, in terms of sensible qualities and relations, the concepts in terms of which we have to think of physical objects, their qualities, and their relations. We may call these respectively the "phenomenalist strand' and the "dualist strand."

      I should not be much ashamed of this, if each tendency occurred only in different writings, and if those writings differed considerably in date and in main purpose. But I must admit that statements exemplifying each tendency are to be found in one and the same book, e.g., Scientific Thought.

      I believe that my main intention at the time was to expound and defend a form of dualism, viz., what I call in Scientific Thought the "Critical Scientific Theory." This may fairly be described as an attempt to refurbish (in the light of criticisms made by later philosophers, and with the help of tools provided by later logicians) the much decried "theory of Representative Perception" or "Causal Theory of Perception," which goes back through Locke to Descartes. Much water (and still more hot air) has passed through the bridges of philosophy since I wrote. But I still think such a theory defensible, and I have not met with any alternative which seems to me less unsatisfactory in view of all the relevant facts. The philosophy of the physical world and of our perception of it becomes analogous in certain respects (though profoundly dissimilar in certain others) to the making and testing of a far-reaching scientific hypothesis. I am well aware that (to parody St. Paul) all this is "to the Phenomenalists a stumbling-block and to the Wittgersnappers foolishness." But I have always thought that Vienna contributed more notably to culture by its Schnitzel than by its Kreis, and ich kann nicht anders.

      There is one further remark that I would make here. Professor Yolton speaks of the "Principle of Isomorphism," and says that it is essential to my way of philosophising about sense-perception and the world of bodies and physical events. I am inclined to think that what I had in mind was something less determinate and more flexible than what Professor Yolton understands by that principle.

      What I would say is this. There is no reason a priori why the locus of those physical events which are the rather remote causal ancestors of the various groups of inter-connected sensations by which various persons ostensibly perceive a certain body, should resemble at all closely the percepta of those persons. There is also no reason a priori why it should not do so. Antecedently, then, we are free to postulate as much or as little resemblance as we choose between

1. the qualities and inter-relations of the hypothetical system of physical things and events, and
2. the qualities and relations which the objects that we ostensibly perceive present themselves to us in sense-perception as having.

      The latter is certainly the only source from which we can derive the empirical concepts, in terms of which we think of the physical system as a whole, and of its detailed contents and their varying states and mutual relations. Undoubtedly our thought of these involves also concepts which I regard as non-empirical, in that they are not and could not possibly be sense-given, in the way in which, e.g., the concepts of colour, of shape, of position, of motion, etc., are so. Examples of such non-empirical concepts, essential to the thought of a system of physical things and events, are the notions of substance, of cause, of disposition, of potentiality and actuality, and so on. But nothing concrete can be thought of wholly in terms of categories; there must be a non-categorial "filling" and specification, and this can come only from what is sensegiven.

      If this be allowed, I am willing to leave to experts to decide

1. in what respects it is necessary to postulate isomorphism, if we are to form any workable and fruitful conception of the actual physical world; and
2. to what degree it is necessary, and within what limits it is permissible, to postulate isomorphism in those respects.