Under this heading I shall deal with a part, but not the whole, of Professor Turnbull's admirably acute and thorough paper, viz., that which is concerned with the account which I have given of what is commonly called "abstraction" and of what I call "descriptive ideas." In discussing this essay I shall not attempt to use the formidable technical terminology which Professor Turnbull deploys at the beginning of it. That is not because I have failed to understand it, or because I think it unsuitable for its purpose. It is because I am sure that all that I have to say here can be stated accurately enough in a simpler and more homely way.

(A) ABSTRACTION.

      This is the process by which, it is alleged, a person forms a dispositional idea of such a characteristic as red or round from perceiving with his senses objects which present themselves sensibly to him as red or as round. I must confess that I took over the account which I gave of this process very uncritically from a tradition which goes back (I suppose) through Locke to the Scholastics. I tried to state it as clearly as I could, but I did not seriously question its presuppositions. When one is forced to consider them, they do seem rather shaky. It is plain that Professor Turnbull rejects the whole traditional story and would offer a very different one in its stead. But he also claims to detect inconsistencies between certain statements within my exposition of the traditional theory. I shall here deal mainly with these alleged internal inconsistencies.

      (1) The first point which I will consider is this. In my account of the origin of the dispositional idea of a determinate colour, e.g., red, I assumed that what a person compares and contrasts are physical objects, e.g., skeins of coloured wool, coloured crayons, the tablets of pigment in a paint-box, and so on. Much later on, when I came to deal with the analysis of sense-perception I suggested that to see a physical object consists in

1. being sensibly acquainted with ("prehending") a certain particular which sensibly presents itself as having a certain determinate colour, shape, etc., and
2. taking it uncritically and non-inferentially as part of the surface of a body, with certain other parts and certain other properties which are not at the moment being sensibly presented to one.

      Now Professor Turnbull seems to think that these latter reflexions about the analysis of sense-perception, and about the interpretation of such sentences as "That body is red," somehow conflict with the original account of abstraction and demand a radical revision of it. If he does think so, I do not agree with him. The account of abstraction is and remains in terms of colours, shapes, etc., of bodies as seen. We are all perfectly familiar with such situations as we should describe as "seeing two skeins of wool of the same shape and different colours" and as "seeing a skein of wool and a postage-stamp which are visibly alike in colour and unlike in shape." According as there are various suggested analyses of "seeing a body as of such and such a shape and of such and such a colour," there will be as many different analyses of the process of comparing and contrasting seen bodies in respect of the colours and shapes which they present to sight. But the process is in any case Just the one which I have indicated, and which I am sure that anyone can recognize from my description.

      (2) I pass now to a much more radical criticism. If I understand him aright, Professor Turnbull holds that my whole account of abstracting (no matter what analysis, if any, be given of sense-perception) is logically circular. The reason alleged is that to see a body as, e.g., red, presupposes that the person who has that experience already has the dispositional idea of red. Professor Turnbull does not argue this point, he seems to find it obvious.

      Now it is not in the least obvious to me. Suppose that a young child (or possibly a cat or a dog) with normal eyesight looks in daylight at a body, such as a ripe tomato, which Professor Turnbull and I (who have acquired the dispositional idea of red and the use of the word "red") would describe as red. Then I assume that the percipient would have a visual experience which differs in a characteristic way from the one which he would have if he were to look, in similar circumstances, at another thing, similar to it in shape and size, such as an unripe tomato, which Professor Turnbull and I (who have acquired the dispositional idea of green and the use of the word "green") would describe as green.

      All that my account of abstraction presupposes is that there are in fact such characteristic unlikenesses and such characteristic likenesses between certain visual perceptual experiences of young children. Obviously a percipient could not describe these differences as "seeing this as red and seeing that as green," and he could not understand such descriptions, unless and until he had acquired the dispositional ideas of red and of green and the proper use of the words "red" and "green." But that is no reason why he should not have experiences which in fact differ in the ways which we so describe, nor is it any reason why he should not come to recognise that fact about his experiences.

      (3) Professor Turnbull raises the question: What do I suppose to be innate in reference to the dispositional ideas of red, of blue, and so on? I suppose that, in order to be able to acquire the idea of red, the idea of blue, and so on, one must have the innate capacity to have visual experiences which in fact differ in all those characteristic ways which we come to describe as "seeing a thing as red," "seeing a thing as blue," and so on. Since it is logically possible (and also causally possible, as the facts of "colour-blindness" show) to have some of these capacities without the others, one must presumably postulate a number of logically (and to some extent causally) independent innate colour-sensation capacities. Given these, I should have thought that one and the same general innate capacity to notice, compare, contrast, and abstract would be involved in regard to each particular colour-likeness and colour-unlikeness experienced. However that may be, I do not think that an valid objection can be made to the traditional doctrine of abstraction on the ground that it may need to postulate a very large number of innate capacities. Since some have to be postulated by everyone, the precise number required by a particular theory seems to be a minor matter.

      (4) The questions so far considered might be called "psychological" or "epistemological." But Professor Turnbull also raises a question which might be called "ontological," viz: What is supposed to be abtracted from what?

      I fear that I cannot offer much positive information on this point. What is supposed to be abstracted is, of course, a quality or a relational property or a relation. What it is supposed to be abstracted from is a particular which sensibly presents itself as qualified by that quality or relational property, or a set of particulars which sensibly present themselves as inter-related by that relation.

      Obviously all expressions which suggest that a particular stands to a characteristic which characterises it as a whole stands to a part of it and that abstraction is analogous to the physical separation of a part from a whole, are hopelessly misleading. When the idea of red has been abstracted from a number of red particulars, they have not thereby lost their redness, as if abstraction were a process of leaching or bleaching! What has happened is that a person has acquired the ability to think of a characteristic, which did (and, for all that concerns the process of abstraction, may still) characterise those particulars, without needing at the time to perceive those or any other particulars as characterised by it. In some sense it is certain that this phrase describes something with which we are all perfectly familiar. The difficulty is to give a satisfactory analysis of this state of affairs. I am quite willing to believe that the traditional account of abstraction fails to do this, and that it may tempt one to ask silly questions or may call up absurd associations. I find little difficulty myself in resisting the temptations and ignoring the associations.

      (5) This leads on to a fundamental objection raised by Professor Turnbull. I said that, when a person has acquired the dispositional idea of, e.g., red, he has ipso facto acquired the ability to "contemplate" the characteristic red without needing at the time to be acquainted, either in sense-perception or in imagery, with a particular which presents itself as red to him. And I said that a person is having an occurrent idea of, e.g., red, whenever he is in fact "contemplating" the characteristic red.

      Now some of Professor Turnbull's comments on this do seem to me to be mainly verbal. He shows quite conclusively that the experience described as "contemplating the characteristic red" is extremely unlike various other experiences which are commonly and more literally described as "contemplating so-and-so." That in itself would show only that the name "contemplating" (which, like all names applied to intellectual operations, is used metaphorically) is not a happy one, and is more likely to mislead than to illuminate.

      But behind this verbal skirmishing there is an attack which I regard as serious, and to which I have no satisfactory defence. The point may be put as follows. I have been inclined uncritically to regard the experience of thinking of a characteristic (e.g., of red), when one is not being presented either sensibly or in imagery with anything that presents itself as red, as analogous to being acquainted with a particular. I have spoken as if the only difference were on the side of the object, viz., in one case a particular and in the other a universal. I have in fact been still more specific, for I have undoubtedly tended to regard the experience as analogous to that acquaintance with a particular which is an essential factor in seeing a body or such a physical event as a flash of lightning. Moreover, I have tended to think of the process of becoming aware of a necessary connexion or disconnexion between two characteristics, e.g., between equilateral and equiangular triangularity, as analogous to prehending two coloured particulars and noting, e.g., that the red one is adjoined to the blue one.

      The use of language drawn from visual perception evidently comes naturally to speakers of Indo-European languages. (I do not know whether it extends beyond them.) We all talk of "seeing" or "failing to see" a logical connexion. If it be a bad habit, it is one that we have inherited from our prehistoric ancestors; for the Latin video and the English wit (with its kith and kin wissen in German, veta in Swedish, witan in Anglo-Saxon, etc.) have a common root, which means "to see." Nevertheless when one is made aware of the habit and begins to reflect on what one has been doing, the analogy suggested by the verbal expressions is found to be faint in the extreme. I am now fully aware of the fact that the experience of thinking of a characteristic, in the absence of any perceived or imaged instance of it, is and must be utterly unlike the experience of seeing in the literal sense. I am sure that, through applying the language of visual perception to intellectual operations, I have often been led unwittingly through its associations to make unjustifiable assertions about them. The driving home of this point is for me the main positive outcome of Professor Turnbull's criticisms.

      (6) As regards the relation between acquiring a dispositional idea, e.g., that of red, and acquiring the power to use and understand correctly the corresponding word, e.g., "red," all that I have to say is this. I take it that Professor Turnbull's view as to the correct analysis of the notion of having a dispositional idea of red is such that it would be logically impossible either

1. to acquire the dispositional idea without acquiring the corresponding verbal ability, or
2. to acquire the latter without the former.

      Now I certainly gave an account of the acquirement of the dispositional idea which made no mention of the acquirement of the corresponding verbal ability. So it may fairly be concluded that I held that the acquirement of the former is logically independent of the acquirement of the latter. It cannot fairly be concluded that I held it to be causally possible for a human being to acquire the dispositional idea without acquiring the corresponding verbal ability. Still less could it fairly be concluded that I held that there are any known instances of a human being acquiring the former without acquiring the latter. I should think it most unlikely that there have been or will be. And I am mclined to think (though without any strong conviction) that it may be causally impossible that there should be such a case.

(B) DESCRIPTIVE IDEAS.

      I developed this notion in reference to a question which Hume raises in Sect. I of Part I of Book I of his Treatise of Human Nature. Suppose that Jones has never seen anything which presented to him any shade of red between the two shades s₁ and s₂. Suppose he has seen things which presented to him the shade s₁, and has seen things which presented to him the shade s₂, and has formed what I called an "intuitive idea" of those shades. Then I alleged that the sentence: "Jones is thinking of the shade of red between s₁ and s₂" means what is meant by the sentence: "Jones is thinking of the proposition that there is one and only one shade of red between s₁ and s₂." Professor Turnbull objects to this, on the ground that to think of a shade of colour cannot be the same as to think of a proposition.

      That certainly seems very plausible at first sight, since a shade of colour and a proposition are obviously entities of two entirely different types. Nevertheless, I do not find the objection conclusive on reflexion. Remember that there may in fact be no shade between s₁ and s₂ (for the series of possible shades may be discontinuous, and these may be immediate successors in the series), or, on the other hand, that there may be more than one; and yet that the fulfilment of either of those possibilities would be quite compatible with Jones having an experience correctly describable as "thinking of the shade between s₁ and s₂." In view of this, is it not somewhat naive to argue, from the premiss that a shade of colour is not a proposition, to the conclusion that "Jones is thinking of the shade of colour between the shades s₁ and s₂" cannot mean what is meant by "Jones is thinking of such and such a proposition"?

      I would make a similar rejoinder mutatis mutandis to Professor Turnbull's in principle similar objection to my account of such experiences as would be described by saying, e.g., "Jones is thinking of a fire-breathing serpent."

      But I agree that his hypothetical example shows that a case is conceivable where an idea which would be simple, according to my criterion, would have as its ideatum a characteristic which could quite properly be called "complex." I would say that, when a person has an idea which is simple, in accordance with my criterion, he has no positive reason to think that its ideatum is complex, in the sense illustrated in Professor Turnbull's example. But a person in that position ought always to be ready to admit the possibility that the ideatum may be complex, in that sense.