CHAPTER I

THE NATURE OF THE PROBLEM

I. WHAT IS MIND?

1.1. The nature of the subject of this study makes its first task the most important and the most difficult: clearly to state the problem to which it will attempt an answer. We shall have moved a considerable distance towards the solution of our problem when we have made its meaning precise and have shown what kind of statement could be regarded as a solution.

1.2. The traditional heading under which our problem has been discussed in the past is that of the ‘relation’ between mind and body, or between mental and physical events. It can also be described by the questions of ‘What is mind?’ or ‘What is the place of mind in the realm of nature?’ But while these expressions indicate a general field of inquiry, they do not really make it clear what it is that we want to know. Before we can successfully ask how two kinds of events are related to each other (or connected with each other), we must have a clear conception of the distinct attributes by which they can be distinguished. The difficulty of any fruitful discussion of the mind-body problem consists largely in deciding what part of our knowledge can properly be described as knowledge of mental events as distinguished from our knowledge of physical events.

1.3. We shall attempt to avoid at first at least some of the difficulties of this general problem by concentrating on a more definite and specific question. We shall inquire how the physiological impulses proceeding in the different parts of the central nervous system can become in such a manner differentiated from each other in their functional significance that their effects will differ from each other in the same way in which we know the effects of the different sensory qualities to differ from each other. We shall have established a ‘correspondence’ between particular physiological events and particular mental events if we succeed in showing that there can exist a system of relations between these physiological events and other physiological events which is identical with the system of relations existing between the corresponding mental events and other mental events.

1.4. We select here for examination the problem of the determination of the order of sensory qualities because it seems to raise in the clearest form the peculiar problem posed by all kinds of mental events. It will be contended that an answer to the question of what determines the order of sensory qualities constitutes an answer to all questions which can be meaningfully asked about the ‘nature’ or ‘origin’ of these qualities; and further, that the same general principle which can be used to account for the differentiation of the different sensory qualities serves also as an explanation of the peculiar attributes of such other mental events as images, emotions, and abstract concepts.

1.5. For the purposes of this discussion we shall employ the term sensory ‘qualities’ to refer to all the different attributes or dimensions with regard to which we differentiate in our responses to different stimuli. We shall thus use this term in a wide sense in which it includes not only quality in the sense in which it is contrasted with intensity, extensity, clearness, etc., but in a sense in which it includes all these other attributes of a sensation.1 We shall speak of sensory qualities and the sensory order to distinguish these from the affective qualities and the other mental ‘values’ which make up the more comprehensive order of ‘mental qualities’.

2. THE PHENOMENAL WORLD AND THE PHYSICAL WORLD

1.6. A precise statement of the problem raised by the existence of sensory qualities must start from the fact that the progress of the physical sciences has all but eliminated these qualities from our scientific picture of the external world.2 In order to be able to give a satisfactory account of the regularities existing in the physical world the physical sciences have been forced to define the objects of which this world exists increasingly in terms of the observed relations between these objects, and at the same time more and more to disregard the way in which these objects appear to us.

1.7. There exist now, in fact, at least two3 different orders in which we arrange or classify the objects of the world around us: one is the order of our sense experiences in which events are classified according to their sensory properties such as colours, sounds, odours, feeling of touch, etc.; the other is an order which includes both these same and other events but which treats them as similar or different according as, in conjunction with other events, they produce similar or different other external events.

1.8. Although the older branches of physics, particularly optics and acoustics, started from the study of sensory qualities, they are now no longer directly concerned with the perceptible properties of the events with which they are dealing. Nothing is more characteristic of this than the fact that we find it now necessary to speak of ‘visible light’ and ‘audible sound’ when we want to refer to the objects of sense perception. To the physicist ‘light’ and ‘sound’ now are defined in terms of wave motions, and in addition to those physical events, which, as is true of certain ranges of ‘light’ and ‘sound’ waves, cause definite sense experiences, he deals with imperceptible events like electricity, magnetism, etc., which do not directly produce specific sensory qualities.4

1.9. Between the elements of these two orders there exists no simple one-to-one correspondence in the sense that several objects or events which in the one order belong to the same kind or class will also belong to the same kind or class in the other order. They constitute different orders precisely because events which to our senses may appear to be of the same kind may have to be treated as different in the physical order, while events which physically may be of the same or at least a similar kind may appear as altogether different to our senses.

1.10. These two orders have been variously described by different authors as the subjective, sensory, sensible, perceptual, familiar, behavioural or phenomenal5 world on the one hand, and and as the objective, scientific, ‘geographical’, physical, or sometimes ‘constructional’ on the other. In what follows we shall regularly employ the pair of terms ‘phenomenal’ and ‘physical’6 to describe the order of events perceived in terms of sensory qualities and the order of events defined exclusively in terms of their relations respectively, although we shall occasionally employ the term ‘sensory’ as equivalent to phenomenal, especially (as in the title of this book) in the phrase ‘sensory order’. We shall later (Chapters V and VIII) also describe these two orders as the ‘macrocosm’ and the ‘microcosm’ respectively. Their relation is the central problem of this book.

1.11. It is important not to identify the distinction between the phenomenal and the physical order with the distinction between either of these and what in ordinary language is described as the ‘real’ world. The contrast with which we are concerned is not between ‘appearance’ and ‘reality’ but between the differences of events in their effects upon each other and the differences in their effects on us. It is indeed doubtful whether on the plane on which we must examine these problems the term ‘real’ still has any clear meaning.7 For the purposes of our discussion, at any rate, we shall not be interested in what a thing ‘is’ or ‘really is’ (whatever that may mean), but solely in how a particular object or event differs from other objects or events belonging to the same order or universe of discourse. It seems that a question like ‘what is x?’ has meaning only within a given order, and that within this limit it must always refer to the relation of one particular event to other events belonging to the same order. We shall see that the mental and the physical world are in this sense two different orders in which the same elements can be arranged; though ultimately we shall recognize the mental order as part of the physical order, a part, however, whose precise position in that larger order we shall never be able to determine.

1.12. Historically the concept of the ‘real’ has been formed in contradistinction to mere ‘illusions’ based on sense deceptions or on other experiences of purely mental origin. There is, however, no fundamental difference between such corrections of one sense experience by others, as we employ, e.g., to discover an optical illusion, and the procedure employed by the physical sciences when they ascertain that two objects which may to all our senses appear to be alike do not behave in the same way in relation to others. To accept this latter test as the criterion of ‘reality’ would force us to regard the various constructs of physics as more ‘real’ than the things we can touch and see, or even to reserve the term ‘reality’ to something which by definition we can never fully know. Such a use of the term ‘real’ would clearly pervert its original meaning and the conclusion to be drawn from this is probably that it should be altogether avoided in scientific discussion.8

1.13. The relation between the physical and the phenomenal order raises two distinct but related problems. The first of these problems presents the task of the physical sciences while the second creates the central problem of theoretical psychology. The task of the physical sciences is to replace that classification of events which our senses perform but which proves inadequate to describe the regularities in these events, by a classification which will put us in a better position to do so. The task of theoretical psychology is the converse one of explaining why these events, which on the basis of their relations to each other can be arranged in a certain (physical) order, manifest a different order in their effect on our senses.

1.14. The problems of the physical sciences arise thus from the fact that objects which appear alike to us do not always prove to behave in the same way towards other objects; or that objects which phenomenally resemble each other need not be physically similar to each other, and that sometimes objects which appear to us to be altogether different may prove to be physically very similar.

1.15. It is this fact which has made it necessary, in order to build up a science capable of predicting events, to replace the classification of objects or events which our senses effect by a new classification which corresponds more perfectly to the manner in which those objects or events resemble or differ from each other in the effects which they have upon each other. But this progressive substitution of a purely relational for a qualitative or sensory order of events provides the answer to only one part of the problem which is raised by the existence of the two orders. Even if we had fully answered this problem we should still not know why the different physical objects appear to us as they do.

1.16. It is because the physical sciences have shown that the objects of the external world do not regularly differ in their effects upon each other in the same way in which they differ in their effects upon our senses that the question why they appear to us as they do becomes a legitimate problem and indeed the central problem of theoretical psychology. In so far as the similarities or differences of the phenomena as perceived by us do not correspond with the similarities or differences which the perceived events manifest in their relations to each other, we are not entitled to assume that the world appears to us as it does because it is like that; the question why it appears to us as it does becomes a genuine problem.9

1.17. It is, perhaps, still true that psychologists in general have not yet become fully aware of the fact that, as a result of the development of the physical sciences, the explanation of the qualitative order of the phenomenal world has become the exclusive task of psychology. What psychology has to explain is not something known solely through that special technique known as ‘introspection’, but something which we experience whenever we learn anything about the external world and through which indeed we know about the external world; and which yet has no place in our scientific picture of the external world and is in no way explained by the sciences dealing with the external world: qualities. Whenever we study qualitative differences between experiences we are studying mental and not physical events, and much that we believe to know about the external world is, in fact, knowledge about ourselves.10

1.18. It is thus the existence of an order of sensory qualities and not a reproduction of qualities existing outside the perceiving mind which is the basic problem raised by all mental events. Psychology must concern itself, in other words, with those aspects of what we naïvely regard as the external world which find no place in the account of that world which the physical sciences give us.

1.19. This reformulation of the central problem of psychology has thus been made necessary by the fact that the physical sciences, even in their ideal perfect development, give us only a partial explanation of the world as we know it through our senses and must always leave an unexplained residue. After we have learnt to distinguish events in the external world according to the different effects they have upon each other, and irrespective of whether they appear to us as alike or different, the question of what makes them appear alike or different to us still remains to be solved. The empirical establishment of correspondences between certain phenomenal and certain physical constellations of events is no sufficient answer to this question. We want to know the kind of process by which a given physical situation is transformed into a certain phenomenal picture.

1.20. Since the peculiar order of events which we have called the phenomenal order manifests itself only in the responses of certain kinds of organisms to these events, and not in the relation of those events to each other, it is natural to search for an explanation of this order in some feature of the structure of these organisms. We shall eventually find it in the fact that these organisms are able within themselves to reproduce (or ‘build models of’) some of the relations which exist between the events in their environment.

1.21. The fact that the problem of psychology is the converse of the problem of the physical sciences means that while for the latter the facts of the phenomenal world are the data and the order of the physical world the quaesitum, psychology must take the physical world as represented by modern physics as given and try to reconstruct the process by which the organism classifies the physical events in the manner which is familiar to us as the order of sensory qualities. In other words: psychology must start from stimuli defined in physical terms and proceed to show why and how the senses classify similar physical stimuli sometimes as alike and sometimes as different, and why different physical stimuli will sometimes appear as similar and sometimes as different.11

3. STIMULUS, IMPULSE, AND THE THEORY OF THE SPECIFIC ENERGY OF NERVES

1.22. Before we proceed farther it is necessary to define more precisely some of the terms we shall have constantly to employ. This applies especially to the terms ‘stimulus’ and ‘nervous impulse’ and more particularly to the sense in which we shall speak of particular ‘kinds’ of stimuli or of the same and of different nervous impulses. It will be convenient also to consider already at this stage the meaning and significance of the famous principle of the ‘specific energy of nerves.’

1.23. The term stimulus will be used throughout this discussion to describe an event external to the nervous system which causes (through or without the mediation of special receptor organs) processes in some nerve fibres which by these fibres are conducted from the point at which the stimulus acts to some other point of the nervous system. It appears that at least some receptor organs are sensitive not to the continuous action of any one given stimulus but only to changes in that stimulus. Whatever it is that is produced in the nerve fibre and propagated through it we shall call the impulse.

1.24. The physical event acting as a stimulus is described as such only with regard to its action on the receptors.12 This leads sometimes to a rather confusing distinction between the stimulus and its ‘source’, sometimes described as the stimulus object. What will here be described as stimulus will always be the proximal stimulus,13 i.e., the last known physical event in the chain which leads to the production of the impulse. In some instances (particularly in the case of odours) this proximal physical stimulus, however, is not certainly known, and we must be satisfied with reference to some more remote event which has then to be regarded as the source of an unknown proximal stimulus.

1.25. It is necessary from the outset carefully to avoid the assumption that to each kind of sensation there will always correspond one stimulus of a particular kind. Not only can several different stimuli produce the same sensation, but it appears that in many instances, and perhaps as a rule, several different stimuli, acting on different receptors, may be required to produce a particular sensation.14

1.26. Since our central problem is the manner in which different stimuli affect our nervous system, or how they are classified by it, we clearly cannot make our starting point that classification of the stimuli which our senses perform. The distinction between different stimuli, or between different kinds of stimuli, must be independent of the different effects they have on the organism. This independence can never be complete, since all our knowledge of the external events is derived from our sensory experience. But it can be independent in the sense that we can classify the stimuli not according to their direct effects on our senses, but according to the effects which they exercise on other external events, which in turn act as stimuli on our senses. This classification of the events which act as stimuli, according to their effects on other events which in turn are classified according to their effects on still others, is, of course the classification of the stimuli developed by the physical sciences; and it is this which we must adopt.

1.27. We shall, e.g., have to regard as the same physical stimulus not all light which appears to us to have the same colour, or all substances which smell alike, but only light waves which in various combinations with other physical objects (usually apparatus designed for the purpose) produce the same effects, or substances which in their chemical composition are identical.

1.28. For our purpose it will also be necessary to regard as different any stimuli which are physically identical but which act on different parts of the body, since it is by no means obvious (or always true) that such stimuli should produce the same sensory qualities. The question why as a rule stimulation of different individual receptors by physically identical stimuli should produce similar sensations is in fact the simplest form in which our problem arises.

1.29. The production of a nervous impulse by a stimulus is usually mediated by the selective action of specific receptor organs which respond to certain kinds of stimuli but not to others. This selectivity of the receptor organs is, however, not perfect. Even the so-called ‘adequate stimuli’ to which a given receptor normally responds, consist as a rule not only of one precisely defined physical stimulus (such as, e.g., waves of a particular frequency) but to a more or less wide range of such stimuli extending, e.g., over a certain band of frequencies. In addition to this, some events other than the adequate stimuli can often set up impulses in a given nerve fibre. An impulse in the visual nerves and the consequent sensation of light may, for instance, be caused by a blow on the eyeball.

1.30. The receptor organs thus already perform a certain sorting out, or classification, of the stimuli, and there will be no strict correspondence between the different stimuli and the different impulses. Moreover, only a small part of the physical events in our environment are capable of acting as stimuli or are recorded by impulses in the nerve fibres. Of the continuous range of electromagnetic waves only a very small band acts on our organs of vision while by far the greater part of this range does not act as a stimulus on our nerves.

1.31. Impulses in a particular sensory nerve fibre may thus be set up by any one of a group of stimuli which physically may be similar or altogether different. But if a given fibre responds to any of these stimuli, the character of the impulse transmitted will always be the same, irrespective of the nature of the stimulus. The effect of the impulse is independent of the nature of the particular kind of stimulus which evokes it, and any characteristic effects which this particular impulse brings about must therefore be due to something connected with that impulse and not to any attributes of the stimulus.

1.32. This is the main contention of the so-called principle of the specific energy of nerves. When it was first enounced by Johannes Müller, it was aimed against the conception that the nervous impulses transmitted some attribute of the stimulus to the brain; and it was intended to emphasize that the sensation produced depended solely on the fibre which carried the impulse and not on what had caused that impulse. The form in which it was stated, however, was not free from ambiguity and soon gave rise to a new misconception.

1.33. The fact that the theory was called the theory of the specific energy of nerves led to its being connected with one particular alternative explanation of the determination of sensory qualities which is no less questionable than the theory which it was intended to displace. On this interpretation it was understood to mean that, if it was not the physical properties of the stimuli which determined the quality of the resulting sensations, it must be some property of the individual impulses proceeding in the different fibres which in some sense ‘corresponds’ to the differences of the sensory qualities.

1.34. Although this is by no means a necessary consequence of the proposition which Johannes Müller had been anxious to establish, it was widely assumed, in fact, that the sensory qualities produced by impulses in different fibres would be different, similar or equal according as the physical properties of the corresponding impulses differed from or resembled each other. This interpretation was to some extent suggested by Müller’s own formulation of the theory in which he asserted more than was necessary to establish his conclusions. In his summary of his theory he stated that ‘the sensation is not the conduction of a quality or state of an external body to the consciousness, but the conduction to the consciousness of a quality or state of our sensory nerves induced by an external cause’;15 and he went on to emphasize that these qualities are different with the different senses.

1.35. The recognition, however, that the difference of the sensory qualities is not due to the communication of a difference in the stimuli does by no means make the conclusion inevitable that it must then be a difference in the properties of the impulses taking place in the different fibres, which accounts for them. To interpret the theory of the specific energy of nerves in this sense is merely to accept at this stage an explanation similar to that rejected at the earlier stage: the specific character of the effect of a particular impulse need be neither due to the attributes of the stimulus which caused it, nor to the attributes of the impulse, but may be determined by the position in the structure of the nervous system of the fibre which carries the impulse.16

1.36. We do not only possess no information which would entitle us to assume that the impulses carried by the different fibres differ qualitatively, but, what is more important, it also seems impossible to conceive of such differences between the physical attributes of the individual impulses that they could be said in any sense to ‘correspond’ to the differences of the sensory qualities. Even if qualitative differences between the impulses were discovered, this would not yet provide an answer to our problem. It would still be necessary to show how these differences in quality determined the different effects which the different impulses exercise upon each other; and while it is conceivable that these latter differences may be connected with differences in their individual physical attributes, this need not be so. The important point here is that no differences of the individual impulses as such would provide an explanation of the differences between their mental equivalents, and that any differences of their causal connexions with each other seem at least as likely to be due to structural connexions as to qualitative affinities. This is important especially because the hope of thereby providing an explanation of the differences in mental qualities appears to have been the prime motive for the persistent and unsuccessful search for ‘specific energies’, and because the same conception seems also largely responsible for the persistence of the belief in a ‘pure core’ of sensation.17

1.37. The evidence which we possess suggests, in fact, that the impulses carried by the different fibres, at least within any one sense modality, are qualitatively identical, so that, if we were to cut two sensory fibres and to re-connect the lower part of each with the upper part of the other, they would still function but exchange the results which an impulse in either would cause. It seems, therefore, that the cause of the specific effects of the impulses in different fibres must be sought, not in the attributes of the individual impulses, but in the position of the fibre in the central organization of the nervous system.

4. DIFFERENCES IN QUALITY ARE DIFFERENCES IN THE EFFECTS

1.38. That the similarities and differences between the experienced sensory qualities do not correspond strictly to the differences and similarities between the physical attributes of the stimuli has become most familiar in connexion with the perception of configurations or gestalts. We all readily recognize as the same tune two different series of tones, or as the same shape or figure structures of different size and colour. In all these instances groups of stimuli which individually may be altogether different do yet as groups evoke the same sensory quality or are classified by our senses as the same gestalt.

1.39. But, though the fact that physically different stimuli produce similar sensory qualities is perhaps most conspicuous in connexion with the perception of ‘wholes’, it is no less present or less important where more simple or ‘elementary’ sensations are concerned. The fact that physically similar stimuli which act on different individual receptors and therefore set up impulses in different fibres evoke the same sensory quality raises a real problem. And the question why different physical stimuli for which different receptor organs are sensitive, and even physically similar stimuli acting on different kinds of receptor organs, should produce different sensations raises a problem of the same character.

1.40. While as a rule the same kind of physical stimuli acting on different receptor organs produce the same sensory quality, this is generally true only if they act on receptors of the same kind and even then not in all instances. The same vibration which, if perceived through the ear, will be experienced as a sound, may be experienced as a vibration by the sense of touch. In other instances ‘the same external agent in one case produces light, in another warmth.’18 The same temperature may be experienced as hot, cold, or pain according as it affects different end organs.19 The same chemical stimulus may produce different sensory qualities according as it affects the mucous membranes of the eye or of the mouth.20 And an electrical stimulation seems to be capable of evoking an even greater variety of different sensations. Moreover, even the same stimulus affecting the same receptors may produce different sensations according as different other stimuli operate at the same time on other parts of the nervous system.

1.41. The same sensory quality, on the other hand, may be evoked by different physical stimuli. This happens not only where a particular receptor organ is excited by several different stimuli. In such a case any one of the different stimuli will, of course, evoke the same impulse. But impulses or groups of impulses set up in different fibres by different stimuli also often produce the same sensory quality. The classical instance is the case of colour vision and particularly the sensation of ‘white’ which can be produced by an infinite variety of different mixtures of light rays. But this same fact that physically different stimuli acting on different kinds of receptors produce the same sensory qualities seem to be of very frequent occurrence.

1.42. There exists, therefore, no one-to-one correspondence between the kinds (or the physical properties) of the different physical stimuli and the dimensions in which they can vary, on the one hand, and the different kinds of sensory qualities which they produce and their various dimensions, on the other. The manner in which the different physical stimuli can vary and the different physical dimensions in which they can be arranged have no exact counterpart in the manner in which the sensory qualities caused by them will differ from each other, or in the dimensions in which these sensory qualities can be arranged. This is the central fact to which we have referred when we insisted that the two orders, the physical order of the stimuli and the phenomenal or mental order of the sensory qualities, are different.

1.43. It has long been believed that, e.g., in the field of vision the three dimensions of the stimulus, wave-length, homogeneity and intensity correspond to the three phenomenal dimensions of visual experience, hue, saturation and brightness, and that similarly in the field of hearing frequency and intensity as physical dimensions of the stimulus correspond to pitch and loudness respectively as the phenomenal dimensions of sensation. Recent work, however, has amply shown that within any given modality a change in one dimension of the stimulus may affect almost any dimension of the sensation. Hue depends not only on wave-length but also on intensity; pitch not only on frequency but also on intensity.21

1.44. The orders or dimensions of the stimuli and of the sensations, moreover, not only show no one-to-one correspondence; they also differ in their general character. Any one of the physical dimensions of light and particularly wave-length which is mainly (though not exclusively) the cause of variation in colour, varies on a linear scale, while phenomenal colours can be arranged in a continuous circle in which the order of the wavelength is preserved, but the gap between the two extremes of the spectrum, yellowish red and violet, is closed by pure (or ‘unique’) red and purple which correspond to no distinct wave-length but can be produced only by various mixtures of different wave-lengths. Moreover, continuous variations of the stimuli often produce discontinuous variations in the sensory qualities22, while in at least one case a continuous variation in the sensory qualities, namely from cold to hot, is brought about by what we must regard as a discontinuous variation of the stimuli, since the objectively continuous variation of temperature acts on the organism through different receptor organs.

1.45. It may be generally said that the organization of the sensory order, as represented by the various geometrical figures (such as the colour octohedron, Titchener’s touch pyramid, Henning’s smell prism and taste tetrahedron) by which psychologists have described the dimensions in which the sensory qualities vary, are by no means identical with the order of the corresponding physical stimuli and often differ very substantially from them. The fact that the two orders resemble each other in some degree must not obscure the fact that they are distinct and different orders.

1.46. When we speak of the physical order we mean by similarity of two events that they will produce the same effects in certain circumstances but not in others. Different physical events can evidently be similar to each other both in different degrees and also in different respects: two events may each be similar to a third but not be similar to each other. In other words, similarity is a non-transitive relation.

1.47. The same is true with regard to mental events. Two sensory qualities will be equal if their effects on other mental events or on behaviour will be the same in all respects. They may be similar in varying degrees and in different respects according as they will evoke the same other mental events or the same behaviour in certain circumstances but not in others.

1.48. It will now be clearer what we mean when we speak of the two orders of events, the physical23 and the phenomenal or mental order. Some events will occupy definite positions in both orders, but the relations between several such events in each of the two orders may be different. Some events in the physical order, such as electrical currents which we can only infer, will have no corresponding events in the phenomenal order; and some events in the phenomenal order, such as images or illusions which are not produced by external stimuli, will have no counterpart in the physical order. While there will thus be some degree of correspondence between the individual events which occur in the two orders, it will be but a very imperfect correspondence.

1.49. What we call ‘mind’ is thus a particular order of a set of events taking place in some organism and in some manner related to but not identical with, the physical order of events in the environment.24 The problem which the existence of mental phenomena raises is therefore how in a part of the physical order (namely an organism) a sub-system can be formed which in some sense (yet to be more fully defined) may be said to reflect some features of the physical order as a whole, and which thereby enables the organism which contains such a partial reproduction of the environmental order to behave appropriately towards its surroundings. The problem arises as much from the fact that the order of this sub-system is in some respects similar to, as from the fact that it is in other respects different from the corresponding more comprehensive physical order. The meaning of the conception of an ‘order’ will be further explained in the next chapter (2.282.30).

1.50. In recent physiological psychology these problems have received attention, mainly owing to the work of H. Klüver, under the headings of equivalence of stimuli and of sensory generalization. Klüver’s original statement of the problem is probably still the clearest exposition of it to be found in the literature.25 Merely another aspect of the same problem is the phenomenon of transfer of acquired responses from a given stimulus to others, which is of course the process through which phenomenal similarity manifests itself in behaviour. Yet, though the central character of this problem is now fairly generally recognized, it is usually mentioned merely to point out that it is ‘one of the most perplexing problems to be faced’26 or ‘the recognized stumbling block to all simple mechanical hypotheses of habit formation.’27

1.51. Equivalence, generalization and transfer are all instances of identity of the effects of different stimuli, while discrimination means a difference in the effect of individual stimuli or groups of stimuli. The qualitative order of the sensations which manifests itself in these phenomena is thus a difference in the order in which the stimuli in various combinations produce different effects; and sensory qualities can be regarded as groups or classes of events which, with respect to the responses of the organism, are identical, similar or different in their effects. The order of sensory qualities thus is identical with the totality of the differences of the effects which the different nervous impulses will produce in different circumstances. If we can explain the process which determines the differential responses of the organism to the various physical stimuli, we have at the same time also explained the qualitative order which is the peculiar characteristic of mental phenomena.

1.52. The significance of this statement, which in its bare form may sound more ‘behaviouristic’ than it is intended, will become clearer when we examine the kinds of different ‘effects’ which have to be considered in this connexion (2.232.26). At this point it need only be pointed out that by the term ‘effects’ we do not mean only, or even mainly, overt behaviour or peripheral responses, but shall include all the central nervous processes caused by the initial impulses, even though we may be able only indirectly to infer their existence.

1.53. Our problem is then to show how it is possible to build from the known elements of the nervous processes a structure of intermediate links between the physical stimuli and the overt responses which can account for the fact that the responses to different stimuli differ from each other in precisely that fashion in which we know the responses to the experienced sensory qualities to differ from each other. We must show that from the known physiological elements a structure can be formed which can differentiate between different impulses passing through it in exactly the same manner in which our sensory experience differentiates between the different stimuli.

1.54. Our problem must therefore be stated in terms of the relationships (of equality, similarity, difference, etc.) existing between the sensory qualities. It can be answered only by showing that a strictly equivalent system of relationships can exist between physiological events so that the effects of any event or any group of events in that system will produce a set of effects strictly corresponding to the effects the corresponding sensory qualities will produce. (The reader should observe already at this stage that this does not imply that any given physiological event will always produce the same effects irrespective of the other physiological events occurring at the same time. On this and on the general danger of a too narrow interpretation of the conception of a one-to-one correspondence between the sensory and the neural order see below 2.102.13.)

1.55. This contention implies that if we can explain how all the different sensory qualities differ from each other in the effects which they will produce whenever they occur, we have explained all there is to explain; or that the whole order of sensory qualities can be exhaustively described in terms of (or ‘consists of nothing but’) all the relationships existing between them.28 There is no problem of sensory qualities beyond the problem of how the different qualities differ from each other—and these differences can only consist of differences in the effects which they exercise in evoking other qualities, or in determining behaviour.

5. THE UNITARY CHARACTER OF THE SENSORY ORDER

1.56. The conclusion to which we have been led means that the order of sensory qualities no less than the order of physical events is a relational order—even though to us, whose mind is the totality of the relations constituting that order, it may not appear as such. The difference between the physical order of events and the phenomenal order in which we perceive the same events is thus not that only the former is purely relational, but that the relations existing between corresponding events and groups of events in the two orders will be different.

1.57. The order of the sensory qualities is difficult to describe, not only because we are not explicitly aware of the relations between the different qualities but merely manifest these relations in the discriminations which we perform,29 and because the number and complexity of these relations is probably greater than anything which we could ever explicitly state or exhaustively describe, but also because, as we shall see, it is not a stable but a variable order. Yet we must attempt here to describe at least certain general characteristics of that order, because our problem is whether we can account for at least the kind of properties which it possesses, even if we cannot explain its detailed arrangement.

1.58. One main point about this order is that, in spite of its division into the different modalities, it is still a unitary order, in the sense that any two events belonging to it may in certain definite ways resemble each other or differ from each other. Any colour and any smell, any tone and any temperature, or any tactual sensation such as smoothness or wetness and any experience of shape or rhythm may yet have something in common, or be at least in some sense akin to or in contrast with one another. Experiments have shown that these experienced similarities extend much further than we are usually aware of and that, e.g., even a person who at first thinks such an attempt nonsense, has no difficulty, once he can bring himself to try, to find a tone whose brightness is the same as that of the smell of lilac.30

1.59. Some qualities, especially those which, like colours or tones, are connected into qualitative continua and which, since Helmholtz, we describe as forming distinct modalities, probably always seem to belong more closely together than others such as, e.g., the sensations of pressure, pain, and temperature, which used to be regarded as belonging to the one sense of touch but do not form one modality in the sense just defined. But when we try to describe the differences between different qualities belonging to the same modality, such as different colours, we find that in order to do so we usually resort to expressions borrowed from other modalities. One colour may be warmer or heavier or louder than another, one tone brighter or rougher or thicker than another. This indicates that, though in some respects one particular colour or one particular tone may be most closely related to other colours or other tones respectively, yet in other respects they may be closer to qualities belonging to different modalities.

1.60. Although within any given modality qualities vary continuously31 they need not vary in a constant direction or dimension. While it is true of tones that if one tone is higher than a second, and a third higher than the first, the third will also be higher than the second, we cannot similarly say that, because orange is yellower than red and green bluer than orange, green is therefore either more yellow or more blue than red. While, with regard to pitch, tones can be arranged in one linear scale, colours do not, in this sense, vary in a single direction.

1.61. It makes sense, on the other hand, to say that two different colours differ in the same manner in which two different temperatures or weights do, or that two tones differ similarly as do two sensations of colour or touch. This means that qualities of different modalities may vary along similar or parallel directions or dimensions, or that the same kind of differences can occur in different modalities. It is, e.g., part of the difference between blue and red that blue is associated with coolness and red with warmth. There exist apparently certain intermodal or intersensory attributes, and with regard to some of the terms which we use for them, such as strong or weak, mild or mellow, tingling or sharp, we are often not immediately aware to which sense modality they originally belong.32

1.62. In our highly developed conscious picture of the sensory order these intersensory and intermodal relations are not very prominent and with the development of conceptual thought and particularly, as a result of the great influence which sensualism has had on it, in scientific thought, they are more and more driven back until they are almost completely disregarded.33 We may become aware of their existence only when we attempt to describe a particular sensory quality and in doing so find ourselves driven to describe a colour as soft or sweet, a tone as thin or dark, a taste as hot or sharp, or a smell as dry and sweet. There can be little doubt that these seemingly metaphorical expressions refer to truly intersensory attributes; and experimental tests have at least in some instances shown that different people tend to equate the same pairs or groups of different qualities.34

1.63. These facts may also be described by saying that relations between different qualities may in turn also possess distinct qualities, and that the relations between different pairs or groups of qualities belonging to different modalities may possess the same qualities. These qualities attaching to the relations between different qualities may in turn be similar to individual sensory qualities. The successive musical intervals from the second to the octave, e.g., have been described as ‘gritty’, ‘mellow’, ‘coarse’, ‘hollow’, ‘luscious’, ‘astringent’, and ‘smooth’ respectively.35

1.64. These intermodal relations may occasionally be so strong that different sensations belonging to one modality may regularly be accompanied by the experience of qualities belonging to another modality, as in the case of colour-hearing and other instances of synaesthesia. There is some evidence that these synaesthetic modes of perception are particularly strong in relatively early stages of mental development, and that our habit of thinking of particular colours as primarily belonging to the range of colours, or of a tone primarily as being one of a range of tones, is the produce of a comparatively late and abstract attitude.36

1.65. More familiar than the facts of synaesthesia is the fact that most sensory qualities are closely associated with certain affective tones and that there exists thus a close connexion between the order of sensory qualities and that of affective qualities. The emotional values attaching to various sensory qualities are well known, and there are indeed few sensory qualities which we do not regard at least as either pleasant or unpleasant, or as simply good or bad. The general relation between sensations and emotions or drives will, however, have to be considered later and cannot be further examined at this point.

1.66. The relations or connexions between different sensory (and affective) qualities find expression in the expectations which their occurrence arouses. A red colour does not merely evoke the image of warmth but we shall be rather surprised if a red objects turns out to be very cold; and a certain smell will not only conjure up certain tastes but we shall be shocked if a deliciously smelling fruit turns out to have a vile taste. In this way certain groups of qualities tend to ‘belong’ together, and particular qualities come to ‘mean’ to us certain other qualities.

1.67. Whether the facts briefly summarized in this section do or do not justify the assertion of a ‘Unity of the Senses’ in such a manner that ‘all senses are alike in respect to their attributive dimensions’,37 they probably entitle us to say that, directly or indirectly, all mental qualities are so related to each other that any attempt to give an exhaustive description of any one of them would make it necessary to describe the relations existing between all.

6. THE ORDER OF SENSORY QUALITIES NOT CONFINED TO CONSCIOUS EXPERIENCE

1.68. We have so far assumed that the reader is familiar with the system of sensory qualities from his own conscious experience of these qualities. This, however, is not to be understood to mean that this particular classification of events appears only in our subjective experience. Of course we know this system of qualities from this source. But just as experience tells us that in their relations to each other things do not always resemble each other or differ from each other in the same manner as they seem to be alike or different to us, so we also learn that what appears alike or different to us usually also appears alike or different to other men. Beyond this, it seems clear that not only other men in their conscious action, but both we and others in unconscious action, and also animals, treat as alike or different not what is so in the physical sense, but more or less what in our own conscious experience appears to us to be so. In other words, the order of sensory qualities, once it is known, can be recognized as present in actions which are not directed by consciousness or by a human mind.

1.69. It would, of course, not be possible to discuss the phenomenal world with other people if they did not perceive this world in terms of the same, or at least of a very similar, order of qualities as we do. This means that the conscious mind of other people classifies stimuli in a manner similar to that in which our own mind does so, and that the different sensory qualities are for them related to each other in a manner which is similar to that which we know. In other words, although the system of sensory qualities is ‘subjective’ in the sense of belonging to the perceiving subject as distinguished from ‘objective’ (belonging to the perceived objects)—a distinction which is the same as that between the phenomenal and the physical order—it is yet inter-personal and not (or at least not entirely) peculiar to the individual.

1.70. Nor is the classification of stimuli in terms of sensory qualities confined to conscious experience. We know that both we and other people classify stimuli in our unconscious responses (or in responses to stimuli of which we do not become conscious) according to roughly the same principles as we do in our conscious action.38 The order of sensory qualities exists therefore also outside the realm of consciousness. If, as we shall suggest, we identify with the realm of mental phenomena the range of events within which a classification in terms of sensory (and similar mental) qualities occurs, this realm extends far beyond the sphere of conscious events which merely constitute a special group within the more comprehensive class of mental events.

1.71. It is possible, finally, to ascertain by various experimental methods that not only other men but also most higher animals classify stimuli according to an order which is similar to that of our own sensory experiences. It has even been shown that some animals, e.g., chicks in the famous Révész experiment,39 are subject to the same optical illusions as men. We must therefore conclude that the general principles according to which the neural system of the higher animals classifies stimuli are, at least in their general outline, similar to those on which our own mind operates.

1.72. While it has been inevitable that in introducing our problem we started from the conscious experience of sensory qualities, this proves now to be only one particular aspect of a wider problem. In the further discussion we shall treat conscious experience as merely a special instance of a more general phenomenon, and speak of mental phenomena whenever we deal with any events which are ordered on principles analogous to those revealed by conscious experience. All further consideration of the peculiar additional attributes which a mental event in this sense must possess in order to be described as ‘conscious’ will be postponed to a later stage (Chapter VI).

1.73. It has undoubtedly been unfortunate for the development of psychology that the distinguishing attribute of its object was so long considered to be the ‘conscious’ character of experience, and that no definition of mental events was available which was independent of this conscious character.40 The sphere of mental events evidently transcends the sphere of conscious events and there is no justification for the attitude frequently met that either identifies the two or even maintains that to speak of unconscious mental events is a contradiction in terms.41

1.74. But although we can agree with the Behaviourists in deploring the exclusive concentration of the older psychology on conscious events, they themselves, in their endeavour to get rid of consciousness have gone to the opposite extreme and with the problem of consciousness have tried to eliminate the problem of the existence of the qualitative order which is peculiar to mental phenomena. This problem, as we shall see, cannot be disregarded even if we want merely to account for observed behaviour.

7. THE DENIAL OR DISREGARD OF OUR PROBLEM BY BEHAVIOURISM

1.75. It will help to bring out more clearly the precise meaning of our problem if we contrast our approach with that of two other points of view which require either less or more of any explanation of sensory perception than our statement of the problem demands. This and the next section will accordingly be devoted to an examination, firstly, of the views of a school of thought which either explicitly denied the existence of our problem, or at least proceeded as if it did not exist; and, secondly, to the consideration of an opposite point of view which would probably maintain that even if a complete answer to our problem were achieved, there would still remain unsolved a significant problem concerning the ‘absolute’ or ‘intrinsic’ nature of sensory qualities.

1.76. The point of view which denies, at least by implication, that ours is a genuine problem is (or was?) represented mainly by the classical behaviourists42 and by similar schools aiming at a strictly ‘objective’ psychology. These schools maintained that psychology can entirely dispense with any knowledge of the subjectively experienced mental qualities, and that it ought to confine itself to the study of bodily responses to physical stimuli.

1.77. All the schools of psychology which thus claim to confine themselves to observed physical facts, are, however, in fact, always and inevitably inconsistent in their procedure: they never really avoid using knowledge which according to their professed principles they have no right to use. They almost invariably describe the external stimuli which elicit behaviour not in terms of their physical properties but in terms of their sensory attributes. They naïvely accept as a fact not requiring explanation that different minds treat as equal, similar, or different, groups of stimuli, which physically are not such but merely appear so to our senses.

1.78. The adherents of these schools, in other words, treat as something not requiring explanation the fact that stimuli, which to their senses appear similar, will also appear so to others; and they do this in spite of our knowledge that physically these stimuli may be very different events and in fact may have nothing in common except that very circumstance that whenever they act on us or other people they will evoke the same sensations (and/or responses). They disregard, in other words, the very phenomenon which raises the problem of the existence of a peculiar mental order.

1.79. It might therefore be said that behaviourism, from its own point of view, was not radical and consistent enough, since it took for its starting point a picture of the external world which was derived from our naïve sense experience, instead of taking, as it ought to have done, one obtained from the physical sciences which describe the objective properties of this world. If the behaviourists had been consistent in their desire to take no notice of the qualitative order of their own sense experience, they ought to have started by studying the effects on the organism of physical events of a certain kind, e.g., of light waves of a certain frequency, and then have proceeded to establish experimentally to which of these different physical stimuli the individual responded in the same and to which he responded in a different manner. Before going any further they ought, in other words, to have built up experimentally that classification of the different stimuli which our senses effect.43

1.80. Behaviourists, however, did not seriously try doing anything of the kind. They uncritically accepted the fact that things which are physically different appear alike to our senses, and that things which are physically the same sometimes appear different, or that different things may appear to differ from each other in a manner which is in no way commensurable with the physical differences which objectively exist between them; and they appeared to see no problem in the fact that other organisms classify stimuli in the same manner as we do ourselves, or in a manner different from it.

1.81. This curious blindness to an important problem does not always show itself as blatantly as in the instance reported by W. Köhler in which a behaviourist insisted on referring to a ‘female’ as ‘a stimulus’ to a male bird.44 The error in this instance does not lie merely, as Köhler suggests, in the fact that it involves ‘closing one’s eye to the problem of gestalt and organization.’ It appears already in the disregard of the fact that physically different stimuli affecting different receptors produce the same or similar sensory qualities and therefore are treated as being the same, and in pretending at the same time that sensory qualities do not enter at all into their considerations. (The language of the behaviourist in this instance could be justified only if he meant to imply that the female was always recognized through the same physical stimulus, such as a certain smell, or rather by the stimulation of certain organs of olfaction by definite chemical substances.)

1.82. It would involve the same disregard of the central problem if, e.g., two red spots reflected on different parts of the retina, or the same temperature affecting different parts of the body, were treated as representing the same stimulus. In treating as the same kind of event all events which appear to us to possess the same sensory qualities, behaviourism tacitly assumes the existence of the whole order of such qualities which at the same time it pretends to ignore.

1.83. This acceptance as data of the sensory qualities as they are known to most men from their subjective experience is indeed inevitable in the study of any complex behaviour. But it is only because, while thus accepting them, the behaviourists at the same time deceived themselves about the true character of their procedure, that they avoided the main problem which psychology has to face. If they had been more radical and more consistent in their efforts to link up psychology with the world of physical science, they would have discovered45 that their attempt to explain behaviour without reference to subjective sensory qualities could not be consistently carried through unless it was first shown what determined that system of sensory qualities.

1.84. Like many of the traditional schools of psychology, behaviourism thus treated the problem of mind as if it were a problem of the responses of the individual to an independently or objectively given phenomenal world; while in fact it is the existence of a phenomenal world which is different from the physical world which constitutes the main problem. Behaviourism merely appeared to avoid the problem of mind by confining itself to the study of man’s behaviour in the phenomenal world and by thus treating the main manifestation of mind as a datum rather than as something requiring explanation.

1.85. Although no behaviourist ever consistently adhered to what are the professed principles of his school, and although, if he had, he would never, in the present state of knowledge, have got on to the phenomena in which he was interested, it will be instructive briefly to consider what a consistently ‘objectivist’ study of behaviour would have to be like. It will then be seen that even if the behaviourists had succeeded in carrying out their programme, there would still remain a problem of mind requiring an answer.

1.86. In the first instance, much knowledge that we undoubtedly possess but which is not derived from experimental evidence—such as the knowledge that we are likely to respond in the same manner to different physical stimuli which produce the same sensation—would have to be strictly excluded from such a study of human behaviour. The first task of such a consistently objectivist approach would therefore have to be to ascertain experimentally what to us is the starting point of all knowledge, namely the phenomenal order in which the different stimuli appear in our mind.

1.87. It is at least not inconceivable, although not likely, that by proceeding thus we might in the course of time succeed in reconstructing approximately that grouping of the stimuli which our senses perform. We might then be able to list all the different physical stimuli which, acting on particular receptors and under particular conditions, produce the same sensations (or have always the same influence on the response), and also to reconstruct all the different conditions under which (and all the different respects with regard to which) the several stimuli produce different effects. In other words we might, starting from the physical order of events, experimentally reconstruct the phenomenal order in which these events are reproduced by our senses.46

1.88. This would be merely the first task which a psychology would have to undertake which took the basic idea of behaviourism literally. Only after completing this task could it at least undertake to link directly observable behaviour and physical stimuli. And in order to be quite consistent it would have to define not only the stimuli but also behaviour in strictly physical terms. We need not inquire at this stage whether it is conceivable that this task should ever be fully completed. (We shall later give reasons why we think that this is impossible.) At this point we are concerned with the question whether, even if this task were achieved, there would still remain a problem of the kind with which we are here concerned.

1.89. A solution of that problem would show us what the apparatus of perception does in response to particular stimuli, but not how it does it. Even if we had established a correspondence between all the observed combinations of stimuli and the resulting sensations, we should still be ignorant of the mechanism by which the one kind of order is translated into the other. Our knowledge would be purely descriptive in the sense that it would be confined to a knowledge of the correspondence between observed stimuli and observed responses. We should not possess a theory from which we could derive new conclusions which could be empirically tested.

8. THE ‘ABSOLUTE’ QUALITIES OF SENSATIONS A PHANTOM-PROBLEM

1.90. A different type of objection to our manner of stating the problem must be expected from a school of thought which, though not formally organized, is fairly widespread and which in some respects might be regarded as the extreme opposite of behaviourism. It would probably be contended by representatives of this point of view that, even if we succeeded in accounting for all the differences between the effects of the different stimuli or impulses, there would still remain an unexplained factor, the ‘absolute’ or ‘intrinsic’ qualities of the sensations which are not exhausted by all the differences in their effects but which must be experienced to be known.

1.91. This conception of the absolute character of sensory qualities derives probably from John Locke’s conception of ‘simple’ ideas. It has found an explicit defender in no less a student than William James.47 It is a contention which raises what to us seems clearly a phantom-problem which cannot even be clearly stated and with regard to which it is impossible to say what kind of statement would provide an answer. It is nevertheless important, not only because of the pervasive influence of this conception, but also because it is probably one of the main roots of the belief in a peculiar mental substance.

1.92. The first point to note is that it is clearly possible that a sense discrimination of which some other person is capable can raise a problem for us though we ourselves may not be capable of it. The problem of colour vision, e.g., can clearly become a problem to the totally colour-blind person as much as it can to us. What we shall have to show is that there are no questions which we can intelligibly ask about sensory qualities which could not also conceivably become a problem to a person who has not himself experienced the particular qualities but knows of them only from the descriptions given to him by others. In other words, that nothing can become a problem about sensory qualities which cannot in principle also be described in words; and such a description in words will always have to be a description in terms of the relation of the quality in question to other sensory qualities.

1.93. Most people will agree that the question of whether the sensory qualities which one person experiences are exactly the same as those which another person experiences is, in the absolute sense in which it is sometimes asked, an unanswerable and strictly meaningless question. All we can ever discuss is whether for different persons different sensory qualities differ in the same way. To establish whether a person is colour-blind we have to find out, not how ‘red’ looks to him in any absolute sense, but whether and how it differs from various other shades of ‘red’ and from ‘green.’ In all such instances we can find out and know only whether, compared with other people, a person discriminates between given stimuli in the same or in a different manner.

1.94. In other words, all that can be communicated are the differences between sensory qualities, and only what can be communicated can be discussed. Such communication does not imply that the qualities perceived by different people are similar in any absolute sense. The problem which is raised, for instance, by the much greater capacity for pitch discrimination possessed by the experienced musician but not by ordinary persons is not fundamentally different from the problem created by the distinctions between the qualities which most of us experience.

1.95. It is instructive briefly to consider how we should proceed if we were to try to give a congenitally blind person an idea of sight and colour. We should probably base our account in the first instance on the fact that the blind is familiar with three-dimensional space, with shape and movement, and attempt to explain to him that, as he can feel radiant heat or sound emitted by a distant source, so the eye enables us to perceive other qualities at a distance. We should then try to explain that these qualities with which he is unfamiliar will vary not only along a single dimension, as temperature does from cold to hot, but that it can also vary like tones from bright to dark, from loud to soft, from sharp to blunt and from pleasant to unpleasant. We shall point out to him that in groups these qualities can form harmonies or may clash as tones do, and so on.

1.96. How far we could get in thus teaching a congenitally blind the relative values of the different colours has never been systematically tested, largely because the required description of the order of those sense qualities in terms of their common dimensions (1.621.67) has not been systematically developed and because we therefore lack the necessary words. That blind persons can at least learn to use the names of colours so that a person who does not know that they are blind may remain unaware of it in hearing their descriptions is shown by the writings of Miss Helen Keller and others. Today, with our greater familiarity of the phenomenon of synaesthesia, it also no longer seems so absurd, as it seemed to John Locke, that the ‘studious blind man’ who thought that he had discovered what scarlet looked like, described it as ‘like the sound of a trumpet.’48

1.97. An illustration given in a recent book may be quoted at length, as its concluding passage raises our problem in a particularly clear manner:

‘The approach of a scientist to the phenomena which he observes may be realized perhaps by means of an analogy. Suppose you enter a room and see a man playing a violin. You say at once that this is a musical instrument and is producing sound. But suppose that the observer were absolutely deaf from birth, had no idea of hearing, and had never been told anything of sound or musical instruments, his whole knowledge of the world having been acquired through senses other than hearing. This deaf observer entering the room where a violinist was playing would be entirely unable to account for the phenomenon. He would see the movements of the player, the operation of the bow on the strings, the peculiarly shaped instrument, but the whole thing would appear to him irrational. But if he were a scientist interested in phenomena and their classification, he would presently find that the movements of the bow on the violin produced vibrations, and these vibrations could be detected by means of physical instruments and their wave form could be observed. After some time, it might occur to him that the vibrations of the strings and violin must be communicated to the air and could be observed as changes of pressure. Then he could record the changes of pressure produced in the air in the playing of a piece of music, and by analysing the record could observe that the same groups of pressure changes were repeated periodically. Eventually he would attain to a knowledge of the whole phenomenon of music—the form of musical composition and the nature of different musical forms—but none of this would give him any approach to the absolute truth in that he would still be unaware of the existence of sound as a sense and of the part that music could play in the mental life of those who could hear.’49

1.98. Except for the last sentence this passage provides an excellent illustration of the distinction we have drawn between the physical and the phenomenal order of events. The last sentence, however, raises two difficulties (apart from the fact that the author speaks of the ‘phenomenon’ of music where he refers to what we would describe as its physical equivalent). In the first instance the impression which this sentence conveys, that a ‘knowledge of the whole phenomenon of music’ can be attained without at the same time attaining some knowledge not only of the physical but also of the sensory attributes of these events is somewhat misleading. A reconstruction of the theory of music in the manner suggested would involve a study not only of the ‘objective’ attributes of sound but also a study of the manner in which the people producing the music deal with it. It would, e.g., have to include the discovery that for the musicians the continuum of sound waves of different frequencies was divided into discrete steps, so that all the waves belonging to certain narrow intervals were treated as alike or indistinguishable, while wave-lengths of intermediate intervals would not be employed at all; further, that of the distinct musical notes thus determined some were treated as resembling each other and some as being related in other ways, that certain combinations of notes were preferred to others, and that certain successions of notes were in some respects treated as equivalents of other such successions, etc., etc.

1.99. The theory of music thus constructed would therefore not really refer to the relations between physical events or to relations between them defined according to the similarity or difference of their action on other physical events, but to elements defined in terms of their similarity or dissimilarity to the persons who wrote, played, or heard the music. It would be a theory, not about the objective (experimentally tested) relations between the different physical events, but about what these events meant to the persons concerned with music.

1.100. The second problem arising from the concluding sentence of the passage quoted is contained in the suggestion that there is an ‘absolute truth’, an absolute quality of sound as a sensory experience, which must forever remain inaccessible to the deaf from birth. The term ‘absolute’ used in this connexion unquestionably refers to some significant aspects of sensory experience. What we are denying is not that sensory qualities may possess attributes which those who cannot hear cannot learn about, but that whatever incommunicable attributes sensory qualities may possess can ever raise a scientific problem.

1.101. One fact which is probably referred to by the use of the term ‘absolute’ in this connexion is that, however far we may go in describing or explaining differences between sensory qualities, there will always remain some further differences which have not yet been enumerated. This is closely connected with a circumstance which we shall have to consider later, namely that, because of constitutional limitations of our mind, we shall never be able to achieve more than an explanation of the principle on which mind operates, and shall never succeed in fully explaining any particular mental act. But the fact that the differences between the different sensory qualities are too numerous and varied for us ever to be able to state them all, does not mean that any one of these differences should not be capable of becoming a problem to which, at least in principle, we may provide an answer.

1.102. It is merely another aspect of the same problem if it is pointed out that the immediate experience of a group of sensory qualities (say a number of sounds and colours) will always convey more to us (will involve a large number of implied distinctions among themselves and from other possible experiences) than any possible description can convey. In other words: the congenitally blind or deaf can never learn all that which the seeing or hearing person owes to the direct experience of the sensory qualities in question, because no description can exhaust all the distinctions which are experienced. This, however, does not mean that there is more than differences from other qualities, and still less that any such ‘absolute’ character of the qualities can raise a genuine problem.

1.103. It seems thus impossible that any question about the nature or character of particular sensory qualities should ever arise which is not a question about the differences from (or the relations to) other sensory qualities; and the extent to which the effects of its occurrence differ from the effects of the occurrence of any other qualities determines the whole of its character.

1.104. To ask beyond this for the explanation of some absolute attribute of sensory qualities seems to be to ask for something which by definition cannot manifest itself in any differences in the consequences which will follow because this rather than any other quality has occurred. Such a factor, however, could by definition not be of relevance to any scientific problem. The ‘absolute’ quality seems to be unexplainable because there is nothing to explain, because absolute, if it has any meaning at all, can only mean that the attribute which is so described has no scientific significance.

1.105. The contention that all the attributes of sensory qualities (and of other mental qualities) are relations to other such qualities, and that the totality of all these relations between mental qualities exhausts all there is to be said about the mental order, corresponds of course, (perhaps we should say follows from) the conception of mind itself as an order of events. And with the recognition that mind itself, and all the attributes of mental events, are a complex of relations, there disappears of course the need for any peculiar kind of things which by themselves have attributes which constitute them a peculiar ‘substance’.

1.106. The abandonment of the phantom-problem of the absolute character of mental qualities, and the recognition of the relative significance of these attributes, is of fundamental importance, because it opens, as we shall see, the way for a general application of a principle which has long been used to explain those attributes of sensory experience which had been recognized to be relative, such as spatial position.

1.107. It also follows from the relative character of all mental qualities that any discussion of these qualities in terms of their relations to each other must necessarily remain within the realm of mental events: it can never provide a bridge which leads from them to physical events. In the next chapter we shall attempt to show how this circle can be broken.