522 Thought direction," and in terms of images (§ 143). The attitude is as symbolic as the word, and the image may be as symbolic as the attitude; all that thought requires is a system of mental symbols. But this very statement sug- gests another reading of the question in discussion. Thought requires symbols; language is a system of symbols; and we have no reason to suppose that, in the history of mind, it supervened upon or took the place of any previous system. Thought and language, in other words, appear to have grown up side by side; each implies the other; and in this sense it is true to say that there is no thought without words; reasoning and language are two aspects of the same phase of mental development. The old conundrum: Why don't the animals talk? Be- cause they have nothing to say — contains a sound psy- chology; if the animals thought, they would talk; since they do not talk, they do not either think.
The use of language as the vehicle of thought has both advan- tages and disadvantages. If we begin with the latter, we note, first, that language has developed under pressure of practical necessity (p. 57), and is therefore, for scientific purposes, incom- plete. The conscious attitudes would have been recognised long ago if we had had a name for them; or, to put the case more ac- curately, now that we do possess a name, we can see that many of the older psychologists had noticed them, but had failed, for lack of fitting words, to make the notice emphatic. Secondly, language tends very strongly to stereotyped forms. The observer in the psychological laboratory must be trained, not only to attend, but also to express himself in words (§ 6); and the training in expres- sion is oftentimes the more difficult part of his education. These two disadvantages may be summed up in the statement that language puts a premium on the stimulus-error; it has been de- veloped for reference, for design ition, rather than for exact descrip- tion (cf. pp. 202 L, 218, and the quotation on p. 511). A third disadvantage lies in its discreteness. Mental processes are continu- ous and interwoven; words are separate and come in single file.
On the other hand, language is extraordinarily flexible. If it was not made for description, at least it can be pressed into the service of scientific record, and so successfully that there is no ex- perience, however transient and subtle, which cannot, with suffi- cient pains, be translated into words. The observers who tell you, despairingly, that they know what it means to 'feel pretty well,' but cannot for all their good will go behind the feeling by analysis, come a few weeks later with pages of verbal description; when once the stereotyped form is broken up, language will follow all the finest ramifications of a mental process; its wealth of distinc- tions seems inexhaustible; and it has within it a principle of growth, so that a new term or a new collocation of terms is at once intelligible. Here, indeed, is a second advantage: language, just because it has developed in response to practical rather than to theoretical needs, is readily understood and easily commanded; it carries the meaning that we wish to convey, and it shapes itself, automatically, under direction of that meaning. Lastly, language is permanent and constant; it fixes, as if in mosaic, the fluidity of mental experience; it frees us from the errors of memory, and thus lays at every stage of science a foundation for further work.
The actual progress of thought and knowledge is sufficient proof that the advantages of language outweigh the disadvantages. Yet the disadvantages are real and serious. We have only to recall terms like ' perception ' and ' association ' to realise the danger of stereotyping: for that matter, the new word ' attitude ' has, as we have seen, stood by its very coining in the way of analytical de- scription. Words, whose only value is to symbolise experience, must never be permitted to take the place of experience; they are indispensable servants, but their mastery is fatal to science.
Regarded for its own sake, as one of the great social in- stitutions (p. 26), language has a twofold psychological in- terest. The origin of language marks an epoch in mental development; and the growth of language embodies the 524 Thought growth of thought. The psychology of language is, how- ever, a subject for itself, which lies beyond the scope of this book.
Since both of these problems — that of the origin of language and that of semantic change, or change of the meaning of words — bear upon certain topics that we have discussed in preceding Sections (§§ 103, 132), we may, nevertheless, devote a brief space to their consideration. The origin of speech is, according to Wundt, bound up with the origin of gesture language. A gesture may express either the feeling side or the ideational side of an emotion. The wince and brace and bitter look of 5 132 are ges- tures of the former sort. Ideational gestures are of two kinds: de- monstrative and representative. Demonstrative gesture points towards, directly indicates, the object that excites emotion: we point our finger at the thing that has frightened us, or shake our fist at the man who has made us angry. Representative gesture depicts the object, whether by a finger-drawing of its outline in the air, or by the reproduction of one of its characteristic features, or by some purely symbolic movement. Thus, a deaf-mute gesture for ' smoke ' is a spiral action of the forefinger from below up- wards; for ' child,' the action of cradling and rocking the right elbow in the left hand; for ' truth,' the movement of the forefinger in a straight line from the mouth. This gesture language has its own syntax, its own laws of semantic change, its own psychological history. But just as rhythm (§ 94) has become predominantly auditory, so has language passed from gesture to speech.
Rhythm, however, seems to retain its kinaesthetic component; and the movement of articulation reminds us of the original nature of speech. The word, heard or seen, which symbolises an idea, is the last term of a long period of development. At the very begin- ning, speech was a gesture; the essential thing about it was not the sound, but the movement. Attempts have been made to read a meaning into the sound that accompanied the movement: there is a theory which traces language to the imitation of natural sounds, and makes it begin with onomatopoetic words like hiss, roar; and there is a theory which traces it to ejaculations and arbitrary, pre- § 143- The Abstract Idea 525 linguistic utterances, and makes it begin with interjections and a sort of childish babble. Neither of these theories will hold water: onomatopoetic words form a very small part of our vocabulary; ejaculatory and exclamatory language is the analogue of affective, and not of ideational gesture, and has had but little development; and the babble of the human infant is not primitive, but corresponds to a stage in the maturing of an inherited speech-mechanism. We must, then, suppose that the sound was, at first, the incidental, meaningless accompaniment of the gesture, the articulatory move- ment; that it derived a meaning from other, concomitant gestures; and that only gradually, under the influence of continued social in- tercourse, did it manifest its superiority to gesture and acquire its independence. We may say, in the large, that the word heard has never had any other than a derivative and symbolic meaning, and that the self-sufficiency of the word-gesture, combined sound and movement is the origin of language.
The course of semantic change should, if interpreted aright, re- veal the laws that have governed the growth of mind; at the same time, its particular phenomena are psychological facts that call for explanation. We may say at once, on this second point, that ex- planation is possible in terms of associative and determining ten- dencies. The general law of development, from the more concrete to the more abstract, may be briefly illustrated by the words that designate perceptive processes. Thus, German ricchen, smell, is ultimately the same with rauchen, smoke; and English smell is connected with smoulder, and with Danish smul, dust. To touch is originally to pull or draw: cf. tug, team, tuck (in the sense of 5 draw together ' of cloth). The Latin sapio, to taste, sapor, taste, are connected with sapa, must, sapo, soap, sebum, tallow, i.e., with names of substances that are readily diluted or liquefied. A further stage of the same process is illustrated by English feel, German fuhlen, which at first meant to touch; cf. Latin palma, English palm of the hand. Not till the eighteenth century were the words reserved for the affective or 'subjective' side of mind; and even now the shift of meaning is by no means complete.
§ 143. The Abstract Idea. — Thought, as we have said, 526 Thought may go on in terms of attitude, of words, or of images. The imaginal complex which is characteristic of thought is known as the abstract or general idea. The name is, psycho- logically, a misnomer; for it is no more correct to speak, in psychology, of an abstract idea, than it would be to speak of an abstract sensation. What is abstract or general is not the idea, the process in consciousness, but the logical meaning of which that process is the vehicle.
It has, however, been maintained that the idea of the ab- stract is itself generalised, in the sense that it is the result- ant of many single memory-ideas. The idea has been compared to what is termed a composite photograph. If we wish to secure a typical face — of a statesman, of a sol- dier, of a consumptive, of a dement — we photograph a number of individual faces upon the same sensitive plate, giving each one a fraction of the normal time of exposure that the plate requires. As a result, we obtain a face in which the resemblances are emphasised and the differences slurred. On this analogy, the general idea, say, of man might be a visual image in which all the points of likeness between men are clear and intensive, while all points of difference are left faint and obscure.
There is no doubt that imager of this composite kind exist. But they are never formed in the way suggested, by a mechanical reinforcement of the like and a mechanical suppression of the unlike elements; they are formed always under the influence of a foregone intention or suggestion. And the range of their formation is necessarily small. If the reader will try to call up an abstract idea of man, as a visual image, he will find — in all probability — that he sees, in the mind's eye, a white man dressed in civilised clothing; if he does not, he will see some similarly individual figure.
§ 143- The Abstract Idea 527 The abstract idea may represent the average height and the average proportions of the human race, but it cannot repre- sent the average colour of skin and hair and eyes. Perhaps the psychologists who use the figure of the composite photograph forget that the photograph is uncoloured.
To illustrate the composite idea, we may quote the following passage from Huxley. " An anatomist who occupies himself intently with the examination of several specimens of some new kind of animal, in course of time acquires so vivid a conception of its form and structure, that the idea may take visible shape and become a sort of waking dream. But the figure which thus pre- sents itself is generic, not specific. It is no copy of any one specimen, but, more or less, a mean of the series." l Huxley, it will be noted, confines himself to form and structure. Moreover, the anatomist is working under the suggestion of a type, of a composite picture that will make a diagram for a monograph or text-book, and this antecedent determination, rather than the mere repetition of specimens, is responsible for the abstract idea.
An abstract idea is any idea whose meaning is abstract. And meaning, as we saw in § 103, appears psychologically as context, as the processes that accrue to the given pro- cess through the given situation, under a particular sug- gestion or instruction. Hence an abstract idea is any idea, however individually pictorial, whose context and deter- mination carry the meaning of abstractness or generality. As a rule, the abstract idea is first presented in verbal form, as what is called a concept, while the context, per- ceptive or imaginal, may or may not be verbal. In minds of the verbal type, the idea remains verbal; in minds of the imaginal type, the word may be supplemented or even replaced by some imaginal process. But this imaginal form of the abstract idea is then secondary, not original.
528 Thought The notion that the abstract idea must itself be abstract harks back to the fundamental error of associationism. If the individual idea is the idea of the individual, the idea that means individual (§ 106), and if the abstract idea is intrinsically the idea that means abstract, then of course the fusion or blending of a number of individual ideas will give us an abstract idea; the individual cat-meanings of our experience, run all together, will issue in a resultant, generalised cat-meaning. But, psychologically, no idea is intrinsically an individual or an abstract idea, just as no idea is intrinsically a memory-idea. An idea becomes abstract by its set- ting; the idea itself may be photographically particular.
In the author's mind, the context that means ' this is abstract ' is usually a visual schema, a closed visual pattern; the particular cases seem literally to fit into this pattern, or into some area of it, without disturbance of the lines. In another mind of similar type the context is a vague image of an overarching dome, which covers the particulars, or to which the particulars are subsumed. We may suppose that the same context, for an observer of the kinaesthetic type revealed in the quotation on p. 517, would be the representation of some large and inclusive gesture; while for a verbally minded observer it might be an echo of the formal definition of the word employed. All this is matter for individual psychology; the point to remember is that any idea, any image, is an abstract idea, if its context is the psychological vehicle of the logical meaning ' abstract.' And in like manner any word what- soever is a concept, if its context gives it the abstract reference.
Most of our abstract ideas come in verbal form; we grow up in an environment of language, and we become familiar with ab- stract words before experience has furnished the data for general- isation. In minds of the verbal type, the abstract idea of honesty or pride is then just the word ' honesty ' or ' pride ' as it appears in internal speech. In minds of the visual type, the verbal ideas are accompanied (or under certain circumstances replaced) by con- ventional pictures: the idea of value, e.g., by the image of a man putting something into a scale, or the idea of pride by the image of a strutting, swelling figure. When the author happens to think § 144- Generalisation and Abstraction 529 of cows, he sees a longish rectangle ending to the left in a sort of exaggerated pout; when he thinks of horses, he sees a double curve and a rampant posture with a suggestion of mane.
While, however, it is true that there is no such thing as an ab- stract idea, or a concept word, in the sense of associationism, it is also true that the dropping out of the context, and the lapsing of the determination from consciousness, leave us with images and words that have the character of abstractness, so to say, incorporated within them. The composite or conventional pictures, once formed, remain abstract in their own right; we do not need the determination or the context, later, to assure us that they are ab- stract. And language, as we have just remarked, comes to us ready made. We learn from the study of semantic change that all the abstract words were originally concrete; but the situations that made them abstract dropped into unconsciousness long ago, and the words themselves now have the stamp of abstractness upon them. So it comes about, not only that the conscious rep- resentation of the idea of honesty or pride is the mere word, ' honesty ' or ' pride,' as it occurs in internal speech, but also that this same internal speech embodies the meaning of abstractness; the verbal image stands both for idea and for context.
§ 144. Generalisation and Abstraction. — We have spoken of the abstract or general idea, as if the adjectives were interchangeable. The processes of abstraction and gener- alisation are, in fact, very closely related. When we ab- stract, we pick out the features of a situation that are relevant to our present determination, and neglect or dis- card its other features. When we generalise, we bring to light likenesses that have been masked by differences: but this statement implies that we neglect the differences, as irrelevant, and pick out the likenesses, as relevant; and that is a special mode of abstraction. We may, perhaps, read a difference into the two experiences by making ab- straction mainly negative, the discarding of the irrelevant.
530 Thought and generalisation mainly positive, the bringing together of the similars which are relevant.
Abstraction, then, is the necessary result of determination. We have seen that all determination is double-faced, inhibitive as well as facilitative (p. 461); and we have seen, further, that the ob- servers in the associative reaction experiment instinctively narrow or specialise their instructions (p. 445). Experiments made with the direct aim of studying the process of abstraction confirm and illustrate these results. Various arrangements of four three-letter nonsense syllables, written in different colours, were exposed in a dark room, by means of lantern projection, for a period of \ sec. The observers were required, in successive series, to devote their attention especially to the following points: the total number of letters visible, the colours and their places in the visual field, the figure formed by the arrangement of the syllables, and the deter- mination of individual letters and of their positions in the field. It was found, as we should expect, that the introspective report was fullest, most nearly correct and most definite when it dealt with that aspect of the visual complex to which attention had been given. On the negative side, it appeared that there were degrees of abstractive difficulty; the letters and their number could be more easily neglected than colour and figure; and as the direc- tion of attention upon the former characters proved to be the more difficult task, we may say that a negative abstraction has the more pronounced effect, the greater the difficulty of the corresponding positive instruction. As regards the twofold nature of determina- tion, it is interesting to note that this negative side of abstraction showed itself either in complete suppression or in indefiniteness of apprehension: a figure might be correctly described, while nothing at all could be said of letters or colours; or letters might be correctly named, while the colours were reported merely as like or different, as dark or light, or as such and such colours in un- known arrangement. When no preliminary directions were given, the observers instinctively set themselves the easier tasks; thus, the letters and their number were less often reported than colour and figure.
§ 144- Generalisation and Abstraction FlG. 65. Group of forms for abstraction of the like. — A. A. Griinbaum, Arch. f.
Experiments upon the elementary process of generalisation, upon the positive abstraction of similars, have been made as fol- lows. Groups of nonsense forms, arranged as in Fig. 65, were shown to the observer, under the conditions just described, for a period of 3 sec. The groups were of varying com- plexity, but always contained one common element; and the instruction was to await the exposure with as even as possible a distribution of at- tention and then, when the figures appeared, to pick out the two that were alike. No less than eight modes of procedure were distinguished. The observer might work through the forms with full knowledge, excluding one by one those that were unlike; this method is laborious, and is employed for the most part only in practice experiments. Or he might travel over the groups, back and forth, until he lit upon a form which appeared familiar; this is the method of simple recognition. Or again he might set out on his journey of exploration, and find himself suddenly ar- rested by some insistent form, some figure that stood out more clearly than its fellows. The insistence itself might be bare and unrelated, a mere subjective accentuation; or it might be accompanied by the conscious attitude of acceptance, — 'that is the figure I want '; or by an attitude of less strong assurance, — ' that may be the right one.' Here are mixed methods, part active search and part passive impression. In other cases, the two like forms stood out in quick succession, as if the one had drawn the other after it; the experience was more markedly passive. In still other cases, there was actual simultaneity of apprehension; the two like forms sprang forth, of their own accord, either im- mediately or after a brief interval, without any active search from the side of the observer. Lastly, in rare instances, passivity reached its maximum: the observer looked at the field, was held at once 532 Thought by some outstanding form, and knew that this was the form re- quired, although he had not at all remarked the presence of its pair.
We cannot here enter into explanatory details; nor, indeed, is the time ripe for a general discussion. It is enough to have pointed out the relationship of generalisation and abstraction, and to have shown that, under experimental conditions, they yield to introspective analysis. But the experimental work is hardly more than begun, and much remains to do before we can connect these simplest experiences with the processes of reflective thought.
§ 145. Comparison and Discrimination. — One of the com- monest tasks assigned to the observers in psychological laboratories is the comparison of two mental processes, the discrimination of quality or intensity of sensation, of the duration of intervals, of the magnitude of spatial forms. In these experiments, stimuli may sometimes be presented together: colours and lines may be shown side by side, a chord may be sounded for resolution into tones. All stimuli, without exception, may be presented in succession, and with different intervals of time between the first and second members of a pair. As the conditions of observa- tion vary, so do the details of the comparative or discrim- inative consciousness.
It is a tradition in psychology that successive comparison requires the intervention of an image. I have a certain experience; if, now, I desire at some later time to compare another experience with it, I must call it up in imaginal terms, and place the idea alongside of the perception. Recognition is explained, in this way, as the comparison of a repeated experience with its own memory-image. Nothing, however, can be more certain than that the pres- ence of an image is unnecessary; comparison may be direct, § 145- Comparison and Discrimination 533 the immediate outcome of a determination; and if it is in- direct, the mediation need not be imaginal.
Moreover, comparison is oftentimes complete — curious as the statement may appear — before the second of the paired stimuli has been presented; we are ready with our answer before the full question has been put. The abso- lute impression made by the first stimulus is, in these cases, the cue to the introspective report. If the first of two tones strikes us as unusually loud, or as ridiculously faint, we are ready to declare the second tone weaker or stronger, before we have heard it.
We may illustrate the first point by reference to the comparison of musical tones. The stimuli, chosen from the middle part of the scale, and differing but very slightly in pitch, were sounded for 1 sec. with intervals of 2 to 60 sec. between the members of a pair; the observers were instructed to record their introspection, with reference always to the second tone, as higher, lower, same or doubtful. Three types of discrimination were found. The first, which constituted the prevailing type for most observers, showed no trace of any imaginal representation of the first tone. In the second, the image appeared; but it was not attended to, and did not serve as basis of comparison. In the third, the image was an essential component of the discriminative consciousness.
We confine ourselves, for the moment, to reports of difference. The imageless comparisons, which formed the great majority, were either direct or indirect. In the former case, the second tone touched off the reaction word, higher or lower, as if automatically. In the latter, comparison might be mediated by explicit imagery: the visual picture of a printed musical scale or of a piano keyboard, the kinaesthetic imagery of striking a piano note a half-tone above or below another, and so on. More often, the mediation was effected by a complex of strain sensations, perhaps with less prominent visual and organic elements, which was described as a tightening or relaxing. The muscles concerned in this complex 534 Thought differed with different observers: the strains might be referred to chest, throat, eyebrows, scalp and the region of the ears: but the meaning was always the same; the tightening meant up, and the relaxing meant down, in the tonal continuum. This symbolic ref- erence is, no doubt, the result of empathic experience, in play- ing or singing or listening to music, reinforced by verbal associa- tion.
If an image was present in consciousness when the second stimulus occurred, it was almost invariably thrust aside, by the direction of attention to this stimulus, and the observer's report was determined by other factors. There were instances, however, — when the conditions were novel, when the second tone failed to touch off a response, when two contradictory impulses were felt, — in which deliberate use was made of the image as a stand- ard of comparison. Recourse was had to this method only when the observer was hesitant and uncertain; and the results were as likely to be wrong as right.
Reports of equality or identity are less frequently based on imageless comparison than reports of difference. The presence of the image is, however, not necessary, and where an image ap- pears, it is not, as a rule, the object of attention. The second stimulus seems, in these cases, to reinforce or flow into the image, and the recognition, while it is still direct, consequently gains in definiteness; the second tone is apprehended, not merely as familiar, but as ' that ' tone, the same tone as the other.