380 Association the present contents of consciousness as one of the determinants of attention (p. 270). Nevertheless, all cases of similarity prove, on examination, to involve contiguity. Caesar suggests Alexander, not — it is true — by way of the class-idea ' great commander,' but simply because some component of the idea of Caesar has previously been together, in a conscious present, with that of Alex- ander. We may, then, if we like, say that all associations reduce to associations by contiguity, — being careful, in our own thought, to translate this law into psychological terms. On the whole, how- ever, it is advisable to drop the traditional laws, and to retain only the formula of the text; there is a risk in pouring the new wine into the old bottle.
§ 108. The Experimental Study of Association. — The ideas of associationism are meanings; and meanings, from our point of view, are conscious contexts or nervous deter- minations of consciousness. Whichever they are, they sadly complicate enquiry into the conditions of association. We have a general law; but we want to know how it comes about, in the particular case, that this and not that idea arises on the recurrence of the other, that Caesar suggests now Alexander and now Napoleon. We want to get to the bare essentials of the association.
Some twenty-five years ago, Ebbinghaus solved this prob- lem by the introduction of nonsense syllables. He made up over 2000 meaningless 'words,' all consisting of a vowel or diphthong between two consonants; the German language uses these combinations far less than the English. Here, then, were pure perceptions, sights and sounds that had no meaning and no associates; here was material so varied and yet so simple, so rich and yet so uniform, that experiments could be made under laboratory conditions, and the results of one experiment could be compared, directly, with the results of another. It is not too much to say that the re- § io8. The Experimental Study of Association 381 course to nonsense syllables, as means to the study of asso- ciation, marks the most considerable advance, in this chapter of psychology, since the time of Aristotle.
It must not be supposed that the nonsense syllables work auto- matically. We are inveterately given to meaning; and the ob- server who sits down to learn a series for the first time shows a terrible ingenuity in reading sense into what by hypothesis is non- FlG. 60. Apparatus for the Serial Exposure of Nonsense Syllables. The syllables are printed upon the periphery of a cardboard disc, which is placed in the box to the right, and are viewed through the radial slit in the lid. The movement of the disc is governed by the metronome; as the pendulum swings, and electrical contact is made between the platinum strips and the mercury pools at the base of the instrument, the disc turns, with a jerk, just so far as to expose a new syllable. At the end of the series, the mechanism may be arrested by opening the key to the left. The two wires at the bottom of the figure are led to a battery. — P. Ransch- burg, Monatsschr. f. Psychiatr. u. Neurol., x., 1901, 321.
sense. A recent writer quotes, as a typical set of English nonsense syllables, the series: leb, rit, mon, yup, kig, des, wer, zam. But rit suggests writ, mon is Scotch for man, yup suggests yap or yelp, kig — if you have a cold — is king, wer is obviously were. And if you know anything of German, leb, des, wer, zam are directly suggestive of meaning. The series is, very certainly, not the kind of series for the beginner, who is constantly hunting about for meaningful connections. Ebbinghaus reports that the syllables dosch pam feur lot were connected by the meaning das Brot 382 Association 382 Association [Fr. pain!] Feuer loscht, the bread puts out fire. Indeed, her^ as in most psychological experiments, the novice is likely to do anything rather than what is required of him; he will search for meanings, stress the position of syllables, mark the rhythm, shift his imagery (§ 114). There are great individual differences; but, in general, it is only after a good deal of practice that the observer be- comes the sheer mechanical associator; and it is therefore neces- sary to make a careful choice of material even within the sphere of the directly meaningless. However, the observer may be trained. And methods have been worked out, for the study of association by means of nonsense syllables, that are as rigorous as those em- ployed for the study of the intensity or quality of sensation.
Nevertheless, we cannot stop short with these nonsense syllables. The results derived from meaningless material, while they are essential to our analysis of the more com- plicated results from words, pictures, etc., must themselves be scrutinised in the light of the ordinary meaningful as- sociations of everyday life; 1 they are fundamental, but they are also artificial; until further test, they hold only for the restricted conditions under which they are obtained. However, this work of comparison, of mutual control, offers no special difficulties; indeed, the rules discovered in the laboratory have already been applied, with success, to cer- tain practical problems.
§ 109. Results: the Conditions of Impression. — Suppose that you read through a list of nonsense syllables, again and again, until you can recite it without error. The read- ing will have established a number of associative connections between the terms of the series. But it has also, plainly, established the terms themselves. These terms have, as stimuli, impressed the nervous system, imprinted themselves on it, stamped it in a determinate way.
lWe return to this part of the subject in § 123 below.
§ 109. Results: The Conditions of Impression 383 The correct recitation depends, then, not only upon as- sociative connections, but also upon impression. A term may be but weakly associated to its preceding term, and yet, if it has made a strong impression, — if, as we may say figuratively, its idea is almost ready to appear of itself, — may be brought to consciousness by the weak associative tendency; and, conversely, a term may be but weakly im- pressed upon the nervous system, and yet may be brought to consciousness by a strong associative tendency. It is impossible, in experiments of the sort under discussion, to separate the two factors in the result; the conditions of impression are also the conditions of association. We may say, however, that impression depends upon the length of the series, the position of the terms within it, the rate of succession of the terms, their grouping into complex units, the number and distribution of the readings, the active participation of the observer, and the mode (total or partial) of repetition.
Attention, we may remind the reader, is presupposed. Impres- sion then depends, first, on the length of the series. While 6 or 7 syllables can be recited correctly after a single reading, a larger number throws the observer into confusion. The first and last terms of the series have the advantage over the others; they may, indeed, be the only terms that can be recited after a single read- ing of a 12-syllable series. The impression is deepest if the syl- lables are first presented at a moderate rate (perhaps 2 in the 1 sec), and if this rate is slowly increased as the readings proceed. It is of assistance to introduce a subjective rhythm. Impression deepens, further, with repetition. The first reading is more im- portant than any single later reading; after that, there is for a while little if any improvement; then the results take a sudden step up; and thenceforward progress is fairly steady until the limit of the experiment is reached. The distribution of the read- 384 Association ings in time is also of great importance: thus, it is better to dis- tribute 24 readings in pairs to 12 days than to take them in fours on 6 days; and it is, again, better to do this than to take them in eights on 3 days.
The same rules hold, with the necessary changes, for meaning- ful material. While 8 or 9 one-syllable words, and 10 or 12 figures, can be recited correctly after a single reading, a larger number brings confusion. We return to this point later (p. 387). The rate of presentation may be much quicker: in the reading of poetry, e.g., 140 to 150 iambic measures in the 1 min. (4 or 5 syllables in the 1 sec). The grouping of the material is given, not only by rhythm, but by the meaning of the successive passages.
There remain the two factors which we have called the active participation of the observer and the mode of repetition. It is found that a recitation is far more effective for impression than a reading. The reason may be, in part, that attention is greater; in part, that the auditory and kinaesthetic stimuli reinforce the visual. It seems to the author, however, that the chief effect of the reci- tation is to equalise the attentions; every term in the series must be brought out sharply and clearly; the observer discovers his weaknesses, and has the opportunity to overcome them. Finally, it is found that connected, meaningful material is most impressive if it is read as a whole, from end to end, in the successive repeti- tions, while nonsense syllables and disconnected meaningful ma- terial (dates of events, words of foreign languages) are best taken discretely, read over and over a little at a time. — The nervous modification which we have here named ' impres- sion ' is, clearly, the first term in the series of nervous changes which condition the process of learning. If we read the list of syllables again and again until we can recite it correctly, we have learned it. Learning, however, is a very complex affair, depending upon im- pression, upon associative tendency, upon the retentiveness of nerve-substance, and upon cortical set. Hence we have avoided the use of the word in the present Section.
§ 1 10. Results: the Conditions of Associative Tendency. — Anything that makes for the impression of two stimuli, §no. Results: Conditions of Associative Tendency 385 during the conscious present, will also serve to establish an associative tendency between them; so that the recur- rence of the one, whether as perception or idea, will be likely to arouse the idea of the other. The impressing of our series of syllables has, accordingly, established certain associative tendencies. The strongest of these, as might be supposed, is that which leads from term to term in the order of presentation; but there is good reason to believe that every term is, in some measure, connected with every other term of the series.
We find in the experimental results cases of association, not only of immediately successive ideas (direct association), but also of ideas separated, within the conscious present, by other ideas (remote association); this latter is relatively weak, but it is still definitely discoverable. To put the matter in terms of the alpha- bet, we find associations, not only of a and b, of b and c, of y and z, but also of a and d, oiv and z. And, what is more important, we find that the associative tendencies may work retroactively; there are associations of z with y, of z with x, and so forth. The series of syllables has therefore been impressed, not as an interrupted series, but as a very complicated meshwork, functionally inter- connected through all its parts.
There is yet a further complication. The series is impressed upon a brain which is already the seat of a vast concourse of as- sociative tendencies; and the incoming stimuli may thus excite to full or partial activity some preexisting tendency whose arousal was neither expected nor intended. The observers not infrequently re- port, e.g., associations of position, of the place of a syllable in the series; the cue may be given, according to the circumstances of the experiment, by a verbal idea of number, by the spatial arrange- ment of the syllables, by inflection of the voice. Under certain conditions, the subarousal of such a tendency determines an as- sociation, while the place-idea does not itself appear in conscious- ness. Suppose that a is associated to b, a to /?, and that both pairs of letters have been associated to the place-ideas first, second.
2C 3^6 Association Then, in the given case, the sight of a may suggest the idea of /?, although the observer has not thought of 'first.'
Mediate Association. — Some psychologists believe that an asso- ciation may be set up, originated, by unconscious (purely physio- logical) intermediaries. I examine a picture, we will say, which the artist has signed; my eyes travel over the signature, but I fail entirely to remark it. At some later time I am examining another picture, signed in the same way by the same artist; again my eyes travel over the signature, but again I fail to remark it. Neverthe- less, the second picture suggests the first; the signature has im- pressed my brain, although it has not aroused a perception; the ideas of the two pictures are connected by this unconscious link. Here is a case of mediate association.
The question has been put to the trial of experiment, and the results are mainly negative; it is, however, so difficult to meet the conditions of a crucial test, that difference of opinion is natural and justifiable. In the author's judgment, association requires attention (p. 379); mediate association, of the kind described, does not occur. The alleged instances may be interpreted as associations whose conscious mediation (odour, organic sensation) has been overlooked, or as remote associations, or as associations due, in the way just described, to the subarousal of associative tendencies already established. A fourth possibility will be dis- cussed later (p. 400).
If, now, a series of nonsense syllables establishes this complicated network of associative tendencies, a stanza of poetry or a paragraph of prose must set up excitations of far greater complexity. And the pattern of the excitations will vary, according as the meaningful material is familiar, and therefore throws into play a total cortical disposition, or is unfamiliar, and therefore starts up only partial and fragmentary associations. We cannot trace these effects in detail; but we have evidence both of the mutual rein- forcement and of the mutual interference of associative tendencies.
§110. Results: Conditions of Associative Tendency 387 We have had an instance of interference in the confusion that fol- lows the first reading of a long series of words or syllables (pp. 383^). So long as we are within the range of attention (§ 80), there is no difficulty. But when we pass beyond it, the law of retroactive inhibition comes into effect. The associative tendencies need a certain time to establish themselves, to settle down; and if this time is not allowed, but stimulus treads on the heels of stimulus, there is no impression of a pattern, and no associations are formed. A recently acquired association may even be abolished — as most of us know to our cost — by intensive occupation with an entirely different topic. You have just got to your point, to the phrasing, the insight, the argument, that will clinch things; you are dis- tracted by some irrelevant business; and when you come back to your work, the point has gone. So nicely balanced and so easily disturbed are the associative tendencies, that you may never re- cover it; try as you will to reinstate the conditions, you cannot get the exact pattern back again. The compensation is that the tendencies, left to themselves, fall into their own patterns. School- boys, with a keen sense for economy of effort, learn their lessons only partly overnight, and trust to a hasty review in the morning; the associative tendencies work while their owners sleep. Here, too, is the secret of the practised speech-maker. Knowing that he has to talk on a certain subject at a certain date, he runs over his present ideas in ten minutes of concentrated attention, and drops them; then, on the appointed day, he finds that the associa- tive tendencies have prepared his address.
With meaningful material, interference may arise in other ways. Take the alphabet, again: a is connected with b through the fre- quent repetition of abc, but is also connected with z by the phrase ' a to z.' Let a appear in consciousness; what happens? It may be promptly followed by b or z; the one of these ideas may, as we have expressed it, be more nearly ' ready ' than the other. If the terms of the two associations are complex, the incoming idea may derive from both of them: a may, so to say, call up an idea that is partly b and partly z. But, if the ^-tendency and the 2-tendency are of approximately equal strength, they will cancel each other, and there will be no association. A question often 388 Association leaves you dumb, not because you have no answer, but because you have so many answers that no one of them can force through to expression.
This sort of interference is known as terminal inhibition, in con- tradistinction to another kind, which is called initial inhibition. If a is already connected with b, then it is difficult to connect it with k; b gets in the way. You have some particular fault of style, or you have fallen into the habit of spelling wrongly some particu- lar word; you want to correct the fault, to spell aright. But every time that you are off guard, the mistake recurs; the existing asso- ciation a-b heads off the desired association a-k. This law, then, cuts across the law of remote association. The first impression of the alphabet sets up, it is true, not only the direct association a-b, but also the remote association a-k. As, however, the alpha- bet is more and more often repeated, the direct association more and more strongly inhibits the remote, so that it presently requires more readings to establish the connection a-k than were needed for the original connection a-b.
On the other hand, the associative tendencies may reinforce one another. Association may be convergent; a whole constellation or complex of tendencies may work together in the interests of a single idea, and this complex, as we have repeatedly seen {e.g., p. 274), need not by any means appear as a whole in consciousness. Chil- dren who are brought up to speak two languages rarely mix their words; as they have begun, so do they continue, a sentence; the first utterance has behind it the directive pressure of a multitude of cooperating tendencies. More than this, association may be at first divergent, and then convergent. A perception may stir into activity a number of tendencies, or a number of constellations; and these may, in their turn, all converge upon a single idea. Here we get a glimpse into the physiological basis of conscious context, of psychological meaning (§ 103). The words of a sen- tence, the sentences of a paragraph, the paragraphs of a chapter, the chapters of a book, arouse innumerable tendencies in the nervous system of the reader. Sometimes a special constellation gains the upper hand, and the reader is sidetracked by his own ideas; in general, the writer has his way, and the divergent ten- § ill. The Associative Consciousness 389 dencies are continually recombined. However, this is not the whole story of meaning; the reader's total attitude is also of great importance (§ 141).
§111. The Associative Consciousness. — There are as many- modes or forms of association as there are ways of being together within a conscious present. What these ways are, we already know; they have been outlined in our discussion of perception. We may say, then, that there are as many forms of association as there are forms of perception and idea; the pattern of the associative conscioussness may be spatial, temporal, qualitative, or mixed. The pure percep- tion is itself an association of sensations, and the idea is an association of images.
But is there no psychological difference between percep- tion and idea, on the one side, and the association on the other? Not, certainly, in composition: the elementary processes that analysis reveals are the same in both, — sensations and images. Not, necessarily, in complexity: many of the instances of association given in the text-books are simpler, contain fewer elementary processes, than the more complicated perceptions. Not, again, in mode of connection of the elements: the laws governing connec- tion are the same throughout. And not, necessarily, in closeness of this connection: the names of familiar things are as closely bound up with their perceptions as the sensory elements in the perceptions are bound together. The differ- ence is, in fact, rather a psychologist's than a psychological difference. We must build up a psychology by stages, in orderly fashion; and it is convenient to distinguish, first, the bare elements, sensations and images; next the con- nection of the bare elements, perceptions and ideas; and then, later, the connections of elements that have already 39° Association 39° Association been connected, the associations of ideas. If we may use a figure, which roughly expresses the truth, we can say that the elements of the perception have never been together before, while the elements of the association have manifold habits of connection already upon them.
The doctrine of association has, however, played so im- portant a part in the history of psychology, and the influence of tradition is so strong, that many psychologists tend, as if instinctively, to differentiate the idea from the association of ideas. The tendency shows itself in two ways: first, in the overemphasis of successive, as compared with simultaneous association; and secondly, in the attempt to classify and tabulate the various forms of association.
Most students of psychology, if they hear the phrase 'association of ideas,' think at once of the successive association; and this, by its very nature as a train or succession, is more complex and more variable in its course than is the idea. Yet it is certain that the simultaneous association is the typical association, and that the successive association, as illustrated in the books, is of rare occur- rence, a limiting case of association at large. " In a discourse of our present civil war," wrote Hobbes in 1651, "what could seem more impertinent than to ask, as one did, what was the value of a Roman penny? Yet the coherence to me was manifest enough. For the thought of the war introduced the thought of delivering up the king to his enemies; the thought of that brought in the thought of the delivering up of Christ; and that, again, the thought of the thirty pence, which was the price of that treason. And thence easily followed that malicious question." Here, no doubt, Hobbes has his finger on the coherence, the context. But there is just as little doubt that the psychology of the situation was widely different from his conception of it; there was no simple sequence of thought upon thought, idea upon idea. Let the reader observe for him- self!
We shall not here enter upon the question of classification. The § in. The Associative Consciousness 391 various authors who have drawn up tables of associations have based their arrangement, almost without exception, upon logical principles, upon the meaning of the associated ideas; they have referred the associations to certain logical categories, such as subordination, coordination, cause and effect, means and end. The results may be of value for a psychology of individual differences; they have no place in a general, descriptive psychology. Moreover, associa- tions vary with circumstances. It has been found, under experi- mental conditions, where a stimulus-word is given and the observer has to reply at once by naming the first word which occurs to him, that substantives are associated to substantives, adjectives to adjectives, numbers to numbers; and that the association is nearly always formed within a single sense-department, so that red calls up green, and hard, soft. But if the conditions are changed, and a little more time is allowed the observer, the character of the asso- ciated ideas is also radically changed. Fatigue, again, may bring out purely mechanical associations (snow — ball, fish — dish) which are ordinarily foreign to the observer's consciousness.
Introspective Analysis. — If a familiar visual stimulus (word, simple picture) is presented to the observer, with the instruction that he shall receive it passively and report the consequent course of his mental processes, the following results appear. First, there is a simultaneous association of stimulus and internal speech; the word or the pictured object is named. Thereupon follows an associative complex, which may assume any one of three principal forms. (1) The stimulus arouses, either at once or very shortly after the named perception has become clear in consciousness, and either as a whole or by way of some part or aspect, an affective process, a feeling in the widest sense (p. 228). The feeling, in turn, calls forth an associated idea, which may subsist for a time alongside of the original perception, but soon replaces it. Thus, a word printed in very small letters upon a large ground aroused the feeling of loneliness; a word printed in red, the feeling of ex- citement; the word 'blinding,' the feeling of a blinding light; and then the feeling itself (or, in the author's view, the kinaesthetic and other organic components of the feeling) brought up an idea which supplanted the meaning of the stimulus-word. (2) In other cases 392 Association the named perception is either supplemented or replaced by the idea of some object or picture previously seen. Thus, an outline drawing of a face may suggest the idea of a friend, whose features are then, so to speak, read into the drawing; the perception is lost or merged in a simultaneous association. Or the word 'Tell,' printed on a blue ground, calls up a familiar picture of William Tell springing from a boat to the rocks • the blue of the background becomes the blue sky of the painting. There are many intermediate forms between these extremes, as there are also between this group of associations and the next. (3) Here the stimulus arouses an idea which, at first thought, appears to be separate and detached; we have the traditional pattern of the successive association. And, occasionally, we cannot go behind that pattern; the named per- ception and the idea seem to be mechanically linked; we are in presence of the limiting case. Usually, however, introspection takes us farther. Thus, the outline drawing of a tent called forth the idea of a certain city market; and for a moment the observer could simply report the bare succession of the experiences. But then he found the cue: he had ridden through the market, on his bicycle, in much the same attitude in which he now sat; a massive complex of organic sensations was common to the two situations. — Observations of this sort make it quite clear that the type of associa- tion is the simultaneous association, and that successive asso- ciations — to put the matter a little paradoxically — are simply simultaneous associations drawn out in time.
If we keep strictly to the observed facts, we can find no psychological difference between the idea and the association of ideas. Still, the practical difference remains that our ideas come ready-made, whereas we can establish new associations. May not introspection show the con- scious mechanism of this novel grouping? and may we not thus* get fresh light on the nature of the mixed per- ception, and of the idea itself?
Wundt has answered these questions in the affirmative. He believes that associations are always established in the § in. The Associative Consciousness 393