Association

Here is a kind of attraction which in the mental world will be found to have as extraordinary effects as in the natural, and to show itself in as many and as various forms.—David Hume

§ 31. The Association of Ideas.—The doctrine of the ‘association of ideas’ is one of the oldest and most influential in the history of psychology. It begins, in a somewhat casual way, with Aristotle. Suppose, Aristotle says, that we are trying to recall something that has slipped our mind; what do we ordinarily do? We hunt through a number of things, beginning with something that is like what we want to recall, or contrary to it, or that was next it in time, or adjacent to it in space. These other things, the like, the contrary, the just before or just after, the adjoining, have the power to suggest what we have forgotten. Aristotle gives the impression that everybody acts in this way, as a matter of course; and no doubt his hearers acquiesced; for the statement sounds reasonable. We want, for instance, to remember a certain picture that we saw ten years ago: how do we set to work? We start from something like it: ‘I remember that it reminded me of Van Eyck’; or from something opposite: ‘I remember smiling to think how a Venetian would have treated it’; or from something next it in time: ‘I remember coming to it after three whole hours of Dutch genre’; or from something next it in space: ‘I remember that it hung beside a Mabuse portrait.’ Seeing how natural and obvious such remarks are, we can understand that Aristotle’s single sentence had tremendous consequences for psychology. It foreshadowed the four ‘laws of the association of ideas,’ the laws of similarity, of contrast, of succession in time, and of coexistence in space. According to the doctrine of association, one idea ‘calls up’ another because it is like that other, or contrasts with it, or was next to it in time or space; likeness and difference, succession and adjacency, somehow give an idea the power to recall, and render it liable in its turn to be recalled. The four laws thus represent an attempt to explain the course of our ideas, and for that reason they have always appealed to common sense.

But, for the same reason, the laws have not proved an unmixed blessing to psychology. Aristotle, it is clear, was simply raising a practical question; and practical questions are answered in terms of meaning, not of process. Moreover, Aristotle was temperamentally a logician, and he could not help throwing even this bit of everyday practice into formal logical shape. Notice the arrangement in pairs: like-contrary, coexistent-successive; that is logical. Notice also the nature of the pairs. Like-contrary is the extreme way of saying like-unlike; and when you mention succession, you mention the only kind of non-coexistence that can come into account for psychology; so that both pairs have the form ‘A and not-A’ (like and not-like, coexistent and not-coexistent); and that is logical again. Aristotle’s four rules are therefore not really empirical, in the sense that they are directly derived from a study of experience; they rather show the inveterate logician, who is bound to schematise and tabulate. Later writers, swayed now by experience and now by logic, have both increased and decreased the number of these ‘laws’ of association; the general tendency has been to reduce them to two, or even to one. Thus, we can make contrast, logically, a case of likeness; the palace reminds us of the hovel, apparently by contrast; yet are not the palace and the hovel alike, as human habitations? We can, still more easily, reduce space to time. If the two pictures hung together on the wall, they were seen at the same time. Simultaneity, however, is one kind of contiguity in time; succession is another; and temporal contiguity thus includes everything. The four laws have become two: similarity, and contiguity in time.

Can we go further? Yes, if we go on arguing. The picture reminded me of Van Eyck; it was like a Van Eyck; the association seems to be an association by similarity. Yet it is practically certain that the picture in question was, at some time or other, present in my mind along with some picture by Van Eyck. It is practically certain, in other words, that the two ideas were in temporal contiguity; and every instance of association by similarity raises the same sort of presumption. That being the case, we may discard the law of similarity; and contiguity stands alone, the sole survivor of the Aristotelian quartet. Only, this is all logic, a matter of meanings, a translation of psychological fact; we have not got to the facts themselves.

We shall come to psychology presently. Meantime you should try to realise how well this doctrine of association works for practical purposes, and how strong is the appeal it makes to the practical side of our nature. It explains the appearance of every single idea that has ever occurred to anybody; it offers to take us to the very heart of psychology without need of training or preparation; it flatters us into the belief that we have all our lives been talking and thinking psychology without knowing it; it covers up the gap that separates common sense from science. Small wonder that Hume compared the law of association in psychology with the law of gravitation in physics! All the great names in British psychology (and the fact throws a good deal of light on the psychology of the nation itself) are connected with the doctrine of association; a whole science has taken its national colour from a single principle of explanation. Association has also played its part, though less dominantly, in France and Germany.

Realise all this; and realise also that the doctrine was of great service in the days when psychology was in the making; it is not only agreeable to common sense, it is not only historically important, but it also did true psychological service. Let us admit all this: and then we must add that the reign of associationism was over as soon as ever psychology became scientific; as soon, that is, as the proper task of psychology was recognised and formulated (p. 18). For let us take an instance: what does the word ‘summer’ suggest to you? Very likely it suggests ‘winter.’ How, then, is this association to be explained psychologically? By contrast? But the ideas of summer and winter may be exactly alike, both of them verbal-auditory-motor, or both of them mental pictures; the contrast is a contrast of meaning, not of mental process or pattern; the real summer, what we mean by the word ‘summer,’ contrasts with the real winter, and not the idea of summer with the idea of winter. By resemblance? But, if the ideas of summer and winter are exactly alike, so are they also like thousands of other ideas, verbal-auditory-motor or visual-imaginal; there is no reason in their psychological likeness why the one should suggest the other; and if they do suggest each other by ‘resemblance,’ the resemblance is again a likeness of meaning (they are both seasons of the year) and not of mental constitution. Try the matter out for yourself, in any concrete case of association, and you will reach the same result; the ideas of associationism are not psychological ideas. James sums things up for us: “Association,” he says, “so far as the word stands for an effect, is between things thought of; it is things, not ideas, which are associated in the mind. And so far as association stands for a cause, it is between processes in the brain; it is these which, by being associated in certain ways, determine what successive objects shall be thought.” The brain associates, and meanings are associated. We have already said something of the psychology of meaning (pp. 26 ff., 117 ff.); what can we now say of the associative functions of the brain?

§ 32. Associative Tendencies: Material of Study.—We want to find out how those processes in the brain which are the correlates of our ideas go together, get connected or associated. The brain is a machine; and it is not only complicated, but it is also plastic, that is, it is subject to change and modification. The complexity of the machine makes it necessary for us to work with simple stimuli and by strict methods; only if we work with simple stimuli shall we get to the bare essentials of the associative functions; and only if we work by strict methods shall we obtain results which other investigators can repeat and verify. Even so, the plasticity of the machine makes it impossible for us to lay down hard and fast laws of connection; we can speak only of connective tendencies or of associative tendencies; what actually happens, in any particular case, is likely to be the joint result of many tendencies, weak and strong, conflicting and concurring.

The task before us is, therefore, not easy; but it is straightforward; and that is the next best thing. We want to find out how associative tendencies in the brain are set up; and to do this we must, evidently, find some way of creating a bond between one nervous process and another; we must devise experiments in which we make or construct brain-connections. We need not look far afield; for we make such connections whenever we learn anything new; so that we have only to learn under experimental conditions, and the task is accomplished. But what shall we learn? what stimuli shall we employ in the experiments? ‘Words,’ you will say; and words have many advantages for learning; but they have, in this case, the supreme disadvantage that they are ingrained meanings. Words therefore will not do; but something very like them will. The question of the stimuli to be employed was, in fact, answered for us, thirty years ago, by the German psychologist Hermann Ebbinghaus, who—by one of those happy thoughts that come after long and intensive occupation with a subject—hit upon the notion of the meaningless syllable. Ebbinghaus made up over 2000 meaningless ‘words,’ all consisting of a vowel or diphthong between two consonants; syllables standing in the same relation to his own language that leb, rit, mon, yup, kig, wes, der, zam, for instance, bear to English. See the advantage of this kind of material for the work we have in view! The syllables are just like words, in that they may be seen, heard, or felt in the throat; they are unlike words, and vastly superior to them, in that they have no habitual associates; they lack context and meaning; every syllable in a series may be considered to have the same chances of making connections as every other. The material is so rich and varied that endless experiments can be made; it is so simple and uniform that the results of one experiment may be compared directly with the results of another; it may be drawn from any language, and so may be used in the laboratories of any country. Moreover, it is absolutely under control; it is just the kind of material that we need when we are tied down to strict and accurate method; we can vary at will the manner of presentation to the learner, the number of syllables in a series, the rate at which they follow one another, and so on; and the report required from the learner himself is easy and natural; there are no long descriptive phrases; he has only to say or to write the syllables he has learned. Lastly, we may proceed from experiments with this meaningless material to experiments with real words, words that mean; and we may hope in that way to pass beyond the bare essentials of the brain’s associative function, and to get a clue to the complex interplay of associative tendencies in real life. All in all, it is not too much to say that Ebbinghaus’ recourse to meaningless syllables, as means to the study of associative tendencies, marks the most considerable advance, in this chapter of the psychological system, since the time of Aristotle.

§ 33. The Establishment of Associative Tendencies.—The use of meaningless syllables has brought with it a whole armoury of technical methods for the study of the associative tendencies. We have here no space to treat of these methods in detail; fortunately, the results that we shall mention speak for themselves; and it may be added that all the methods of experiment are, in principle, changes rung upon one simple model, in which the observer sits down before a series of syllables, reads them through, so-many times over, in a state of attention, and then, either immediately or after an interval of time, repeats them ‘from memory.’ We proceed, then, to answer the question: How are associative tendencies established in the brain?

Their establishment depends, first and most obviously, upon the number of syllables in the series presented to the observer. While he can recite correctly, after a single reading, a series of 6 or 7, a longer series simply throws him into confusion. The first and last terms have a definite advantage; they may, indeed, be the only syllables that can be repeated after a single reading of a 12-term series. Secondly, the tendencies are strengthened by repetition. The first reading is more important than any other single 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. Thirdly, the tendencies are furthered by a grouping of the syllables. The observer learns a series more quickly if, for instance, he throws it into a rhythm. Fourthly, it is important to distribute the readings in time. Two readings a day for 12 days give better results than four a day for 6 days, or eight a day for 3 days, although the total number remains the same. Fifthly, the rate of reading has its effect; the syllables must not follow one another too fast or too slowly. There are great differences between individual learners; but we may say in general that the syllables should at first be presented at a moderate rate (perhaps two in the second), and that the rate should be slowly increased as the readings proceed. Sixthly, not only repetition itself, but also the manner of repetition, makes a difference. Meaningless syllables are learned somewhat better if the whole series is read through, over and over, from end to end, than if they are taken a few at a time, in small lots. Lastly, recitation or reading aloud is ordinarily more effective than silent reading; largely, perhaps, because the separate pronouncing of every syllable equalises attention; every term of the series is brought out sharply and clearly, and there is no chance to slur.

Here, however, we must remember the differences of imaginal type (p. 138); and it is true that a markedly visual learner will profit less by recitation than an auditory-motor learner. These experiments have, indeed, revealed other typical differences between individuals, such as those of slow and quick, and of receptive and ingenious learning. Some of us, it seems, are naturally quick, and some are naturally slow learners, just as some work best at night and others in the morning. Some observers, again, accept the series of syllables, passively and without question; others embroider and interpret the meaningless forms in all manner of ways; mon becomes man, and kig king, and wer where, and so on. We know nothing at present of the correlated differences in the nervous system.

The results just given may be compared with those obtained when meaningful stimuli are employed. Thus, 8 or 9 one-syllable words, and 10 to 12 one-place numbers, can be recited after a single reading. Meaningful material, which is grouped or unified by its topic, may be learned ten times as quickly as meaningless syllables. It may also be presented more rapidly; iambic and trochaic verses, for instances, may be taken at double the rate of the syllables. Dates of historical events, and the words of a foreign language, are best learned like the meaningless syllables; and connected meaningful material, like a poem or an oration, should very decidedly be read as a whole, from end to end, in the successive repetitions. If there are brief passages of unusual difficulty, they may, of course, be gone over by themselves, in the intervals between the total readings; the general rule, however, is to learn by wholes. This appears, in fact, to be the procedure generally followed by bards and tellers of folktales; and actors who play many rÔles in quick succession are able to ‘wing a part,’ as the phrase goes, by reading it through several times over at brief intervals. Children who memorise a poem in sections, a stanza now and a stanza to-morrow, waste a great deal of time.

Let us now come back to the meaningless syllables, and ask what is the net result of all the influences that we have listed. Suppose, in other words, that a series of syllables has been presented at a certain rate, thrown into a certain rhythm, repeated a certain number of times with fitting distribution in time, recited at every repetition: what is the final outcome, as regards the establishment of associative tendencies in the brain? It is this: that a strong connection has been set up between the successive terms of the series, in the order of their presentation; and that weaker connections have been set up between every term and every other term, whether the terms are near or remote in the series, and whether they are taken forwards or backwards. Let us illustrate by reference to the alphabet. If the alphabet represents a series of meaningless syllables, then there is a strong connection between a and b, b and c, ... y and z; but there are also weaker connections between a and d, ... v and z; and further, there are connections backward between z and y, z and x, ... d and a. The series of syllables has thus impressed the brain with a very complex meshwork of associative tendencies, stronger in some places (direct forward connection) and weaker in others (remote and backward connection), but still functionally interconnected through all its parts.

§ 34. The Interference and Decay of Associative Tendencies.—If a set of associative tendencies, such as we have just described, is left to itself, and neither disturbed nor renewed, it gradually disappears; the loss is at first very rapid, then proceeds more slowly, and thereafter goes on only at a snail’s pace. To make the matter concrete, we may think of the meshwork of tendencies as a meshwork of channels, deeper and shallower, in the substance of the brain; then the rule is that the channels tend to fill up,—the shallow ones speedily, the deeper ones at first quickly and then more and more slowly,—until everything is smooth again. This is a mere figure, but it carries the meaning that we desire. The same thing happens with the tendencies set up by meaningful material; they too slowly die away; but it is doubtful if they ever wholly disappear; in their case the brain, if it has been thoroughly impressed, seems never wholly to ‘forget.’ Ebbinghaus learned some stanzas of Byron’s Don Juan, for experimental purposes, and did not look at them again for 22 years; yet he relearned those stanzas in 93 per cent. of the time required to learn new stanzas; a saving of 7 per cent. Some stanzas that he had learned more thoroughly were not read again for 17 years; these were relearned with a saving of nearly 20 per cent. He had no memory whatever of the verses formerly learned; but his brain ‘remembered’; the associative tendencies had not completely disappeared.

As a rule, however, a particular set of tendencies is not allowed to die a natural death; it is interfered with by others. All associative tendencies need a certain time to establish themselves, to settle down; and if this time is not granted, but stimulus treads on the heels of stimulus, there is no impression of the meshwork, and no connections are formed; we have seen that a series of excessive length simply throws the learner into confusion. A recently acquired connection may even be abolished, as most of us know to our cost, by interruption of the train of thought; you have just got to your point, to the insight, the phrasing, the argument, that will clinch things; you are distracted by some irrelevant matter; 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 recover it; no wonder that the constructive worker, in literature, in science, in affairs, ‘hates to be interrupted’!

With meaningful material, interference may arise in other ways. Take the alphabet again; a is connected with b through the frequent repetition of abc, but is also connected with z by the phrase ‘a to z.’ If, then, a appears; and if the b-tendency and the z-tendency are of approximately equal strength; then there may be no connection at all; the two tendencies cancel or inhibit each other. A question may leave you dumb, not because you have no answer, but because you have so many different answers that no one of them can force through to expression. This sort of interference, which comes at the end of the associative process, is called terminal inhibition; there is another kind, coming at the beginning of the process, which we may call 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 particular word; you want to correct the fault, to spell aright. But every time that you are off guard, the mistake recurs; the existing connection a-b heads off the desired connection a-k.

Fortunately, there are compensations. If a group of tendencies, for instance, does escape interference, then the brain settles down of itself. Schoolboys, with a keen sense for economy of effort, learn their lessons only partway overnight, and find that a hasty review next morning is enough to fix them; the associative tendencies work while their owners sleep. The practised speaker, knowing that he has to talk on a certain subject at a certain date, marshals his present ideas in half-an-hour of concentrated attention, and then drops the whole thing; his brain incubates it for him; and when the appointed day comes near, he finds that his associative tendencies have practically prepared his address. Besides, the tendencies may converge, as well as interfere; we have seen how continued attention opens the mind to relevant facts and closes it against the irrelevant (p. 98). If they did not, it would be impossible for us to follow the thread of a paragraph, to say nothing of a chapter or of a whole book. Convergence thus offsets interference. We shall meet it in various forms later (§§ 42, 45, 65); meantime we leave the brain, and pass to the mental processes themselves. How are they connected?

§ 35. The Connections of Mental Processes.—So far as the elementary processes are concerned, this question has already been answered in our discussion of perception. We found that there were two modes of sensory connection, two ways in which sensations may go together. In qualitative perceptions, such as the perception of a musical note, there is a blend or fusion of qualities; we can, to be sure, analyse the compound tone, after practice, into fundamental and overtones; yet it still comes to us as unitary, as a single impression; it stands only at one remove, so to speak, from the simplicity of sensation itself. The tastes of coffee and lemonade, with their blending of taste and smell, of touch and temperature; the organic feels of hunger and thirst and nausea; the kinÆsthesis aroused by grasping and pulling, by lifting the arm and swinging the foot; all these experiences are fusions, more or less intimate, more or less complex, of sensory qualities. They too can be analysed; but the analysis is not easy; the qualities cling together, seem in a way to merge into one another. In spatial perceptions, on the other hand, in such perceptions as the sight of my desk with its litter of writing materials, the elementary processes stand out side by side; brown contrasts with blue, dark with light; here, we might say, is no confluence, but rather concourse. In the perception of rhythm we have the same separateness of sensations, only that it is now temporal instead of spatial; and in the perception of change (p. 132) we find both modes of connection, separate qualities or intensities passing into one another by that peculiar blur or fusion which we have called the index of change. This second type of connection, whether it is the side-by-side of space or the end-to-end of time, may be named conjunction.

The associative tendencies which we have been more recently discussing are set up by series of meaningless syllables, that is to say, by discrete stimuli. It is clear, then, that the connection of the correlated mental processes is of the conjunctive type; we have said nothing of the brain-processes which underlie sensory fusion. We can, indeed, say nothing of them; we have no knowledge of their nature. It has been suggested that qualitative perception is correlated with a synergy of the brain-processes, that is, with a cooperation so close that every process taking part in it loses something of its individuality. That is possible; we cannot say more.

When we leave the elementary processes for complex experiences, for perceptions and ideas, and ask how these are connected, we cannot return any completely satisfactory answer. Experiments may be made; thus, a familiar visual stimulus (word or simple picture) may be shown for a few seconds to the observer, with the instruction that he receive it passively and report the consequent course of his mental processes. Under these circumstances, it invariably happens that the stimulus is immediately named. After that, apparently, any one of three typical things may happen. First, the named perception is supplemented by a sense-feeling. A word printed in very small letters on a large background aroused the feeling of loneliness; a word printed in red, a feeling of excitement; the word ‘blinding,’ the disagreeable feeling of a dazzling light. Then the feeling gives way to an idea, which supplants the meaning of the stimulus. Secondly, the named perception is resolved into the idea of some object previously seen. An outline drawing of a face may be replaced by the idea of a friend, whose features are, so to say, read into the drawing; or the word ‘Tell,’ printed on a blue ground, may be replaced by the idea of the familiar picture of William Tell springing from a boat to the rocks; the blue of the background becomes the blue sky of the painting. Thirdly, and only occasionally, the named perception is followed by an idea which comes separate and detached; we have the traditional pattern of the ‘successive association.’ These three types of connection (there are, of course, intermediate forms) do not furnish a satisfactory answer to our question, mainly because the experiments are not properly under control; the observer comes to them with all sorts of associative tendencies at work; and unless we make a very large number of observations, we cannot be sure that our results are either representative or exhaustive.

At the same time, such experiments help us; they show, for instance, that the doctrine of association—quite apart from its logical leanings, or perhaps just by reason of them—regarded the course of ideas in too ‘intellectual’ a way; the sense-feelings, and other feeling-blends that we shall mention later, play a larger part in our thinking than the associationists dreamed of. They show, too, that the ‘successive association’ is not the commonest, but rather the least common, form of mental connection. Listen to a quotation from Hobbes! “In a discourse of our present civil war,” he writes, “what could seem more impertinent [less to the point] than to ask, as one did, what was the value of the 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: and all this in a moment of time, for thought is quick.” Hobbes has worked out the logical coherence, the coherence of meaning; but he is very far from a psychology of the situation. What actually took place in the mind of the questioner we shall never know; we may be very sure, however, that his mental processes did not follow one another in logical order, as Hobbes imagines. There was a convergence of associative tendencies, which expressed itself in the question; there need not have been any succession of ideas at all.

§ 36. The Law of Mental Connection.—We have spoken at some length of the establishment of associative tendencies in the brain, of their decay with time, and of their mutual interference. Can we sum up our knowledge of them in a single general statement? And can we then translate this general statement into psychological language, and so reach a formula of mental connection that may stand in place of the logical laws of association? Let us try.

We must proceed very carefully, even if our care drives us into clumsiness of expression. We cannot, for instance, leave out the fact that the meaningless syllables are given in the state of attention. It appears, indeed, that attention is necessary to association; we may doubt if any amount of repetition—to take that example—would set up an associative tendency, were it not for attention. Repetition, we remember, is one of the determinants of attention (p. 94); so that the repeated experience is likely to become vivid in the very nature of the case; but if it does not, if for any reason our attention is diverted or we fail to notice the stimulus, repetition has no associative power. How many of us would like to recall the carpet or wall-paper of the room we slept in as children! Thousands of times we saw the colours and the patterns; but our adult memory is an absolute blank; those repeated stimuli never ‘impressed’ us.

We cannot either leave out the fact that the meaningless syllables are bracketed all together, so to speak, by a certain situation, namely, the situation created by the experiment. The observer comes to them, in accordance with this situation, intending to learn them, to memorise them: a fact of very great importance!—and a fact that needs to be dwelt on for a little, if we are to see our way clearly in what follows. We said on p. 149 that meanings are associated. Yet we have been studying the formation of associative tendencies in the brain, the associating organ, by the help of—meaningless syllables! Is there not a flat contradiction here between theory and practice? No, that is really not the case; and the key to the riddle lies in this fact of the ‘situation’ which we are now discussing. The syllables are meaningless as syllables; they are thus set apart from ordinary syllables that are meaningful; and it is this difference from words, combined with their likeness to words in other respects, that makes them useful to the experimenter (p. 151). For since they are themselves meaningless, we can put upon them a constant meaning of our own; we can introduce them into any situation of our own making; and the meaning that we give them, in the study of the associative tendencies, is the meaning of ‘an experimental series to be learned under certain instructions’: a meaning which is definite, and which remains the same throughout the experiments. You see, then, that the ‘situation’ is important.

Attention, as we know, means reinforcement of certain nervous processes and inhibition of others (p. 107); and the intention to learn implies the activity of directive nerve-forces (p. 96), the existence of a special set or disposition of the brain. Let us keep these things in mind; and let us call the brain-processes that are correlated with mental processes ‘psychoneural’ processes. Then we may say: When a number of psychoneural processes, all of which are reinforced and all of which stand alike under the directive influence of a nervous disposition, occur together under certain favourable conditions, then associative tendencies are established among them, such that the recurrence of any one tends to involve, according to circumstances, the recurrence of the others. The phrase ‘under favourable conditions’ refers to the effect of repetition of the series, of their distribution in time, and so forth; and the phrase ‘according to circumstances’ means that heed must be paid to the lapse of time since learning, to the working of initial or terminal inhibition, and so forth.

So much for a generalised law of associative tendency, derived from the work with meaningless syllables! That is a law of nervous action; now let us turn to psychology, and see if we can formulate a law of mental connection. We shall be dealing with perceptions and ideas; and we shall be dealing with them as experiences, made up of core and context (p. 117).

Attention is again necessary. Intention, on the other hand, seems not to be necessary; there need be no special purpose behind the experiences, as the intention to learn is behind the experiments with meaningless syllables; attention is enough. The idea of a surgical operation, for instance, may be permanently connected with the idea of the surgeon who performed it, although the intervention of that particular surgeon was quite casual and unexpected. The reason is that attention brings a situation, its own situation, with it; the determinants of primary attention are, as we put it on p. 97, the ‘great biological stimuli,’ things that an organism must take notice of, if it is to persist as a living organism at all; and the determinants of derived primary attention are also what we may call ‘situational’ affairs, things that appeal in certain circumstances to certain sides of our nature, things that interest or ‘impress’ us. So attention, too, implies a set or disposition of the nervous system; common sense is so far in the right—though its words are misleading—when it talks of a ‘concentration of the mind,’ of ‘pulling oneself together,’ and the like; and this general set is sufficient, without the presence of a distinct purpose. Our law will read, then, somewhat to this effect: If a number of vivid perceptions or ideas, whose situational context is the same, occur together under favourable conditions, then the later appearance in the same situational context of any one will tend to be accompanied, according to circumstances, by the reappearance (as ideas) of the others.

That is correct, so far as it goes; though, as we shall see in a moment, it does not go quite far enough. Meanwhile, you must clearly realise that the processes which compose the perceptions and ideas are extremely variable. We have already discussed this matter; we have seen that the perception of an object and the idea of the same object do not by any means correspond, term for term, like original and copy; the form of our ideas depends, in the first instance, upon our imaginal type, and secondarily upon the special circumstances under which they appear (pp. 139 f.). When, therefore, we speak of ‘the later appearance of an idea in the same situational context,’ we really mean the appearance of that complex of mental processes which, under the law of imaginal type and under the special circumstances of the moment, has taken the place of the original complex. In the next chapter we shall be discussing the ‘memory-image,’ and you will then be shown how radically an idea may be transformed; so radically, that it may be likened rather to a translation than a copy of the perception, rather to a rendering into another language than a reproduction. If you want a catch-phrase, to hold this fact of change in mind, think of association as a marriage by proxy; the marriage-bond, the situational context, remains the same, but the parties are represented by very variable mental complexes.