§ 3. Different Forms of the Association of Ideas. — (a) Contiguity (Continuity of Interest). The law of Contiguity, as ordinarily understood, may be stated as follows, — If B has been perceived or thought of together with A or immediately after A, then, on a future occasion, the perception or idea of A will tend to call up the idea of B. In other words, the sequence of ideas follows the order in which their objects have been attended to in previous experience. The underlying principle is that mental activity when partially revived tends to repeat itself; it can only repeat itself if its original direction and order are reproduced.

This perhaps is the best form which can be given to the law of Contiguity, as ordinarily understood. But even in this form it requires qualification. It is by no means true that association of this kind connects only those objects which occupy attention in immediate succession. This has been shown experimentally. Professor Ebbinghaus found that after learning by heart a series of disconnected words, which we may denote by A, B, C, D, etc., it cost him a much shorter time to learn the same series with regular gaps in it, e.g. A • • D • • G • • , etc. Repetition of the series A, B, C, D served to establish associative links not only between A and B, B and C, C and D, etc., but also between A and D, D and G, G and II, etc. The same point is more conspicuously illustrated in ordinary experience. In recalling a train of events we usually pass from one salient occurrence to another, leaving out the relatively unimportant details which actually intervened between them. Similarly, in describing an object, I do not mention all the details which I actually observed in the exact order in which I noticed them. On the contrary, I pass from one characteristic and distinctive feature to another, oblivious of much which is not characteristic and distinctive. The dominant interest of the original experience and the dominant interest at the time I recall it, determine a selection of items which is by no means tied down by the condition that objects which introduce each other in the train of ideal revival must have been attended to in immediate succession.

The truth is that the fundamental principle of association is not contiguity in the strict sense of the word, but rather continuity of interest. The stronger the dominant interest, the cognitive tendency guiding the whole process, the more selective is the revival apt to be, links being dropped out which are relatively unimportant to the general trend of mental activity. This is well brought out in the special case in which some process having continuity of interest is carried out with interruptions occurring at intervals. If the interruptions are not themselves of a specially interesting kind, we tend to omit them altogether in recalling the main activity. The gaps, so to speak, close up.

It may be asked why in any case we should remember the interruptions, even when they are specially interesting. For where there is interruption, there is not continuity, but rather discontinuity of interest. The answer is that at the moment at which the interruption takes place there is continuity between the two processes which are otherwise disconnected. The interruption is itself an experience which belongs to both equally and serves to link them together.

Immediate succession, then, is not the fundamental condition of the association which is called association by contiguity. Ideal revival may and often does proceed by leaps and bounds. But it must be conceded that the immediacy of the succession does count as a very important factor. Other things equal, the direct transition of attention from A to B will be repeated in ideal revival, rather than a transition from A to C which originally took place through the intermediate link B. In proportion as the control of a dominant interest is weak and intermittent, the tendency is to exactly repeat the original order without omissions and inclusive of interruptions. This is well seen in the conversations of feebleminded persons. Of course the original order will be exactly repeated, where there is an interest in exact repetition, as in learning by heart.

(b) Similarity. The characteristic feature of reproduction by similarity is, as Dr. Bain observes, that it is opposed to routine and counteracts its effects. In such reproduction one object may recall another with which it has never been connected in previous experience. I see a man who reminds me of the Duke of Wellington by some resemblance in his personal appearance. I have never had occasion before to think of this man and the Duke in any kind of connexion with each other. The ideal revival seems to give rise to a completely novel combination instead of reproducing a past combination.

If this were really so, we could not properly speak of association as having anything to do with the matter. Association must at least imply that revival depends on objects having somehow come together in previous experience.

In fact a closer analysis shows that this actually is so in the example chosen and in all instances of socalled association by similarity. What is really operative in calling up the idea of the Duke of Wellington, is the personal appearance of the man in so far as it resembles that of the Duke. The experience I have now in looking at the man is partially the same in character as the experiences which I have previously had in looking at the Duke's portraits. The mental disposition left behind by these experiences is partially reexcited, and in consequence it tends to be reexcited as a whole. But this reexcitement of the whole in consequence of the reexcitement of the part is due to continuity of interest, and not to any essentially distinct principle. The principle of continuity alone is operative, but it operates in a very different manner and produces a very different result: reproduction by similarity and reproduction by contiguity respectively. Reproduction by similarity is most aptly described by reference to its effect. It ought to be called reproduction of similars rather than reproduction by similarity. Reproduction by contiguity may be called by way of distinction repetitive reproduction. Both repetitive reproduction and the reproduction of similars are in a sense cases of reproduction by similarity. Neither involves complete identity. Smoke reminds one of fire because of preformed associations. This is repetitive reproduction. But the smoke I now see may have features of its own in which it differs from previous experiences. It may be more voluminous, lighter or darker in colour, and so on. In other words, there need only be similarity, not complete identity. The points of difference do not contribute to bring about the reproduction. The partial identity is alone operative in this. But the specific differences none the less play a positive part in the process. Though they do not help to bring about the reproduction, they modify the nature of what is reproduced. A thin thread of smoke suggests a small fire; a large volume suggests a big fire. Smoke on a moorland and smoke rising from a house in London both suggest fires, but with very important differences. The reproduction is due to their identical character; the difference in what is reproduced is due to their different mental setting.

Now in the reproduction of similars the points of divergence between the reproducing presentation and that which is reproduced play no positive part in determining the reproduction. The partial identity of personal appearance between a man whom I meet casually and the Duke of Wellington calls up in my mind the idea of the Duke. But this idea is not transformed in a special manner by the divergent characters which distinguish the man before me from the victor of Waterloo.

We must carefully distinguish between the actual reproduction of similars and the processes which frequently follow on it. When one presentation has called up another similar to it, the mind may proceed to compare them, and it may make the partial identity which is discernible between them the basis for working out a parallelism in other respects by means of repetitive reproduction. The relation of an apple to the earth reminds Newton, according to the familiar legend, of the relation of the moon to the earth. But he does not stop here. Fixing attention on the partial identity, he strives to enlarge it by tracing identity in other respects also. This takes place by trains of thought in which the effect of repetitive association is profoundly modified, but not arrested by the difference of the two.