particular object which may answer to a certain description, or be found to present certain characteristic attributes or traits. Or, as the logicians express it, a general idea is a representation of a class of things.

The reader must be careful to distinguish the meaning of the term class as here used from its meaning when applied to a definite number of objects viewed as a collection, as a class of children in a school. In thinking of man as a (logical) class, I do not represent a definite number at all; nor do I represent men as a collection. It would be more correct to say that I am representing, in a more or less distinct way, the fact that this, that, and an indefinite list of other things are related as like or answering to one description. How this mode of representation is effected will appear presently.

Now it is evident that general ideas as thus defined are reached slowly and by degrees. It is exceedingly doubtful whether any of the lower animals acquire them. The baby does not possess them, and even after attaining to speech remains for a long time with only the rudiments of them. In their perfected articulate form, as required for exact scientific thought, they are confined to a few highly-trained minds.

§ 14. Generic Images. The first stage in the formation of such general ideas is the welding together of a number of concrete images into what has been called a generic image. The idea tree' or 'house' may be taken as an example. Such generic images may be supposed to be formed by a process of assimilative cumulation. Let us imagine that a child, after observing one dog, sees a second. In this case the strong resemblance in the second to the first effects a process of assimilation analogous to the automatic assimilation already described. That is to say, the percept corresponding to the second animal is instantly fused with the surviving image of the first by reason of easily-apprehended points of likeness. By such successive assimilations a cumulative effect is produced which has been likened to that of the superimposing of a number of photographic impressions taken from different members of a class (e.g., criminals) whereby common features become accentuated, and so a typical form is produced.1

1 For an account of such composite photographic pictures and their analogy to generic (mental) images, see Mr. F. Galton's Inquiries into Human Faculty, Appendix, "Generic Images".

THE GENERIC IMAGE.

415

This process, the deepening and accentuating of common traits and the effacing of individual or variable ones, can only be looked on as a tendency never perfectly fulfilled. Interesting differences would in all cases tend to reinstate themselves. Thus my generic image of a church is a building with a tall spire, because the main church in my native town was of this form. Recent examples would also tend to contribute variable peculiarities. Thus the baby's generic image of a dog would have the special characters of the dog last observed.

Such a process of cumulative assimilation would be largely passive, and independent of those active processes of comparison just described. It would further be capable of being carried forward (to some extent at least) independently of language. Hence we may, with some degree of confidence, attribute generic images to the child before he comes to the use of words as well as to many of the lower animals. Thus it is highly probable that a baby of six months forms a generic image of the human face out of the percepts answering to that of its mother, its nurse, and the like, and that when suffering from loneliness it has this image in its mind. Similarly, a predatory animal may be supposed to compound a generic image out of the percepts gained from this, that, and the other specimen of his prey, so that when seized with hunger this typical image is recalled.

In order to illustrate what is meant by a generic image, it is important to take the case of a pure representation detached from a presentation. Thus we cannot say that because a diving bird recognises a new sheet of water it must have at the moment a generic image answering to water. As was pointed out above in dealing with automatic assimilation, the recognition of a thing does not imply a distinct representation of the thing as previously seen. The presentative and representative ingredients are fused in this case, or, to express it otherwise, the image is latent and undeveloped. Similarly with respect to such rudimentary processes of conception or general ideation as those here considered. We can only attribute a developed and detached generic image to baby or animal when we have reason to think that these occur in the absence of the percept, e.g., in states of desire, in dreaming, and so forth.1

§ 15. Relation of Generic Image to General Idea. The question still remains how far such generic images are, properly speaking, general ideas in the sense defined above. Is, for example, the typical face that is pictured by the lonely infant

1 The argument in support of the proposition that generic images or (as the writer calls them) "recepts are actually reached by the lower animals is ably set forth by Romanes, Mental Evolution in Man, p. 51 ff.

thought of as something common to this, that, and the other concrete object? Does it carry with it any clear consciousness of a class of things? There is no evidence to show that this is so. As has been pointed out above, the mental image corresponding to one and the same individual object, as the infant's mother, is composite also and in the same way as the generic image. Thus the baby forms an image of its mother out of a number of partially unlike percepts, corresponding to varying appearances of the object in different positions, different lights, different dress, and so forth.1 Generic images accordingly differ not in kind but only in degree (viz., proportion of common to variable features taken up and accentuated) from particular or concrete images. And so long as they remain merely pictorial images there seems no reason to attribute to them any general function or import.

The true process of conception, as generalisation or general ideation, that is, a conscious representation of something as common to many as distinguished from one, involves the active processes of thought, analysis and synthesis, abstraction and comparison. It is only when the child begins consciously to break up its images, to distinguish this element or feature from that, and by help of such analysis to recognise and mark off common features, that general thought, properly so called, begins. In this way it reaches a distinct idea at once of an individual thing, and of general or common aspects among individuals. We have now to examine into this true thoughtprocess.

§ 16. Transition to Conception Proper. The transition from merely imagining to thinking proper is effected by processes of reflective attention in which abstraction and comparison play a chief part. In order to understand how this occurs, we may suppose the process of automatic assimilation checked by the introduction of some impressive difference. Thus a child proceeds to play with a visitor's dog, and finds it wanting in the friendly sentiments of his own pet. Here difference, which in the earlier stages of automatic assimilation remained indistinct in the background of consciousness, is brought forward. The unlikeness of morale in spite of the likeness of physique is forced

1 Cf. above, p. 260 f.; also Taine, On Intelligence, pt. i. bk. ii. chap. ii.

FROM IMAGE TO CONCEPT.

417

on his attention, the present percept is separated from and opposed to the image, and a step is taken in marking off likeness from surrounding difference.

nesses.

As differences thus come into distinct view and impress themselves on the mind as the constant accompaniment of likenesses, a new and explicit grasp of likeness-in-difference ensues. This starts from a mental separation of the several perceptual constituents of the generic image, and a reflective comparison of these one with another so as to demarcate common features or likenesses from peculiar features or unlikeSuch comparison, or series of comparisons, begins with incomplete analysis and vague apprehension of likeness, and ends in a more complete analysis and more definite apprehension of likeness. In this way, for example, the child waking up to differences, say among apples, goes back on his various experiences, and by noting and setting aside variability of taste, size, and so forth, gets a clearer grasp of the common essential features. Such a conscious active separation of definite points of resemblance from among a confusing mass of difference is what psychologists and logicians more especially mean by Abstraction.

§ 17. Differentiation of Notions of Individual and Class. As was pointed out just now, the co-existence of likeness with unlikeness in the child's experience may mean one of two things, viz., persistence or identity of one individual object, in spite of certain changes, or a general similarity among a number of different individuals. The process of conception is sometimes described as if the child started with a definite knowledge of individuals, and then proceeded to generalise or form a class-idea. There is, however, every reason for saying that the two modes of interpretating likeness-in-difference are reached concurrently and by processes largely similar. Thus it seems most reasonable to suppose that the baby which 'da-das' every bearded person it sees is as yet clearly conscious neither of individuality nor of generality. In other words, we must not assume that it is stupidly confounding its sire with a stranger, or, on the other hand, forming an idea of a general class. At this stage the child merely recognises certain interesting similarities, and proceeds to express the fact. We have to suppose that the clear apprehension of individual sameness is reached but slowly and

in close connexion with the first clear consciousness of different things attached by a bond of likeness.

To say that the child's knowledge begins with the concrete individual is not to say that it attains a clear consciousness of what we mean by an individual thing, persisting and the same (in spite of change), before it begins to generalise. As was pointed out in the brief sketch of the process of identification in connexion with sense-intuition, the cognition of "thing" as persistent and continuous is the result of lengthy and complex processes of comparative reflexion. To individualise is thus to think, just as to generalise is to think.1 In truth, the psychological development of the idea of individuality proceeds along with that of generality, each being grasped as a different way of interpreting partial similarity among our percepts.'

§ 18. The Process of Generalisation. When once this differentiation of the individual-idea from the class-idea has advanced far enough the process of generalisation proper, or the grasp of common or general qualities, is able to be carried out in the way usually described by psychologists. That is to say, a number of individual things, represented as such, are now compared, the attention withdrawn by a volitional effort from points of difference and concentrated on points of likeness (abstraction), and so a true process of generalisation carried out.

The common account of conception here followed, as made up of a sequence of three stages, comparison, abstraction and generalisation, rather describes the ideal form of the process as required by logic than the mental process actually carried out. As we saw above, a vague analysis or abstraction precedes that methodical comparison of things by which the abstraction becomes precise and perfect, that is to say, definite points of likeness (or unlikeness) are detected. With respect to generalisation, it has already been pointed out that this is to some extent involved in abstraction. To see the roundness of the ball is vaguely and implicitly to assimilate the ball to other round objects. It is to be added that an imperfect grasp of general features as such commonly precedes the methodical process here described. The child realises in a measure the general function of the name 'horse' before he carries out a careful comparative analysis of the equine characters. At the same time the use of the word "generalisation" is important as

1 Cf. above, p. 233 ff. Hence the logician can speak of the idea answering to a proper name as a singular concept. (See Lotze, Logic, p. 34.)

2 "A singular individual is as much conceived when he is isolated and identified away from the rest of the world in my mind as is the most rarefied and universally applicable quality he may possess. (Sh. Hodgson, quoted by James, op. cit., i. p. 479.) The question of the priority, in the development of the child, of the knowledge of the individual or of the general class, the question known as the primum cognitum, has been much discussed in connexion with the linguistic problem whether names are first used as singular or as general names.