and the union of ideas. The primitive phenomenon is impression, or, as it is commonly called, sensation; idea is a feebler copy of this; then ideas associate themselves, unite, and there result complex or aggregate phenomena. Mr. James Mill admits only sensations, ideas, and associations of ideas.1

He classes our sensations under eight heads,-Smell, Hearing, Sight, Taste, Touch, Sensations of disorganization in some portion of the body, Muscular sensations, Sensations of the alimentary canal. As we shall see hereafter, contemporary psychologists generally reduce the last three groups to two,—muscular sensations, organic sensations; the former relating to the muscles, and which reveal tension or effort, the latter relating to the good or bad condition of the organs. But it is important to remark, that our author has seen more clearly than the Scotch school,2 which, adhering to the traditional five senses only, could not achieve more than a curtailed analysis of the sensations. Thence came the impossibility of any scientific explanation of exterior perception; for had not this school neglected the analysis of the muscular sense, which reveals to us resistance; that is to say, the fundamental sensation of exteriority? Thus James Mill is right when he says, 'there is no element of consciousness which demands more attention than this, though until of late it has been deplorably neglected.'

It is a peculiarity of our constitution that when our sensations cease through the absence of their objects, something remains. After having seen the sun, if I shut my eyes I no longer see it, but I can think of it. That which thus survives sensation I call ' a copy, an image of the sensation, sometimes a representation or a trace of the sensation.' This copy is the

The general faculty of having sensations is called sensation : the general faculty of having ideas is called by the author Ideation. As the idea is the copy of the sensation, and as

1 See Essays 2 and 3.

2 That of Reid, Dugald Stewart, and their contemporaries.

there are eight groups of sensations, there are eight groups of ideas of which it is easy to find examples.1

We know the simple sensations and those secondary sensations, which are their images. These are the two primitive states of consciousness. From thence all those combinations whose varieties are innumerable result; they are produced by the association of ideas.

All the philosophers of whom we treat regard the phenomenon of association as one of the most general laws of psychology, and even as the fundamental fact to which they endeavour to bring back everything in our mental life. This doctrine, known in England by the generic name of Association-Psychology, is only in its beginning in James Mill's works, but supported by the preceding studies of Hume and Hartley, it is presented in a clear and decided form, as we shall presently see.

Association is so general a fact that our whole life consists in a succession of sentiments (train of feelings). Can an order be discovered in this? Let us remark, in the first place, that association is produced as well between sensations, as between ideas.

Association between the sensations ought to take place conformably to the order established between the objects of nature; that is to say, according to a synchronic order or according to a successive order. Synchronic order, or that of simultaneous existence, is order in space; successive order, or that of anterior and posterior existence, is order in time. The taste of an apple, its resistance in my mouth; the solidity of the earth which carries me, etc., this is synchronic association. I see a bombshell thrown, I follow it with my eyes, I see it fall, and cause destruction, this is successive association.

As our ideas are derived, not from the objects themselves, but

1 Mr. Stuart Mill calls attention in Note 24 to the fact that the idea, being the copy of the sensation, it may be asked whether there is not also a copy of the copy, an idea of the idea. My idea of Pericles, or of an existing person whom I have never seen, corresponds to a real existing object, or one which has been existent in the world of sensation. Nevertheless, as my idea is derived not from the object, but from the words of another person, my idea is not a copy of the original, but a copy of the copy of another; it is the idea of an idea.

D

from our sensations, we may expect from analogy that their order shall be derived from that of the sensations, and this most frequently occurs. 'Our ideas are born or exist in the order in which the sensations, of which they are the copies, have existed.' Such is the general law of the association of ideas.1

When sensations are produced simultaneously, ideas are also awakened simultaneously; when sensations have been successive, ideas spring up in succession. The causes of association seem to be two in number: the vivacity of the associated sentiments, and the frequency of the association.

Association takes place not only between simple, but between complex ideas, which melt together so as to form an idea which appears simple. Such are our ideas of most familiar objects; the idea of a wall is a complex idea resulting from the already complex ideas of bricks and lime.

Hume, as we know, had said that our ideas associate themselves on three principles: contiguity in time and space, resemblance, and causality. The author, who admits the first principle only, contiguity in space (synchronic order), and contiguity in time (successive order), endeavours to bring the two others into this one, an attempt at simplification which, in the judgment of Mr. John Stuart Mill, 'is perhaps the least happy in the whole work' (Note 35).

II.

Before approaching imagination and memory, which, it would seem, ought immediately to follow, we shall find a study of words, parts of speech, the act of naming, which appears to us the most antiquated portion of the book.

It is remarkable that English contemporary psychologists, who have profited so largely by the recent progress of physiology, have borrowed nothing from linguistics. It may be main

1 Vol. i. ch. iii.

2 In the tribunals, says the author, it is observed that ocular and auricular witnesses always follow the chronological order in their narratives; that is to say, the order of their sensations, whereas those who invent seldom observe that order.

tained that that science is as yet neither sufficiently mature nor sufficiently well co-ordinated; but it is incontestable that it has much to reveal to us concerning the constitution, and above all the development, of the human soul. It will become one of the elements of that objective and inductive method which tends to prevail in psychology. Maupertuis, in his Réflexions philosophiques sur l'origine des Langues, speaks of the utility of studying the languages of the savages, 'which are conceived on a plan of ideas so different from our own.' It has been done, and we can readily believe that comparative philology will reveal things to us, of much more intimate and delicate bearing upon the mechanism of the soul and its variations, than physiology.

From the time of Aristotle, who said, 'We do not think without images, and words are images,' until the almost contemporary group of the ideologists, the sensualist school has always understood the importance of language. James Mill is of their school on this point; his general Grammar resembles that of Condillac or of Destutt de Tracy. His authorities are Horne Tooke and Harris. A long exposition of doctrines which have been left far behind since the author's time, would be useless here. A few words will suffice.

After having spoken of the simple states of consciousness, we must pass, he says, to the complex states. But all these imply, in some manner, the 'process of naming.' We must, therefore, first see in what this 'artifice' consists. It consists of 'inventing' signs or marks which we impose upon sensations and ideas. 'Substantive' names are marks of ideas or of sensations; adjective names are marks placed upon substantive names, or marks upon marks, in order to limit the signification of the substantive, and instead of marking one great class, to mark a subdivision of that class. Example: a 'great' man. The verb is also a mark upon a mark.

Three different sorts of marks render predication or affirmation possible. 'I have the name of the individual, John, and the name of the class, man. I can place in juxtaposition my two names, John, man. But it is not sufficient to effect the communication which I desire to make, that the word 'man' is a mark of the idea of which 'John' is a mark, and a mark of other ideas

with those; that is, those of which James, Thomas, etc., are the marks. In order to execute my design completely, I invent a mark, which, placed between my marks John and man, fixes the idea that I wish to express, and I say, "John is a man." In every language the verb which denotes existence, has been employed to respond to the design of adding the copula in the affirmation.'

The method of the author, which is that of the eighteenth century, is unacceptable on several points, and is now generally rejected. It has the primary defect of explaining natural things artificially, of believing in too much regularity in the march of the human mind, of not allowing a sufficient place to its spontaneity. It has no feeling for that which is primitive, for that far distant epoch when the senses and the imagination predominated, and when the mind seized only upon living and concrete things. It treats language after the fashion of logic, and not of psychology. A second defect is, that these explanations are at most applicable only to the family of Aryan languages. We cannot see how the theory of 'marks of marks' can be applied to the agglutinative or monosyllabic languages.

Thus Mr. A. Findlater makes important reservations in the name of comparative philology (Note 53). This theory of affirmations, he says, is in conformity with the phenomena of the family of languages known as Indo-European. Logicians, in fact, in treating this subject, have never taken into consideration any other languages than Greek, Latin, and the modern literary languages of Europe. It may then be presumed that this theory would not apply to languages of a totally different structure. The mental process must, no doubt, be the same in all, but the means are new and without precedent. If the naturalists had wished to construct a type of animal organism without having ever seen anything but vertebræ, their theory would certainly have failed in its generalization. In the same way, the current theory of affirmation, considered by the light of a more and more profound knowledge of the organism of speech, seems to attach

1 On this point see Renan, De l'origine du Langage; Max Müller, Science of Language, vol. ii. chiefly.