In conclusion: it is to be regretted that linguistic psychology attracts so few people, and that many recent treatises on psychology, excellent on all other points, do not devote a single line to language. Yet this study, especially if comparative, from the lowest to the most subtle forms, would throw at least as much light on the mechanism of the intelligence as other highly accredited processes. _Physiological psychology_ is much in vogue, since it is rightly concluded that if the facts of biology, normal and morbid, are being studied by naturalists and physicians, they are available to psychologists also, under another aspect. So too for _languages_: comparative philology has its own aim, psychology another, proper to it. It is incredible that any one who, with sufficient linguistic equipment, should devote himself to the task, would fail to find adequate return for his labors. CHAPTER III. INTERMEDIATE FORMS OF ABSTRACTION. Having thus acquainted ourselves with this new factor—speech—which as an instrument of abstraction becomes steadily more and more important, we can take up our subject from the point at which we left it. In passing from the absence to the presence of the word, from the lower to the intermediate forms of abstraction, we must again insist on our principal aim: viz., to prove that abstraction and generalisation are functions of the completely evolved mind. They exist in embryo in perception, and in the image, and at their extreme limit involve suppression of all concrete representation. This conclusion will hardly be contradicted. The difficulty is to follow the evolution step by step, stage after stage, and to note the difference by _objective_ marks. For intermediate abstraction, this operation is very simple. It implies the use of words; it has passed the level of pre-linguistic abstraction and generalisation. We may go farther, and—always _with the aid of words_—establish two classes within the total category of mean abstraction: 1. The lower forms, bordering on generic images, whose objective mark is the feeble participation of the word: it can indeed be altogether foregone, and is only in the least degree an instrument of substitution. 2. The higher forms, approximating to the class of pure concepts, and having as their objective mark the fact that words are indispensable, since these have now become an instrument of substitution, though still accompanied by some sensory representation. The legitimacy of this division can be justified only by a detailed comparison of the two classes. I.