Beaunis has proposed a classification of pleasures and pains which also has as its basis a single principle: the various modes of motion. He discriminates three classes of pains: the nervous centres may be inactive through insufficiency of motion; their activity may be in excess through exaggerated motion; or their activity may be suddenly checked by arrest of motion. The same classification is adopted for pleasures: inaction, activity, arrest. I am inclined to think this division preferable to the other. He has also attempted a detailed classification of physical (p. 176) and moral (p. 235) pains; but he gives none for pleasures.[91] Footnote 91: Beaunis, _Sensations internes_, chap. xxi. For my part, I am inclined to believe that a classification (in the exact sense of the word) of pleasures and pains is an impossible task. As these characters are very general, one can only establish exceedingly general divisions. As soon as we go beyond this, we are, in reality, classing internal or external sensations, percepts, images, concepts, modes of action, accompanied by a pleasurable or painful state, positive or negative, due to activity, overactivity, or arrest; but the modalities of the pleasurable and painful, which, besides, are infinite, are not classed in and for themselves. The varieties of physical pain, the simplest, commonest, and best studied kind, the easiest to isolate, and the most free from concomitant representation, have never yet been subject to a fixed classification, from Hahnemann, who reckoned them as 73 in number, to Beaunis, who enumerates 83. In short, the “science of pleasure and pain,” as L. Dumont somewhat emphatically calls it, belongs to the category of sciences which do not proceed by way of classification, since they do not as yet possess the material. We can only lay down extremely general divisions, and then proceed by _incomplete enumeration_. II. The emotions, at least the simplest and best defined, present themselves as psychic states having their own specific characteristics. They differ among themselves, not as one mode of pleasure or pain differs from another mode, but as one thing differs from another thing; in this way, they appear as _objects_ susceptible to classification. We have already said that two methods have been adopted. (1) The first strongly resembles the so-called artificial classifications, which might also be called concrete or synthetic. It takes the emotions as realities and places itself before them as the zoologist and the botanist place themselves before the varieties of animals and plants. It is empirical—_i.e._, it has no guiding principle; it classifies according to observation only, following external resemblances and differences.