There are other things in the photographic art remarkably similar to intellectual thinking. The gelatine film behaves very like the mental plasma: only one other physical object (so far as I am aware) is a better image of the plasm. In theory the object or phenomenon has no importance. Even when it has the quality we call 'beauty,' that is not a property of the bare object, for it is not seen by every person or animal with good eyesight; it is a sentimental effect associated with the object. Hence we might, if it were possible, ignore all objects except those which have value to us as signs of sentimental effects. But in practice we cannot do this. Objects are thrust upon our notice which we cannot avoid, and which have no sentimental interest for us. These objects are necessarily classified according to their phenomenal appearance only, and such ideas lack an essential characteristic of true general ideas. But we cannot prevent their formation in the mind, for generalisation is merely a kind of abbreviated memory, and, objects being once perceived, their recollection is to a great extent beyond our control. Artificial and adventitious utilities produce the same kind of one-sided generalisation. If society pays a man in fame or money to observe and describe certain things, his classification of them will be purely phenomenal. He will classify dogs with wolves and nightshade with potato, and will lump together the whole population of a country in one class, although it consists of the most divers elements--fools and philosophers, rogues and righteous, saints and sinners, patricians and plebeians. These are differences much more important than sameness of nationality, colour, race, or language. This practice, no doubt, gives symmetrical classifications. The greater classes are subdivided into subordinate classes, and these again into lower classes in a many-stepped series. Gradation occurs also in true generalisation, but not to the same extent. If we confine our observation to things that are much like each other, the average idea will not be greatly different from a particular idea: this is called _lowness_ in generality. If we run together quadrupeds, bipeds and fishes, we shall have a much higher general idea: the average will be very unlike any concrete animal. The higher we generalise the smaller becomes the content of the idea, but the wider its extension, that is, the realm of objects from which it has been drawn, or which it is considered to represent. The usual practice is to generalise by fine gradations. Get the general idea of sheep, then of cow, then of horse; then average the averages. The result is much the same if we run all the objects together and average them in one operation, but the slower process gives the neater results. The gradations of generality are distinguished by names such as (beginning from below) variety, species, genus, class, family, kingdom.