yet untouched by anything which we can understand as science, the child is taught that this dog is called “Phylax” (without needing to learn the ground for this name), just as it is taught that both this animal and those which accompany its neighbour on the chase “are dogs,” i.e. have this name in common. The difference is, that in order to apply this name rightly the child must learn to know the ground for it. We do not call all four-legged animals “dog,” but these with the grave looks, which attract attention by their “barking”; other larger quadrupeds with manes are called “horse,” while both dogs and horses are called “animal”. Upon the ground of this easy discrimination by rough universal ideas, there begins with the learning of characteristics which do not force themselves upon immediate perception, the more specific naming of particular groups within an already established whole, and the comprehension of several wholes within the limits of larger wholes; for at first it is true that the more universal the idea the more indefinite. But while all practical knowledge consists and develops in the knowing of these specific universal ideas and names, theoretical interest depends much more on generalisations and their more accurate grounding and determination by actual characteristics. Thus side by side with the universal ideas, such as horse, dog, animal, which are elaborated into concepts, there arise new concepts, which afterwards become universal ideas, such as mammal, vertebrate mollusc, and finally concepts of living beings of which not only do the common characteristics remain unknown without study, but which are themselves imperceptible for the natural senses, e.g., the concept “bacillus”. But in all these actual constructions of concepts nothing more takes place than the connecting of many presented objects into a single new apperception-mass, which possesses fewer characteristics in proportion as it is more universal. There is no essential difference when the objects or concepts are not things, but qualities or events. They are always just particular—sensuous or non-sensuous—impressions, to which there is attached a name, which now shows itself to be applicable to many such impressions. None of these concepts, any more than the natural universal ideas, are “abstract” concepts in our sense, but the names attributed to them denote many concrete objects in reference to certain characteristics common to them all. Of course it makes a great difference whether we intend to denote objects or ideas; the universal is not in the objects, but in the ideas.
52. We do not form an “abstract” concept until with the
