Truth and Convenience 303 we express ourselves in one way we find our ideas in confusion and our action impotent: if in another our ideas cohere har- moniously, and our action is edifying. The convenience of least disturbing vested ideas, and at the same time rearranging our views in accordance with new facts that come to our know- ledge, this is our proper care. But it is idle to say we do not know anything about things—perhaps we do, perhaps we don't—but we at any rate know what sane people think and are likely to think about things, and this to all intents and purposes is knowing the things themselves. For the things only are what sensible people agree to say and think they are. vii The arrangement of our ideas is as much a matter of con- venience as the packing of goods in a druggist's or draper's store and leads to exactly the same kind of difficulties in the matter of classifying them. We all admit the arbitrariness of classifications in a languid way, but we do not think of it more than we can help—I suppose because it is so inconvenient to do so. The great advantage of classification is to conceal the fact that subdivisions are as arbitrary as they are. Classification There can be no perfect way, for classification presupposes that a thing has absolute limits whereas there is nothing that does not partake of the universal infinity—nothing whose boundaries do not vary. Everything is one thing at one time and in some respects, and another at other times and in other respects. We want a new mode of measurement altogether; at present we take what gaps we can find, set up milestones, and declare them irremovable. We want a measure which shall express, or at any rate recognise, the harmonics of resemblance that lurk even in the most absolute differences and vice versa. Attempts at Classification are like nailing battens of our own flesh and blood upon our- selves as an inclined plane that we may walk up ourselves more easily ; and yet it answers very sufficiently.