Suppose, however, he is content to strip the "real" thing of what are commonly called the secondary qualities of matter, and to leave to it what are known as the primary. He will follow the example of the wholly unreflective man and speak of it in such a way as to suggest that the thing itself is something apart from its "real" qualities. A tree, he will say, _has_ qualities. It would certainly sound odd to hear him say it _is_ qualities. And he will very possibly go on to justify the use of the language he employs by distinguishing between the "real" qualities represented by his mental picture of the tree and an obscure something which he assumes as underlying them; thus embracing the Lockian distinction of ideas, "real" qualities, and substance. He may conclude, it is true, that substance in this sense of the word is chimerical, and that the belief in it arises out of a misunderstanding of the significance of language; but if he has gone so far as to assume duplicates of things immediately known, in the form of "real" qualities, it is more probable that he will be inclined to complete his classes of beings by adding the third.[4] [4] There is, of course, no reason why he may not add as many more classes as he pleases, and justify the additions as he justifies this. Men do not do this, as a matter of fact, but that is no reason. Now, it does not concern me to consider whether this change of view is to be regarded as a real progress in reflective knowledge or as a progressive decline and fall of the unreflective man. The point which concerns me is this: The unreflective man talks as if but one tree were under discussion. The man who reflects uses the same forms of speech: and even when he believes that he must distinguish between the tree immediately known and the obscure something which he has come to look upon as its cause, or between the tree immediately known, the bundle of "real" qualities inferred, and the obscure something that he connects with these, he still goes on talking as if he had only one thing to talk about. The danger of such a proceeding is obvious. If I talk about two or three things as though they were one, it is but natural that I should sometimes confuse them with each other. Should proof be forthcoming for one of these, it would be but natural for me to fall occasionally into the error of supposing that it somehow applies to the others. If I go on saying "the tree" when I mean one tree and something else, two trees, or two trees and something else, it is only to be expected that I sooner or later come to grief in my reasonings.