It will be seen that Mill here draws a line between sameness in sense first and the samenesses in which there is an element of duality. He also draws attention to the fact--a fact to which I have already referred--that successive mental elements, considered in themselves, are more likely to be confounded than material things, though these last may be quite as closely similar. Language shows how men overlook the duality of two similar feelings which differ only in time. They may speak of two similar objects as the same, as they frequently do, and yet they will not usually lose the sense of their twoness. They say _these_ objects are the same. But when they compare a feeling experienced to-day with one experienced yesterday, they say _this_ is the same feeling I had yesterday. There is nothing in the language used to indicate duality at all.
I have said that, in the extract given, a line is drawn between samenesses which imply duality and the sameness which does not; and yet such illustrations are used to represent the latter as a man, a table, and a ship--objects which are the same in sense third as well as in sense first, and which consequently imply duality, in some sense of the word. But, if one is not considering single members of the chain of experiences which, taken together, we call a man, but is considering the whole group as a unit, this difficulty disappears. This is evidently what Mill has in mind, and he cannot be taxed with inconsistency. One may, however, object to the statement that it is an improper use of the word same to speak of things merely similar as the same. The word has many meanings, and we can hardly say that any one of them is illegitimate. It is merely illegitimate to confound them. And one should not take quite literally the description of resemblance in the highest degree as "amounting to undistinguishableness." Strict undistinguishableness removes all duality, and consequently makes impossible what we call resemblance or similarity. To be similar, things must be distinguished as two. Finally, one may object to a treatment of samenesses which merely groups them into two classes, when there are at least seven kinds that should, in the interests of clear thinking, be kept separate. It is only by carefully marking such distinctions that fallacious reasonings are to be avoided. As, however, this discussion of samenesses is merely a side issue where it occurs in the Logic, it would perhaps be unjust to blame Mill for not going into it more fully.