Now for two things to differ numerically only is to be sheerly different, that is, for one to be different from the other while not differing from it in any character or relation except that of being other than it. Otherness, in such a case, would not depend on dissimilarity, but could still exist when the characters of the two things were even in theory indistinguishable. I must confess that to me this makes no sense. It is an attempt to maintain a difference where there is not even a possible ground of distinction. If A is to be other than B, there must be some element of content, some attribute or relation possessed by one which is not possessed by the other; if there is none, on what ground could it be called other? We do not first note two things to be two, and then note, perhaps with surprise, that they are at some point dissimilar; it is because they observably differ in some respect-in quality, in location, in time-that we call them two. We never ascribe twoness or otherness in the absence of any perceived distinction at all. If, looking at a lamp-post, I say 'Both lamp-posts are green,' you may well answer, 'I see only one lamp-post; where is the other?', and if I insist that there are two, though oddly enough they do not differ in colour, size, shape, or location, you would begin to harbour suspicions about my health of mind. For otherness is based on dissimilarity, not dissimilarity on a logically prior otherness.

Brand Blanshard, Reason and Analysis.