John Stuart Mill was the most notable of those who had argued that our knowledge of the inner states of others depends on analogical inference....We have good reasons to reject Mill's account over and above those that Stein cites. Were the situation such as Mill pictured it, all such inferences would be unwarranted. For we would be entitled to make such inferences only if we already know that others resemble us, both in having an inner life of thoughts and feelings and in expressing it in much the same manner that we do. But this, if our situation in relation to others is indeed as Mill pictures it, is something that we do not and cannot know.
[Alasdair MacIntyre, Edith Stein, Rowman & Littlefield (Lanham, Maryland: 2006) 78-79.]
It is surely right to reject the view that our knowledge of the mental states of others wholly depends on analogical inference. But MacIntyre's further problem for Mill is not, I think, a problem. It is false to say that "we would be entitled to make such inferences only if we already know that others resemble us" in having a mental life expressed as ours is; if this were true, all analogical inference would be worthless. Analogical inference, however, is often a method of extrapolation, and it is perfectly reasonable to use it as such. When we do so, it yields not knowledge but defeasible presumption; but there's nothing wrong with that.