What I meant to point to by saying “part of what bachelor means” is the containment relationship between the concepts, but I think that you are right that the way I formulated is confusing.

Thanks for explaining the reasons why math couldn’t be analytic in Kant. I knew that for him it was synthetic, and that it is one of the differences with later analytic/synthetic divide in logical empiricists, but I wasn’t aware that those were the reasons.

I’m not sure this is quite right. After all, part of what 12 means is 7+5; but that, as Kant insisted, is not analytic at all, but synthetic. The reason is that Kant did not invent the sense of ‘analytic’ with which he is working — he gets it from Wolff’s scholasticized Leibnizianism. And in that system, there are two crucial elements that constrain what concepts can be analytically contained in other concepts: the concepts can’t overlap, they can’t repeat, and they must together constitute an exhaustive division. That is, two concepts can only be contained in another concept analytically as a single genus in a species; the only way two concepts can both contain another concept is if they share it as a single genus; and they must be distinguished in such a way that all the species-level concepts give all the possible differentia that can attach to the genus-level concepts. As Kant recognized, it is impossible to place most mathematical concepts in an analytic scheme like this. 12 = 7+5, but it also = 6+6, etc.; to handle this you’d have to break all three rules. So addition statements must be not analytic but synthetic, not an analysis of concepts but a construction upon them.

This, of course, is not the traditional analytic/synthetic divide as we usually think of it, so we perhaps need to keep in mind that there is more than one possible account of what the divide even is.