In most contexts, when people talk about Kant's Critique of Pure Reason, they mean to denote any of the versions which combine both the A and B edition. The history of this usage is pretty interesting. According to (I'm pretty sure it was) Safranski (but it might be Beiser and/or Förster), by the mid nineteenth century nearly everyone would have taken themselves to denote the B edition with the book's title. But Schopenhauer became a vocal defendant of the A edition, which was actually only brought back into print because he championed it. At some point the combined version was released, which is the base for all current translations (as far as I know).
Contrast this with Calvin's Institutes of the Christian Religion. From 1536 to 1559 it went through five Latin editions.* In this transition it evolved from a much smaller question and answer type catechism to an over 1000 page treatise of four books. I'm not a scholar of these things, but I think that it would be impossible to do for Calvin's text what we have done for Kant's. With Kant's, enough of the material is the same and in the same order in both editions, so you just have to add the A stuff that's not in B and note the B stuff that's new. With Calvin's text, if you just mention the book out of the blue people are overwhelmingly likely to think that you intend the latest edition. But why is this the case?
I think we generally automatically defer to the intentions of the author later in life about what should be intended by uses of the books title. If Calvin wanted to rework the original catechism so much, then he must know best. My impression is that this is how people took the B edition of the First Critique to be before Schopenhauer had the audacity to argue that the later time slice of Kant did not know best. So we came up with the A/B edition as a compromise between Schopenhauer and Kant.
I'd be interested in other cases where an artist's intentions about either the identity criteria of their work or (probably equivalently) the proper denotation of the title of the work are later discarded. I don't know that much of the literature about "the intentionalist fallacy" other than that Wimsatt and Beardsley went too far in their original article. But if I remember right, the relevant bits of authorial intent involved interpreting the text, which is quite a different question from what religious scholars call "establishing the text."
The Schopenhauer/Kant thing presents an obviously important issue with respect to canonical texts such as the Bible or the Presbyterian Book of Confessions. The latter text is actually established democratically by the church as a whole, though not via simple majority. New confessions can be added and previous one's deleted or altered only if the majority of delegates at the General Assembly vote to do so, and 2/3ds of all Presbyteries reaffirm the vote within the next two years. Interestingly, this kind of democratic church governance traces back to John Calvin's reforms. But we don't have the same freedom with respect to the Bible itself, which is odd because: (1) scholars disagree about what the original texts might have included (we just have inconsistent versions which have to be redacted), (2) during the Reformation there was a lot of disagreement about which books should be included, (3) reformed Christians have a lot of latitude about the interpretation of scripture (anti-intentionalism), but none about the constitution of scripture (intentionalism). But why should I have to accept that the more conservative epistles of Saint Paul, the ones most scholars don't think were actually written by him, are actually in the Bible?
That is, I'm not sure that (3) is a very coherent position. I just don't know. To the extent that there's a problem, it might be a problem for any theology (to be clear, which I accept) that can be summed up with "scripture alone, faith alone, grace alone."
*A semantic anomaly. When people talk about the canonical "fourth edition" they mean to reference the fifth Latin one. I'm pretty sure that this is because the second Latin one was the first to be translated into French.]