One question, five answers.
And you? want do you think? What was Kant’s main mistake?
Henry Allison: Obviously, there are many things I wish that Kant had never said (for example, some of his statements about women, sexual morality, and the like) and many arguments which leave much to be desired. In addition, there are the areas (such as mathematics) where Kant`s views have been rendered obsolete by subsequent developments. But none of these count as “mistakes” in my view. If I have to select one such mistake, I believe that the chief is his regrettable decision to publish his “Ueber ein vermeintes Recht aus Menschenliebe zu luegen.” Not only does Kant here misrepresent his own views on the matter, but it has made it all too easy for critics to attack a distorted picture of his moral theory.
Beatrice Longuenesse: I think Kant’s main mistake lies in the way he thinks about the relation between his theoretical and his practical philosophy. He thinks he can salvage the notion of freedom indispensible to his moral philosophy only by reintroducting as objects of faith the metaphysical truths he had denied as objects of knowledge. I think this is a confusing and unnecessary move, which sends him straight back into the pre-critical, metaphysical age he had wanted to break away from.
Lourdes Flamarique: The “anthropologization of philosophy”: In the Kritik der reinen Vernunft Kant builds metaphysics on the basis of the natural inclinations of reason and the recognition of the limits of the intellect. Accordingly, Kant outlines an ontology with consciousness as the central point.
To put it in Aristotelian terms, I think that Kant’s biggest mistake is his hylemorphic approach to human knowledge. For this amounts to think knowledge in terms of “poiesis” instead of “praxis” (in the Aristotelian sense). This is a huge mistake, whose only justification is Kant’s own interest in explaining the method of physical science against the background of an empirist tradition.
Paul Guyer: In my view, Kant’s chief mistake was to assume that we can have complete certainty about the most fundamental principles of nature, and to base his distinction between appearance and reality on this assumption, which distinction then allowed him to treat freedom of the will as something that automatically exists in the unknowable realm of noumenal reality rather than something that must be painstakingly realized within the limits of nature. One of Kant’s chief worries was that if the freedom of the will could not be demonstrated then people could use determinism as an excuse for the immoral actions they are all too naturally disposed to perform; but transposing freedom of the will from the sensible world of appearances to the supersensible realm of reality could have precisely the same effect. Kant did realize that people must learn to discipline their inclinations in the natural world, but he should have recognized that our freedom and rationality can be achieved only in the natural world and within the limits of nature, and that our chief task is precisely to learn how to do this.
Patricia Kitcher: One major mistake was assuming that scientific problems were amenable to philosophical solution. Despite his careful distinction between method in mathematics and in philosophy, he still offered transcendental idealism as as solution to problems such as the infinite divisibility of lines. Assuming that the problems raised by Newtonian mechanics could never be resolved by science, he again tried to offer philosophical solutions to questions about the nature of space and time.
Kant's philosophy 
RE: Henry Allison:
I don’t think he misinterprets his own view in “supposed right to lie”, I think rather that his argument is on a much more narrow subject than readers usually assume.
RE: Beatrice Longuenesse:
Longuenesse shows here her fundamental misunderstanding of Kant’s Practical Philosophy, and thereby also her misunderstanding of Transcendental Philosophy generally. When she says, “He thinks he can salvage the notion of freedom indispensible to his moral philosophy only by reintroducting as objects of faith the metaphysical truths he had denied as objects of knowledge.” She shows her suspicion (granted a common one) about Kant’s moral project – that somehow he is cheating or forcing an opinion of his own.
RE: Paul Guyer:
Coming from a translator of Kant that I respect I’m rather horrified to see Guyer’s statements here. I’m not even sure exactly what he is talking about. Once again, more suspicion: “One of Kant’s chief worries was that if the freedom of the will could not be demonstrated then people could use determinism . . .”
RE: Patricia Kitcher:
The questions that Kitcher mentions as being resolvable by science are not the same questions that Kant is dealing with. Kitcher seems to not understand the project of Transcendental Philosophy.
* * *
For my own purposes, there is no sense in looking for errors in Kant. Looking for errors in philosophy is an interpretive heuristic that leads nowhere. It isn’t that we can’t find errors, but finding an error doesn’t mean you can’t also resolve it with further reading. However, if you take finding an error as the telos of your interpretation, when you find errors you will be satisfied in having found them and come to the end of your pursuit.
Maybe I would rather ask: what would I like Kant to have spoken about? The first thing that comes to mind is the anatomy of moral disagreement – something that didn’t have a place in the second critique (though he could have put it in the Doctrine of Method there – instead he discussed pedagogy, which is also noble), and which he doesn’t detail in his Anthropology.