Our decisions need not be seen as resting on procedures that are merely instrumental in making judgments that are reliably truth-tracking. The procedures might be more directly related than that to truths about what is right or good, or about what we ought to do, or to principles that tell us what is true about these matters. And I have no metaphysical theory about the truth-conditions of such truths, except to say that as objective truths, they must be independent of the attitudes, decisions or actions that they are supposed to justify or for which they are to offer reasons.

Kant is not saying - about freedom or any other subject - anything of the form: "Not-p but we must assume that p." That's close to self-contradictory, like Moore's paradox: "p, but I don't believe that p".

We can never prove that we are free or integrate our freedom in any way into our objective conception of the causal order of nature.

From the beginning, there has been a tension in the reception of the Kantian idea of autonomy. If you emphasize the 'nomos' (the law), then you get one picture: the objectivity of ethics. If you emphasize the 'autos' - the self - you get the idea that we make the law. Kant never hesitated in his choice between the two emphases. He emphasizes the nomos (the universal and objective validity of the law).

I think the term "Kantian constructivism" as an oxymoron. Kant was a constructivist about mathematics, but not about ethics.

As Kant says, the contribution of any common laborer would be greater than that of the greatest philosopher unless the philosopher makes some contribution to establishing the rights of humanity.

Kant's system of duties constitutes a Doctrine of Virtue because the duties also indicate what kinds of attitudes, dispositions and feelings are morally virtuous or vicious.