I don't think Kant's approach to religion is any longer viable in its original form. But that does not mean it is simply wrong or that we cannot learn from it.

We can treat human responses to cognitions as involving law-like connections grounded on free choices which show themselves in our character.

Kantian ethical theory distinguishes three levels: First, that of a fundamental principle (the categorical imperative, formulated in three main ways in Kant's Groundwork); second, a set of duties, not deduced from but derived from this principle, by way of its interpretation or specification, its application to the general conditions of human life - which Kant does in the Doctrine of virtue, the second main part of the Metaphysics of Morals; and then finally an act of judgment, through which these duties are applied to particular cases.

When people think that moral problems can be solved by some simple strategy of calculation, that sets them up for ghastly overreaching. They think they can turn everything into a "science" the way mechanics was turned into a science in the seventeeth century. They want to turn everything over to technocrats and social engineers. They become shortsighted or simplistic about their ends, and they disastrously overestimate their ability to acquire the information they need to make the needed calculations.

When Marx, in the Theses on Feuerbach, says that only idealism up to now has understood the active side of material Praxis, what he says is more true of Fichte than of any other philosopher in the classical German tradition.

It is often difficult to know about one's own era which philosophers in it will be remembered as the most important ones, but I think it is already clear that John Rawls is the greatest moral philosopher of the twentieth century.

Fichte thinks that the mutual recognition of one another as free beings belongs among the transcendental conditions of self-consciousness itself.