The Transcendental Deduction of Kant never really worked.

  Mainly, I believe that the Leonard Nelson approach makes the most sense because the transcendental deduction of Kant never really worked to be able to combine reason and perception. As Kelley Ross put it,-- both must  have a deeper source.  That is non-intuitive immediate knowledge. But as Michael Huemer pointed out, that there is no reason to believe implanted knowledge has any relation to reality.  Therefore, immediate non-intuitive knowledge (-the axioms by which knowledge starts) must be open to correction. It is not infallible. And if one asks-- that empirical facts ought not to be able to correct a priori knowledge, -that is where Hegel comes in handy. [Hegel would not have wanted empirical facts to correct an a priori, but even so even in Hegel, logic and reason flow through everything. Every fact is partly a priori.] There is a point where knowledge itself gets to a plateau, and gets above it by contrasting two points where separate series of reasoning led to and end up in a contradiction by which one starts the whole process again. Something like Electromagnetism and Newton that contradicted until you got Special Relativity,-- and Relativity and Quantum Mechanics contradicted until you got Quantum Field Theory; and to sew up gravity, you get String Theory. [In STRING THEORY, I think there is a way forward with Hashimoto Flow in differential geometry, i.e. to see how the closed strings develop in time.]

The issue here are the two points of Jacob Fries: empirical facts can not tell us anything about  how to make logical deductions. And logical deductions can't tell us if there is a continent between Europe and Asia.  

[There is no problem here from General Relativity because to Kant, space is a formal intuition.--Not a priori.]

There is another problem in Kant's transcendental deduction. That deduction says that the knowledge of transcendental subject depends on the knowledge of the transcendental object;-  and knowledge of the transcendental object depends on knowledge of the unity of the transcendental subject. The circularity here does not bother me, even though it probably should. What bothers me is that how can knowledge of object A be dependent on knowledge of object B? Or more precisely: Why is it that Kant says I can not know that I exist unless I know that something else exists?  I assume people in Europe knew that Europe existed before they knew that America existed. Therefore I think that Fries and Leonard Nelson were right that there has to be a deeper source of knowledge: immediate non intuitive knowledge.

I realize that there is a certain amount of stretching things to combine Fries and Hegel in the way that I am doing it here. But what I am doing here I think is somewhat like Plotinus when he combined Plato and Aristotle.] 

I might mention here that I have thought a lot about the Fries-Leonard Nelson approach for years and how it compares with the Hegel- McTaggart approach. [Though there are other approaches to Kant and Hegel but these seem the best of both. But in the long run I think I have to go with the Leonard Nelson non intuitive immediate knowledge idea.]