What does the Transcendental Deduction prove, and when does it prove it? Henry Allison on Kant's Transcendental Deduction

Henry Allison's recent book on the Transcendental Deduction complements his long-term emphasis on the "discursivity" thesis: having separated intuitions and concepts, Kant intends the Transcendental Deduction to re-connect them by demonstrating that the categories must apply to all intuitions that are possible for us. In Allison's view, Kant successfully argues for this conclusion in the final stage of the second-edition Deduction, when he shows that our representations of space and time have a synthetic "unicity" that embraces all the intuitions we may ever have and depends on the use of the categories. I question whether Kant can justifiably introduce a synthetic representation of space or time that comes between the pure forms of intuition and the empirically-dependent representation of a world of objects in space and time.