According to his own testimony, the transcendental deduction cost Kant "the greatest labour – labour, as I hope, not unrewarded" (B xvi). After at least three preliminary attempts, he was still not entirely satisfied with the fourth version which he published in the first edition of the Critique in 1781, and he undertook a fifth version for the second edition of 1787. This effort at re-formulation alone reveals how long and hard he struggled with a question that had been central since Descartes. Even the final version of the B edition hardly succeeds in presenting the kind of lucid overview which would permit the reader to follow the argument step by step, to grasp its various ramifications, or to consider and respond to plausible looking objections. The winding intricacies of the text, the disorienting references back and forth which mark the course of the argument, raise considerable difficulties for any attempt to present a clear and coherent interpretation of this section of the work. Thus Schopenhauer (Werke, II: 529), Heidegger (1929:  Chapter 6) and most recently Kitcher (1990: 61–90) have all preferred the version of the deduction provided in the A edition.