There is no consensus concerning how to understand the ‘two-step proof structure’ (§§15–20, 21–7) of the Transcendental Deduction in the B-edition of the Critique of Pure Reason. This disagreement invites a closer examination of what Kant might have meant by a ‘transcendental deduction’. I argue that the transcendental deduction consists of three tasks that parallel Kant’s broader project of a ‘critique’ of pure reason; first, an origin task to justify reason’s authority to use them; second, an analytical task that determines (...) the conditions under which this authority can be legitimately exercised; and third, a dialectical task to determine the conditions under which this authority cannot be legitimately exercised. So long as we continue to read the B-Deduction solely in terms of its two-step proof structure, we overlook how Kant’s notion of ‘critique’ constitutes the real grounds for his argumentative strategy there. (shrink)