2 Paton, H. J., The Categorical Imperative: A Study in Kant's Moral Philosophy (New York: Harper and Row, 1967).Google ScholarRoss, W. D., Kant's Ethical Theory: A Commentary on the Grundlegunfi Zur Metaphysik der Sitten (Oxford: Clarendon Press, 1954).Google ScholarLiddell, B. E. A., Kant on the Foundation of Morality (Bloomington: Indiana University Press, 1970).Google ScholarAmeriks, Karl, Kant's Theory of Mind (Oxford: Clarendon Press, 1982).Google ScholarHenrich, Dieter, “Die Deduktion des Sittengesetzes. Über die Gründe der Dunkelheit des letzten Abschnittes von Kants Grundlegung zur Metaphysik der Sitten” in Denken im Schatten des Nihilismus, Weischedel-Festschrift, ed. by Schwan, Alexander (Darmstadt: Wissenschaftliche Buchgesellschaft, 1975), 55–112Google Scholar.

3 This formulation of Kant's central problem was previously argued for in my “Kant's Application of the Analytic/Synthetic Distinction to Imperatives”, Dialogue 18/3 (1979), 373–391.CrossRefGoogle Scholar The present paper shows in §6 that Kant's overall aim must be emended. It is to show that “Every imperfectly rational being which conceives of itself as having a will is subject to the principle of autonomy as an imperative”. That Kant holds negative and positive freedom to be reciprocal concepts is again stated on G.M.S., 450 (Akademle pagination).

4 Ross asserts that Kant attempts to show the reality of freedom in this paragraph (73). Paton asserts that Kant attempts to show the necessity of presupposing freedom (217). Liddell claims that Kant shows that rational beings possessed of a will believe in free agency (217). Ameriks claims that Kant attempted a practically sufficient proof of the claim that we are transcendentally free (196, 201f.). Henrich sees this paragraph as a remnant of a justification of transcendental freedom that Kant worked on for a time but later abandoned as he realized its futility (74). See also his “Der Begriff der Sittlichen Einsicht und Kants Lehre vom Faktum der Vernunft” in Die Gegenwart der Griechen in neuren Denken. Festschrift fur Hans Georg Gadamer (Tubingen: J. C. B. Mohr, 1960), 197fGoogle Scholar.

5 Paton, Categorical Imperative, 224; Ross, Kant's Ethical Theory, 74; Ameriks, Kant's Theory of Mind, 204. Paton claims that Kant believed himself to have argued in par. 4 from our subjection to the moral law to freedom, but that Kant did not in fact make such an argument. Liddell, Kant on Morality, 220, on the other hand, takes these paragraphs as indicating Kant's awareness that the crucial second premise in the par. 4 argument, to be discussed later, has not been shown. Henrich, 70, contends that these paragraphs show Kant's awareness that whereas par. 4 showed what Henrich calls freedom of judgment, transcendental freedom has not been shown. That the authors hold Kant to be producing an independent warrant for freedom in pars. 10–16 is evident from Raton, Categorical Imperative, 225; Ross, Kant's Ethical Theory, 75; and Lid-dell, Kant on Morality, 222f., 229, 230.

6 Paton, Categorical Imperative, 225; Ross, Kant's Ethical Theory, 75; Liddell, Kant on Morality, 223.

7 See D. Henrich, “Der Begriff”, 108. Also my “Kant's Rejection of the Argument of Groundwork III”, Kant-Studien, 73 (1982), 169–190Google Scholar.

8 The four texts are: (I) “We can enquire whether we do not take one standpoint when by means of freedom we conceive ourselves as causes acting a priori …” (G.M.S..450); (2) “We now see that when we think of ourselves as free, we transfer ourselves into the intelligible world as members and recognize the autonomy of the will …” (G.M.S., 453); (3) “And in this way categorical imperatives are possible because the Idea of freedom makes me a member of an intelligible world …” (G.M.S., 454); and (4) “This better person he believes himself to be when he transfers himself to the standpoint of a member of the intelligible world. He is involuntarily constrained to do so by the Idea of Freedom …” (G.M.S., 455f.). For a non-standard interpretation of the circle see: Pierre Laberge, “L'espece de cercle dont, a ce qu'il semble, il n'y a pas moyen de sortir”, Dialogue 21/4 (1982), 745–753.CrossRefGoogle Scholar Paton is aware of these texts, but takes them as an indication of Kant's confusion (225f.). Ameriks is also aware of them (206).

9 Some, e.g., Ross, Kant's Ethical Theory, 75, 86, hold that the argument of pars. 10–16 shows the existence of freedom. Kant denies this (C.M.S., 459), and further says: “It [freedom] holds only as a necessary presupposition of reason in a being who believes himself to be conscious of a will…” (see also: G.M.S., 461). Ameriks presents the view that negative or transcendental freedom has been shown in par. 4, but that what the argument requires is positive freedom (205). He fails to note that since Kant has established in pars. 1–3 that negative and positive freedom are reciprocal concepts, a proof of the reality of the one entails the reality of the other (cf., G.M.S., 450).

10 See “Kant's Rejection” (cited above) and my “The Objection of Circularity in Groundwork III”, Kant-Studien (forthcoming).

11 To show subjection to the principle of autonomy as an imperative more than is mentioned here is required. See my paper, “The Objection of Circularity”. Observe that freedom—as the capacity of an agent to have done otherwise than what he in fact did—is not mentioned here or, for that matter, anywhere else in the Groundwork.

12 Textual justification of this point will follow.

13 “All beings which must act with a subjective awareness of acting as free agents are beings which, in a practical sense, are really free agents'” (Liddell, Kant on Morality. 215)Google Scholar.

14 “Consequently if freedom of the will is presupposed, morality, together with its principle, follows by mere analysis of the concept of freedom” (G.M.S., 447).

15 Ameriks interprets the argument for KP(2) in much the same way as Paton. They differ, however, concerning how the conclusion to Kant's entire par. 4 argument, namely, P(3), is to be construed. Paton correctly takes the conclusion to be PP(3): “…a rational agent can act, just as he can think, only on the presupposition offreedom: he must think and act as if he were free” (Paton, Categorical Imperative, 219). Ameriks incorrectly takes the conclusion to be AP(3): “Every rational being with a will is free” (Ameriks, Kant's Theory of Mind, 201f.). Ameriks argues that there is a contrast in Kant's Lectures on Ethics between a proof which is merely practically sufficient and one which is theoretically sufficient (Ameriks, Kant's Theory of Mind, 196). A proof of freedom would be theoretically sufficient if it showed how freedom is possible, i.e., explained the “noumenal mechanics” (202) of freedom; a proof would be practically sufficient if it showed that we are free, but was not theoretically sufficient. I believe this distinction has application to the second Critique, but not to the Groundwork. In the Groundwork Kant says: “Every being … is—in a practical respect (in praktischer Riicksicht)—really free; that is to say, for him all the laws inseparably bound up with freedom are valid just as much as if his will could be pronounced free in itself on grounds valid for theoretical philosophy”. Ameriks' interpretation requires Kant to have said something of the sort: “I have given a practically sufficient proof that every rational will is free”. Kant did not say anything of this sort. “In praktischer Riicksicht” modifies “free”, not “proof”, a word which Kant does not use here. Ameriks also hypothesizes that Kant might have “practical freedom” in mind (Ameriks, Kant's Theory of Mind, 201; K.r.V., A802=B830). But this hypothesis does not fit the text either. Kant says: “Every rational being … is—in a practical respect—really free …”. He does not say: “Every rational being has the property of practical freedom”. Kant himself explains “in praktischer Rikkskht” by the “that is” clause. Kant is, therefore, saying that to say a being is free—in a practical respect—is: (1) not to say that he is free, but (2) to say that whatever laws are entailed by freedom are valid for him. Suppose for some reason I have to think my hands as dirty and suppose that I desire clean hands. Then, the imperative: “Wash your hands”, is valid for me. If my hands are not in fact dirty, but I—for some reason—still have to think of them as dirty, then the imperative is still valid for me. But, if I know that my hands are not dirty, then the imperative is not valid. Similarly, if I have to think myself as free, then the principle of autonomy is valid for me. If I do not know whether I am in fact free, but I still have to think my self as free, then the principle of autonomy is valid forme. But. if I know that I am not free, then the principle of autonomy is not valid for me. In order to exclude the last case Kant argues in the Solution to the Third Antinomy that I can't know that I am not free

16 G.M.S.. 420; cf., Log. Einl. Anh. (IV 96).

17 Uh. e. Entdeck. I. Abs. A. Anm. (V 3, I If.).

18 Kant uses the phrase “practical judgment” in K.p.V'., 67f.

19 Henrich, “Die Deduktion”, 64f.

20 Henrich, “Die Deduktion”, 67.

21 Henrich, “Die Deduktion”, 68.

22 Cf.. Henrich, “Der Begriff” 108.

23 Paton. Categorical Imperative, 220.

24 G.M.S., 450. 454, 453, 455f.

25 Kant neither asserts the truth nor the analyticity of: “A rational being, who conceived of himself as having a will, cannot conceive of himself as having been directed from outside in regard to his practical judgment”. Yesterday. I thought I acted and I conceived of myself a s having a will, but today when I sit at my desk and reflect on the thesis of determinism, I think that yesterday when I conceived of myself as having a will and thus as an agent, I was wrong. What I took to be my practical judgment was really the result of factors outside my reason. Kant has no objection to such reflection on the thesis of determinism. What he is saying, however, is that when I conceive of myself as having a will, I cannot conceive of my practical judgment as the result of factors outside my reason. If I were to do this, I would cease to conceive of myself as having a will.

26 If we replace “who conceives of himself as having a will” with “who conceives of himself as having theoretical reason”, then we will get an argument of the same form, but which will show the necessity of the presupposition of the freedom of theoretical reason. Each of these arguments is distinct from the argument I presented under the title: A Final Attempt Along the Lines of Paton ami Henrich, for that argument presupposed the objective validity of rational principles. In the final analysis what Kant appears to be doing in his argument for P(2) is the following. He is arguing that any rational being who conceives himself as making a judgment (either theoretical or practical) must regard himself as free. Folglich, if a rational being conceives itself as making a practical judgment (which it does when it conceives of itself as having a will) it then must conceive of its practical reason as free.

27 Henrich, “Die Deduktion”, 67.

28 Henrich, “Die-Deduktion”, 67.

29 In fact, Henrich gave me the clue to the correct formulation of P(2). In the “Review of Schulz” (last par.) Kant writes: “For I can be indifferent to where the situation in which I am now to act originates; I only ask what I have to do now, and there freedom is a necessary practical presupposition …”. In Kant's claim: “Every rational being which conceives of himself as having a will must conceive of himself as free”. I take “which conceives of himself as having a will” as equivalent to “which conceives of himself as having causality in regard to action”, and as also equivalent to “which conceives of himself as an agent”.

30 Kant's argument, as I have construed it, is for the claim: P(2) “Every rational being which conceives of itself as having a will must regard itself as free”. To this some will object that Kant has not shown that every rational being must conceive of itself as having a will. I do not think that Kant can show this, since it is not analytically true. The concept of a rational being which only reflects is, as Henrich has pointed out, a perfectly consistent notion. There is, however, no reason for Kant to show that every rational being must conceive of itself as having a will. Kant's aim in Groundwork III is t o show the objective validity of the categorical imperative. If Kant can show (by the par. 4 argument and by the subsequent argument) that every imperfectly rational being which conceives of itself as having a will, i.e., which conceives of itself as an agent, is subject to the principle of autonomy as an imperative, then Kant can show a great deal, namely, that everyone who thinks he acts is obligated by the principle of autonomy. To object to this that there may be some rational beings who only reflect and, thus, never conceive of themselves as agents, and, therefore, that Kant's argument does not show such beings as obligated, is, it seems to me, to require too much of Kant.