1 Kant Immanuel. Werke, ed. Weischedel Wilhelm (Wiesbaden: Insel Verlag, 1960). IV. p. 82 = G.M.S., B98f. Hereafter, I shall use the standard German abbreviations of Kant's works and adopt the common practice of letting the letters A and B stand for the first and second editions respectively.
2 On G.M.S., B98f. Kant, to be sure, appears to state this proposition. I shall contend that this is in fact not the proposition he sets out to justify. H. J. Paton also finds this statement of the problem puzzling. Compare Paton H. J.. The Categorical Imperative, (5th ed.; New York: Harper and Row, 1967, p. 201.)
8 Prol., § 5n. = A41f.; Logik. § 117 = A230
9 This is how Kant puts it in the Preface. A student of the third Section would. I dare say, characterize it differently.
10 Analytic arguments may be constructed, however, with either analytic premises or analytic conclusions, or both.
12 Yet, in some of the analytic arguments there is such a connexion.
15 Following suggestion (1), one could alternatively say that the step which can elliptical argument makes is justified if, and only if, the omitted necessary premise is justified.
16 If the synthetic inference-rules are regarded as premises of the argument, then the argument would proceed on an analytic inference-rule. For example, the argument: Human actions are events./Therefore, human actions have temporally prior causes, employs a synthetic inference-rule. Yet, if the inference-rule is supplied as a premise, then the argument becomes: Every event has a temporally prior cause. Human actions are events./Therefore, human actions have temporally prior causes. Such an inference employs an analytic rule of inference. Kant is committed to this by calling the laws of logic analytic.
17 I am taking combustion as a chemical process defined in such a way as to involve the presence of oxygen.
18 This representation greatly conflates the argument.
20 G.M.S., B89. B95. Hence, it is also analytic that 'The principle of autonomy is the sole principle of morality.'
21 I wish to make two points in regard to the manner in which I have presented Kant's argument. First, I do not claim that the manner in which I have presented arguments (1) — (10), namely as arguments — the premise and conclusion of which are categorical propositions, mirrors the manner in which Kant presented them. What I have presented as arguments (1) — (3), for example, are not presented by Kant as arguments, but rather as propositions. Nevertheless. Kant does offer arguments that connect the conceptions in these propositions. Kant's presentation, however, is not always in the propositional form. He presents (9). for example, as an argument (B51f.). My reason for putting Kant's thought always as arguments, rather than as a mixture of arguments and propositions, is so that the distinction between analytic and synthetic argument, which Kant intended to clarify the structure of the Groundwork, can actually do so. Second, putting Kant's thought as arguments has necessitated my use of the indefinite article in the premises and conclusions of the arguments. Consider argument (I). Kant would hold the proposition: 'Every imperfect will is good, 'as false; yet he holds that imperfectly good wills, if there are any, act for the sake of duty. Hence, I have used the indefinite article in my formulation of the argument. When the argument is expressed as the proposition: ‘If an imperfect will is good, then it acts for the sake of duty,’ it may be put in universal form. Consider argument (9). Kant holds, prior to the argument of the third Section, that the proposition: ‘Every imperfect will is subject to a categorical imperative,’ may be false. Yet, he holds that imperfect wills subject to a categorical imperative, if there are any, are subject to the just formulation of the principle of categorical imperatives. Hence, I have used the indefinite article in my formulation of the argument. When the argument is expressed as the proposition: 'If an imperfect will is subject to a categorical imperative, then it is subject to the first formulation of the principle of categorical imperatives,' it may be put in universal form: 'Every imperfect will is subject to a categorical imperative only if it is subject to the first formulation of the principle of categorical imperatives.' Had Kant been consistent in his manner of presentation, had he presented us entirely with universal propositions in hypothetical form, and had he said that the first two sections contain analytic propositions, rather than analytic arguments, then we would not have so many difficulties. In point of fact, he presented a mixture of arguments and propositions and only said that the first two sections contain analytic arguments. Autonomy of the will is the property the will has of being a law to itself (independently of every property belonging to the object of volition). Hence the principle of autonomy is 'Never to choose except in such a way that in the same volition the maxims of your choice are also present as universal law.' That this practical rule is an imperative — that is. that the will of every rational being is necessarily bound to the rule as a condition — cannot be proved by mere analysis of the concepts contained in it, since it is a synthetic proposition. For proof we should have to go beyond knowledge of objects and pass to a critique of the subject — that is, of pure practical reason — since this synthetic proposition, as commanding apodeictically. must be capable of being known entirely a priori. This task does not belong to the present chapter. Nonetheless, by mere analysis of the concepts of morality we can quite well show that the above principle of autonomy is the sole principle of ethics. For analysis finds that the principle of morality must be a categorical imperative, and that this in turn commands nothing more nor less than precisely this autonomy.
22 G.M.S., B87f. Paton's translation.
24 Nor, of course, does it follow that it is synthetic. That would require a separate argument.