The ideas of this paper have been fruitfully discussed with Ricardo Caracciolo. He has not always agreed with my points of view, so responsibility remains with me. A research stay as visiting scholar at Purdue University, due to a Fulbright scholarship, also allowed me to talk over some of the ideas of this paper with Patrick Kain, whom I thank. Cristina Redondo has also read this paper with extraordinary detail, her suggestions were very helpful. Allen Wood, Hugo Seleme, Hernán Bouvier, Gustavo Cosacov, Diego Papayanis and Rodrigo Sanchez Brígido have carefully read this paper as well. Their suggestions have been extremely useful. A Spanish version was published in Analisi e Diritto, 2011.

1. Rousseau, JJ, “The Social Contract” (1762) in The Social Contract and the First and Second Discourses, ed and translated by Dunn, Susan (New Haven and London: Yale University Press, 2002 Google Scholar) at 166 (Book I, Chapter VI).

2. Kant, Immanuel, Philosophical Correspondence, 1759-99 (Chicago: Chicago University Press, 1986 Google Scholar) at 132.

3. Kant, Immanuel, Critic of Pure Reason (1781-1787) ed and translated by Guyer, Paul & Wood, Allen (Cambridge: Cambridge University Press, 1998 Google Scholar) at 397 (B. 373).

4. See Smith, MBE, “Is There a Prima Facie Obligation to Obey the Law” (1973) 82 Google Scholar Yale LJ 950; Raz, Joseph, The Morality of Freedom (Oxford: Clarendon Press, 1986 Google Scholar) especially ch 4 [Raz, The Morality of Freedom]; also Raz, Joseph, The Authority of Law (Oxford: Clarendon Press, 1979 Google Scholar) especially ch 12 [Raz, The Authority of Law]; and Raz, Joseph, Ethics in the Public Domain (Oxford: Clarendon Press, 1994 Google Scholar) especially ch 15; Simmons, John, Moral Principles and Political Obligations (Princeton: Princeton University Press, 1979 Google Scholar); Caracciolo, Ricardo, “Existencia de Normas” (1997) 7 Google Scholar Isonomía, ; also “Realismo Moral vs. Positivismo Jurídico” (2000 Google Scholar) Diritto, Analisi e; and specially “La Relevancia Práctica de una Autoridad Normativa. El Argumento de las Razones Auxiliares” (1998 Google Scholar) Analisi e Diritto at 6; Gans, Chaim, Philosophical Anarchism and Polítical Disobedience (Cambridge: Cambridge University Press, 1992 CrossRefGoogle Scholar); Greenawalt, Kent, Conflicts of Law and Morality (New York: Oxford University Press, 1987 Google Scholar); Wolff, RP, In Defense of Anarchism (New York: Harper Torchbooks, 1970 Google Scholar).

5. For an analysis of the relationship between legitimate authority and (general) obligation to obey the law see Edmundson, William, “State of the Art: The Duty to Obey the Law” (2004) 10 Google Scholar Legal Theory.

6. For differences in anarchist doctrines, (a priori vs a posteriori; autonomy, community or equity based; philosophical vs political), see Simmons, John, “Philosophical Anarchism” in Justification and Legitimacy—Essays on Rights and Obligations (Cambridge: Cambridge University Press, 2001 Google Scholar). For a critical analysis of his classification see Iosa, Juan, “Anarquismo Filosófico y Anarquismo Político” (2013) 4 Google Scholar Revista de la Facultad.

7. Wolff, , supra note 4 at 71–72 Google Scholar.

8. Ibid at 18-19, 70.

9. See Nino, Carlos, Etica y Derechos Humanos, (Buenos Aires: Astrea, 1970 Google Scholar) at 370. Also Bayón, Juan Carlos, La Normatividad del Derecho. Deber Jurídico y Razones para la Acción (Madrid: Centro de Estudios Constitucionales, 1991 Google Scholar) at 604.

10. “No one has brought out the problematic aspect of authority better than Robert Paul Wolff in his In Defense of Anarchy…. Wolff’s insight was to see that the problem is not in the right to rule directly, but in the duty to obey the ruler which it brings in its wake. The duty to obey conveys an abdication of autonomy, that is, of the right and the duty to be responsible for one’s action and to conduct oneself in the best light of reason. If there is an authority which is legitimate, then its subjects are duty bound to obey it whether they agree with it or not. Such a duty is inconsistent with autonomy, with the right and the duty to act responsibly, in the light of reason. Hence, Wolff’s denial of the moral possibility of legitimate authority. This is the challenge of philosophical anarchism.” Raz, Joseph, “Introduction” in Raz, Joseph, ed, Authority (Oxford: Basil Blackwell, 1990 Google Scholar) at 4. “[T]he principle of autonomy entails action on one’s own judgment on all moral questions. Since authority sometimes requires action against one’s own judgment, it requires abandoning one’s moral autonomy.” Raz, , The Authority of Law, supra note 4 at 3 Google Scholar. In the same vein Bayón, ibid at 618-19.

11. Against this position see Ripstein, Arthur, “Authority and Coercion” (2004) 32 Google Scholar Phil & Public Affairs 1.

12. In 2.3.2 I will specify two ways of understanding the idea of a change in others’ reasons for action and, so, two ways of understanding the practical difference thesis. Here I am only interested in highlighting that there are conceptions of authority that do not understand the right to rule as the ability to make a normative change or a relevant practical difference in the categorical-practical reasoning of another. Among the most discussed are the conception of authority as morally justified use of power (see Ladenson, Robert, “In Defense of a Hobbesian Conception of Law” in Raz, Joseph, ed, Authority (Oxford: Oxford University Press, 1990 Google Scholar) at 32-54) and the conception of practical authorities as theoretical authorities over practical issues (see Hurd, Heidi, “Challenging Authority” (1991) 100 Google Scholar Yale LJ 7). I will not analyse these conceptions here for reasons of space and because, doubtless, the practical difference conception is the standard one.

13. Kant, Immanuel, “Metaphysics of Morals” (1797) in Practical Philosophy, The Cambridge Edition of the Works of Immanuel Kant, ed by and translated by Gregor, Mary (New York: Cambridge University Press, 1996 Google Scholar) at 379 (MS 6:224).

14. Wolff, , supra note 4 at 4 Google Scholar.

15. Lucas, John, The Principles of Politics (Oxford: Oxford University Press, 1966 Google Scholar) at 16 [emphasis in the original].

16. Raz, , The Authority of Law, supra note 4 at 12 Google Scholar. Although the Razian formula is presented in terms of reasons, these reasons are specifically duties, i.e., categorical reasons (for an equivalent definition but in terms of duties see Raz, , The Morality of Freedom, supra note 4 at 23 Google Scholar). So, it is also true to assert that X has authority over Y if from its directive ‘Y ought to do φ’ it follows or results that Y ought to do φ. Later I will present the particular Razian mode of accounting for the idea of duties dependent on the emission of commands in terms of protected reasons ( Raz, Joseph, Practical Reasons and Norms (Oxford: Oxford University Press, 1999 CrossRefGoogle Scholar) at 191 [Raz, Practical Reasons and Norms]) or exclusionary reasons ( Raz, , The Morality of Freedom, supra note 4 at 41–42 Google Scholar). Here I am merely interested in highlighting that the Razian definition underlines that “[w]hat one ought to do depends on who has authority in a non-relativized sense. That a person has authority according to some system of rules is, in itself, of no practical relevance. Just as one can draw no conclusions as to what ought to be done from any such conclusions from the mere fact that according to some rules authority is vested in Parliament.” Raz, , The Authority of Law, supra note 4 at 10 Google Scholar. In other words, to assert authority is a form of making non relative the relative statements of reason. When we have a theory of legitimate authority (and it is also true the contingent assertion that “the Legal System X, which states that φ ought to happen, has legitimate authority”) then we can pass from “legally or from the point of view of the legal system X, φ ought to happen” to “φ ought to happen.” Only the latter type of statements is entitled to participate in our practical reasoning.

17. Categorical duties are a type of reason for action that is not subject to the existence of a desire or interest to perform the action concerned. For altering our instrumental or prudential calculation, i.e., our hypothetical-practical reasoning, authority is not required.

18. First, notice that an authority can undoubtedly give a particular command (like “do X”) to an individual and thus generate a specific duty, perhaps a final duty for that person. I am not interested in analyzing here these types of situations given that the most important are those which intend to generate general and abstract duties. Second, I do not ignore the efforts of understanding the performance of authorities in terms of modification or specification of the factual premises (or the auxiliary reason in terms of Raz). For an effort in this direction see Nino, Carlos, Derecho, Moral y Política (Barcelona: Ariel, 1994 Google Scholar) at 122. Bayón, , supra note 9 at 605, 646 Google Scholar also refers to a similar idea when he asserts that “[i]t is not radically ruled out that in some areas or matters (…) an agent understands that there can be reasons for doing what the political authority orders, precisely on the basis that it has a more reliable knowledge of the facts relevant for determining what requires the balance of underlying reasons that he accepts” [translated by author]. The analysis leads Bayón to contend that the best knowledge of facts which are relevant for the determination of the concrete normative situation merely allows justifying theoretical authorities. As he regards the last one as the proper way to account for authoritative directives, he rejects the idea that authorities have the ability to make a practical difference (at 652). Here I will not offer an analysis of the feasibility of an understanding of authorities, within the practical difference thesis, in terms of auxiliary reasons. Even though I have no definite position on the matter, I do want to leave settled that I do not find such conception entirely implausible. Let’s consider the judicial authority. The courts have the normative premise of their practical reasoning (the sentence) given by the normative order. Problems of normative interpretation aside, judicial authority has primarily the power to authoritatively set the facts to be taken as true at trial—the factual premise of the argument—and to derive the solution from the premises. And indeed, it is plausible to understand the sentence, in its operative part, as a new reason for the parties, independent or exclusionary of the reasons that ground it. If we understand that judges make some difference in the practical reasoning of those under its jurisdiction, a plausible way to analyze it is in terms of a power for setting the factual premises or auxiliary reasons that are deemed to be true. Finally, if, as Raz claims ( Raz, , The Morality of Freedom, supra note 4 at 42–46 Google Scholar), it is plausible to use the model of judicial authority for the analysis of legislative authority, then it seems plausible (without abandoning the practical difference thesis) to analyze the legislative authority in terms of auxiliary reasons.

19. By “final duty” I mean the one that is the conclusion of a practical reasoning, what the agent must do once all things have been adequately considered; i.e., the duty which precedes action.

20. For a conception of authority as creating prima facie reasons see Reiman, Jeffrey, In Defense of Political Philosophy (New York: Harper Torchbooks, 1972 Google Scholar). For the idea of pro tanto reasons see Kagan, Shelly, The Limits of Morality (Oxford: Oxford University Press, 1989 Google Scholar) at 17.

21. Hobbes, Thomas, Leviathan (1651), Truck, Richard, ed (Cambridge: Cambridge University Press, 1996 Google Scholar) at 176 (Ch XXV). Certainly, Hobbes’s sentence can be interpreted in a conceptual, explanatory, or justificatory sense. Here I use the second. When presenting the voluntarist theory of justification I could use it in its justificatory sense. We can see another voluntarist conception of the nature of commands in the Hartian idea of a peremptory reason. See Hart, HLA, Essays on Bentham (Oxford: Clarendon Press, 1982 Google Scholar) at 253.

22. Authoritative directives are issued for reasons, which are deemed to justify them. Typically, these are reasons which show that the prescribed act is one which those subjects to the authority have good reason to perform” Raz, , Practical Reason and Norms, supra note 16 at 191 Google Scholar. Notice that this quotation from Raz suffers from the same ambiguity observed in the sentence from Hobbes quoted above. This is no more than a reflection of the relationship posited here between theses concerning the nature of commands and theses concerning their justification (or the justification of authority in general). On the other hand, I do not contend here that the epistemic theory of commands should deny that they also express an intention of the authority that the normative subject behaves according to the content of her beliefs. But this intention or will does not count here (as it does within the voluntarist conception) when determining the content of the authoritative command. Such content is determined by the underlying normative reasons that (the authority considers) apply to the situation that is to be regulated by the command.

23. The idea at issue should not be confused with analyses of practical authority in terms of theoretical authority on practical matters. For the latter conception, authoritative statements are not intended to change but merely to identify what the normative situation is, a situation that their directives do not alter nor even purport to alter. In contrast, for the rationalist or epistemic conception of authority within the practical difference thesis, the belief of authority reflected in its directives may change, in a way not yet specified, the agent’s normative situation. In other words, once inserted in a command, the commander’s belief about the reasons—independent of the authoritative command—that agents have, can work (or claim to work) as a reason for a belief about what reasons for action there are (without trying to change these reasons) or as a new reason for action. Here I am simply trying to account for the second idea. It may well be that after analysis we are compelled to abandon the idea that practical authority can be understood as the ability, based on epistemic grounds, to make a practical difference. Remember that here I am not intending to defend its feasibility but merely its presence in the contemporary debate. In fact, in my opinion a clear case of an epistemic theory of authority within the practical difference thesis is the conception of authority as service of Joseph Raz. Raz believes that one of the main services that a practical authority can offer, a service which also legitimates it as such, is an epistemic service (see Raz, Joseph, “Introduction”, supra note 10 at 6 Google Scholar, and Raz, , The Morality of Freedom, supra note 4 at 75 Google Scholar). Raz understands that in its directives, the authority must reflect the agent’s underlying reasons, those reasons for action that the agent has independently of the commands of authority. See Raz, , The Morality of Freedom, supra note 4 at 47 Google Scholar. As long as the authority knows those reasons best (its beliefs about this issue are more likely to be true and justified), individuals must follow an indirect strategy, guiding their actions by the directives of the authority in order to increase their conformity with such underlying reasons. However, as the commands will allow increasing compliance with the underlying reasons on condition that agents refrain from evaluating on their own, the commands are what Raz calls exclusionary reasons. See Raz, , The Morality of Freedom, supra note 4 at 46–48, 75 Google Scholar. Such reasons make a difference in the practical reasoning of others. So, Raz thinks that practical authority is both an epistemic authority and an authority able to make a practical difference. And it has the capacity to make a difference because it is an epistemic authority. See Raz, , The Morality of Freedom, supra note 4 at 48, 59–60 Google Scholar.

24. For reconstructing the relations between cognitivism, non-cognitivism, realism and antirealism, I found extremely useful the work of Bouvier, Hernán, Particularismo y Derecho (Madrid: Marcial Pons, 2012 Google Scholar).

25. See Gauthier, David, “Why Contractualism?” in Vallentyne, Peter, ed, Contractarianism and Rational Choice, Essays on David Gauthier’s Morals by Agreement (New York: Cambridge University Press, 1991 Google Scholar); Wolff, RP, The Autonomy of Reason, A Commentary on Kant’s Groundwork of the Metaphysics of Morals (New York: Harper & Row, 1973 Google Scholar); Reath, Andrews, “Legislating the Moral Law” (1994) 28 Google Scholar Noûs 435.

26. Gauthier, , supra note 25 at 23 Google Scholar.

27. Certainly it is not clear whether a theory like the one I am describing here can fairly be attributed to any of the classical contractualists. In any case, I’m not thinking of a specific philosopher but on the common sense that classic contractualists generated in their readers. For example Hart, : “on Hobbes’ view, the Sovereign in giving his commands which are law is exercising a right arising from the subject’s contract.” Hart, HLA, supra note 21 at 253 Google Scholar. Clear examples of the thought I have in mind are a posteriori anarchists. For (some) a posteriori anarchists there are no legitimate states because in fact there is no (and there is highly unlikely to be) the kind of universal consent that would constitute them as such. But if such consent would arise, if all of us would promise obedience to an authority, then it would have the right to rule. On the issue see Simmons, , supra note 6 at 122–57 Google Scholar.

28. By the way, for a contractualist theory as the one I try to sketch here, it makes little sense to distinguish between moral and legal duties. See Wolff, , supra note 25 at 224 Google Scholar.

29. From now on, and for simplicity in the exposition, when I refer to non cognitivist antirealism I will have in mind this type of theory. But I do not think that this is the only possible type of non-cognitivist antirealism. For example, a theory that conceives that the source of morality is God’s will, is non-cognitivist and antirealist regarding the divine commands. In other words, God does not command what he commands because it is good regardless of His will; on the contrary, the good is good because God commands it. Notice that such moral theory would keep the type of relationship postulated here with a voluntaristic theory of justification of authoritative commands and with a voluntaristic theory of the explanation of its content. It may seem that this is not the case. For even if what we must do is determined by the will of God, it may well be that there would be someone who would better know its content. Wouldn’t we have, in this case, a duty to obey that person as an epistemic authority? But that person would not be authority unless God wanted it. To put it in another way, suppose that we must do what God wants. X knows what God wants. But it is not part of God’s will that we follow the commands of X. Should we obey X? In my view this is not the case. I, at least, even knowing the divine will regarding the conduct of His subjects, would refrain from claiming to represent Him without His consent. I understand that the same caution, well that from the side of obedience, should show the normative subjects.

30. It is true that this type of theory seems to assume the existence of an objective duty to respect promises. This duty would result from a requirement of consistency inherent in our nature as rational agents. It seems that this is the minimum possible rationalism, necessary to account for the normative character of the will, i.e., its capacity to be a source of categorical reasons.

31. I am certainly not claiming that for realism the moral world is completely static and that the will of agents is not able to make any modification. But for realists, as Raz points out, “[r]easons precede the will. Though the latter can, within limits, create reasons, it can do so only when there is a non-will-based reason why it should.” Raz, , The Morality of Freedom, supra note 4 at 84 Google Scholar. For antirealists the adequate relation is the reverse: “The relevance of the agent’s concerns to practical justification does not seem to me in doubt. The relevance of anything else, except insofar as it bears on the agent’s concerns, does seem to me very much in doubt.” Gauthier, , supra note 25 at 20 Google Scholar.

32. Schneewind, Jerome, The Invention of Autonomy (Cambridge: Cambridge University Press, 1998 Google Scholar) at 4.

33. See Bayón, , supra note 9 at 643 Google Scholar.

34. For a defense of an epistemically grounded practical authority whose directives must be followed even if mistaken see Raz, , The Morality of Freedom, supra note 4 at 60–61 Google Scholar.

35. According to Raz (see Raz, , Practical Reasons and Norms, supra note 16 at 39 Google Scholar) a first-order reason is a reason for doing something or for refraining from doing something. A second-order reason is a reason to act for a reason or to refrain from acting for a reason. An exclusionary reason is a second-order reason to refrain from acting for some reason. In The Morality of Freedom, Raz characterized norms dependent on commands as exclusionary reasons (see Raz, , The Morality of Freedom, supra note 4 at 60 Google Scholar). In the postscript to the second edition of Practical Reason and Norms(see Raz, , Practical Reasons and Norms, supra note 16 at 191 Google Scholar). Raz rectifies himself and starts to consider them as protected reasons: “a systematic combination of a reason to perform the act … required by the rule, and an exclusionary reason not to act for certain reasons (for or against the act)”. Against the possibility of normative change from an epistemic conception of authority see Bayón, , supra note 9 at 690 Google Scholar, quotation 632.

36. Reath, , supra note 25 at 455 Google Scholar.

37. Ibid at 436.

38. For an analysis of the idea of content-independent reasons as a typical feature of the norms dependent on the emission of commands see Hart, , supra note 21 at 254 Google Scholar. This Hartian conception of norms relies on a voluntarist conception of authoritative commands (see Hart, , supra note 21 at 252 Google Scholar).

39. Therefore, I leave aside the analysis of other conceptions of autonomy such as the idea of personal autonomy. This idea refers to the ability of people to direct their own lives. Usually it is understood that a person has personal autonomy if she is free to be the kind of person she wants to be. But, as pointed out by Gerald Dworkin, it is clear that someone might want to be a good citizen, i.e., a law-abiding person. Therefore there is no conceptual contradiction between personal autonomy and authority. See Dworkin, Gerald, Theory and Practice of Autonomy (New York: Cambridge University Press, 1995 Google Scholar) at 27-28. For an analysis of the various conceptions of autonomy present at the contemporary debate see Iosa, Juan, “Concepciones de la autonomía” (2010) 234 Google Scholar Revista Brasileira de Filosofía.

40. Wolff, , supra note 4 at 14 Google Scholar.

41. Ibid at 13.

42. Scanlon, Thomas, “A Theory of Freedom of Expression” (1972) 1 Google Scholar Phil & Pub Affairs 2 at 216.

43. When stating the conceptual incompatibility thesis (leaving aside for a moment his oscillating use of the two conceptions of autonomy set out above) Wolff refers to the second idea. When we act because the authority has required some action, what happens is that we leave aside our own moral judgment. However, in his view, we remain morally responsible because we still have a duty of exercising that judgment. See Wolff, , supra note 4 at 14 Google Scholar.

44. I don’t intend to account for this procedure here. But it is clear that it must be reliable and accessible to all those who have the duty of self-judgment.

45. The idea of presenting this problem as a dilemma belongs to Cristina Redondo.

46. If moral realism and autonomy as self-judgment are compatible doctrines one would have a duty to conform to the real, existent moral reasons and also a duty to judge for oneself. In a case where the agent’s acts are based on beliefs that are false, the agent could have complied with the duty of autonomy but would not with the duty of acting in accordance with real, existent reasons. Though this is possible, the two doctrines are logically compatible because as norms they both can be complied with at the same time. I am grateful to the Editor of CJLJ for pressing me to clarify this point.

47. Kant, for instance, states that everyone has equal access to the moral law (G 4:391, 404, 411), and that, as rational agents, we have an obligation to act rationally, i.e., morally on a practical level (G 4:412). For quoting Kant’s, Groundwork of the Metaphysics of Morals (1785)Google Scholar, [abbreviated G] I use the canonical numbering of the Academy. I have used Kant, Immanuel, Practical Philosophy, The Cambridge Edition of the Works of Immanuel Kant, ed and translated by Mary Gregor, , (New York: Cambridge University Press, 1996 Google Scholar). If, moreover, as stated by Wood, Allen (Kantian Ethics (New York: Cambridge University Press, 2008 Google Scholar) at 112) and by Kain, Patrick (Self-legislation in Kant’s Moral Philosophy (2004) 86 Google Scholar Archiv für Geschichte der Philosophie 266), Kantian moral theory must be read in realist terms, then it would meet the three conditions set. Raz also counts among the realists who believe that, at least prima facie, there is an obligation of self-judgment. His theory of authority is precisely an attempt of exposition of the conditions under which it is justified to give up judging for oneself.

48. Reath, , supra note 25 at 435 Google Scholar.

49. Schneewind, , supra note 32 at 3 Google Scholar.

50. Reath, Andrews, Agency and Autonomy in Kant’s Moral Theory (New York: Oxford University Press, 2006) 122 CrossRefGoogle Scholar.

51. Reath, , supra note 25 at 435 Google Scholar.

52. Schneewind, , supra note 32 at 6 Google Scholar. See also Wood, , supra note 47 at 106 Google Scholar: “Kant’s ethical theory is grounded upon the idea that the moral law binds me only because it is regarded as proceeding from my own will.”

53. Today, most interpreters of Kant understand that autonomy is the source of categorical normativity in general. Those who, like Katrin Flikschuh, deny that this is the case will hold that, according to Kant, “autonomy as self-legislation is simply irrelevant” in the legal field. (See Flikschuh, Katrin, “Justice Without Virtue” in Denis, Lara, ed, Kant’s Metaphysics of Morals, A Critical Guide (New York: Cambridge University Press, 2010 Google Scholar) at 53.) This is to understand Kant as a supporter of the split of practical reasoning, i.e., as offering a ground other than autonomy (as the idea of external freedom) for legal obligation, so that legal and moral obligation do not stem from the same source. On my part, I am not committed here to any of the possible readings of Kant. It is enough for me to show the alternatives.

54. Wood, , supra note 47 at 107 Google Scholar.

55. Wood, , Autonomy as the Ground of Morality (1999 Google Scholar) online: http://www.stanford.edu/~allenw/webpapers/Autonomy.doc at 1.

56. Wood, , supra note 47 at 106 Google Scholar [emphasis in original]. The apparent internal inconsistency of the Kantian concept of moral autonomy is something now widely accepted on studies over the subject. See Reath, , supra note 25 at 435 Google Scholar; Kain, , supra note 47 at 264 Google Scholar; Dworkin, , supra note 39 at 39 Google Scholar.

57. See Wood, , supra note 47 at 110 Google Scholar.

58. “Consider this: Is the pious being loved by the gods because it is pious, or is it pious because it is being loved by the gods?” Plato, , “Euthyphro” in Cooper, J & Hutchinson, DS, eds, Plato Complete Works (Indianapolis/Cambridge: Hackett, 1997 Google Scholar) 10a. Likewise we could ask as follows: What we approve as correct, we do approve it for being—independently—correct, or is it correct because we approve it?

59. Wolff, , supra note 25 at 26 Google Scholar.

60. Bittner, Rüdiger, What reason demands (New York: Cambridge University Press, 1989 Google Scholar).

61. Reath, , supra note 50 at 164, n17 Google Scholar.

62. For the idea of “categorical imperative procedure” see Rawls, John, “Themes in Kant’s Moral Philosophy” in Beiner, Ronald & James, William Booth, eds, Kant & Political Philosophy. The Contemporary Legacy (New Haven: Yale University Press, 1993 Google Scholar).

63. Someone could certainly object that it is not necessarily true that justifying practical authority on epistemic grounds requires justifying the abandonment of autonomy as self-judgment. Raz himself can be seen as someone who justifies authority while denying that this implies forgoing autonomy. The issue is clarified by distinguishing between first-order autonomy (self-judgment on first order reasons) and second-order autonomy (self-judgment about the relevant reasons, including second order reasons). For Raz, and I think for any cognitivist, justifying authority requires justifying surrender of the first order autonomy. Authority does not require, however, giving up second order autonomy. Regarding this point see Raz, , The Authority of Law, supra note 4 at 27 Google Scholar.