I:\28962 Indiana Law Rev 47-2\47masthead.wpd PERMISSIVE JUSTIFICATION ERIC J. MILLER* INTRODUCTION Often, practical decision presents us with a zero-sum choice, so that whatever is gained by choosing one option is lost by choosing the others. A parent must choose which child gets to ride first on the bicycle; a doctor must determine which among some range of courses of treatment to follow; an architect must pick which style of structure best suits a site; a judge must decide which side wins in a dispute before her. In each case, the decision-maker must select one from a menu of options and reject the others. Where the results are all-or-nothing in this way, an agent does not have the option of endorsing multiple outcomes all at once. Her choice separates the options into two categories: winners and losers. Sometimes reason is decisive, so that the reason for favoring the winner is also the reason for rejecting the loser. But sometimes reason is indecisive: all the reasons prevail (every option is a winner), or none do (every option is a loser). Rational indecision presents a problem for zero-sum decisions: the range of available reasons for decision is greater than the range of available options or outcomes.1 Reason alone does not select the winners and losers. The decision- maker, rather than reason alone, is ultimately responsible for the outcome. Agent responsibility for decision-making raises problems for practical justification. In zero-sum decisions, if all the reasons are winning reasons, it is difficult to justify to the loser why he or she lost. When all the reasons are losing reasons, justifying losing is easy; justifying why one of the losers gets to win is hard. In either case, when called upon to justify her decision, the decision-maker cannot simply point to some decisive reason as requiring the outcome. In such cases, reason is indecisive, and under-determines the outcome of practical conflicts.2 Much easier, we might think, are those conflicts in which the answer is clear- cut because reason is decisive. Decisive reasons provide the decision-maker with a unique justification for their decision. Call this the decisive justification thesis. Some theorists, most notably Ronald Dworkin, go further and endorse what he calls the “bivalence thesis.”3 For Dworkin, a decision is justified only if the * Professor, Loyola Law School, Los Angeles. My thanks to John Gardner, Dan Markel, Ekow Yankah, Chad Flanders, Nicolas Terry, Carlton Waterhouse, Kerry Ryan, Matthew Bodie, Anders Walker, Sam Jordan, and Molly Wilson for reading and commenting on this draft. 1. See Ronald Dworkin, No Right Answer, 53 N.Y.U. L. REV. 1, 3 (1978). 2. Id. 3. Id. at 2 (introducing his “bivalence thesis”). The problem does not arise when there are multiple winning parties or winning options for the decision-maker to choose among. In such circumstances, the decision-maker may be able to favor them all; or where there are multiple losing parties or losing options, the decision-maker might be able to reject them all. See id. at 2-4. Furthermore, if the stakes of practical choice are trivial, the decision-maker may be absolved from justification. See id. Trivial choices, we might think, do not need much, or any, justification, and 690 INDIANA LAW REVIEW [Vol. 47:689 reason the winner wins also provides the loser with a reason why she lost.4 When reasons conflict, Dworkin believes, they always separate outcomes into winners and losers.5 The bivalence thesis holds that justification is decisive because the rational world is decisive, too.6 In the professional sphere, two features of institutional decision-making may drive the demand for decisive justification. First, a variety of professions, including law, medicine, and so on, typically throw up the sorts of non-trivial, high-stakes circumstances that press decision-makers to seek decisive justifications to insulate them from charges of arbitrariness or bias.7 Second, some professions, including the law, may require decision-makers to present certain of their choices as all-or-nothing.8 For example, legal justification sometimes demands that the judge act as if the parties can be neatly separated into winners and losers. These professional pressures might be thought to recommend a defeat model of rational conflict, one that entails decisive justification.9 The defeat model holds that our practical choices are justified just in case they are supported by reasons that prevail over all the eligible competing reasons for action. The defeat model preserves the zero-sum aspect of decisive justification: there are only two categories of reasons, those that win and those that lose.10 Among legal theorists, Ronald Dworkin famously adopts a version of the defeat model—his “one right answer” thesis.11 A strong version of the defeat model, comparativism,12 goes one step further, to hold that justification requires not only that the prevailing reasons defeat their competitors, but that they do so by outweighing them on some unitary scale of value.13 The defeat model responds to the worry that justification is a very serious business.14 Decision-makers often enter justifications because losers demand an so it may suffice to resort to a variety of more or less arbitrary procedural devices to deal with indecisive options (such devices as tossing coins, splitting differences, declining to decide, or changing goals). 4. Id. at 2. 5. Id. 6. Id. 7. See id. 8. Id. 9. See, e.g., JOSEPH RAZ, ENGAGING REASON: ON THE THEORY OF VALUE AND ACTION 46- 66 (2002) [hereinafter RAZ, ENGAGING REASON]. 10. Id. 11. See Dworkin, No Right Answer, supra note 1, at 2. 12. See RUTH CHANG, INCOMMENSURABILITY, INCOMPARABILITY, AND PRACTICAL REASON 1-22 (Ruth Chang ed., 1997) [hereinafter CHANG, INCOMMENSURABILITY]; RUTH CHANG, MAKING COMPARISONS COUNT 43-45 (2002) [hereinafter CHANG, MAKING COMPARISONS]; Ruth Chang, Comparison and the Justification of Choice, 146 U. PENN. L. REV. 1569, 1569-98 (1998). 13. CHANG, INCOMMENSURABILITY, supra note 12, at 9-10; CHANG, MAKING COMPARISONS, supra note 12, at 46-48. 14. See, e.g., RAZ, ENGAGING REASON, supra note 9, at 46-66. 2014] PERMISSIVE JUSTIFICATION 691 accounting of the reasons why they lost, and so justifications stave off charges of wrongdoing or claims that the decision is rationally unintelligible.15 To satisfy the losers, the defeat model contends, justification must be “unqualified”16 and decisive, rather than ambiguous, permissive, or indecisive. Only decisive reasons, ones that defeat competing reasons, are sufficiently weighty (so the story goes) to demonstrate that reason required the loser to lose.17 In competition with the defeat model is the one this Article endorses, which shall be called (following Raz18) the eligibility model of rational conflict, one that introduces the possibility of permissive justification. It holds that our practical choices may be justified even if supported by indecisive reasons, that is, ones that though not defeated themselves, do not prevail over all competing reasons. In such circumstances, this Article claims, the decision-maker has a normative permission to select any one of the competing options and would be rationally justified in so doing. An account of permissive justification, however, only gets one so far. It could be that, even if the demand for decisive justification is inappropriate in the ordinary course of practical action, it nonetheless applies in the law. The law could be a sufficiently high-stakes enterprise that only decisive justifications count,19 or the law could be the sort of enterprise in which the outcomes are always all-or-nothing:20 either way, there will be a winner and a loser. Or it could be that the law is a zero-sum game all the way down: legal argument is just structured as a battle between premise and counter-premise, such that asserting a premise necessarily rejects the paired counter-premise.21 In any of these scenarios, the defeat model asserts that the judge needs some prevailing reason to justify her argument or decision, so as to make it intelligible to the parties and 15. See, e.g., id. 16. RONALD DWORKIN, JUSTICE FOR HEDGEHOGS 89-96 (2011) [hereinafter DWORKIN, JUSTICE]. 17. See, e.g., id. 18. RAZ, ENGAGING REASON, supra note 9, at 47 (Raz describes what I call the “eligibility model” as the “classical” conception of human agency). 19. See William Lucy, Adjudication, in THE OXFORD HANDBOOK OF JURISPRUDENCE AND PHILOSOPHY OF LAW 206-67 (Jules L. Coleman & Scott Shapiro eds., 2002); see generally THOMAS NAGEL, MORTAL QUESTIONS (2000). 20. See TIMOTHY A. O. ENDICOTT, VAGUENESS IN LAW 58-75 (Timothy A.O. Endicott ed., 2000) [hereinafter ENDICOTT, LAW] (calling this phenomenon “juridical bivalence”); JOHN M FINNIS, NATURAL LAW AND NATURAL RIGHTS 280 (1980) [hereinafter FINNIS, NATURAL LAW]; Timothy A. O. Endicott, Vagueness and Legal Theory, 3 LEGAL THEORY 37, 61-62 (1997) [hereinafter Endicott, Theory]. 21. See, e.g., Duncan Kennedy, A Semiotics of Critique, 22 CARDOZO L. REV. 1147 (2001); Duncan Kennedy, A Semiotics of Legal Argument, 42 SYRACUSE L. REV. 1 (1991) [hereinafter Kennedy, Argument]; Duncan Kennedy, Freedom and Constraint in Adjudication: A Critical Phenomenology Contemporary Legal Theory, 36 J. LEGAL EDUC. 518 (1986) [hereinafter Kennedy, Adjudication]; Karl N. Llewellyn, Remarks on the Theory of Appellate Decision and the Rules or Canons about How Statutes are to be Construed, 3 VAND. L. REV. 395, 401-06 (1949). 692 INDIANA LAW REVIEW [Vol. 47:689 avoid wronging the loser. This Article shall suggest that though, on occasion, the law may require judges to select one or another outcome, the law may do so despite the underlying reasons proving indecisive. Furthermore, indecisive reasons are enough for full- blown justification, even in high-stakes enterprises such as legal adjudication. On this permissive model, demands for decisive justification set the bar too high. When faced with a range of indecisive reasons for decision, a decision-maker may be permitted to pick one among them as a reason for decision, and be fully justified in whichever one she picks (even if she is, on occasion, rationally precluded from presenting her decision in this fashion). According to the eligibility model of rational conflict, indecision turns out to be a virtue for liberal professions (like the law) that value a diversity of professional perspectives and styles in working through the various options that the world throws at us.22 Yet the dominant understanding of the practice of adjudication is that judges are not allowed the normative space afforded to other liberal professionals.23 Even if lawyers can adopt different practice styles—black-letter lawyer, legal reformer, commonsense practitioner24—judges, so the dominant position goes, cannot—or at least, cannot do so expressly, and so must present their reasons as producing decisive justifications.25 The permissive model suggests that matters are yet more complex at the level of adjudication, and raises important questions about judicial styles and institutional structures of legal justification.26 22. See RAZ, ENGAGING REASON, supra note 9, at 47 (Raz describes what I call the “eligibility model” as the “classical” conception of human agency). 23. See Kennedy, Adjudication, supra note 21, at 519-20. 24. See Introduction to Law Precedent Lecture 4 2000, http://users.ox.ac.uk/~lawf0013/ PrecedentLecture2.htm, archived at http://perma.cc/74CA-93GP (quoting Lord Reid). 25. See Kennedy, Adjudication, supra note 21, at 519-20. 26. See id. (describing the complexity at the adjudication level). The point of this Article is to provide a formal rather than substantive account of rational justification. Substantive questions are primarily addressed by the theory of value. See, e.g., id. at 56-57 (describing value). For example, one claim about the relationship between reasons and value is that decision on the basis of a reason should be distinguished from decision on the basis of reason. This seems to equivocate between two different meanings of “reason”: one in which reasons, no matter how weak, can provide rational justification for a given action; and another that considers actions or beliefs rational and justified only if they pass some more-or-less weighty threshold for the justification of an action. See, e.g., RAZ, ENGAGING REASON, supra note 9, at 73 (breaking down reason into categories: substantive and procedural). These two meanings are compatible if the former is a formal description of the relation between reasons and rational justification, or if the standards of rationality are sufficiently low. The two meanings conflict if it is not true to say that just any reason satisfies the standards of rationality. See id. Thus, on the latter picture, a decision on the basis of a reason may not be reasonable: the reason may not satisfy the relevant standards of rationality. My interest is a different one. I propose to describe the ways in which permissions and reasons interact and justify action, and identify two different senses of “complete” justification. One sense defines a “complete” justification as a decisive justification, such that reasons fully 2014] PERMISSIVE JUSTIFICATION 693 I. DECISIVE JUSTIFICATION Suppose (to use an example suggested by William Lucy27) that a patient is suffering from a medical condition that requires immediate treatment. Each available treatment has grave consequences and requires her doctor to undertake some serious and invasive medical procedure. The doctor, however, can pursue the treatments only one at a time, rather than in concert. In such a circumstance, one might think, a doctor ought to produce some decisive reason to justify whichever treatment she decides to pursue.28 More difficult yet is the case in which a doctor must choose which of two patients, A and B, should receive some life-saving treatment. In that case, if one patient receives the treatment, the other does not (and so dies). One limit on justification might be this: if the doctor’s decision is to count as justified, she must be able to provide some reason to select A to receive the treatment that is also a reason why B ought not to receive it. That is, we might suppose that justification, in its central case, is all-or-nothing, so that the reason for favoring the winner is also a reason for rejecting the loser. A comparable legal example is contained in the thought experiment suggested by Ronald Dworkin: Imagine a judge sending an accused criminal to jail, perhaps to death, . . . and then conceding in the course of his opinion that other interpretations of the law that would have required contrary decisions are just as valid as his own. Or a friend who insists that you keep a burdensome promise though he concedes that a different interpretation of what you said, which contains no promise, would be an equally successful report of your meaning.29 Dworkin thinks that indecision undermines justification (and that we can see that it does by imagining what it would be like to offer a permissive justification for the judge’s or the friend’s choice).30 Justification, Dworkin believes, is aimed at consoling losers. Only a decisive reason, he believes, provides the sort of reason that could count for the losing party.31 Where reasons conflict and are in equipoise, what the judge needs is some reason that could tip the scales to provide her with some decisive reason for sentencing the offender. The challenge presented by the decisive justification thesis is thus a narrow determine what an agent ought to do or to believe. Another sense identifies justification as “complete” even if indecisive, so long as the justifying reason is a rationally adequate one, that is, undefeated. In each case, some theory of rationality could hold that a formally complete justification proves substantively inadequate because of insufficient weight. 27. Lucy, supra note 19, at 244. 28. Id. 29. See DWORKIN, JUSTICE, supra note 16, at 125-26. 30. Id. 31. Id. 694 INDIANA LAW REVIEW [Vol. 47:689 one. It does not address, for example, substantive questions regarding the grounds of legal justification.32 Nor does it particularly address the circumstances that call for justification.33 It is simply that any attempt to justify a practical decision must provide a reason not only for the winner winning but also for the loser losing.34 The thesis does, however, have a significant payoff: it requires rejection of permissive justification.35 The decisive justification thesis states a claim about the sort of reasons rationally sufficient to support some act. The claim is that, not only does an agent have undefeated reasons for some act, but also that those reasons defeat any and all reasons another agent might have to not act.36 Accordingly, we might define decisive justification in the following terms: Decisive Justification: an act is rationally justified only if the reasons to act are undefeated and the reasons to not act are all defeated.37 The defeat model of practical conflict which undergirds decisive justification does not require that the reason for action be absolute.38 As Joseph Raz explains, reasons may be decisive but not absolute, and vice versa.39 Reasons are absolute if there is no possible world in which they could be overridden by some other conflicting reason. Reasons are decisive if they prevail in some conflict situation.40 Suppose I have an absolute reason to avoid meeting the person who will kill me on Saturday. I have promised to meet Alan in the park on Saturday, but Belinda has asked me to go to the movies at the same time. So long as Alan is not the man who will kill me, I have a decisive reason to meet Alan in the park. My reason to meet Alan defeats the reason I have to go to the movies with Belinda. However, the absolute reason I have not to meet my killer does not 32. See Joseph Raz, Permissions and Supererogation, 12 AM. PHIL. Q. 161, 161 (1975) [hereinafter Raz, Permissions]; see also, e.g., David Lyons, Derivability, Defensibility, and the Justification of Judicial Decisions, 68 MONIST. 325, 325-46 (1985) [hereinafter Lyons, Defensibility]; see generally David Lyons, Justification and Judicial Responsibility, 72 CAL. L. REV. 178 (1984) [hereinafter Lyons, Justification]. 33. See, e.g., JOHN GARDNER, HARM AND CULPABILITY 1 (A. P. Simester and A. T. H. Smith eds., 1996) [hereinafter GARDNER, HARM]; John Gardner & Timothy Macklem, Reasons, in THE OXFORD HANDBOOK OF JURISPRUDENCE AND PHILOSOPHY OF LAW 440-75 (Jules L. Coleman & Scott Shapiro eds., 2002) [hereinafter Gardner & Macklem, Reasons]. 34. See, e.g., Lyons, Justification, supra note 32, at 181 (discussing choosing one answer at the exclusion of all others). 35. RAZ, EMERGING REASON, supra note 9, at 28. 36. See, e.g., Lyons, Justification, supra note 32, at 181 (discussing choosing one answer at the exclusion of all others). 37. The decisive justification thesis entails that our choices to act or believe thus and so are justified only if the belief or act chosen is itself justified. 38. JOSEPH RAZ, PRACTICAL REASON AND NORMS 27-28 (1999) [hereinafter RAZ, PRACTICAL REASON]. 39. Id. 40. Id. 2014] PERMISSIVE JUSTIFICATION 695 apply here. Although it is absolute, it does not operate as a ground for decision in this case. Furthermore, as the Alan-and-Belinda example illustrates, the sort of reason that provides a decisive justification need not be an overriding reason. An overriding reason requires that the reason for action outweigh some competing reason: it requires that the winning and losing reasons are comparable on some scale of value. Decisive justification does not: it requires only that the reason that wins defeats the reasons that lose, whether the do so comparatively or non- comparatively. My reason to go to the park to meet Alan defeats my reason to go to the movies with Belinda, not because it outweighs my Belinda-based reasons (it may not; these reasons may be quite weighty), but because the act of promising excludes my Belinda-based reasons, and so non-comparatively defeats them. Even if my reasons to go to the movies with Belinda are, all things considered, stronger than my reasons to go to the park to meet Alan, I do not get to evaluate them in that way, because my promise precludes and replaces them as reasons for action.41 Decisive justification does not entail comparison: the decisive justification thesis holds that practical decisions are justified just in case they are based upon some reason that defeats the loser’s reasons for action.42 Otherwise, so the thinking goes, losing is arbitrary because irrational or a- rational, and those types of practical decision are merely an expression of fiat rather than principle.43 Ronald Dworkin proposed such a standard, which he termed the principle of bivalence, in an early article defending his “one right answer” account of adjudication.44 While the bivalence argument is primarily a linguistic one, Dworkin’s argument applies more broadly to reasons for action as well.45 His claim is that, for a decision to count as justified, the reasons that support the winner must also discredit the loser.46 The decisive justification thesis entails that justification contains no gaps in reason: if an option is justified, every reason either defeats or is defeated. There is no need for some middle term “undefeated”: from the perspective of decisive justification, the world of practical accountability is rationally determinate.47 Our acts are either justified or unjustified, with no hemming or hawing in between. This position is, perhaps, quite radical. It certainly has disastrous consequences 41. See id. at 35-40 (discussing exclusionary reasons and promises). 42. See Dworkin, No Right Answer, supra note 1, at 2; see also Lon L. Fuller, Reason and Fiat in Case Law, 59 HARV. L. REV 376 (1945) [hereinafter Fuller, Reason]; Herbert Wechsler, Toward Neutral Principles of Constitutional Law, 73 HARV. L. REV 1, 25 (1959). 43. Wechsler, supra note 42, at 11. 44. See generally Dworkin, No Right Answer, supra note 1. 45. The principle of bivalence is often associated with the principle of non-contradiction, which asserts that there are no gluts in truth values: no sentence is both true and false. See, e.g., Endicott, Theory, supra note 20. 46. Dworkin, No Right Answer, supra note 1, at 2. 47. Dworkin’s bivalence thesis is thus an application of his “no gaps” thesis: the world is rationally determinate, such that there is always one right answer to any practical problem. See id. 696 INDIANA LAW REVIEW [Vol. 47:689 for any theory of permissive justification. If justification is successful only when based on decisive reasons, then many types of normative permission, and many of our practical decisions or actions, are unjustified. Decisive justification thus sets the bar quite high when advancing standards by which to evaluate the rational acceptability of our actions. For example, even if they do not accept the existence of incommensurable or incomparable reasons for action, most people think reasons can be of equal weight and so undefeated, so that where reasons A and B conflict, A does not override B nor is it overridden by B, and vice versa. If reasons can be of equal weight (and Dworkin for one, recognizes that they can be, even in the law),48 then it is possible that the world of practical accountability (or the law) is rationally indeterminate.49 Decisive justification thus establishes an especially high standard by which we are to justify our actions in the world. The English or Commonwealth tradition of legal philosophy tends to adopt an institutionally limited version of the demand for decisive justification. For example, Timothy Endicott50 and John Finnis51 have each endorsed “juridical bivalence” as (in Finnis’s terms) an accurate “exegesis of the judgment inter partes.”52 Finnis and Endicott argue that it is in the nature of legal justification to represent judicial decisions as a series of zero-sum choices among the available options, whether or not the arguments underlying the decision are so limited.53 Finnis, for example, approvingly cites Dworkin’s “one right answer” thesis as accurately describing the “momentary legal dogma” that,54 in justifying the outcome of the case to the parties, the judge must act as if there is a winner and a loser if she is to respect the losing party’s appeal to the law.55 Even if the law is indecisive, the judge must act as if it is not. Accordingly, the law presents a judge engaged in the practice of legal justification with fewer resources than those available to her than her non-legal counterparts. Where reasons are indecisive, the judge cannot toss coins or split differences, but must instead provide some decisive-looking reason to justify her decision. In the American context, the demand for decisive justification is often more expansive. Where Endicott and Finns identify a constraint on the way in which 48. See generally Ronald Dworkin, Hard Cases, 88 HARV. L. REV 1057 (1975). 49. Unless one thinks that equally weighted reasons instantiate the same value. See, e.g., MICHAEL STOCKER, PLURAL AND CONFLICTING VALUES 165-68 (1990). 50. ENDICOTT, LAW, supra note 20, at 72-73. 51. JOHN M FINNIS, IV THE COLLECTED ESSAYS OF JOHN FINNIS: PHILOSOPHY OF LAW 14 (2011) [hereinafter FINNIS, COLLECTED ESSAYS]. 52. Id. 53. Id.; ENDICOTT, LAW, supra note 20, at 72-73. 54. See FINNIS, COLLECTED ESSAYS, supra note 51, at 13-14 (“the judgment that prevails in such a case (and in any hard case, as in any easy case) includes as part of its legal content or entailment the proposition that, just as the losing party’s appeal to legal rules or principles is (to the relevant extent) legally erroneous.”). 55. On this point, see, e.g., ENDICOTT, LAW, supra note 20, at 72-73; FINNIS, COLLECTED ESSAYS, supra note 51, at 13-14. 2014] PERMISSIVE JUSTIFICATION 697 judges represent the outcome of decisions, an American jurisprudential tradition adopts the defeat model to articulate a claim about the nature of practical or legal argument. For example, Dworkin’s contemporary discussion of adjudication as a process of “principled” decision-making56 is just the most recent contribution to a liberal American tradition that seeks to preclude the operation of judicial “fiat”57 in adjudication, and with it, the suggestion that the law is gappy.58 Thirty years before Dworkin proposed his bivalence thesis, Lon L. Fuller forcefully argued that the stringent requirements of decisive justification are appropriate because adjudication is: a device which gives formal and institutional expression to the influence of reasoned argument in human affairs. As such it assumes a burden of rationality not borne by any other form of social ordering. . . . We demand of an adjudicative decision a kind of rationality we do not expect of the results of [other types of social organization].59 In this tradition of decisive justification, the decision-maker is presented with a series of choices among arguments, and at each stage of the decision-making process, the judge is supposed to find some decisive reason to support her claim that one or other argument wins out.60 In all its various guises from Fuller, through Wechsler, to Dworkin, the idea that judicial decision must respect the rights of the parties through principled adjudication requires that the decision-maker demonstrate to the losing party that the state’s exercise of power over them is non-arbitrary because decisive. The reason they lost is the reason the winner won. Accordingly, the appropriate standard for legal (or political or rational) justification is that decisions be justified, not on the basis of an undefeated or indecisive reason, but on the basis of a decisive one. 56. See, e.g., RONALD DWORKIN, A MATTER OF PRINCIPLE 29-31 (1985) [hereinafter DWORKIN, PRINCIPLE]; Ronald Dworkin, The Forum of Principle, 56 N.Y.U. L. Rev. 1 (1981). 57. Lon L. Fuller is particularly influential in this regard. See, e.g., Fuller, Reason, supra note 42, at 378; Wechsler, supra note 42, at 11 (citing Fuller, Reason, supra note 42, at 378); NEIL DUXBURY, PATTERNS OF AMERICAN JURISPRUDENCE 226 (2001) (citing Fuller, Reason, supra note 42, at 378). 58. For example, Dworkin’s selection of Riggs v. Palmer to expound his theory of principled adjudication is strongly reminiscent of Cardozo’s discussion of the same case. See DUXBURY, supra note 57, at 217-19 (citing BENJAMIN N CARDOZO, THE NATURE OF THE JUDICIAL PROCESS 41-43 (1921)). 59. Lon L. Fuller, The Forms and Limits of Adjudication, 92 HARV. L. REV. 353, 367 (1978) [hereinafter Fuller, Forms]. 60. See, e.g., Llewellyn, supra note 21, at 401-06 (1949) (describing a series of canons and counter-canons that are interrelated as a series of doctrinal thrusts and parries). For a modern variant of Llewellyn’s argument, see Kennedy, Argument, supra note 21 (discussing legal argument as a process of selection among opposed “argument bites”). 698 INDIANA LAW REVIEW [Vol. 47:689 A. Tipping Decisive justification, if it is to get off the ground, requires the ready availability of decisive reasons.61 That is, the decisive justification demands not only that decisive reasons are required but also that they are generally available to justify our practical choices. Arguments supporting the availability of decisive reasons are often associated with some form of “comparativism”:62 the claim “that comparability is a necessary condition for justified choice.”63 Comparativism entails that an action or belief is justified only when the reasons supporting that action or belief are comparable under some scheme of value, and outweigh or override competing reasons. Where reasons are comparable in this way, some reason, such as personal commitment to some project or goal can almost always tip the scale where reasons or values appear to be of equal weight.64 On this view, small differences in value will always be able to defeat competing options.65 The tippability argument is perhaps the most famously encapsulated by the example of Buridan’s ass,66 which must decide which of two equally appetizing, but equally distant bales of hay to choose.67 If the bale of hay on the left is slightly larger or (if the same size) slightly closer than the one on the right, the ass has a decisive reason to choose the one on the left. Contrariwise, if the bale on the right is larger or closer, then the ass has a decisive reason to choose that one. Where the bales of hay are equidistant and of equal size, then there is a tie and neither bale is decisively more appealing. In the usual telling of Buridan’s tale, if the reasons are in complete equipoise, the ass, because indecisive, dies.68 61. RAZ, ENGAGING REASON, supra note 9. 62. CHANG, INCOMMENSURABILITY, supra note 12. 63. Matthew Adler, Law and Incommensurability: Introduction, 146 U. PENN. L. REV. 1169, 1172 (1998). 64. As Joseph Raz explains, on the defeat model, “[t]here are always factors—we call them reasons—that guide the agent’s choices and decisions.” RAZ, ENGAGING REASON, supra note 9, at 49. 65. As Joseph Raz points out, this is not a feature of conflicts among incommensurable or incomparable reasons for action, hence the requirement of comparability and reasons of equal value if the tipping argument is to work. Raz argues that “the rationalist conception, if not committed to complete commensurability, is committed to the view that incommensurabilities are relatively rare anomalies.” Id. at 48. Comparativists abhorrence of incommensurabilities” leads them to argue that “[t]here are always factors—we call them reasons—that guide the agent’s choices and decisions.” Id. at 49. 66. On Buridan’s ass and similar problems, see generally Nicholas Rescher, Choice Without Preference: A Study of the History and of the Logic of the Problem of “Buridan’s Ass,” 142 KANT- STUDIEN 51 (1960). 67. See, e.g., MICHAEL BRATMAN, INTENTION, PLANS, AND PRACTICAL REASON 11-12 (1987) (discussing the Buridan’s ass example); Edna Ullmann-Margalit & Sidney Morgenbesser, Picking and Choosing, 44 SOC. RESEARCH 757, 758-59 (1977) (same). 68. Another version of the problem proposes that the choice between bales of hay is trivial: 2014] PERMISSIVE JUSTIFICATION 699 The traditional, Leibnitzian solution to the Buridan’s ass problem claims that “there is never an indifference of equipoise, that is [situations of choice] where all is completely even on both sides, without any inclination towards either. Instead, Leibnitz thinks, there are likely to be “petittes perceptions”69— “unperceived impressions, which are capable of inclining the balance.”70 Further scrutiny will reveal which reason defeats the others by tipping the balance one way or the other. The comparativist version of the Leibnitzian solution is to claim that an agent’s goals or commitments provide a way to solve problems of choice among equally weighty options. A goal or commitment can provide an agent with reasons that decisively sway the balance. When an agent chooses between satisfying her hunger or her thirst she can do so by considering her goals and commitments to determine how these might give her an additional, balance- tipping reason. If the agent is committed to the enjoyment of food, she will have a decisive reason to eat the hay before drinking the food; if she is committed to the enjoyment of drink then she has a decisive reason to choose the water first. Since such differences are usually available at the level of our personal goals and commitments, we can always find some decisive reason—for ourselves at least—for choosing one action over another. Ronald Dworkin, though not a comparativist, has provided a neat example of the Buridan’s ass phenomenon.71 He suggests we consider a horse race in which the management of the track has purchased equipment for deciding among apparent ties, but that this equipment is somewhat imprecise. Though the equipment will narrow the cases in which a tie is a possible outcome, nonetheless on some occasions “it cannot be clearly established which horse has won, [and] they shall be deemed to have tied, in spite of the fact that superior equipment might have shown a winner.”72 For Dworkin, if the initial scale is not sufficiently precise to balance the options and produce some ultimate reason for decision, that may count as a reason for switching to a more fine-grained scale. Dworkin’s point is that real ties are extremely rare, so that with enough effort a decision-maker can almost always find tipping reason. Thus, while the “instruction [to eliminate ties] does not deny the theoretical possibility of a tie . . . it does suppose that a judge will, whichever bale the ass chooses, she gets the same thing. See, e.g., STOCKER, supra note 49. Confronted with equally appetizing, equally handy bales of hay, the ass should simply pick one. Things might be a little more complex if the ass has to choose between two equally balanced but significantly different options; for example, a bale of hay and a bucket of water. Here, we might think, the ass satisfies a different value in choosing either option: the value of hydration from water and of nourishment from the hay. Not so, a comparativist might argue; each of those sub-values in turn serves the further over-arching value of sustenance, and her choice can be tipped through determining how well each contributes to the overriding value. 69. Rescher, supra note 66, at 161. 70. Id. 71. RONALD DWORKIN, TAKING RIGHTS SERIOUSLY 285-86 (1977). 72. Id. at 286. 700 INDIANA LAW REVIEW [Vol. 47:689 if she thinks long and hard enough, come to believe that one side or the other has, all things considered and marginally, the better of the case.”73 If we suppose the judge does not merely engage in post-hoc self-deception, the we might think that the judge could, from within her set of commitments or goals, cast around for reasons to think that her preferred outcome is the best one that she can defend.74 These reasons will in fact tip the scales, such that one horse is revealed as the winner and the other as the loser, and so the outcome can be rationally defended even to the loser. Ruth Chang, who is a comparativist, provides a more complex example: that of two musicians in a music competition who are each equally (or incommensurably) good on some evaluative scale of musical talent.75 A central feature of this scale is that it is robust: it provides the grounds for justification. The competition judges cannot separate the musicians by turning to some non- musical-talent-based scale of value because the musicians demand to be judged based on the values under which they competed.76 Changing the scale of evaluation misrepresents the range of justifications on which the parties relied. A judge cannot justify her decision by telling the competitors that, for example, one had been chosen because she was more attractive. They did not ask to be judged on that alternative scale of value, and would justifiably criticise the move to some alterative, arbitrary or personal reason for breaking the tie. Dworkin similarly proposes non-arbitrary limits on the grounds of legal adjudication: justice and fit.77 On Dworkin’s “one right answer” account, legal justification requires the judge (or anyone engaged in legal reasoning) to produce the “best” reconstruction of the law given the judge’s theory of political morality in light of the case’s “fit” with pre-existing law.78 Fit excludes one source of purely personal tipping reasons: it requires the decision-maker to rely on the legal materials of a given community (rather than their own views about what the law should be) to develop some threshold criteria explaining what counts as law by incorporating as many of the uncontroversial legal cases as possible, explaining why the controversial ones are controversial, and providing some way of 73. Id. at 285. This injunction to think long and hard is strikingly similar to Brian Bix’s antidote to the sort of paralysis presented by incommensurability. See BRIAN BIX, LAW, LANGUAGE, AND LEGAL DETERMINACY 105 (1993) (“after long consideration of the options . . . the decision-maker slowly begins to identify with one alternative rather than the others”). This after- the-fact form of identification does not deny incommensurability, but is consistent with it. 74. That is, what matters is the judge’s ability to choose. And this ability exists, Dworkin believes, whether the judge is faced with equal or incommensurable reasons for decision. See Ronald Dworkin, On Gaps in the Law, in CONTROVERSIES ABOUT LAW’S ONTOLOGY 84-90 (Paul M. Amselek ed., 1991). 75. See CHANG, INCOMMENSURABILITY, supra note 12, at 7-9. 76. This is also Lon L. Fuller’s point. See Fuller, Forms, supra note 59. 77. RONALD DWORKIN, LAW’S EMPIRE 255 (1986) [hereinafter DWORKIN, LAW’S EMPIRE]. 78. Id. at 230-31 (using the metaphor of a chain novelist’s interpretation of a preceding chapter in deciding how to write his portion of the text). 2014] PERMISSIVE JUSTIFICATION 701 resolving them.79 Political morality provides another set of constraints upon judicial decision- making. It requires the judge to ground her decision in ‘some . . . set of principles about people’s rights and duties [and] . . . the political structure and legal doctrine of their community.”80 Furthermore, these criteria—the community’s pre- existing rules of law and political morality (or justice)—are also the grounds that parties to a lawsuit would themselves propose as the applicable standards for legal decision. So the values at play non-arbitrarily limit the scope of the reasons upon which a judge can rely if she is to tip the scales one way or another.81 B. Relaxing Standards of Justification A significant challenge to the defeat model arises if reasons or values are incommensurable or incomparable. In that case, scales of values may prove robustly impervious to tipping. Where reasons or values are incommensurable or incomparable, agents cannot engage in the sort of comparative weighing that is a necessary precondition if reasons are to tip the scale of value.82 In law, as in many other institutional settings, the range of tie-breaking reasons or procedures are more limited than in everyday practical reasoning. A feature of the law as an institutional system (or games such as cricket or chess) is its prohibition on certain grounds for altering the normative status of the parties (personal predilections and so on) or certain procedures (coin tossing and so on) for doing so.83 Where the available institutional reasons conflict and are in equipoise, the judge cannot turn to such procedures to break the tie. Some versions of the decisive justification acknowledge this problem and address it head on. They do so by adopting a variety of approaches to relaxing the demand for decisive justification as the sole standard for assessing the rationality of an agent’s practical choices. One important approach is to retain decisive justification as the gold standard, but to propose that there are plural rather than singular standards for evaluating practical action. 1. Fragmentation of Value.—Some comparativists acknowledge that reason is conflicting and gappy, and so insert a third category of rational choice between fully justified and fully unjustified choices. Among these comparativists are value pluralists who nonetheless retain a comparativist approach to justification. That is, they accept that justification is decisive and based upon relative rankings of values, but nonetheless think that sometimes reasons or values are incommensurable or incomparable, and so our choices among reasons or values 79. See DWORKIN, LAW’S EMPIRE, supra note 77, at 255; see also id. at 235-37, 244-45, 250. 80. Id. at 255. 81. Id. at 256-68. 82. See, e.g., JOSEPH RAZ, THE MORALITY OF FREEDOM (1986) [hereinafter, RAZ, MORALITY]; J. J. C. SMART & BERNARD WILLIAMS, UTILITARIANISM: FOR AND AGAINST (1973). 83. In the case of games, the available procedures may be somewhat more expansive not much more so. For example, many games (cricket, football) begin with a coin-toss to determine who goes first, but not to resolve who is out or when a goal is scored. 702 INDIANA LAW REVIEW [Vol. 47:689 are incapable of justification.84 According to Thomas Nagel, for example, where values are plural, our choices among value cannot be fully justified. At most, we can only advance a second-best standard of rational acceptability, which is that our choices reflect our good judgment. Nagel implicitly invites us to distinguish between justifications that are complete and standards for evaluating rational action that are incomplete.85 Complete justifications are simply comparativist rankings according to some scale of value: what Nagel calls “a single, reductive method or a clear set of priorities.”86 Nagel’s view of what counts as total or complete justification fits the defeat model’s demand for all-or-nothing reasons for action: justification is complete when reasons fully determine the outcome, so that the reasons that the winner wins are also the reasons that the loser loses. Justification is incomplete, according to Nagel, where reason simply narrows down some range of rationally eligible undefeated reasons for action,87 so that there is some “slack that remains beyond the limits of explicit rational argument.”88 Nonetheless, he thinks, there remain standards of rational judgment provide “good” enough grounds for rational action.89 In such cases, the reason that the winner wins does not determine that the loser loses. Reason is indecisive, and the standard of rational decision is incomplete because of the inability to mollify the losers.90 What Nagel calls “good judgment” provides a secondary standard for unjustified (that is, not completely justified) but rationally supported reasons for action.91 Good judgment is necessary, Nagel claims, because values are not always commensurable on some common scale: Nagel thinks that incommensurable values “fragment” into five formally distinct types.92 None of 84. Thomas Nagel, for example, appears to hold onto a more-or-less comparative approach when he argues that, “there can be good judgment without total justification.” NAGEL, supra note 19, at 134. 85. Id. at 134. 86. Id. 87. Id. at 135. That is, he thinks that what Raz calls the classical model of rational action cannot provide a complete justification for action. In one sense (in the sense that reason does not completely determine what to do), Raz agrees. See RAZ, ENGAGING REASON, supra note 9, at 236- 37 (discussing complete and incomplete reasons). However, I shall add that differently understood, Nagel is right to contend that, “Provided one has taken the process of practical justification as far as it will go in the course of arriving at the conflict, one may be able to proceed without further justification, [and] without irrationality either.” NAGEL, supra note 19, at 135. In other words, the eligibility model of permissive justification is sufficient, even without determining rational action. 88. NAGEL, supra note 19, at 135. 89. Id. 90. Id. 91. Id. 92. The five formally incommensurable categories he identifies are (1) “specific obligations to other people or institutions,” (2) “constraints on action deriving from general rights that everyone has, either rights to do certain things, or not to be treated in certain ways,” (3) utility, that is “the effects of what one does on everyone’s welfare,” (4) what Nagel calls “perfectionist ends”, that is 2014] PERMISSIVE JUSTIFICATION 703 these values or their entailed points of view is homogenous with any of the other values or points of view and so each value is irreducible to any one of the other values.93 Though the reasons for action cannot be compared and ranked, they nonetheless provide some reason for action (if not one that can separate options or outcomes into winners and losers). For instance, our doctor may be in the position of determining which patient should receive an organ transplant, and face a variety of conflicting concerns. On one scale of value, for example, the doctor may decide based upon the ability of each patient’s age or lifestyle to promote long-term health; on another scale of value, she might decide based upon the contribution of the transplant to increasing each patient’s quality of life. Each scale of value demands the decision-maker’s full rational attention, such that decisions made from “inside” any particular scale render appeal to the other scales of value unjustified.94 Cross- scalar appeals are as meaningless as (to use John Finnis’ example) ‘sum[ming] up the quantity of the size of this page, the quantity of the number six, and the quantity of the mass of this book.”95 Because there is no super-scale from which to rank and compare the five individual scales of value, Nagel’s position is that there cannot be full justification when choosing among incommensurable points of view, but at best some lesser standard of rational adequacy.96 From Nagel’s perspective, however, our doctor can exhibit good judgment in choosing among incommensurable treatment options, even if she cannot justify any particular outcome to the patient.97 Each treatment option is one that is supported by some undefeated reason, and is to that extent rational to pursue.98 It may not be the best course of action, because not justified by a decisive reason.99 But it is a good or “correct”100—that is rationally “eligible”101 or “adequate”102 course of action because justified by an undefeated reason.103 It is certainly no worse than any of the other available treatment options. 2. High Stakes and Low Stakes.—The core thought shared by Nagel and the comparativists is that reasons guide action just in case they require an agent to act in certain manner, and that guiding reasons justify only if some one action is “the intrinsic value of certain achievements or creations,” and, finally, (5) “private commitments to one’s own projects or undertakings.” Id. at 128, 129-33. 93. Id. 94. See, e.g., SAMUEL SCHEFFLER, THE REJECTION OF CONSEQUENTIALISM: A PHILOSOPHICAL INVESTIGATION OF THE CONSIDERATIONS UNDERLYING RIVAL MORAL CONCEPTIONS (1982). 95. FINNIS, NATURAL LAW, supra note 20, at 115. 96. NAGEL, supra note 19, at 128; 129-33. 97. Id. at 135. 98. Id. 99. Id. 100. Id. at 128, 129-133. 101. RAZ, ENGAGING REASON, supra note 9, at 24. 102. John Gardner, The Mark of Responsibility, 23 OXFORD J. LEGAL STUD. 157, 157-71 (2003) [hereinafter Gardner, The Mark]. 103. NAGEL, supra note 19, at 135. 704 INDIANA LAW REVIEW [Vol. 47:689 comparatively required. Nagel relaxes the rational demands of comparativism by proposing two standards of rationally eligible104 or adequate105 action: decisive justification (high standards) and good judgment (low standards). One way of further undoing the comparativist demand is to claim that high-standards decisive justification applies to high-stakes decisions and that low-standards good judgment applies to low-stakes decisions. This is the approach favored by William Lucy. Lucy’s distinction between high-stakes and low-stakes circumstances places low-stakes decisions “within the personal sphere.”106 Low-stakes circumstances include “deciding whether to take a hiking or a skiing holiday . . . or even more important choices such as whether or not to have children or which career to pursue.”107 High-stakes decisions are decisions outside the personal sphere that “confer [upon some decision-maker] . . . significant power and authority over [another].”108 Lucy thus restates the familiar distinction between personal and impersonal reasons for action and links them to high- and low-stakes decisions. Decisions that call for impersonal reasons are always high-stakes, and so demand decisive justification; low-stakes decisions may be resolved using personal reasons, and so do not call for decisive reasons for action (or are easily tippable).109 The doctor example, Lucy thinks, is a central case of decision calling for impersonal reasons and so decisive justification. On such occasions, Lucy proposes: it might be thought desirable to have decisions and actions compelled by reasons. It therefore shows situations in which either the [eligibility model] adopts an uncharacteristically stringent account of rationality (weighing reason(s)) or in which the [defeat model] of practical reason, agency, and the will operates.110 Here, the problem is whether some circumstances (such as grave medical operations or denying one patient a transplant in favor of another patient) are so high-stakes that the decision-maker must eschew indecisive decision and secondary standards such as good justification. Instead, the doctor must provide the sort of decisive reasons mandated by comparativist-style complete justification. Lucy’s underlying normative proposition—that we aspire to high-standards 104. RAZ, ENGAGING REASON, supra note 9, at 24. 105. See Gardner, The Mark, supra note 102, at 158. 106. Lucy, supra note 19, at 244. 107. Id.. 108. Id. at 245. 109. Id. at 244-45. Joseph Raz makes a similar distinction, but rather than personal and impersonal, he distinguishes between self-interested and moral considerations. See Joseph Raz, Hart on Moral Rights and Legal Duties, OXFORD J. L. STUD. 123, 130 (1984) [hereinafter Raz, Hart]. 110. Lucy, supra note 19, at 244. 2014] PERMISSIVE JUSTIFICATION 705 (that is, decisive) justification in high-stakes situations111—appears broadly correct. Even in high-stakes situations, however, such justifications may not exist.112 Lucy is no monist: he implicitly accepts that values and reasons are incommensurable or incomparable.113 Accordingly, the high-stakes, high- standards position, like comparativism more generally, comes in stronger and weaker versions. Weak comparativism, which bifurcates the standards upon which to ground rational action into full justification and some lesser standard, might rest content with practical disappointments.114 That is, weak comparativism might simply accept that when reasons are indecisive in high stakes circumstances, no (full) justification is possible. Another tack that weak comparativism might take is, however, just plain wrong: when reasons are indecisive in high stakes circumstances, we cannot just wish away indecision and fix upon a (non-existent) decisive reason for action.115 To avoid lumping Lucy with this mistaken view, we could reduce his position to the judicial bivalence thesis advanced by Endicott and Finnis: that high-stakes decisions must be presented as high-standards decisions (that is, based on a decisive reason) even if they are not.116 So it may be the case that a decision maker must act as if justification is high-standards even if such standards are unavailable. Her decision would (using the standard of decisive justification) be not justified, though it may, given the exigencies of judicial bivalence, be presented as such to the losing party. That is not to deny what the first version of the high-stakes, high-standard theory accepts: that where the stakes are high, or where we wish that justification could be all-or-nothing, then it would be nice if the decision-maker could rely on some decisive reasons to justify her action. In that circumstance, the possibility of justification would match the necessities of decision in a straightforward manner, and the decision-maker could represent her decision in a transparent and rationally complete manner. However, reason is not always decisive: whether or not incommensurability exists, reasons (and decisions and justifications) may be of equal or incomparable weight. The aspiration to decisive justification, though itself (under certain circumstances) rational, is sometimes incapable of satisfaction. Nagel proposes one way around dissimulation. He thinks that the law deals primarily in one of these scales of value; that is, the law assesses our reasons and actions from the point of view of general rights.117 Though he identifies four other scales, they will not count for purposes of legal justification. Nagel claims that: 111. Id. at 245. 112. Id. Waluchow makes a similar argument against Dworkin. See Wilfrid J Waluchow, Strong Discretion, 33 PHILOSOPHICAL Q. 321, 321-39 (1983). 113. Lucy, supra note 19, at 206-67. 114. See Waluchow, supra note 112, at 321-39. 115. Id. 116. See supra note 20 and accompanying text. 117. NAGEL, supra note 19, at 136. 706 INDIANA LAW REVIEW [Vol. 47:689 Sometimes a process of decision is artificially insulated against the influence of more than one type of factor. . . . The example I have in mind is the judicial process, which carefully excludes, or tries to exclude, consideration of utility and personal commitment, and limits itself to claims of right. Since the systematic recognition of such claims is very important (and also tends over the long run not to conflict unacceptably with other values), it is worth isolating those factors for special treatment.118 Nagel believes, then, that some schemes of value do not count in legal evaluations of action and that others inevitably tend to produce the same outcomes as general rights do.119 The first claim—that some scales do not count—fits with standard accounts that seek to prevent arbitrariness by demanding that legal decisions rest on something more than the judge’s personal interests or commitments. It also jibes with Ronald Dworkin’s claim that legal decisions must rest on rights or on “integrity,” understood as limiting the sorts of reasons that a government can use to justify its actions and decisions.120 On this view, the law can provide justified answers to practical problems, but only by narrowing down the types of reasons a court may rely upon in adjudicating the questions before it. Nagel thus argues that the law’s ability to justify its outcomes depends upon excluding incommensurable points of view to produce decisive reasons from within one order of value; that of rights or (in Dworkin’s terms)121 the domain of principle.122 Having narrowed down the required sources of value to a unitary scheme, comparative rankings of reasons for action can, Nagel believes, produce decisive justification.123 Nagel gives no reason for thinking that rights are unitary in this way.124 Competing rights may be incomparable or incommensurable, and so irreducible to some single scale of value. For example, equality and liberty are often thought to be incommensurable and so irreducible to each other.125 While we could engage in some ordering of the values, such an ordering would be non- comparative and so (for the comparativists at least) arbitrary. Accordingly, we might think, “rights” does not so much provide a ‘scale” of value as a mode of 118. Id. 119. Id. 120. See DWORKIN, LAW’S EMPIRE, supra note 77, at 225; see also Jeremy Waldron, Pildes On Dworkin’s Theory of Rights, 29 J.L. STUD. 301, 301-08 (2009). 121. See DWORKIN, PRINCIPLE, supra note 56. 122. NAGEL, supra note 19, at 131-37. 123. Id. 124. In fact, he identifies a problem with this approach that applies to his vision of law: “the danger of exclusionary overrationalization, which bars as irrelevant or empty all considerations that cannot be brought within the scope of a general system admitting explicitly defensible conclusions.” Id. 125. Elinor Mason, Value Pluralism, STANFORD ENCYCLOPEDIA OF PHILOSOPHY (June 20, 2006), http://plato.stanford.edu/entries/value-pluralism/, archived at http://perma.cc/E6VL-SAGP. 2014] PERMISSIVE JUSTIFICATION 707 valuation: honoring values rather than promoting them.126 II. PERMISSIVE JUSTIFICATION In those circumstances in which we must select winners and losers, it would be ideal if our practical decisions could be justified by some decisive reason. The world, however, is not ideal in this way: sometimes we have no way to distinguish winners from losers, but must distinguish them nonetheless. Up to this point, this Article’s focus has been on the decisive justification, which holds that the reasons that are sufficient to justify some action must be rationally adequate to console the loser of some practical conflict. This Article has suggested that the defeat model, with its emphasis on decisive reasons for action, is unsatisfying. The defeat model’s version of the relation between reasons and justification demands more from reason than reason can deliver.127 It demands, for example, that reason provide grounds for justifying decisions to winners and losers alike. It rejects the claim that indecisive reasons—and so certain types of moral and legal permissions—could count as grounds for decision and so provide adequate reasons for action. In other ways, it demands less from reason than reason can offer. It shuts down the opportunities reason gives us for embracing the alternative possibilities that the world presents to us. The eligibility model of permissive justification is sufficient. The permissive model is more liberal: it does not require that our undefeated reasons for action also defeat all comers.128 Instead, as John Gardner puts it, “[u]nder the heading of justification . . . we claim [only] that the reasons in favour of what we did were not all defeated by conflicting reasons, and that our action was performed on the strength of some or all of the undefeated reasons in its favour.”129 Permissive justification does not start from the premise that justification is called for to console the losers of practical conflicts. Instead, permissive justification takes the narrower position that, in its central case (or “strictly speaking”),130 justification is called for where reasons conflict.131 On the 126. Philip Pettit, Consequentialism, in CONSEQUENTIALISM (Stephen Darwall ed., 2003). 127. Proponents of the defeat model are thus the rationalist equivalent of what Hart calls “disappointed absolutists.” H.L.A. HART, THE CONCEPT OF LAW 139 (2d ed., Penelope A. Bulloch & Joseph Raz eds., 1994). 128. See Gardner, The Mark, supra note 102, at 158. 129. Id. 130. See GARDNER, HARM, supra note 33, at 107 (discussing stricter or what I call more interesting sense of justification). 131. See id. Absent a conflict of reasons, either anything goes (when there are no reasons) or only one thing goes (when all the reasons are on one side). In the latter case, even a very weak reason will justify an action or outcome. We might think that weak justifications provide little claim on our rational attention, and so at the margins, the difference between anything goes and only one thing goes is slight. Most of the interesting cases of permission fall somewhere in the middle, in this ‘strict” or central case of justification. It is precisely with these cases that permissive justification is concerned. 708 INDIANA LAW REVIEW [Vol. 47:689 permissive model, even though a loser may properly demand justifications when reasons conflict,132 the response may often be unsatisfying. By removing losers from their central place in theories of justification, the permissive model narrows the scope that standards of rational adequacy must meet.133 Reason is not always decisive, and the rational world is neither transparent nor gapless. Reason, when indecisive, is permissive. On the model of permissive justification, the presence of an undefeated reason provides all the justification we need when choosing among plural options.134 Permissive Justification: an act is rationally justified only if some reason to act is undefeated. The model is permissive because, when reasons conflict and there is no decisive reason to recommend one option as uniquely supported by reason, there is a normative permission to choose either option. A normative permission to act exists when there is no eligible reason to not act.135 Normative permissions most obviously exist where there are no reasons at all: reason is absent. Normative permissions, however, also exist where reasons are indecisive: where there is no decisive reason not to undertake the permitted action. In such circumstances, multiple (indecisive) reasons are available to justify multiple actions or outcomes. For example, John Gardner argues that “where there is a bare conflict of weight between reasons for action, it is in my view justifiable to do whatever is supported by a reason that is not outweighed, whether or not it also outweighs.”136 Although Gardner has not made the connection himself, we can characterize the conflict as resulting in a permission. Where reasons are undefeated, there exists an absence of a certain sort of reason: there is no decisive reason for action and so, as Gardner argues in the previously quoted passage, there is a permission to do whatever is supported by an undefeated reason. The claim that permissions can justify action might appear paradoxical. After all, permissions indicate the absence of reasons to act, and justification depends 132. See Gardner, The Mark, supra note 102, at 158 (arguing that “we cannot but want our lives to have made rational sense, to add up to a story . . . of . . . whys. We cannot but want there to have been adequate reasons why we did (or thought or felt) what we did (or thought or felt).”). 133. See id. Justification is called for where reasons conflict, not where reasons are indecisive. Where there are no reasons, full stop, against doing the action there is a bare permission to act. When reasons are available and conflict, we may demand justifications, and enter reasons to assess which prevail and which do not. Sometimes, one reason decisively prevails over all the others. Sometimes, none prevail. What matters is that actions or beliefs call for justification where there is conflict, not simply indecision. 134. See id. 135. See, e.g., Raz, Permissions, supra note 32, at 161. See also RAZ, PRACTICAL REASON, supra note 38, at 85. 136. John Gardner, Justification under Authority, 23 CAN. J. L. & JURISPRUDENCE 71, 78 (2010). 2014] PERMISSIVE JUSTIFICATION 709 upon reasons.137 As Gardner argues elsewhere: Unlike a duty, neither a permission nor a power is a reason (or entails or even suggests the existence of a reason) to act as one is permitted or empowered by it to do. Is one to get married just because one’s marriage would be legally valid? Is one to become a vegetarian just because it is morally unexceptionable to be one? Of course not. As they stand these are not intelligible explanations of one’s actions. Something else—a reason for the action—is still needed.138 There is, however, a gap between a permission that indicates the absence of a duty (a “no-duty”) and a permission that indicates the absence of any reason whatsoever (a “norm-absence”139 or a bare permission). Gardner’s discussion of undefeated reasons depends upon this feature of permissions, and forcefully argues that permissions are more than just the absence of reasons.140 Indeed, much of the discussion of permissions advances a simplistic understanding that acknowledges only two forms of normative permissions: strong or express permissions, supported by some rule; and weak or bare permissions, that operate in the absence of reasons.141 I shall suggest, following Gardner and Joseph Raz, that the universe of permissions is much more complex than this dualism suggests, and that some permissions come with reasons attached: what I call supported permissions to complement Raz’s category of exclusionary permissions. Consider again the two medical examples sketched out so far: the doctor advising a patient who needs some invasive medical procedure, but faces a choice as to which to pick; or the doctor who has a limited supply of organs, and must pick which patient to save. In each case, the available reasons for favoring one or the other outcome may be in equipoise: of equal value, or incommensurable or incomparable as to value. In each case, the doctor has no decisive reason guiding her decision, but she does have plenty of undefeated reasons. If she is permitted to pick, and it is sufficient (under the model of permissive justification) that she acts for an undefeated reason, her decision is justified whichever option she chooses. In each case, the final decision over which option to pick might (using the standard of decisive justification) appear whimsical or arbitrary. Without a decisive reason to defeat the competing options, the decision looks more or less non-rational. The doctor’s ultimate decision, whichever it is, however, is not made in the complete absence of reasons. If there were no reasons (and so just a bare permission) there would be no need for justification. Whatever she did 137. Id. 138. Id. 139. John Gardner & François Tanguay-Renaud, Desert and Avoidability in Self-Defense, 122 ETHICS 111, 120 (2011). 140. See generally id. (arguing that permissions are more than just the absence of reasons). 141. Cf. Joseph Raz, Permissions, supra note 32, at 161. 710 INDIANA LAW REVIEW [Vol. 47:689 was, in Gardner’s terms, unobjectionable.142 Indeed, in the absence of reasons, justification is neither required nor possible. Lacking reasons, the doctor could just pick an outcome with a blithe “why not?”143 At the very least, then, in the doctor’s situation there is some reason, albeit an indecisive reason, for decision that would justify the outcome. The indecisive reason fully or completely justifies the outcome even if some other undefeated reason justifies a different outcome. This relation of reasons to permissions both complicates the picture and requires us to add to the taxonomy of permissions to explain how reasons support permissions. Normative permissions come in different forms, and are not simply one-size- fits-all. In organizing permissions by their relation to reasons, this Article shall argue that two types of relationships are of central importance. The dimension of support refers to the presence or absence of first-order reasons: these I call supported or unsupported reasons. The dimension of protection refers to the presence of second-order reasons. The relation of justification to normative permissions is thus complicated by the fact that permissions are a distinctive normative category, covering a variety of normative relations, and so some taxonomy of permissions is required to separate out those that are in need of justification from those that are not. A. A Taxonomy of Permissions Normative permissions are properly included in this list of things that are logically related to actions, but not all normative permissions are capable of justifying action. Indeed, as described above, the claim that permissions could justify action appears paradoxical. A normative permission to act entails the absence of some (decisive) reason to not act. Thus, a bare permission to act exists when there is no reason to act or to not act—where there are no reasons at all. In this case, no justification is required or possible. Anything goes, because there is nothing rationally to object to. 1. Supported.—For permissions to justify actions, they must be supported by first-order reasons to do the act that the agent is permitted to do.144 More formally, if an agent is justified in acting, that is because, in addition to the permission to act, she has a reason to act. Permissions can be supported by first- order reasons in various ways, of which this Article has identified two: a permission is decisively supported if there is a decisive reason to act that defeats any and all reasons to not act. In that case, there a permission because there no reason to not act, and the permission is justified because of the (decisive) reason to act.145 For example, if an agent considers whether she is justified in going to the 142. GARDNER, HARM, supra note 33, at 107. 143. Id. 144. See John Gardner, Justifications and Reasons, in HARM AND CULPABILITY 103, 124 (A. P. Simester & A. T. H. Smith eds., 1996). 145. See supra Part II. 2014] PERMISSIVE JUSTIFICATION 711 park or staying at home, and a decisive reason to go the park (she is helping a friend from out of town who is lost in the city) defeats all her competing reasons (she wants to laze around the house), then there is no reason not to go to the park and every reason to go. Reason completely determines the outcome. Nonetheless, there is also a permission to go to the park, even though the permission is also the only rationally eligible course of action. In the case of indecisively supported permissions, this Article has suggested, there is a positive reason to act. There is a conflict of reasons to act and to not act, such that the reasons are in equipoise: they are of equal, incommensurable, or incomparable value. The reason is indecisive because it does not prevail over all competing reasons. But, along with the other undefeated reasons, the permission to act depends upon the absence of a decisive reason to not act. The permission is also supported by an undefeated reason to act. Since there is more than one rationally eligible option, in the absence of a decisive reason to act, there is also a permission to not act that is supported by an undefeated reason.146 In the park example, if the agent has reasons to go to the park (she would like the fresh air) or stay at home (she needs some rest at the end of a busy week) and these reasons are in equipoise, the reason to go to the park does not defeat all comers, and so does not provide a reason why the losing option (staying at home) lost. The eligible reasons to go to the park remain undefeated, and so reason is indecisive. Nonetheless, on the account of justification sketched out, above, either choice is fully justified, because each is supported by an undefeated reason for action. 2. Exclusionary.—Exclusionary permissions depend upon the existence of a second-order permission to disregard (or rely upon) conflicting first-order reasons.147 An exclusionary permission to act entitles the agent to ignore at least some of the reasons for not acting.148 Exclusionary permissions, Raz argues, “always require a justification,”149 and so require some reason or “consideration” to “establish that one may disregard conflicting reasons.”150 Accordingly, a full analysis of exclusionary permissions is quite complex. These permissions depend upon a hierarchy of reasons in which higher-level reasons justify treating lower- level reasons as optional.151 And, by making these reasons optional, they operate to undermine the power of reasons within their scope, and so may affect the outcome of practical inferences.152 Consider the following example: in chess, there is a rule that a pawn may move one or two squares as its first move of the game. The rule establishes an exclusionary permission: the one-or-two-square rule is a second-order reason that 146. See supra Part II. 147. RAZ, PRACTICAL REASON, supra note 38, at 89-90. 148. See Raz, Permissions, supra note 32, at 163; RAZ, PRACTICAL REASON, supra note 38, at 89-90. 149. RAZ, PRACTICAL REASON, supra note 38, at 90. 150. Id. 151. Id. 152. Id. 712 INDIANA LAW REVIEW [Vol. 47:689 undercuts the usual rule for moving pawns one square at a time and replaces it with a permission to move it one or two squares as its first move.153 The rule is justified by reasons of speed and strategy: it makes the openings of the game move faster and it opens up a range of possible moves that make the game more interesting. The exclusionary permission does not, however, provide the player with a reason for moving one square or two squares. But nor does it preclude the giving of further reasons. The permission does not justify either move; but either move may turn out to be justifiable. In the chess example, either choice, moving one square or two, may in turn be justified by some further set of reasons, usually reasons of strategy (for example, it puts the pawn or some other piece at risk of being taken), and is justifiable if supported by some further reason. Such reasons are positively invited to fill the rational gap left by the permission to make the move. Peremptory challenges provide another legal example of an exclusionary permission, and as with the chess permission, one that is unsupported. When selecting the members of a jury, the lawyer may often possess a number of peremptory challenges that permit her to strike a juror for any reason or no reason. That is, the lawyer may exclude a juror from the petit jury “without showing any cause . . . without reason or for no reason, arbitrarily and capriciously.”154 The challenge is an exclusionary permission to strike jurors that provides a permission to disregard all the competing first-order reasons for seating the jurors.155 When a lawyer uses a peremptory challenge to strike a juror, she is not limited to choosing between a range of eligible reasons; she is entitled to disregard reasons altogether, and act on hunch or intuition or some such thing. While the existence of the permission itself is justified by considerations of fairness or autonomy,156 individual strikes need not be supported by reasons. In that case, “[t]he very essence of a peremptory challenge is that its exercise requires no justification or explanation.”157 Exclusionary permissions may thus be unsupported or supported, depending upon whether the permission is a permission to choose among reasons or includes the permission to disregard reasons altogether. Preclusionary permissions are second-order reasons to disregard first-order reasons that defeat those first-order reasons that are within their scope. They are 153. See GARDNER, HARM, supra note 33, at 117. (This type of exclusionary permission may be distinguished from the sort of exclusionary permission that John Gardner calls “cancelling permissions.” That sort of permission cancels, rather than defeats, the conflicting second-order reasons, permitting the decision-maker to turn outside the system of practical reasons and rely on a set of reasons that are normally excluded.) Here, the decision-maker does not turn outside the institutional system of the laws of chess, but rather replaces one norm with another. 154. Pointer v. United States, 151 U.S. 396, 408 (1980). 155. See id. 156. See, e.g., United States ex rel. Darcy v. Handy, 351 U.S. 454, 462 (1956). See also Batson v. Kentucky, 476 U.S. 79 (1986). 157. R. v. Gayle, 54 OR (3d) 36, 59 (2001) (emphasis added). 2014] PERMISSIVE JUSTIFICATION 713 thus better known as exclusionary reasons.158 This Article characterizes them as a form of permission, however, to emphasize that the operation of exclusion does not create reasons, but defeats them, and without some supporting reason, exclusionary reasons leave a rational gap. Examples of supported preculsionary permissions are relatively easy to find, however, for supported preclusionary permissions are just “the coincidental conjunction of a reason to act and an exclusionary reason not to act for certain countervailing reasons.”159 As Raz argues, exclusionary reasons “are almost invariably tied to first- order reasons and their combined application normally leads to a certain action being required.”160 Thus, if I have promised to meet Jane in the park, I now have a second-order reason (keeping the promise) not to act for certain conflicting reasons—a permission to meet Jane in the park—supported by the first-order reason (the fact of the promise) to meet Jane. Preclusionary permissions, however, may be compatible with there being no reason or a range of unexcluded, but indecisive and conflicting reasons. In that case, the unexcluded reasons may prove indecisive, and so the preclusionary permission does not result in a normative requirement, but retains the character of a permission.161 Only supported permissions justify doing some act. Unsupported permissions may license doing the act, if there is no undefeated reason that counsels against so doing. But, as the chess example demonstrates, merely having a rule granting a permission to move two squares does not justify making the move, unless there is some supporting reason that can fill in the rational gap left open by the permission. 3. Bare Permissions.—Finally, permissions may be both unprotected by some second-order reason and unsupported. In that case, in the absence of reasons, there is what this Article has called a bare permission. For example, there may be no reason for Sam to choose the black shoes or the brown ones (or any shoes at all). Here, she is permitted to pick because there is no rational conflict, no reason that could challenge or defend her selection. Her choice, though not supported by reason, is permissible. However, it neither justifies her selection nor is capable of justification. The relation between permissions, reasons, and justifications is thus not one- size-fits all. Rather, the relationship may be quite complicated, dependent upon the manner in which reasons are present or absent, supporting or protecting, conflicting or not. This brief taxonomy is not designed to identify every permission present in morality, politics, or law. It does, however, provide enough groundwork to consider in more detail the manner in which permissions make room for indecisive reasons to operate and to justify rationally eligible courses of action. 158. See RAZ, PRACTICAL REASON, supra note 38, at 89. 159. Gardner & Macklem, Reasons, supra note 33, at 440, 465. 160. RAZ, PRACTICAL REASON, supra note 38, at 89. 161. Gardner & Macklem, Reasons, supra note 33, at 440, 465. 714 INDIANA LAW REVIEW [Vol. 47:689 B. Permissive Justification Is Complete Justification This Article has suggested that a taxonomy of permissions is required to develop an adequate account of permissive justification. Justification, as previously explained, depends upon reasons for action or belief: “To claim that one has justification for doing or believing as one does is to claim, at the very least, that one has reasons for so doing or so believing.”162 Where there is no reason not to act, in other words, justification is not called for because “[t]he unobjectionable . . . is in no need of justification.”163 If there is one reason not to go to the park on Saturday, then one may go on a whim—one need not justify her choice, but just do it. If there is no reason not to take up stamp collecting, one need not justify selecting this choice of hobby, to others or to herself. One can simply decide to collect stamps or not, as she chooses. Accordingly, in its central case (or ‘strictly speaking”),164 justification is called for where reasons conflict, that is, “when one also has some reason not to act, believe, etc. as one does.”165 One could perhaps consider another, even weaker, candidate for justification: justification as freedom from criticism. In this case, one is justified in acting if one is permitted to act, even if one has no reason to act. Justification simply blocks criticism. In this case, justification primarily has a negative function: “why not?” would become the central case of justification. In that case, exclusionary permissions and bare permissions would be sufficient to justify action. One reason for not going down this weaker path is that it removes the centrality of reasons from practical life. As John Gardner puts it, “explanation in terms of reasons is what a rational being aspires to.”166 An advantage of both the defeat and eligibility models is that they preserve the importance of rationality for the justification of practical action. A second reason for not going down the criticism-blocking path is that it is unnecessary. Supported permissions can answer the call for rational justification, and do so in a central, rather than peripheral, way. Where reasons conflict and are undefeated in a manner that produce rationally supported permissions, that is sufficient for “full” or “complete” justification. Since the view this Article defends entails that there is a fully adequate reason to act so long as the action is rationally eligible,167 all that is required for justification is some undefeated reason for action. If an agent has an undefeated reason to go to the park on Sunday that conflicts with an undefeated reason to stay at home, each is rationally eligible to operate as a ground of decision. Since each is eligible, each is sufficient for justification and either choice—to go to the park or to stay at home—is fully justified. 162. GARDNER, HARM, supra note 33, at 103. 163. Id. at 107. 164. See id. (discussing stricter or a more interesting sense of justification). 165. See id. 166. Gardner, The Mark, supra note 102, at 159. 167. RAZ, ENGAGING REASON, supra note 9, at 24. 2014] PERMISSIVE JUSTIFICATION 715 Accordingly, under the eligibility model of permissive justification, there is a complete justification for acting just in case there is an undefeated reason to act.168 Unlike decisive justification, permissive justification does not go the extra step to require a decisive reason to act because, on the eligibility account, the standard is simply that “one has reasons”169 for what one does and those reasons are undefeated. At the same time, we should be careful to distinguish this definition of complete justification from the variety of other ways in which to measure completeness. One way is to define complete justification in terms of the extent to which reason determines the outcome.170 On this account, indecisive reasons are incomplete, but only because they do not fully determine the agent’s course of action. However, they are not incomplete in the sense of being less than whole171 or otherwise provisional. Supported permissions have sufficient rational heft to justify an agent’s decision to act or not act. Furthermore, each justification provides as much reason to act as any other eligible justification. In this sense, it is as “full” of justification—as a rationally acceptable basis for action—as any of the competing, eligible justifications.172 The justification may not be determinative of the agent’s choice of one option over another.173 Nonetheless, the justification is as rationally complete as we could expect (and as the world allows). III. JUSTIFICATION IN A LIBERAL PROFESSION Consider once more an example introduced by William Lucy and which I have relied upon to this point: a doctor, confronted with indecisive reasons for selecting between one or other serious and invasive medical procedure, must justify her choice option to her patient. In such a circumstance, Lucy thinks, the doctor ought not to “resort to [the doctor]’s will as the driving force of decision and action.”174 Instead, Lucy thinks, the doctor ought to produce some decisive 168. GARDNER, HARM, supra note 33, at 103. 169. Id. 170. See generally GARDNER, HARM, supra note 33; RAZ, ENGAGING REASON, supra note 9. 171. See generally Gardner & Macklem, Reasons, supra note 33; RAZ, ENGAGING REASON, supra note 9, at 24. 172. RAZ, ENGAGING REASON, supra note 9, at 47. 173. In this sense of “complete,” the decision-maker need not turn outside reason to determine her choice, because there is a decisive reason for action. In this case, a (rationally) incomplete justification is one in which reasons only partly explain the agent’s choice of one option over another. Reason is indecisive, and the agent must do more than simply ratify the verdict of reason that a particular action is required. Something in addition to reasons (for example, “will”) is required to help the agent’s selection. See, e.g., RAZ, ENGAGING REASON, supra note 9, at 47. There are other versions of the distinction between complete and incomplete reasons for action as well. See, e.g., Gardner & Macklem, Reasons, supra note 33; GARDNER, HARM, supra note 33; RAZ, PRACTICAL REASON, supra note 38. 174. Lucy, supra note 19, at 245. 716 INDIANA LAW REVIEW [Vol. 47:689 justification for her decision.175 Here, the doctor faces a high-stakes choice between competing significant medical procedures. In such a circumstance, Lucy asks: [i]f it can be said that [the doctor] has adequate reason to do either [one treatment or another], would it be acceptable that [the doctor’s] will—which is “informed and constrained by reason but plays an autonomous role in action”—determines which treatment is adopted? It seems unlikely that it would.176 But does it? It is unlikely only on the assumption that (1) the patient cannot participate in the decision-making process; and (2) the patient has no choice among doctors—that she cannot seek a second opinion. But an ordinary understanding of the doctor-patient relationship is not so rigid and authoritarian. Rather, contemporary liberal medical practice operates under a norm of shared or devolved choice: while a doctor may choose her preferred treatment based on her style of practice, patients can choose their doctors. There may be a separate duty of candor from the doctor in a doctor- patient relationship, so that while the doctor may recommend one treatment first, and counsel waiting until after trying that treatment before trying the others, the doctor should nonetheless also advise the patient that she could try them the other way around. On this view, doctors are not just blind appliers of reasons independent of ideological investment. Instead, doctors—and other professionals, from architects and accountants to web-designers and zookeepers—have choices among the different ways of practicing their profession and honoring the conflicting values that underlie those choices. The values are sometimes of equal weight, or incommensurable or incomparable. A liberal profession is one that includes these different values, ideologies, and the plural styles of practice they produce.177 To ensure that the patient can make her own choice and is not simply forced to accede to the doctor’s decision, the duty of candor may also require counseling the patient to seek second opinions. Accordingly, the doctor should encourage the patient to shop around and try out other advice and other doctors. In that case, some of the worries raised by permissive justification disappear. The patient is not at the mercy of the doctor’s whims, but gets to decide herself which among different doctors adopting plural treatments to select. The patient has an option to decide which style of treatment she prefers, and so participates in the decision. 175. Id. Here, the real problem is at what point to resort to the will or some other deadlock- breaking device. That is, the point is not one about haste or deliberation. Proponents of indecision, such as myself, and of will-based theories of decision, such as Joseph Raz, would, along with Lucy and along with decisive-justification theorists, reject any “hasty resort” to either the will or reason to break the deadlock among indecisive reasons. 176. Id. at 244. 177. See Joseph Raz, Liberalism, Skepticism, and Democracy, 74 IOWA L. Rev 761, 780-81 (1989). 2014] PERMISSIVE JUSTIFICATION 717 A. Professionalism: Liberal and Dogmatic Permissions help us distinguish between liberal and dogmatic approaches to professions. It is part of the ethics of many professions that there is more than one permitted way to perform them. There is more than one way to treat the patient, to account for one’s profits, or to design a building. Accordingly, in professional settings, permissions might create a route for diverse approaches to be reflected within the range of reasons that operate to justify discrete, rationally eligible courses of action. Part of the justification for permissions may be precisely to fit these diverse ways into the profession.178 Permissions provide ways in which second opinions and appeals to other authorities and ways of doing things enter the professional realm. Accordingly, permissions can liberate us from the threat of rational domination, where (because of the high-stakes nature of the choices) agents demand decisive justifications before engaging in action.179 Given the right institutional structure, permissions allow a plurality of reasons and projects to flourish as eligible justifications for action. Permissions, however, do not liberate agents from responsibility for their commitments, and it is this type of liberation that can appear decisive. For example, if the doctor can produce a decisive reason, one that defeats all competing reasons, then she can demonstrate that her choice is the best, rationally, that she could do.180 Reason requires the doctor to act this way rather than another way; therefore, she is decisively justified in selecting a particular course of treatment. On this view of rationality, reasons liberate by freeing one from moral blame for one’s rationally justified decisions.181 On this view, decisive justification gets one off the hook for the various choices she makes.182 When one asks, in this blame-avoiding mode, what is an adequate reason for action, it is one that will prevail over all other reasons. On this view, the negative consequences of one’s actions are not “hers to bear” but reason’s alone.183 Lacking decisive reasons, the doctor is back on the hook. No matter what she does, she is personally answerable for any harm she does.184 178. See supra Part III. 179. William Lucy, Adjudication, in THE OXFORD HANDBOOK OF JURISPRUDENCE AND PHILOSOPHY OF LAW, 206, 244 (Jules L Coleman et al. eds., 2002). 180. Douglas N Husak, Conflicts of Justifications, 18 L. & PHIL. 41, 41-68 (1999). 181. Id.; RONALD DWORKIN, SOVEREIGN VIRTUE: THE THEORY AND PRACTICE OF EQUALITY 287-90 (2000) [hereinafter DWORKIN, SOVEREIGN VIRTUE]. 182. DWORKIN, SOVEREIGN VIRTUE, supra note 181, at 287-90. 183. Gardner, The Mark, supra note 102, at 157. 184. A different explanation, mentioned and rejected in Raz’s, Incommensurability and Agency, which is that the worry is rational bewilderment. The agent lacks the means to explain why she undertook one rather than any other course of action. RAZ, ENGAGING REASON, supra note 9, at 49. 718 INDIANA LAW REVIEW [Vol. 47:689 The “deeper” view, this Article suggests, following Gardner, is that rational justification rests upon arguments about the intelligibility of our lives as rational beings.185 Because we wish our lives to make rational sense, we want to follow rationality rather than simply avoid its unwelcome consequences.186 However, reason is not always decisive, and the world is neither transparent nor rationally complete. Reasons, as well as decisions and justifications, may be of equal, incommensurable, or incomparable weight. The aspiration for decisive justification, though rational and psychologically compelling, is sometimes incapable of satisfaction. 1. Practice and Projects.—One way of thinking about liberal practice is as a practice including different styles, perspectives, or projects as a means of maximizing the range of available or eligible reasons for decision. Styles of practice—doctoring, accounting, and so on—are much like personal goals.187 These styles of practice are rationally underdetermined when selected.188 However, once the professional has committed herself to a particular style of practice, she answers to reason, and further, actions in line with that style of practice are determined in part by the initial choice because the professional’s project-related commitments and goals constitute reasons for action.189 In this way, we might think of a style of practice as one form of what Bernard Williams calls projects.190 Projects, according to Williams, are goal-oriented and temporally extended practical or intellectual enterprises, such as an idea, value, or activity, that agents can generate or participate in, either individually or in groups.191 One way in which individuals pursue projects is by committing themselves to those projects. Commitments are thus attitudes or activities that express an agent’s adoption and endorsement of certain projects as her own. Williams emphasizes the value of practical commitments to individuals’ understanding of the structure and value of their lives.192 A commitment is thus something more than a whimsical or faddish engagement with some project.193 A commitment requires dedication. For Williams, projects and commitments are interrelated concepts. Projects are enterprises that require the sort of extended engagement characteristic of commitments: They may not be treated as “dispensable” or as simply “one satisfaction among others.”194 Instead, Williams understands the pull of our projects in terms of continuity or stability.195 185. Gardner, The Mark, supra note 102, at 158. 186. Id. 187. Gardner & Macklem, Reasons, supra note 33, at 440-75. 188. Id. 189. Id. (“goals do not merely reflect reasons but also constitute them 190. SMART & WILLIAMS, supra note 82, at 100, 103-06. 191. See id. While imprecise, the definition of project is serviceable for current purposes. 192. Id. at 100. 193. Id. (calling this type of engagement “tastes” or “fancies”). 194. Id. 195. See Peter Railton, Alienation, Consequentialism, and the Demands of Morality, in 2014] PERMISSIVE JUSTIFICATION 719 However, continuity or stability alone is insufficient to explain the normative and rational force of commitments. Continuity and stability are consistent with a non-normative account of commitment as habitual behavior.196 Without more, Williams’ concept of commitment merely records, describes, or predicts the agent’s conduct in terms of regularly repeated acts,197 without seeking to include the normative reasons why the participant acts as she does.198 For our commitments to be rationally goal-oriented practical enterprises or to engage the intellect, they must do more than identify stable and regular patterns of conduct. They must provide reasons that do or would support the agent’s claim that she is rationally justified in pursuing her projects such that others must respect her commitments (even if not adopt them for themselves).199 Commitments, like practices more generally, provide second-order reasons for action that pre-empt some of the agent’s other reasons for action.200 This feature helps explain the categorical character of commitments: they “are not hostage to the [other] prevailing personal goals of the agent to whom they apply [and so do] not bend to the changing winds of one’s ambitions.”201 Given the categorical and exclusionary nature of our commitments, we might characterize them, following Gardner and Macklem202 and Raz,203 as generating a protected reason for action. This conclusion, that our commitments are normative and provide reasons for action, fortifies Peter Railton’s critique of Williams’ criteria of stability and seriousness.204 Railton suggests that some projects should be defeasible.205 An agent may have misjudged the value of her projects, or their value may have changed; therefore, she should immediately abandon them. In that case, Williams’ criteria of seriousness and stability misrepresent the nature of commitment: if the agent’s reasons for commitment are cancelled or defeated,206 the act of committing to the project no longer provides her with rationally eligible reasons for continuing to pursue her project or goal.207 The commitment reasons CONSEQUENTIALISM AND ITS CRITICS 93-135 (Samuel Scheffler ed., 1988). 196. HART, supra note 127, at 89. 197. Id. at 54-56; 89; see also NEIL MACCORMICK, H.L.A. HART 30-34 (1981) (The major impact of the external perspective is to avoid characterizing the observed conduct as rule- following). 198. HART, supra note 127, at 89. 199. Gardner & Macklem, Reasons, supra note 33, at 440-75. 200. Id. 201. Id. 202. See generally id. 203. See generally RAZ, PRACTICAL REASON, supra note 38. 204. Railton, supra note 195, at 93-133. 205. Id. 206. Id. at 101. Railton appears to equate defeasibility with various sorts of cancelling conditions: either internal to one of competing reasons or due to changes in the world. Id. Here, it is not the case that one reason defeats the other. Where reasons are cancelled, one reason drops out of the rational calculus. 207. Id. 720 INDIANA LAW REVIEW [Vol. 47:689 are no longer justified, even by the relaxed standard of permissive justification.208 Discarding a commitment may happen relatively quickly. Tomorrow, the doctor’s style of professional practice may no longer be rationally defensible, given a change in medical practice today. This change does not mean that, when she adopted the particular mode of treating patients, she was not seriously committed to that treatment. Despite the parlous impact on her patients’ health from persisting with such treatments, the doctor may still be irrationally committed to that style of practice moving forward. Although she ought not to be committed to this treatment in the future, the fact that she may desist using the treatment tomorrow does not negate her commitment to the treatment today. The normative force of commitments and projects can also explain how one’s choices can generate reasons and goals. It may be that one’s adoption of projects is—from the perspective of the defeat model—arbitrary: an agent picks a project, for no reason or for no decisive reason, such as a career in law over a career in teaching209 She may reject professionalism altogether and head off to a kibbutz.210 In any case, picking that project is a rationally permissible course of action, one that is more or less justified because it is supported by reasons. Having picked a project, her commitment to that project now gives her further reasons to continue.211 Her choices create reasons that operate to justify her subsequent actions. Any of these permissible human projects—the practice of law, the vocation of teaching, working on a kibbutz, or even taking up a hobby such as stamp collecting—is an activity that has or can generate value. Any of these occupations could be personally meaningful, financially profitable, a form of recreation, or so on. While the agent may initially have equivocal or no reasons for taking up one of these projects, the way in which she commits herself can provide reasons that justify her project-relevant actions both to herself and to others.212 Once normatively committed to a permissible human project, the agent cannot just abstain from the project absent some good reason. If her commitment is genuine, she has placed herself under an obligation to persist.213 So long as her reasons for engaging in stamp collecting remain undefeated or un-cancelled, her choice of career or hobby is one that others are bound to respect.214 That is, her 208. See id. 209. See RAZ, MORALITY, supra note 82, at 341-45 (suggesting that having to choose between a career in law or a career in teaching would be an incommensurable decision). 210. Ronald Dworkin, Objectivity and Truth: You’d Better Believe It, 25 PHIL. & PUB. AFF. 87, 136 (1996). 211. RAZ, MORALITY, supra note 82, at 388. 212. Id. 213. Gardner & Macklem, Reasons, supra note 33. 214. One way of developing this thought would be to consider Joseph Raz’s discussion of social forms, which are socially developed, valuable activities. Social forms depend, however, for their creation and sustenance, on practical and affective features of social interaction that go beyond the individual or even the small group. Accordingly, not just any attempt to commit to a project 2014] PERMISSIVE JUSTIFICATION 721 hobby provides reasons for others to give her stamps instead of coins as a gift. As an additional example, the agent’s choice to pursue a career as a lawyer is one that her parents are bound to respect, even if they think that there are reasons why she should have followed the family profession and become a doctor. In this case, reason and choice215 interrelate. The agent’s will is just the ability to choose among projects and is expressed by selecting a project or goal. The initial act of choosing may be more or less supported by reasons. Permissions may exist where reasons are absent or conflicting. In any case, having normatively committed herself to that project or goal, she now has reasons to continue down her chosen path, and these are reasons that others should respect.216 In this way, our choices can create reasons, both for ourselves and for others. But if others have reason to respect an agent’s project, even to the extent of acting to support it, they are not bound to choose it for themselves. The stamp- collector’s decision does not commit other agents to become stamp-collectors themselves or even to taking an interest in her hobby.217 Her decision to become a stamp-collector generates reasons for others, not obligations.218 Reasons are advisory only, and they can be overridden, ignored, and so on.219 Obligations cannot. They are mandatory (and categorical) and, therefore, are not so easy to dismiss.220 Stamp collecting (or the practice of medicine), unlike law, does not depend upon the exercise of a normative power over others. That is, the doctor’s and the stamp collector’s decisions do not put others under some duty to obey. Thus, when a patient seeks out a doctor and asks why this treatment rather than that, the doctor can perfectly reasonably say: “[the reason is] because I’m that kind of doctor.” The patient has reason to respect the doctor’s choice, even though the patient is not bound to choose it for herself. In other words, the doctor may justify her choice of treatment for the patient by pointing to her commitment to this style of practice.221 But the commitment reasons only operate as a justification for the doctor to engage in this style of practice. They do not require the patient to accede to having the doctor treat her in this way. In some circumstances, for example, where no other doctors are available, the treatment cannot wait, and the doctor is able to treat in multiple styles, the patient’s commitment to a particular form of treatment may trump the doctor’s commitment and may require the doctor to give up her insistence on a particular style of treatment. 2. Egalitarian and Inclusive Professionalism.—So far, I have argued that is socially valuable, and not just any project can demand the rational respect of other members of the community. See RAZ, MORALILTY, supra note 82, at 388. 215. Lucy, supra note 19 (calling choice “will”). 216. RAZ, MORALITY, supra note 82, at 388. 217. See Gardner & Tanguay-Renaud, supra note 139, at 120. 218. See id. 219. See id. 220. See id. 221. See RAZ, MORALITY, supra note 82, at 388. 722 INDIANA LAW REVIEW [Vol. 47:689 practical choices can produce reasons and that an agent’s commitments to certain projects can generate reasons for action, both for herself and for others.222 Nonetheless, this Article suggests that the relation between reasons and projects that proves compelling for one agent may or may not be compelling for another.223 Often these commitments express themselves as a matter of personal and professional style. In the institutional context, these commitments and projects may figure as a style of professional practice. A liberal professionalism encourages a multiplicity of styles to flourish. Permissive justification thus provides a different vision of the relation between liberation and responsibility than does decisive justification.224 On the one hand, permissive justification keeps the decision-maker on the hook for her choices.225 In the doctor example, it shifts much of the responsibility for the consequences of a decision onto the patient. When the doctor presents the patient with a range of undefeated options, the doctor may also identify the style of practice she pursues, and the patient decides whether that treatment style fits the patient. If not, the patient should shop around. Two features of liberal professionalism stand out here. First, this shifting of responsibility for choosing among rationally eligible practice styles works best when the patient can shop around; therefore, it works best if the patient has the opportunity or ability to access different styles. Second, this sort of decision only works if the patient has the information she needs to choose. Accordingly, liberal professionalism imposes an independent duty of candor on the doctor to make clear that there is a range of treatment options and to disclose which style the doctor prefers.226 In this way, under liberal professionalism, responsibility for ensuring rationally adequate decision-making becomes shared among decision- makers. This view stands in contrast to Lucy’s account of the doctor faced with a decision between incommensurable and conflicting courses of treatment.227 One of Lucy’s justifications for his high-stakes, high-standards positions is that, unless justification is decisive, the doctor will engage in some “hasty resort to [her] will as the driving force of decision and action.”228 However, the eligibility model also rejects a too-quick turn to the first available reason. Permissive justification also counsels deliberation to ensure that plural practice styles are included in the decision-making process. Permissions to choose, as this Article has described 222. See supra Part III.A.1. 223. This feature of agent-centered projects forms the basis for one of Joseph Raz’s critique of Hart. See Raz, Hart, supra note 109, at 129-30. 224. See supra Parts I, II. 225. Gardner, The Mark, supra note 102, at 158. 226. I do not suggest that liberal professionals always, or ever, live up to this duty. For example, lawyers might rarely suggest to their clients that another firm across the way would better serve her needs. Nonetheless, the duty of candor operates as a standard by which to criticize the profession even when absent. 227. See Lucy, supra note 19, at 245. 228. Id. 2014] PERMISSIVE JUSTIFICATION 723 them, counsel seeking second opinions and devolving decisions onto those most directly impacted by their choices. Furthermore, egalitarian and inclusiveness reasons support the independent duty of candor in urging the patient to shop around. In such circumstances, we might hope that doctors adopt a personal tone, advising a patient that the options are equally balanced, but that if the doctor in the patient’s situation, the doctor would pick one option over the other. This personal tone emphasizes that the patient and the doctor are on an equal footing in determining what treatment is appropriate. The final decision is not the doctor’s alone (though she may insist on her style of treatment); it also rests with the patient to choose among styles and therefore among doctors. The patient ought to be included in the decision- making process as an equal, not excluded as a subordinate. If Lucy’s claim is that the doctor will (or should) dictate some course of action to the patient, then he has placed himself in the dogmatist’s camp. Here, Lucy supposes, the doctor retains ultimate decision-making authority, and so justification must be decisive in “contexts in which it might be thought desirable to have decisions and actions compelled by reasons.”229 Liberal professionalism, under the more egalitarian model of shared and deferential decision-making outlined above, generally seeks to avoid this form of authoritarianism. Diversity among professional approaches or styles allows a multiplicity of professional practice styles to compete for our rational attention. James Bohman suggests that what he calls “perspectives”230—“different social positions primarily emerging from the range and type of experience” afforded by different practical points of view—are necessary for the deliberative process, because they open up the decision-maker’s deliberative possibilities, expose individuals to competing points of view, and enable practical decision-makers to more accurately assess the range and weight of reasons open to them.231 These perspectives make room for a panoply of different ideologies of professional practice that may not be reducible to simple proxies, such as conservative or liberal agendas.232 That is, rather than assert that professional ideologies (and perhaps political ideologies more generally) fall into to two or three broad categories, we could consider that judges, doctors, and other professionals are as ideological as the wind. Their arguments are likely to be loosely formed and shifting, drawing upon a number of different sources dependent upon the judge’s background, reasoning style, and the circumstances. The concept of a perspective thus permits us to describe more complex relations 229. Id. at 244. 230. James Bohman, Deliberative Democracy and the Epistemic Benefits of Diversity, 3 EPISTEME 171, 171-91 (2006); see also IRIS MARION YOUNG, INCLUSION AND DEMOCRACY (2002). 231. Bohman, supra note 230, at 171-91. 232. See generally CASS R. SUNSTEIN, ARE JUDGES POLITICAL?: AN EMPIRICAL ANALYSIS OF THE FEDERAL JUDICIARY (2006) [hereinafter SUNSTEIN, POLITICAL]. See also Cass R. Sunstein, Deliberating Groups Versus Prediction Markets (or Hayek’s Challenge to Habermas), in SOCIAL EPISTEMOLOGY: ESSENTIAL READINGS 315, 327 (Alvin I Goldman & Dennis Whitcomb eds., 2011) [hereinafter Sunstein, Deliberating]. 724 INDIANA LAW REVIEW [Vol. 47:689 between values, reasons, and experience, and thus accommodates less predictable approaches to law (or medicine, or zoo-keeping). My position contrasts with at least one prominent explanation of judicial ideology in legal decision-making. Cass Sunstein, for example, has argued that ideological factors can skew judicial deliberation towards extremist positions.233 His arguments extend more broadly to identify a major problem infecting group deliberation: that deliberators often respond to interpersonal signals rather than to the merits of the available reasons, and so lack the sort of rational independence necessary to arrive at the best solution.234 Collegiality among deliberators can cause them to second-guess their evaluations of the weight of the available reasons,235 to avoid the stigma of dissent,236 to over-emphasize shared reasons,237 and so on. These interpersonal factors, Sunstein argues, produce ideological patterns of voting among the American federal judiciary that can, if not checked, lead to polarizing and overconfident positions.238 However, Sunstein too hastily equates ideology, outcomes, and reasons. To a large extent, Sunstein’s view of ideology is party political, rather than perspectival. He identifies the judge’s political ideology as adequately expressed by that of the appointing president, characterized as liberal or conservative.239 But such a broad characterization ignores the culture of the legal profession, in which a perspectival account of judicial ideology could provide a way of thinking about adjudication otherwise than as the acts of mere political functionaries. What Sunstein misses, then, is the reason-production aspect of professional opinions, and the manner in which the profession allows for the expression of a range of complex understandings of the law. Consider once more the medical example. Though the appointment of doctors can be quite political (though not predominantly party political), we do not normally think of the doctor as a mere political functionary, following the ideology of particular departments. Rather, an individual doctor’s medical decisions are rarely ideological in that crude sense, but instead differ in ways that are connected to kinds of doctors that they are. A physiotherapist and a surgeon may have a different set of responses when confronted with the same set of symptoms. And doctors may disagree over whether conservative or invasive treatment options are warranted. In that case, in medicine and in the law, we may want a plurality of practitioners advancing a plurality of perspectives—temperamental, stylistic, educational—though that may make the range of opinions more volatile and more 233. SUNSTEIN, POLITICAL, supra note 232, at 9. 234. Id. at 14. 235. Id. at 338; see also Sunstein, Deliberating, supra note 232, at 317. 236. SUNSTEIN, POLITICAL, supra note 232, at 338; Sunstein, Deliberating, supra note 232, at 317. 237. SUNSTEIN, POLITICAL, supra note 232, at 341-42; Sunstein, Deliberating, supra note 232, at 317. 238. See SUNSTEIN, POLITICAL, supra note 232, at 10, 148. 239. Id. 2014] PERMISSIVE JUSTIFICATION 725 fertile. Furthermore, mutual engagement between these different opinions may generate temporary alliances that release different problem and solutions into professional discussion, freeing up topics for debate. In that case, the public good of the law, of medicine, or even of zoo keeping needs to feed of an interplay of different strands of reasons. And this is a justification, pro tanto, for wanting less than decisiveness among categorical reasons. In the face of plural styles, values, reasons, and so on, dogmatism risks overlooking or misrepresenting the range and nature of our reasons for decision.240 Liberal professionalism, on the other hand, respects the diversity of perspectives or practice-styles rationally eligible for adoption by professionals. It counsels inclusiveness as essential to accurately assessing the nature and weight of all the reasons that apply. That is particularly the case if, as Bohman suggests,241 perspectives shape what reasons people find compelling. Dogmatism among professionals over styles of practice risks discounting the rational force of competing perspectives of practice styles. The permissive approach to liberal professionalism requires professionals to acknowledge that others who adopt conflicting perspectives may also be justified in recommending different options, or adopting different practice styles and points of view.242 Including the full panoply of diverse perspectives allows a decision-maker to recognize the range of competing eligible undefeated reasons, to respect other options as permissively justified and so rationally eligible for selection, and (when appropriate) to defer to others’ choice among permissibly justified options.243 The danger of dogmatism lies in promoting an authoritarian and narrow view of professional decision-making.244 In terms of the doctor-patient relationship, Lucy’s version of the doctor-patient relationship presents too judicial a view of doctors. It enshrines the professional’s point of view over the client’s, and sponsors a fruitless search for decisive justifications even when the world is indecisive. Permissive justification, by contrast, permits different professionals’ discrete practice styles or perspectives to tame each other’s excesses and check each others’ views.245 Liberal professionalism thus makes the case for permissive justification. Permissive justification is inclusive, ensuring that all the reasons, values, and 240. See, e.g., FINNIS, COLLECTED ESSAYS, supra note 51, at 252 (discounting that dogmatism is credible). 241. Bohman, supra note 230, at 171-91. 242. See Joseph Raz, Liberalism, Skepticism, and Democracy, 74 IOWA L. REV. 761, 780 (1989). 243. See, e.g., id.; Elizabeth Anderson, The Epistemology of Democracy, 3 EPISTEME 8, 8-22 (2007); Hélène Landemore, Deliberation, Cognitive Diversity, and Democratic Inclusiveness: An Epistemic Argument for the Random Selection of Representatives, 190 SYNTHESE 1209, 1209-31 (2012). 244. See, e.g., FINNIS, COLLECTED ESSAYS, supra note 51, at 252 (rejecting dogmatism as credible). 245. See generally Sunstein, Deliberating, supra note 232. 726 INDIANA LAW REVIEW [Vol. 47:689 perspectives that ought to count do figure in decision-making.246 Not only do they count, but they are properly weighted—as undefeated, incommensurable, incomparable, and so on—rather than being squeezed into some single scale of value. Finally, the spreading of responsibility for decisions and deference to the full range of available justifications promotes an egalitarian understanding of professionalism, one that allows professional and client to look the other in the eye in a non-authoritarian manner. B. Law as a Liberal Profession Like the practice of medicine, the practice of law, for the most part, adheres to the values of liberal professionalism. If one thinks of law as a facilitative247 or as a problem-solving profession, then lawyers operate as liberal professionals in much the same way as I have described a general practitioner in medicine. That is, the lawyer is more or less often in the position of having to explain to a client the range of different incommensurable or incomparable options. The various members of the bar adopt different perspectives and practice styles. Lawyers are expected to suggest that the client shop around to find someone else who will defer to the client’s wishes or adopt a suitable practice style if this lawyer will not. In this way, clients may find the sort of representation that they find most congenial, both for themselves and for their problems, and so participate in decision-making in an egalitarian and inclusive manner. While accepting that law, at this level, is a liberal profession, at a more fundamental level, we might think that the law is dogmatically authoritarian. While the majority of legal practice may occur in the lawyer’s office, the courtroom and the judge’s chambers deservedly occupy an outsized place in legal theory. Courts are the definitive institutions of the legal system.248 The liberal legal practice of solicitors and advocates is derivative of the adjudicative practice of the courts. Plausibly, even if at the level of legal practice law is a liberal profession, at the level of litigation and of judicial decision-making the law is a zero sum game. On that view, judges must announce decisive reasons when deciding against a party. In adjudication, so the claim goes, for every winner there must be a loser, and the judgment of the court must speak to both.249 At the level of judicial decision, the sorts of stylistic reasons that are good enough in the non-legal sphere or at the level of the solicitor are not good enough for the judge. The attorney can tell the client, “I think the law grants a right here, but others do not,” or “You might win here, but then again, you might not”, and so on. Lawyers may adopt particular practice styles, so that they may permissibly say, “My goal is to avoid going to trial and negotiate on your behalf for the best solution, whereas another firm will seek to try the case and adopt a highly adversarial style of practice.” These modes of candor about practice do not 246. See generally Raz, supra note 242. 247. HART, supra note 127. 248. RAZ, PRACTICAL REASON, supra note 38, at 159. 249. See Dworkin, No Right Answer, supra note 1, at 2. 2014] PERMISSIVE JUSTIFICATION 727 appear to be available to the judge. Were the judge to say: “I’m plumping for this result because I am a strict constructionist, but a living constitutionalist could permissibly decide a different way,” one would think she has failed to articulate the right sort of reason for her decision.250 Because the judge is an authority with (on occasion) normative power over the parties, we want categorical reasons that apply to the parties as well, not judge-dependent reasons that apply only to her.251 What we want, following Dworkin (and Finnis, Endicott—and Raz252) is the judge to say that strict constructionism is the best justification of her decision, not simply the judge’s preferred practice style. In most contemporary legal systems, when judges decide cases they must do more than provide reasons of the sort that would justify a private commitment to some personal project (following Raz, we can call these private reasons “prudential” reasons for decision).253 Instead, Raz thinks, the judge must provide impersonal reasons for her decision that apply, not only to the judge, but also to the parties affected by it.254 Impersonal reasons permit agents to make claims on one another,255 by asserting that the agent has some reason to do the act mandated by the decision-maker. What the argument so far establishes is that not any reason will do. The judge, if she is to satisfy the institutional demands of legal justification, must avoid relying on reasons of practice style if she is to adequately justify her decision. One straightforward way of providing style-independent reasons is to find some reason for decision that is both impersonal and decisive. Here, the worry is that permissive justification turns impersonal reasons into prudential ones. As Raz suggests, prudential reasons cannot provide a reason for others. His rather quaint example involves the purchase of sweets: That it may be to my advantage if I refrain from having sweets is a reason for accepting that I ought not to buy them. But that it is to my advantage that you refrain from buying sweets is not a reason for me or anyone else for accepting that you ought not to buy them (unless you ought to promote my interest).256 If judges could advance only prudential reasons for others to obey the law, then it follows that the parties may reject the law as the enforced imposition of another’s will. Prudential reasons thus undermine a central “function” of law, 250. Raz, Hart, supra note 109, at 130. 251. Id. 252. Joseph Raz, Legal Principles and the Limits of Law, YALE L.J. 823, 823-54 (1972). 253. JOSEPH RAZ, THE AUTHORITY OF LAW: ESSAYS ON LAW AND MORALITY 154-55 (1979) [hereinafter RAZ, AUTHORITY]; Robert Brandom, What Do Preferences Express?, in PRACTICAL RATIONALITY AND PREFERENCE 11-12 (Christopher W Morris & Arthur Ripstein eds., 2007); Raz, Hart, supra note 109, at 130. 254. RAZ, AUTHORITY, supra note 253, at 154-55; Raz, Hart, supra note 109, at 130. 255. See, e.g., CHRISTINE M KORSGAARD, THE SOURCES OF NORMATIVITY 8 (1996); Brandom, supra note 253, at 11-12. 256. Raz, Hart, supra note 109, at 130. 728 INDIANA LAW REVIEW [Vol. 47:689 which is to provide compelling reasons for people to do what they judge is not in their interest.257 If the judge is to provide some reason that can bind the parties, then (so it appears) what is required here is not only an impersonal reason for action that agent A could use to demand that agent B act,258 but also one that is decisive. The problem of prudential justification is raised in its most pointed form by the Critical Legal Studies movement.259 Frederick Schauer neatly summarizes a powerful version of that position: [T]he Critical Legal Studies perspective is usefully understood not as the often caricatured claim that “law is politics,” but rather as the broader but nevertheless more plausible claim that “law is almost everything.” . . . [From that perspective, judges] typically have goals—vocations, to put it more grandly—that are describable independently of legal norms, that are more salient for them than legal norms, and that are more important to them than the enforcement of legal norms qua legal norms.260 I have suggested that agents often have project-oriented goals or values, and that these can operate to tip reasons so that, where reasons are finely balanced, one or other justifications become decisive. The Critical Legal Studies point, as interpreted by Schauer, is that these goals are prudential ones, and so do not bind the parties.261 Imposing such goals upon the parties, without providing an opportunity to shop around and without adhering to the duty of candor, would be arbitrary. Dworkin’s turn to decisive justification attempts to eliminate these types of goal, at least, from counting.262 Dworkin’s one-right-answer thesis, like Lucy’s high-stakes, high-standards claim, argued that litigants may properly demand impersonal, decisive, and so judge-independent reasons for decision in the context of legal adjudication.263 In Lucy’s terms, the decision is sufficiently high-stakes that judicial justification must somehow live up to higher, decisive, standards than those applicable to the doctor, accountant, or zoo-keeper’s forms of liberal professional justification.264 The central claim, characteristic of the decisive model of decisive justification, is that judges are not allowed the normative space afforded to other liberal 257. JOSEPH RAZ, ETHICS IN THE PUBLIC DOMAIN: ESSAYS IN THE MORALITY OF LAW AND POLITICS 344 (1994) (“[T]he law’s direct function is to motivate those who fail to be sufficiently moved by sound moral considerations. . . . The law is . . . for those who deny their moral duties. It forces them to act as they should by threatening sanctions if they fail to do so.”). 258. Because agent A believes there is a reason, independent of A and B’s self interest, for agent B to act. 259. See, e.g., Frederick Schauer, The Limited Domain of the Law, VA. L. REV. 1909, 1925 (2004). 260. Id. 261. Id. 262. See generally Dworkin, No Right Answer, supra note 1 263. Id. at 2. 264. Lucy, supra note 19, at 245. 2014] PERMISSIVE JUSTIFICATION 729 professionals. Accordingly, judges either cannot rely on perspectival or stylistic reasons, or they have to lie about their reasons when presiding over zero sum conflicts. Where there are losers, we might think, judges have to look mechanistic: that is what the judicial, adjudicatory form of justification seems to demand. They must become more mechanistic and authoritarian decision-makers than they were before they went to the bench. That simply is what it means to assume the role of the judge. But is this so? Take an example pulled from a criminal law casebook.265 In People v. Rideout,266 the defendant, a drunk driver, stopped in the middle of the road, caused a car crash in which the eventual victim, Keiser, and the car-driver, Reichelt, initially exited their vehicle unharmed.267 Keiser, worried that Reichelt’s car now posed a traffic danger, re-entered the road to turn on the car’s hazard lights.268 A third driver, Welch, did not see the car or Keiser, and collided with both, killing Keiser.269 The question presented to the court was whether Rideout caused Keiser’s death.270 The court determined that three causation tests applied—the response/coincidence test; the apparent safety doctrine; and the voluntary human intervention doctrine.271 The first would render Rideout criminally liable; the other two would identify Keiser’s autonomous act of leaving a place of safety and reentering the roadway as a supervening cause, and so let Rideout off the hook.272 The court acknowledges that under the response/coincidence test, in which the question is whether it is reasonably foreseeable that Keiser would seek to render the car safe for other traffic, Keiser’s actions are foreseeable, as are Welch’s, and so Rideout would be a proximate legal cause of Keiser’s death.273 However, the court prefers the other two doctrines. Here is what the court says: Whether the intervening cause is responsive or coincidental in the case at bar is arguable at best. . . . In our view, Keiser’s decision to reenter the roadway renders the foreseeability factor of little value to the analysis. Rather, that decision directly involves the two remaining factors identified by Dressler that are present here. Those two factors, we believe, compel the conclusion that the intervening cause of the second accident was also a superseding cause.274 The court’s decision is hardly a compelling rejection of the state’s argument. 265. JOSHUA DRESSLER, CASES AND MATERIALS ON CRIMINAL LAW (5th ed. 2006). 266. 727 N.W.2d 630 (Mich. App. 2006), judgment rev’d in part 728 N.W.2d 459 (Mich. 2007). 267. Id. at 632. 268. Id. 269. Id. 270. Id. 271. Id. at 634-35. 272. Id. at 635. 273. Id at 634-35. 274. Id. at 635. 730 INDIANA LAW REVIEW [Vol. 47:689 It acknowledges that the state’s argument competes for selection along with the defendant’s, even though the court attempts to minimize the force of that argument. But a fair reading of the causation tests and of the decision suggests that things are not so clear-cut as the court attempts to suggest. Indeed, the Michigan Supreme Court reversed the Rideout decision on appeal.275 In Rideout, then, the court appears just to acquit the defendant, not based on some decisive reason, but on based upon an indecisive policy decision about where to cut chain of causation.276 Or consider Morrison v. Thoelke,277 a case of first impression concerning formation of contract.278 The problem addressed by the court in Morrison is that, under the general rules of contract formation, an offer may be revoked at any time before its acceptance is communicated to the offeror, but not after acceptance.279 Communication may, however, be a temporally extended process, and where there is a lapse of time between the sending of a revocation and its receipt, the offeree may accept the offer. That is in fact what happened in Morrison. Morrison mailed Thoelke an offer for the sale of property; Thoelke, on receipt of the offer, sent his acceptance of the contract back through the mail to Morrison.280 After mailing the acceptance, but prior to Morrison’s receipt thereof, Thoelke attempted to withdraw his acceptance of the offer.281 The question is whether the acceptance had legal effect once it had been deposited in the post or only upon receipt.282 The judge faced a clear conflict between two legally supported choices, neither of which was decisive.283 In Morrison, there were two conflicting rules, each of which provides persuasive legal authority for the alternative choices.284 The “deposited acceptance” rule stipulates that depositing the letter in the post signifies acceptance; the “acceptance on receipt” rule conceives of the post as the agent of the sender, and delays acceptance until it is received by the offeror.285 Whichever rule was selected would fill a gap in the revocation-of-contract doctrine.286 There was, however, no decisive legal answer.287 The judge acknowledged 275. People v. Rideout 728 N.W.2d 459 (Mich. 2007). 276. Rideout, 727 N.W.2d at 635. 277. Morrison v. Thoelke, 155 So. 2d 889 (Fla. Dist. Ct. App. 1963). 278. Id. at 891. 279. See, e.g., CAL. CIV. CODE § 1586 (West 2007) (“A proposal may be revoked at any time before its acceptance is communicated to the proposer, but not afterwards.”). 280. Morrison, 155 So. 2d at 890. 281. Id. 282. Id. at 891. 283. Id. at 891, 905. 284. See generally id. (discussing the merits and supporting progeny of two conflicting rules: the rule that a contract is created when acceptance is received through the mail, and the rule that a contracts exists only when the offeree mails acceptance). 285. Id. at 891, 897-98. 286. Id. at 891-94. 287. See generally id. at 891-95 (explaining the valid reasons and support for the “acceptance 2014] PERMISSIVE JUSTIFICATION 731 that he faced a case of first impression288 and that other courts were split on the issue, as was the relevant academic literature.289 Here, the jurisdiction had no authoritative reasons guiding the court, and there was no tie-breaking reason to settle which should win out.290 Faced with such a gap, the judge provided somewhat fuzzy reasons for picking one option over the other. As the court put it: [B]oth advocates and critics muster persuasive argument. As [an influential treatise] indicated, there must be a choice made, and such choice may, by the nature of things, seem unjust in some cases. Weighing the arguments with reference not to specific cases but toward a rule of general application and recognizing the general and traditional acceptance of the rule as well as the modern changes in effective long- distance communication, it would seem that the balance tips, whether heavily or near imperceptively, to continued adherence to the [deposited acceptance rule]. This rule, although not entirely compatible with ordered, consistent and sometime artificial principles of contract advanced by some theorists, is, in our view, in accord with the practical considerations and essential concepts of contract law.291 The Morrison decision is at best a tepid recognition of the demands292 for decisive justification that legal authority raises. C. Decisive Formalism and Indecisive Merits The law, because it presents high-stakes scenarios, may make lawyers, the parties, and the public especially willing to tolerate content-independent justifications that do not deal with merits of the legal problem.293 Accordingly, a variety of burdens, presumptions or closure rules may operate to produce decisiveness where the underlying merits of the case do not.294 In such cases, judges may be encouraged to swing from content-independent and formalistic decisiveness to the permissive merits and back again, as the process of justification permits.295 For example, the presumption of innocence operates in a criminal case, along with the burden of proof beyond a reasonable doubt, to set high evidentiary on receipt” and “deposited acceptance” rules). 288. Id. at 891. 289. Id. at 892. 290. Id. at 891-92. 291. Id. at 904. 292. Whether normative or merely psychological. 293. JOSEPH RAZ, ETHICS IN THE PUBLIC DOMAIN: ESSAYS IN THE MORALITY OF LAW AND POLITICS 340 (1995). 294. Id. 295. Id. 732 INDIANA LAW REVIEW [Vol. 47:689 standards that the state must meet if it is to convict an accused of a crime.296 In another case taken from the criminal law casebook,297 State v. Rose,298 a driver non-negligently drove through an intersection, hitting and killing a pedestrian.299 Rose continued over 600 feet before stopping, and then fled the scene.300 The evidence was in equipoise: it was equally (or roughly equally) likely that the victim died on impact, as it was that he died some time after impact (as a result of being dragged down the street under the car’s wheels).301 If the former, the criminal act of killing occurred before Rose had formed any criminally culpable mental state, and so there was no concurrence of the elements of the crime of homicide.302 If the latter, Rose would have had both the request mental state and have engaged in the requisite act at the time the victim died.303 In this case, though the reasons for conviction and acquittal were in equipoise, the court acquitted.304 It did so because the standard of proof in criminal cases—the beyond-a-reasonable-doubt standard—operated to determine that roughly equal reasons were not enough for conviction.305 Instead, some substantial certainty is required.306 Here, a content-independent and formalist closure rule operated instead of the merits to determine what happened when the available options were in rational equipoise. Perhaps the broadest default rule is one that states that “everything that is not clearly prohibited, is permitted.”307 Such a rule, Michael Moore, suggests, is embodied in the “principle of legality backed by its substantive presumptions in favor of liberty and its procedural concern that a would-be criminal have the opportunity to know that the action he contemplates is prohibited.”308 For example, in the California criminal case, Keeler v. Superior Court,309 an estranged husband killed his wife’s unborn while assaulting her.310 The court held that the defendant could not be convicted of homicide, because the unborn child did not fit the statutory definition of a human being, which was an attendant circumstance of the crime.311 The principle of legality, so the court held, required 296. In re Winship, 397 U.S. 358, 359-63 (1970). 297. DRESSLER, supra note 265. 298. 311 A.2d 281 (R.I. 1973). 299. Id. at 282-83. 300. Id. 301. Id. at 284-85. 302. Id. 303. Id. 304. Id. 305. Id. 306. Id. 307. RAZ, ENGAGING REASON, supra note 9; see also Michael S. Moore, Moral Reality Revisited, 90 MICH. L. REV. 2424, 2465 (1992). 308. Id. 309. 470 P.2d 617 (Cal. 1970). 310. Id. at 618. 311. Id. at 632-34. 2014] PERMISSIVE JUSTIFICATION 733 it to eschew broadening the definition of human being, even though the substantive moral merits might demand it, because those reasons conflicted with reasons of fair notice, liberty rights, and the powers of courts to interpret criminal statutes under California law.312 The problem for the court where the available legal rules are indecisive is that no single right answer determines the outcome of the case.313 Indecision presents a problem for a judge who would simply balance the rules or rely on some formalistic type of reasoning that “screens off” consideration of extra-legal reasons and instead requires the judge to rely upon the extant legal rules.314 Joseph Raz, for one, has endorsed these formalistic and content-independent reasons of legal doctrine as a means of settling otherwise indecisive cases.315 The problem, he recognizes, is that by choosing which, among the indecisive reasons, will operate to explain and so justify action, courts act on their own rather than the parties” reasons for action.316 Raz thinks that formalism provides better grounds for decision, because it affords some institutional, shared, and predictable standard of decision, while arbitrary decisions risk being sufficiently random to undermine this feature of law.317 Raz appears to believe that, in such circumstances, if the judge cannot be decisive, she might at least be orderly.318 Rather than rely on the indecisive merits, Raz suggests, that judges ought to turn back to the law to seek some form of determinate outcome.319 Doctrinal or formalist legal values, he believes, provide the only proper source of decisive reasons upon which to base institutional choice. A slightly different option is to acknowledge that what Raz elsewhere calls “pragmatic conflict” in the law320 which operates as a source of liberal legal adjudication. Whatever the larger consequences of pragmatic conflict in undermining the legal values of predictability and stability,321 there is an 312. In Keeler, the substantive moral merits of the case conflicted with reasons of fair notice and the powers of courts (as opposed to legislatures) to interpret criminal statutes under California law. See generally id., see also CAL. PENAL CODE §§ 4-6. 313. Frederick Schauer, Formalism, 97 YALE L.J. 509, 531 (1988). 314. See, e.g., id. at 510 (1988) (“Formalism is the way in which rules achieve their “ruleness” . . . by . . . screening off from a decisionmaker factors that a sensitive decisionmaker would otherwise take into account.”). 315. See RAZ, ETHICS IN THE PUBLIC DOMAIN, supra note 293, at 195. 316. See id. at 201. 317. See id. at 312 (discussing the conflict of choosing the “best” choice over one that adheres to precedent). 318. Id. at 313 (explaining the importance of adhering to a choice once it is made). 319. “Doctrinal reasons, reasons of system, local simplicity and local coherence, should always give way to moral considerations when they conflict with them. But [the doctrinal reasons] have a role to play when natural reason runs out.” Id. at 335. That role is to take the place of personal predilection and to provide an institutional reason for decision when both legal and moral reasons are incommensurable. 320. See id. at 201. 321. See id. 734 INDIANA LAW REVIEW [Vol. 47:689 additional problem for justification that is relatively little noticed in discussions of adjudication. This is the problem of explaining or justifying permissibly indecisive decisions. While it is generally accepted that judges must decide cases, different legal systems, courts, and judges may adopt different standards of explanation. Although a detailed discussion of this phenomenon goes beyond the scope of my argument, it is worth at least indicating the importance of investigating this phenomenon contours that such an investigation might take. D. Institutional Structure and Judicial Reason-Giving It may be that where judges are normatively permitted to select among competing outcomes, they face a choice among styles of legal rationalization. Some may opt for a more dogmatic and doctrinal style at the expense of candor. Others may opt for a fuzzier style, one that expresses rational equipoise, but perhaps at the cost of clarity. Not only judges, but also whole courts may adopt different styles of justification. Some courts may encourage plural opinions, leaving themselves on later occasions (and the lawyers and the parties) to pick and choose among the competing justifications. Others may insist upon unified majority opinions, so that the reasons supporting the judgment are clearly expressed. Each style may render concurring and dissenting opinions more or less important. There may be a host of other ways in which to structure the production of legal opinions and the ways in which the profession and the parties use them to deal with permissive justification in the face of rational indecision. And each institutional arrangement may have different costs and benefits for the status of adjudication as authoritarian and dogmatic or as a form of liberal professionalism. Most important of all, perhaps, is the ability of all the reasons, values, and perspectives that apply to a given legal problem to find their home in the law and their proper acknowledgment or weighting. I have already suggested that, in the American context, an emphasis on principled decision-making has repeatedly led to demands that the standard for settling points of law is that of decisive justification.322 From Herbert Wechsler to Ronald Dworkin, a repeated refrain has been that the resolution of legal arguments, not merely the outcome of legal cases, must turn on decisive justifications.323 At the opposite end of the scale, is what Oliver Wendell Holmes called the judge’s “instinctive preferences and inarticulate convictions.”324 It appears, the outcome of the case is up to the judge: She can decide whichever way she wishes. Not only does reason fail to require a particular outcome, but the judge cannot choose between the options on the basis of reason at all. The available reasons are indecisive as to which to choose. All that is left is her rationally 322. DUXBURY, supra note 57 (citing Fuller, Reason, supra note 42). 323. Chad M. Oldfather, Writing, Cognition, and the Nature of the Judicial Function, 96 GEO. L.J. 1283, 1298-99 (2008). 324. OLIVER WENDELL HOLMES, THE COMMON LAW 35-36 (2009). 2014] PERMISSIVE JUSTIFICATION 735 arbitrary taste or inclination.325 Somewhere in-between, perhaps, is a recognition of various institutional substitutes for judicial ‘second opinions.” One is the possibility of appeal, which is sometimes a matter of right, and sometimes so structured as to depend upon there being some significant issue in dispute.326 In this way, two or three courts can have a look at the issue and determine the outcome.327 A second is the practice of certain trial courts to deny a defendant’s motion for summary judgment where the issues are indecisive, and instead permit the case to go to trial, with the proviso that the court will reconsider the issues at a later point if the plaintiff fails to develop specific facts.328 Motions to dismiss at the end of the moving party’s case based on insufficiency of the evidence may serve the same function.329 All of these types of second opinion or reconsideration provide institutional opportunities for a second-opinion or second look, and may speak to equivocation about the underlying issues. Standard descriptions of indecisive decision identify a familiar range of psychological sources for the resulting judicial choice. These include the “judicial hunch” (or what the judge had for breakfast) as well as political ideology, whether conscious or not.330 Whatever the psychological basis for the resulting decision, having picked a particular option, the judge can only try to render her decision acceptable post hoc by operation of the “characteristic judicial virtues . . . impartiality and neutrality in surveying the alternatives; consideration for the interest of all who will be affected; and a concern to deploy some acceptable general principle.”331 None of these virtues are decisive; rather, they express the values of liberal legal adjudication. If these are the only two options—decisive reasons or whimsy—then decisive reasons are necessary to avoid capricious decision-making. But the range of adjudicatory postures is not binary in this manner, and what is often called judicial will is not exhausted by judicial caprice. Choices may be supported by 325. See SMART & WILLIAMS, supra note 82, at 100. 326. See, e.g., Ross v. Moffit, 417 U.S. 600, 610-14 (1974). 327. See id. 328. See FED. R. CIV. P. 56. 329. See FED. R. CIV. P. 12(b)(6). 330. See, e.g., JEROME FRANK ET AL., LAW AND THE MODERN MIND 108 (1930) (“The process of judging, so the psychologists tell us, seldom begins with a premise from which a conclusion is subsequently worked out. Judging begins rather the other way around—with a conclusion more or less vaguely formed; a man ordinarily starts with such a conclusion and afterwards tries to find premises which will substantiate it.”); Joseph Hutcheson, Jr., Judgment Intuitive The Function of the Hunch in Judicial Decision, 14 CORNELL L. Q. 274, 287 (1928) (“This hunch . . . takes the judge vigorously on to his decision; and yet, the cause decided, the way thither, which was for the blinding moment a blazing trail, becomes wholly lost to view.”); Max Radin, The Theory of Judicial Decision: Or How Judges Think, 11 A.B.A. J. 357, 358-59 (1925) (illustrating how judges struggle to frame the issue that sits before them and therefore must they must ‘select the [framing] that seems to them to lead to a desirable result”). 331. See HART, supra note 127, at 205. 736 INDIANA LAW REVIEW [Vol. 47:689 reasons, albeit indecisive ones. Where reasons are permissively justified, each reason fully justifies some option (in the sense of providing all the reason needed to render the option a rationally adequate candidate for choice), but no reason uniquely justifies the option. In that case, a choice of one among the options can be defended for reasons, rather than as no more than an instinctive preference or inarticulate conviction. Necessity may, however, be only one available posture to judges who may also be able to promote plural outcomes, or default from the merits to formalism, or provide weak support for picking a particular outcome. And these diverse explanatory modes of judicial justification may in turn be related to the institutional structure of judicial decisions, in which multi-judge courts publish multiple, divided opinions authored by each judge on the court, “allowing the next generation to pick what they like”;332 or single majority and dissenting opinions, requiring the court to speak with one voice. These are the problems of the conflict between permissive pluralism and dogmatic authority that are a perennial feature of liberal legal systems. The claim that law as a liberal profession containing diverse judicial styles, though important, does not answer the impulse animating the demand for decisive justification in high-stakes circumstances (such as legal adjudication). In private life, it may be acceptable to just do what we like and rely upon our personal intuitions. In public life, and especially in the life of a public official, such reasons may be unacceptable if forced upon others. The worry is, not that we lack reasons, but that we lack sufficiently neutral ones. Decisive justification is not the answer. It, like permissive justification, is powerless to address that worry. Decisive justification provides a reason that comprehensively defeats competing reasons. It does not provide a reason that is independent of the personal interests of the agent or decision-maker. The sorts of reason that may prove decisive may thus prove decisive for me even if not for you, or for others. The problem of impersonal justification is certainly presented in starker terms by permissive justification than decisive justification. The eligibility model of practical conflict, in which undefeated reasons are sufficient to justify some action or outcome, invites the impersonal question when there is pressure to select one among the eligible options. However, while decisive justification may repress the problem of personal justification, it does not solve it, for the sorts of tipping reason prove decisive may indeed be personal ones. Furthermore, the demand for impersonal justification itself invites a variety of different responses. Is the sort of reason we seek one that is universalizable across agents or circumstances, or one that, though not universalizable, is morally well grounded? In the legal context, we might demand that the agent’s reasons be institutional ones, not moral ones. In that case, agents cannot turn outside the law to find decisive justifications, but must make do with the law as they find it, gaps and all. All this is a matter of some controversy. 332. WILLIAM D. POPKIN, EVOLUTION OF THE JUDICIAL OPINION: INSTITUTIONAL AND INDIVIDUAL STYLES 31 (2007) (quoting Lord Reid in ALAN PATERSON, LAW LORDS 185 (1982)). 2014] PERMISSIVE JUSTIFICATION 737 The idea that universalization can quiet the worries associated with permissive justification is a strong theme in Sir Neil MacCormick’s jurisprudence.333 Universalization provides a distinctive form of rationalization adequate to the justificatory gap left by permissive justification. It does not require the judge to draw upon the reasons she already has, but instead allows the judge to put new reasons into currency, albeit reasons that apply to everyone, and are so neutral in that sense. Universalization respects the fact that the judge is leaving legal reasons for posterity, and so that she is choosing not only for herself, but for others as well. What permissive justification does suggest is that, on occasion, reason is gappy, and that rational gaps need not indicate a rational failure. I have suggested that, in addition, we have a range of procedures to help us deal with those gaps, even if, on occasion, these procedures prove unsuccessful. Some of the procedures invite participation in, and so ownership of, the decision process. Others turn to tie-breaking rules to solve the need for some decisive reason. In any event, the presence of indecisive reasons does not indicate that the sort of permissive justification that goes along with them is somehow substandard, or does not live up to the demands of rational justification. CONCLUSION We began by examining the decisive justification as a means of explaining Dworkin’s demand for bivalent justifications: justifications that not only provide a reason why the winner won, but also why the loser lost.334 Dworkin’s demand for decisive justification and one right answer” is important precisely because it brought to the forefront something lurking in discussions about the responsibility of judges: the idea that justification requires consoling the losers at each point in argument, whichever way it goes, so that the reason that is dispositive for one party must be dispositive for the other.335 While others in the legal profession may have the liberty to take different perspectives on the practice of law, judges engaged in the practice of adjudication are much more narrowly constrained. Dworkin has given names to such objections: the “one right answer” thesis336 is a very definite way of formulating thought that judges are not like other professionals and, given their responsibility to winners and losers, must be categorically decisive with reasons independent of judicial temperament. Permissive justification challenges the decisive justification. It suggests that there is room within the practice of law for different perspectives and styles, not only of lawyering, but of adjudication too. The extent to which concrete legal systems incorporate diverse perspectives and practices, may render some aspects of legal decision-making unpredictable on a day-to-day basis, nonetheless 333. See generally NEIL MACCORMICK, LEGAL REASONING AND LEGAL THEORY (1994). 334. See supra Part I. 335. Dworkin, No Right Answer, supra note 1, at 2. 336. Id. 738 INDIANA LAW REVIEW [Vol. 47:689 creating opportunities for legal agents, including judges, to generate reasons as a feature of a capacious, liberal legal system. What distinguishes judicial decision-making from its lay equivalents is that the judge is, in virtue of her role, peculiarly required to explain how her reasons for decision apply, not only to herself, but also to everyone else. The judge thus has a special duty to provide universal, if not unique, reasons for decision. This Article has all-too-briefly sketched one way in which universalization can make you honor my commitments. Universalization respects the fact that, in committing herself to a particular course of action, the judge is committing the rest of us to that project as well. In doing so, the judge claims, the rest of us cannot refuse to acknowledge her commitments as ones we too have reason to pursue, even if (aside from the judge’s decision) we would have chosen to prioritize other eligible options.