Does Dworkin Commit Dworkin’s Fallacy?: A Reply to Justice in Robes

In an article entitled ‘Dworkin’s Fallacy, Or What the Philosophy of Language Can’t Teach Us about the Law’, I argued that in Law’s Empire Ronald Dworkin misderived his interpretive theory of law from an implicit interpretive theory of meaning, thereby committing ‘Dworkin’s fallacy’. In his recent book, Justice in Robes, Dworkin denies that he committed the fallacy. As evidence he points to the fact that he considered three theories of law—‘conventionalism’, ‘pragmatism’ and ‘law as integrity’—in Law’s Empire. Only the last of these is interpretive, but each, he argues, is compatible with his interpretive theory of meaning, which he describes as the view that ‘the doctrinal concept of law is an interpretive concept’. In this Reply, I argue that Dworkin’s argument that he does not commit Dworkin’s fallacy is itself an example of the fallacy and that Dworkin’s fallacy pervades Justice in Robes just as much as it did Law’s Empire. In an article published in 2003, I described a common fallacy among philosophers of law, in which jurisprudential conclusions are misderived from theories of meaning. I argued that the fallacy can be found in Ronald Dworkin’s classic book Law’s Empire—both in his derivation of an interpretive theory of law from an implicit interpretive theory of meaning, and in his argument that H.L.A. Hart’s conventionalist theory of law was a consequence of Hart’s conventionalist (or, as I put it at the time, traditional) theory of meaning. Since Dworkin is the most famous philosopher of law to have committed the fallacy, I named it after him. In his recent book, Justice in Robes, Professor Dworkin kindly responds to my article. He argues, however, that Dworkin’s fallacy ‘needs a new name’. Rather than committing the fallacy in Law’s Empire, he ‘took pains to warn against it’. * Cabell Research Professor of Law, College of William & Mary, Ph.D. (Philosophy), Yale University, 1990; J.D., Yale Law School, 1996. Email: [email protected] 1 Michael Steven Green, ‘Dworkin’s Fallacy, Or What the Philosophy of Language Can’t Teach Us About the Law’, 89 Virginia L Rev 1897–1952 (2003). 2 Ronald Dworkin, Law’s Empire (Cambridge, Massachusetts: Harvard University Press, 1986). 3 See Green, above n 1 at 1915–1927. 4 Ibid at 1915, 1927–29. 5 Ronald Dworkin, Justice in Robes (Cambridge, Massachusetts: Harvard University Press, 2006) at 226. 6 Ibid. The Author 2008. Published by Oxford University Press. All rights reserved. For permissions, please e-mail: [email protected] at C olege of W iliam nd M ry on N ovem er 9, 2010 ojls.oxjournals.org D ow nladed fom True, he accepted an interpretive theory of meaning—which he describes as the view that ‘the doctrinal concept of law is an interpretive concept’. He also accepted an interpretive theory of law, namely ‘law as integrity’. But he did not think that the latter followed from the former, for he spent much of Law’s Empire arguing against two non-interpretive theories of law (‘conventionalism’ and ‘pragmatism’), each of which is compatible with his interpretive theory of meaning. He suggests that I arrived at my ‘surprising misreading’ because I misunderstood the theory of meaning that he adopted in Law’s Empire. According to Dworkin, I accused him of deriving law as integrity from the view that the concept of law is interpretive. Law as integrity is his theory of law and the view that the concept of law is interpretive is the theory of meaning from which this theory of law is misderived. This is false. In my article I argued that the view that the concept of law is interpretive is itself the theory of law that he misderived from an (implicit) interpretive theory of meaning. Indeed, the point of my critique was that Dworkin committed his fallacy precisely in treating the view that the concept of law is interpretive as if it were a theory of meaning. Because Dworkin continues to make this mistake, his argument that he does not commit Dworkin’s fallacy is itself an example of the fallacy, and Dworkin’s fallacy pervades Justice in Robes just as much as it did Law’s Empire. Or so I shall argue. 1. What is Dworkin’s Fallacy? Dworkin’s fallacy arises when one takes a theory of law to follow from—or to be equivalent to—a theory of the meaning of the word ‘law’ (or of the content of the concept of law). Let us understand a theory of law to be a general account of the conditions that must be satisfied for something to be the valid law of a jurisdiction. The account must be general in the sense that it applies to any jurisdiction (the United States, Uzbekistan, the Roman Empire). Because the criteria for valid law seem to change as one moves from one jurisdiction to the next, a plausible theory of law must offer a higher-order criterion that allows one to identify the criteria, specific to each jurisdiction, that in turn identify the laws of that jurisdiction. In H.L.A. Hart’s theory of law, the higher-order criterion is acceptance by officials in the jurisdiction: Something is the law of a jurisdiction if it satisfies the criteria that the jurisdiction’s officials (judges, legislators, sheriffs and the like) have accepted for enforcing norms. For example, the Securities Exchange Act is valid law in the United States because it satisfies the criteria, such as promulgation in accordance with the U.S. Constitution, that American officials have accepted for identifying norms that may be backed up by 7 Ibid. 8 Dworkin, above n 2 at Chs 4–6. 9 Dworkin, above n 5 at 226–227. 34 Oxford Journal of Legal Studies VOL. 28 at C olege of W iliam nd M ry on N ovem er 9, 2010 ojls.oxjournals.org D ow nladed fom governmental power. Hart calls this official practice of enforcing norms on the basis of these accepted criteria a rule of recognition. An alternative, but not incompatible, description of a theory of law—one favoured by Dworkin—is to speak of the theory as an account of the truth conditions of propositions of law. Consider the proposition ‘The Securities Exchange Act is valid American law’. Hart’s theory of law can be understood as describing the conditions under which that proposition is true—namely if the Securities Exchange Act satisfies the enforcement criteria accepted by American officials. Hart’s theory of law can be understood as conventionalist, because the criteria picking out the valid law of a jurisdiction are those accepted by officials in the jurisdiction. Accordingly, when there is disagreement among officials about how to resolve a hard case, and the disagreement stems from differing views about enforceability criteria, there is no law answering the case, for the rule of recognition is silent on the matter. In contrast, under a non-conventionalist theory of law, the criteria identifying the valid law of a jurisdiction can diverge from the currently accepted views of any group in the relevant jurisdiction. An extreme example of a 10 In fact, sometimes Hart speaks of the rule of recognition not as a social practice but as a proposition specifying the criteria for enforceability within the legal system. See Benjamin C. Zipursky, ‘The Model of Social Facts’ in J. Coleman (ed.), Hart’s Postscript: Essays on the Postscript to the Concept of Law (Oxford: Oxford University Press, 2001) 219 at 227–228. But he also uses the term to refer to the social fact that a certain rule of recognition (in the propositional sense) is practised by officials. The proposition is practised in the sense that officials agree to enforce only that which satisfies the criteria in the proposition. On the distinction between a rule of recognition in the propositional sense and the practice of that rule, see Jules L. Coleman, Practice of Principle: In Defence of a Pragmatist Approach to Legal Theory (Oxford: Oxford University Press, 2001) at 77–78. 11 Dworkin equates a theory of law with an account of the truth conditions of propositions of law at Dworkin, above n 5 at 230, 244, 246–247. Sometimes, however, he treats a theory of law as being articulated at the ‘jurisprudential stage’ of legal theory and an account of the truth conditions of propositions of law as occurring later in the ‘doctrinal stage’. See especially ibid 12–14. Dworkin’s theory of law as integrity, which occurs at the doctrinal stage, would therefore be an account of the truth conditions of propositions of law, not a theory of law. I shall argue below that law as integrity (as well as the two other positions Dworkin entertains at the doctrinal stage, namely Dworkinian conventionalism and pragmatism) are actually applications of Dworkin’s theory of law to particular jurisdictions. They should therefore be understood as accounts of the truth conditions of propositions that describe the law of a particular jurisdiction. A theory of law, in turn, should be understood as an account of the truth conditions of all propositions of law, no matter what particular jurisdiction they happen to be about. 12 H.L.A. Hart, The Concept of Law (Oxford: Oxford University Press, 2nd edn, 1994) at 94. The existence of a rule of recognition in a legal system also requires that the primary rules that are valid according to the rule of recognition are generally—although not necessarily always—obeyed by the population. Ibid at 116–117. 13 Of course, officials might disagree about how a case should turn out even though they agree on enforcement criteria, for they might disagree about whether those criteria are in fact satisfied. They might agree, for example, that the case could be resolved only by statutes enacted by the legislature but disagree about whether there was a quorum required for enactment. 14 In speaking of Hart’s theory as conventionalist, I mean only and vaguely that the rule of recognition rests upon agreement within the jurisdiction. I do not take a stand on whether it should be understood as the solution to a coordination problem. See David K. Lewis, Convention: A Philosophical Study (Cambridge, Massachusetts: Harvard University Press, 1969). For such an account of the rule of recognition, see Gerald J. Postema, ‘Coordination and Convention at the Foundations of Law’, 11 Journal of Legal Studies 165–203 (1982); Jules Coleman, ‘Incorporationism, Conventionality, and the Practical Difference Thesis’ in J. Coleman (ed.), Hart’s Postscript: Essays on the Postscript to the Concept of Law (Oxford: Oxford University Press, 2001) 99 at 114–121. Nor do I take a stand on whether it should be understood as a shared cooperative activity in Michael Bratman’s sense: Michael E. Bratman, ‘Shared Cooperative Activity: Three Features’, 101 Philosophical Review 327–341 (1992). For such an account of the rule of recognition, see Coleman, above n 10 at 96–99. SPRING 2008 Does Dworkin Commit Dworkin’s Fallacy? 35 at C olege of W iliam nd M ry on N ovem er 9, 2010 ojls.oxjournals.org D ow nladed fom non-conventionalist theory of law would simply use the same set of criteria for identifying the laws of all jurisdictions. Such a theory might say, for example, that God’s word, as expressed in the Bible, is the law of every jurisdiction in the world—no matter what people in the jurisdiction might think. But a theory of law can be non-conventionalist even if it is sensitive to the current views of people in a jurisdiction, provided that it does not simply equate the criteria identifying the laws of the jurisdiction with these current views, the way a conventionalist theory of law does. As we shall later see, Dworkin’s theory of law is non-conventionalist in this more moderate sense. Let us now turn to theories of the meaning of the word ‘law’ (or theories of the content of our concept of law). Although such theories might be described as semantic, the word ‘semantic’ is ambiguous. According to one sense of the word, to make a semantic claim is simply to identify the meaning of a word—or the content of a concept associated with the word. To say that ‘bachelor’ means unmarried male (or that the content of the concept of bachelor is unmarried male) is to make a semantic claim in this sense. So understood, theories of law can indeed follow from semantic theories. Hart, for example, understands his theory of law to be an account of the content of the concept of law. Dworkin too understands theories of law, including his own, to be attempted revelations of the content of the concept of law. For both Hart and Dworkin, a theory of law is semantic in the sense that it identifies what the content of the concept of law is. But there is a different sense of the word ‘semantic’, which refers not to claims about what the content of a concept (or the meaning of a word) is, but rather to claims about how this content or meaning is generated. Sometimes the term ‘metasemantic’ is used to identify semantic claims in this second sense. The semantic position that ‘bachelor’ means unmarried male says nothing about why ‘bachelor’ has this meaning and so is not a metasemantic claim. 15 See n 19 below. 16 See H.L.A. Hart, ‘Jhering’s Heaven of Concepts and Modern Analytical Jurisprudence’ in Essays in Jurisprudence and Philosophy (Oxford: Oxford University Press, 1983) 265 at 274; Nicos Stavropoulos, ‘Hart’s Semantics’ in J. Coleman (ed.), Hart’s Postscript: Essays on the Postscript to the Concept of Law (Oxford: Oxford University Press, 2001) 59, 63–88. 17 For example, in Justice in Robes, Dworkin notes that ‘[i]n one respect . . . [Hart and I] are in the same boat. We both believe that we will understand legal practice and phenomena better if we undertake to study, not law in some particular manifestation, like the law of product liability in Scotland, but the very concept of law.’ Dworkin, above n 5 at 145. In Justice in Robes, Dworkin argues that a general theory of law should be understood as investigating the doctrinal concept of law, that is, the concept that one employs when one makes claims about what is valid or invalid law in various jurisdictions. Ibid at 9. He argues that some confusion has resulted from the failure to distinguish the doctrinal from other concepts of law. Ibid at 2–5. I will accept, at least for the purposes of argument, Dworkin’s position that a theory of law reveals the content of the doctrinal concept of law. 18 David Kaplan, ‘Afterthoughts’ in J. Almog et al. (eds), Themes from Kaplan (Oxford: Oxford University Press, 1989) 565 at 573–576; Green, above n 1 at 1905 n 21; Jules L. Coleman and Ori Simchen, ‘‘‘Law’’’, 9 Legal Theory 1–41 at 12, 18 (2003). 36 Oxford Journal of Legal Studies VOL. 28 at C olege of W iliam nd M ry on N ovem er 9, 2010 ojls.oxjournals.org D ow nladed fom Likewise, Hart’s and Dworkin’s theories of law are not metasemantic, for in saying what the content of the concept of law is, they do not say how it came to have this content. Dworkin primarily uses the term ‘semantic’ in this second, metasemantic, sense. For example, in Justice in Robes, he describes as ‘semantic’ the view that a concept is ‘criterial’. To say that a concept is criterial is to say how the concept gets its content. If a concept is criterial, its content is fixed by agreement ‘on a definition—rough or precise—that sets out the criteria for the correct application of the associated term or phrase’. The concept of a bachelor is an example of a criterial concept: ‘People share the concept of bachelorhood only when they know that a bachelor is an unmarried male’. The theories of meaning to which I referred in my article were metasemantic, not semantic, and in this reply I will use the term ‘metasemantic’ to refer to them. The argument of my article, therefore, was that Dworkin’s fallacy arises when one takes a theory of law to follow from (or to be equivalent to) a metasemantic theory of the meaning of the word ‘law’ or of the content of the concept of law. One such metasemantic theory is the view that the concept of law is criterial. (In my article, I described this metasemantic theory as the traditional theory of meaning.) The metasemantic view that a concept is criterial can be understood as conventionalist in the following sense: The concept is shared only to the extent that there is current agreement concerning the criteria for the concept’s use. Other people share my concept of a bachelor only if they also currently accept the criteria unmarried and male for its use. If they don’t, they are thinking about something else when they use the concept of a bachelor. In contrast, under a non-conventionalist metasemantic theory, two people can share a concept with the same content even if they disagree concerning the criteria for the concept’s use. An example of an extreme (and extremely implausible) non-conventionalist metasemantic theory is the view that God determines the content of our concepts. Whatever criteria I might accept for using the concept of a bachelor, the concept has whatever content God says it does. Right now my concept of a bachelor might have the content married female, in which case I am thinking about married females whenever I think about bachelors. Such a metasemantic theory is non-conventionalist, because we can be thinking about the same thing (namely whatever God says we are) even if we disagree about the criteria for using a concept. 19 I follow him in using the word ‘semantic’ in this way in Michael Steven Green, ‘Halpin on Dworkin’s Fallacy: A Surreply’, 91 Virginia L Rev 187–201 (2005). 20 Dworkin, above n 5 at 9. In Law’s Empire, Dworkin describes as ‘semantic’ the view that ‘[w]e follow shared rules . . . in using any word: these rules set out criteria that supply the word’s meaning’. Dworkin, above n 2 at 31. 21 Dworkin, above n 5 at 9. 22 For a fuller description of the metasemantic conventionalist approach, see Green, above n 1 at 1899–1903. SPRING 2008 Does Dworkin Commit Dworkin’s Fallacy? 37 at C olege of W iliam nd M ry on N ovem er 9, 2010 ojls.oxjournals.org D ow nladed fom But a metasemantic theory can be non-conventionalist even if it is sensitive to the current views of people using concepts, provided that it does not simply equate the content of their concepts with these current views, the way a conventionalist metasemantic theory does. As we shall see, Dworkin’s implicit metasemantic theory is non-conventionalist in this more moderate sense. Dworkin’s fallacy is the belief that the level of conventionalism of a theory of law follows from the level of conventionalism of the metasemantic theory that applies to the concept of law. To see why this is a fallacy, consider the nonconventionalist metasemantic theory that God establishes the content of our concept of law. Such a theory is compatible with a conventionalist theory of law, because God might have decreed that the content of our concept of law is that identified in H.L.A. Hart’s conventionalist theory of law. Conversely, conventionalist metasemantics is compatible with a nonconventionalist theory of law, since someone might accept the following as the criterion for using the concept of law: God’s word, as identified in the Bible, is the law of all jurisdictions. Under conventionalist metasemantics, the content of this person’s concept of law would be properly identified by a non-conventionalist theory of law. Dworkin’s is another example of a nonconventionalist theory of law that could be the content of a criterial concept. All that is required is that the relevant person accepts that the concept should be applied in the manner identified in Dworkin’s theory of law. In my article, I argued that Dworkin’s fallacy takes two forms in Law’s Empire. On the one hand, Dworkin misderives his non-conventionalist theory of law from an implicit non-conventionalist metasemantics. But Dworkin’s fallacy also occurs when Dworkin argues that H.L.A. Hart arrived at his conventionalist theory of law because of his attachment to conventionalist metasemantics. Dworkin embraces the same interpretation of Hart in Justice in Robes. Legal positivists like Hart, he argues, assume ‘that all concepts depend on a convergent linguistic practice’. In fact, Hart denied that he was a metasemantic conventionalist. He did not think that the reason the concept of law had the content identified in his theory of law was because language-users agreed that it did. As he put it, ‘the criteria for the application of a concept with a constant meaning may both vary and be controversial’. As we have seen, there is nothing incoherent about this position. A philosopher of law is free to hold a conventionalist theory of law and a non-conventionalist metasemantics. 23 Dworkin, above n 2 at 31–35. 24 Dworkin, above n 5 at 30–33, 165–166, 225–226. 25 Ibid at 225. I explore this example of Dworkin’s fallacy more fully in Michael Steven Green, ‘Dworkin v. The Philosophers: A Review Essay on Justice in Robes’, U of Illinois L Rev 1477–1503 (2007). 26 Hart, above n 12 at 246. 38 Oxford Journal of Legal Studies VOL. 28 at C olege of W iliam nd M ry on N ovem er 9, 2010 ojls.oxjournals.org D ow nladed fom Some have argued, however, that Hart was actually a metasemantic conventionalist. But the fact remains that, even if he was, it is still a fallacy to conclude that his conventionalist theory of law was the consequence of his metasemantic views, since a non-conventionalist theory of law is compatible with metasemantic conventionalism. It is also worth noting that, even though Hart’s theory of law did not follow from his metasemantics, this metasemantics might have had some influence on his views about whether there are determinate answers to hard cases. Let us assume that, according to the enforceability criteria agreed upon by officials in a jurisdiction, someone riding a bicycle in a park should be sanctioned only if his action falls under a particular statute, which prohibits ‘vehicles’ in a park. Under Hart’s conventionalist theory of law, this official agreement means the question of whether riding a bicycle in a park is illegal is answered solely by the statute. If Hart was a metasemantic conventionalist, the fact that officials disagree about whether bicycles are vehicles might mean that it is indeterminate whether the statute refers to bicycles. As a result, the case would be legally indeterminate. But even if metasemantic conventionalism would have this effect assuming Hart’s conventionalist theory of law, it could not force Hart to adopt a conventionalist theory of law in the first place. Hart’s metasemantic conventionalism would leave him free to argue that the question of whether riding a bicycle in a park is illegal should be answered by reference to sources other than the statute. Hart’s criteria for using the concept of law would allow him to consider these other sources, if he accepted a theory of law, like Dworkin’s, that looked beyond official agreement. Why does Dworkin succumb to his fallacy? As I noted in my article, one reason is that he treats a rule of recognition (which is an official legal practice particular to a jurisdiction) as if it were a practice of applying the concept of law (which is a linguistic practice in which anyone can participate—including those in situations of anarchy). Once the two practices are conflated, Dworkin’s misinterpretation of Hart seems to follow: Hart’s alleged metasemantic demand for agreement among participants in the linguistic practice (on pain of their using different concepts of law) starts looking like his jurisprudential demand for agreement among participants in a rule of recognition (on pain of there being no law in that jurisdiction). Conversely, Dworkin’s derivation of his own non-conventionalist theory of law from non-conventionalist metasemantics also seems to follow once one 27 Green, above n 25 at 1496–1497; Stavropoulos, above n 16. 28 For arguments that metasemantic theories would not necessarily have such consequences, see Brian H. Bix, ‘Can Theories of Meaning and Reference Solve the Problem of Legal Determinacy?’, 16 Ratio Juris 281–295 (2003); Green, above n 1 at 1946–1948. 29 Ibid at 1919–1929. I explore this source of Dworkin’s fallacy in greater detail in Green, above n 25 at 1491–1495. SPRING 2008 Does Dworkin Commit Dworkin’s Fallacy? 39 at C olege of W iliam nd M ry on N ovem er 9, 2010 ojls.oxjournals.org D ow nladed fom conflates the two practices. The absence of any metasemantic demand that participants in the linguistic practice agree on criteria for using the concept of law starts looking like the absence of any jurisprudential demand that officials agree concerning criteria for enforcement. Dworkin rejects metasemantic conventionalism because he believes it makes meaningful disagreement about a concept’s content impossible. Such disagreement would simply mean that different concepts were being used and that people were talking past one another. Although Dworkin admits that conventionalist metasemantics rightly applies to certain concepts, such as the concept of a bachelor, he argues that it is implausible concerning essentially contested concepts. For example, according to the criteria accepted by most white Southerners in 1850, slavery should be judged just. If conventionalist metasemantics applies to the concept of justice, then slavery was indeed just according to their concept of justice, even though it is not according to ours. But this, Dworkin argues, misdescribes our disagreement. We think that they were wrong about justice in a shared sense, not right about justice in some different sense. The inability of conventionalist metasemantics to explain disagreement about the content of a concept is the ‘semantic sting’. Dworkin also believes that conventionalist metasemantics cannot apply to the concept of law. Consider, for example, his debate with Hart. Neither he nor Hart wants to conclude that they are talking past one another simply because they offer different theories of law. Their disagreement is genuine. It appears, therefore, that one must give up conventionalist metasemantics with respect to the concept of law. Articulating the content of the concept of law cannot involve simply identifying the criteria people currently accept for using the concept. Dworkin’s criticisms of conventionalist metasemantics may be correct. But Dworkin commits his fallacy in assuming that his rejection of conventionalist metasemantics gives support to his non-conventionalist theory of law. That this is a fallacy should now be clear: Non-conventionalist metasemantics is compatible with a conventionalist theory of law, because even if the content of the concept of law is not simply read off from the criteria that we accept for using the concept, it may nevertheless follow from the true content of the concept of law that the law of a jurisdiction is whatever is identified by the criteria for enforcement agreed upon by officials in that jurisdiction. 30 Dworkin, above n 5 at 9. 31 Dworkin, above n 2 at 73. 32 Ibid at 43–46; Dworkin, above n 5 at 223–26. 33 At times, however, Hart suggests that he and Dworkin are indeed talking past one another. Hart, above n 12 at 239–244. 40 Oxford Journal of Legal Studies VOL. 28 at C olege of W iliam nd M ry on N ovem er 9, 2010 ojls.oxjournals.org D ow nladed fom 2. The View that the Concept of Law is Interpretive Is a Theory of Law, Not a Metasemantic Theory In his response to my article, Dworkin denies that he misderived his nonconventionalist theory of law from a non-conventionalist theory of meaning. Although he accepted an interpretive theory of meaning (which he describes as the view that ‘the doctrinal concept of law is an interpretive concept’) and accepted an interpretive theory of law (namely law as integrity) he did not think that the latter followed from the former, for he spent three chapters of Law’s Empire arguing against two non-interpretive theories of law, each of which was compatible with the view that the concept of law is interpretive: One of these is a form of legal pragmatism, the theory I discuss at greater length earlier in this book. Another I called conventionalism, which is a version of legal positivism offered in the spirit I described in Chapter 6, that is, as itself an interpretation of contemporary practice. I argued that my own interpretive answer to the question posed at the doctrinal stage of legal theory is best because it provides the best such interpretation not because the doctrinal concept [of law] is itself an interpretive concept. This sounds persuasive. If Dworkin thinks that a number of theories of law, including a conventionalist theory (in which the law is limited to what officials accept are enforceable norms) are compatible with his interpretive metasemantics, how can he possibly be accused of Dworkin’s fallacy? He left open which theory of law—conventionalism, pragmatism or law as integrity—follows from his metasemantic presuppositions. According to Dworkin, I accused him of deriving law as integrity from the view that the concept of law is interpretive. But I never said that law as integrity was the interpretive theory of law that he misderived from metasemantic premises. I argued that the misderived theory of law was the view that the concept of law is interpretive. Indeed, the point of my critique was that Dworkin committed his fallacy precisely in assuming that the view that the concept of law is interpretive is metasemantic in nature. Since Dworkin continues to describe it as metasemantic, his argument that he does not commit Dworkin’s fallacy is itself an example of the fallacy. The following is Dworkin’s description of the view that the concept of law is interpretive: We share [the concept of law] as actors in complex political practices that require us to interpret these practices in order to decide how best to continue them, and we use 34 Dworkin, above n 5 at 226. 35 Dworkin, above n 2 at Chs 4–6. 36 Dworkin, above n 5 at 226. SPRING 2008 Does Dworkin Commit Dworkin’s Fallacy? 41 at C olege of W iliam nd M ry on N ovem er 9, 2010 ojls.oxjournals.org D ow nladed fom the doctrinal concept of law to state our conclusions. We elaborate the concept by assigning value and purpose to the practice, and we form views about the truth conditions of the particular claims that people make within the practice in the light of the purposes and values that we assign. In this process, ‘the interpreter settles on some general justification for the main elements of the practice’ and then reforms it by ‘adjust[ing] his sense of what the practice ‘‘really’’ requires so as better to serve the justification’. In both Law’s Empire and Justice in Robes, Dworkin describes this view as if it were a theory of meaning in competition with metasemantic conventionalism. But it is actually a theory of law compatible with metasemantic conventionalism. The problem, once again, is Dworkin’s conflation of the linguistic practice of talking about the law with legal practices specific to a jurisdiction. Consider a metasemantic conventionalist, that is, someone who believes that the content of her concept of law is determined by the criteria she accepts for using the concept. She is still free to argue that the concept of law is interpretive in Dworkin’s sense of the term, for she could accept the following as the criterion for using the concept of law: The law of a jurisdiction is whatever is identified by criteria that result from moral reflection upon the legal practices of the jurisdiction. Rather than offering a metasemantic account that explains why the content of the concept of law is what it is, Dworkin has simply assigned a content to the concept of law. He has, in short, offered a theory of law. Dworkin tries to make the view that the concept of law is interpretive look like a metasemantic theory by situating it within a more general account of ‘interpretive concepts’, which include the concepts of justice, liberty, equality and democracy. But the assignment of content to a concept is not transformed into a metasemantic account simply by being conjoined with the assignment of content to other, similar, concepts. For example, an account of the content of marital-status concepts (such as the concepts of bachelor, husband, wife, and widow) is not metasemantic simply by being general, for to describe the content of these concepts is to say nothing about why these concepts have the content they do. In short, by discussing interpretive concepts generally, Dworkin is simply offering theories of a number of things—e.g. justice, liberty, equality, democracy and law. All these things are similar, according to Dworkin, in the sense that they involve moral interpretation of human practices. This moral interpretation concerns the point of these practices and attempts to make these practices the best that they can be. 37 Ibid at 12. 38 Dworkin, above n 2 at 66. 39 Ibid at 87; Dworkin, above n 5 at 226–227. 40 Green, above n 1 at 1921. 41 Dworkin, above n 5 at 11. 42 Oxford Journal of Legal Studies VOL. 28 at C olege of W iliam nd M ry on N ovem er 9, 2010 ojls.oxjournals.org D ow nladed fom But in offering these theories, Dworkin does not answer the metasemantic question of why these concepts have the content that they do—for example, why they do not look at human practices non-morally the way Hart thinks the concept of law does or, indeed, why they are about human practices at all, rather than being about, say, unmarried males or drainpipes. In contrast, metasemantic conventionalism can answer these questions. According to metasemantic conventionalism, the concept of law is about human practices rather than drainpipes, because this follows from the criteria that people accept for using the concept of law. 3. Dworkinian Conventionalism, Pragmatism, and Law as Integrity Are Applications of Dworkin’s Theory of Law to Jurisdictions, Not Theories of Law on Their Own But if the view that the concept of law is interpretive is Dworkin’s theory of law, what are conventionalism, pragmatism, and law as integrity? I suggested in my article that these positions should be understood as applications of Dworkin’s interpretive theory of law to particular jurisdictions. My suggestion was only in a footnote, however – apparently missed by Dworkin—and I am grateful to have the opportunity to make this point more fully here. (Because what Dworkin calls ‘conventionalism’ is not a theory of law, I will describe it as ‘Dworkinian conventionalism’, to distinguish it from conventionalist theories of law like Hart’s.) According to Dworkinian conventionalism, the law of a jurisdiction is identified by the criteria for enforcement currently accepted by officials in that jurisdiction. Dworkin rightly insists that Dworkinian conventionalism is compatible with the view that the concept of law is interpretive. After reflecting critically upon the underlying moral purposes of the legal practices of a jurisdiction, one might conclude that the law of that jurisdiction is best limited to what is currently accepted as enforceable by its officials, because it is part of the moral purpose of these legal practices that people’s settled expectations not be upset. For example, imagine a jurisdiction that takes the following approach to a hard case: Although a judge is obligated to announce a rule that would resolve the case, the announced rule will not apply to the parties in the hard case itself. The rule will be binding law only going forward. Instead, the judge must dismiss the plaintiff ’s complaint for failure to state a claim, on the grounds that the plaintiff can point to no law that would entitle her to relief. Despite the dismissal, however, if the judge’s announced rule would have led the defendant to be liable had it been applied to the case, the plaintiff will receive the 42 Green, above n 1 at 1920 n 63. SPRING 2008 Does Dworkin Commit Dworkin’s Fallacy? 43 at C olege of W iliam nd M ry on N ovem er 9, 2010 ojls.oxjournals.org D ow nladed fom monetary relief she requested (or, if she was asking for injunctive relief, the monetary equivalent of that injunctive relief) from the state treasury, thus providing plaintiffs with an incentive to bring suits in hard cases and so to contribute to the creation of new law. In this hypothetical jurisdiction Dworkinian conventionalism is true, because a moral interpretation of its legal practices shows that respecting settled expectations is paramount. But the legal practices of another jurisdiction might have different moral purposes incompatible with Dworkinian conventionalism. For example, according to Dworkin, a moral interpretation of Anglo-American legal practices reveals that what is most important is not respecting the settled expectations of citizens, but upholding the moral goal of integrity. According to this goal, ‘rights and responsibilities flow from past decisions and so count as legal, not just when they are explicit in those decisions but also when they follow from the principles of personal and political morality the explicit decisions presuppose by way of justification’. The law should be identified, not by the criteria accepted by officials, but by the criteria that result from ‘the best interpretation of the bulk of pertinent standing law’. There can be law answering a hard case in law-as-integrity jurisdictions such as ours, because there can be a best interpretation of the bulk of pertinent standing law even when judges disagree about criteria for enforcement. But this means that Dworkinian conventionalism and law as integrity are not theories of law. Dworkinian conventionalism is not a theory of law, for it tells us nothing about the law in jurisdictions in which respecting settled expectations is not paramount. And law as integrity is not a theory of law, for it tells us nothing about the law in jurisdictions that do not seek to uphold the moral goal of integrity. Dworkin’s real theory of law is the position that generates Dworkinian conventionalism and law as integrity when applied to jurisdictions. What is this theory of law? His view that the concept of law is interpretive, that is, that the law of a jurisdiction is identified through a moral interpretation of the legal practices of that jurisdiction. This means that in law-as-integrity jurisdictions, the law is doubly interpretive. It is interpretive, first of all, in a way that applies to every jurisdiction. In every jurisdiction the scope of the law depends upon a moral interpretation of the legal practices of that jurisdiction. But the law is also interpretive in law-as-integrity jurisdictions in a way not shared by jurisdictions to which Dworkinian conventionalism (or pragmatism) apply. In lawas-integrity jurisdictions, the moral interpretation of the legal practices of 43 Dworkin, above n 2 at 96. 44 Dworkin, above n 5 at 226. 45 Dworkin, above n 2 at 226. 44 Oxford Journal of Legal Studies VOL. 28 at C olege of W iliam nd M ry on N ovem er 9, 2010 ojls.oxjournals.org D ow nladed fom that jurisdiction recommends further interpretation. The law should be identified through ‘the best interpretation of the bulk of pertinent standing law’. The fact that a non-conventionalist theory of law stands in the background of Dworkinian conventionalism helps explain why no positivist has ever been remotely attracted to it. Under the conventionalist theory of law that positivists advocate, the law must be identified by the enforcement criteria that are currently accepted by officials in the relevant jurisdiction. This is essential to the law. And the theory of law standing behind Dworkinian conventionalism rejects this. Under Dworkin’s theory of law, the law may or may not be limited to what is accepted by officials as enforceable, depending upon how the moral interpretation of legal practices turns out. In short, Dworkin is wrong to assume that the interpretive theory of law that I claimed he misderived from metasemantic premises is law as integrity, that is, a theory that looks to the ‘the best interpretation of the bulk of pertinent standing law’. The theory of law that I attributed to Dworkin was the view that the concept of law is interpretive, that is, that the law is what is identified by a moral interpretation of legal practices. I admit that it is easy to confuse the interpretation of legal practices with the interpretation of the bulk of pertinent standing law, but since Dworkin himself first drew this distinction, he of all people should be sensitive to it. 46 Dworkin, above n 5 at 226 (emphasis mine). At times Dworkin suggests that the laws of other jurisdictions can be understood only in the light of the moral interpretation of the legal practices of one’s own jurisdiction. Having determined that law as integrity is the proper approach to our jurisdiction, the laws of all other jurisdictions would be seen in the light of law as integrity as well. See Dworkin, above n 2 at 102 (useful theories of law are ‘interpretive of a particular stage of a historically developing practice’ and ‘are by their nature addressed to a particular legal culture, generally the culture to which their authors belong’). But see Dworkin, above n 5 at 163 (he seeks to answer the question of what makes a claim of law true ‘not just for a particular legal system, like English law, but for law in general, whether in Alabama or Afghanistan’.); ibid at 185 (his account ‘aims at very great generality’); ibid at 230–231 (same). One problem with the suggestion that the laws of other jurisdictions are seen in the light of the moral purposes of the practices of one’s own jurisdiction is it makes Dworkin vulnerable to the charge of parochialism. If we must see all legal systems in the light of law as integrity, wouldn’t people in my hypothetical jurisdiction see us in the light of Dworkinian conventionalism? But setting this problem aside, it would still be the case that law as integrity is the application of Dworkin’s interpretive theory of law and thus not a theory of law in its own right, for the fact would remain that the criteria that identify the laws of our jurisdiction (and, through it, all other jurisdictions) are whatever result from the moral interpretation of our legal practices. 47 See Hart, above n 12 at 241–242, 248–250. 48 Indeed Dworkin himself admits that the argument for Dworkinian conventionalism is ‘self-defeating’ for the positivist. Dworkin, above n 5 at 180. 49 Ibid at 226. 50 At Green, above n 1 at 1916, for example, I characterized Dworkin’s interpretive theory of law in the following way: According to Dworkin’s interpretive jurisprudence, the law consists of the set of norms that would be accepted after a process in which ‘the interpreter settles on some general justification for the main elements of [legal] practice’ and then reforms it by ‘adjust[ing] his sense of what the practice ‘‘really’’ requires so as better to serve the justification’. The quoted passage is from Law’s Empire, Dworkin, above n 2 at 66, where Dworkin describes interpretive concepts. I repeatedly spoke of Dworkin’s theory of law as involving moral interpretation of legal practices, and never as involving moral interpretation of the bulk of pertinent standing law. Green, above n 1 at 1916, 1919, 1920 n 62, 1921, 1925, 1928 n 76. SPRING 2008 Does Dworkin Commit Dworkin’s Fallacy? 45 at C olege of W iliam nd M ry on N ovem er 9, 2010 ojls.oxjournals.org D ow nladed fom 4. Conventionalist, Realist and Interpretive Metasemantics But if the view that the concept of law is interpretive is Dworkin’s theory of law, what is Dworkin’s interpretive metasemantics, from which he misderived this theory of law? I argued in my article that this metasemantic theory is latent in Dworkin’s writings. It is because Dworkin’s metasemantic theory is latent that I presented my account of his metasemantic views in my article with some hesitancy. That Dworkin’s metasemantic views are veiled in Law’s Empire is itself a consequence of Dworkin’s fallacy. Because he presents his theory of law as if it were a metasemantic theory, his descriptions of his own opinions on metasemantic issues are unreliable. And because he makes the same mistake in Justice in Robes, nothing Dworkin says in this recent book makes the project of uncovering his true metasemantic views any easier. Indeed, some of his comments in Justice in Robes manage to muddy the waters even more. When addressing the question of Dworkin’s metasemantic views, one must keep in mind that a metasemantic theory needs to answer the following problem: We know that many people disagree about the content of the concept of law. If conventionalist metasemantics applied to the concept (that is, if the concept were, as Dworkin puts it, ‘criterial’), we would have to conclude that people like Hart and Dworkin are simply talking past one another. But we don’t want to come to such a conclusion. So we must provide an alternative explanation of why there is a right answer to their debate, an answer that does not simply read off the content of the concept of law from current beliefs and attitudes. In my article I spelled out three answers to the question of what determines a concept’s content, the last of which, I suggested, was Dworkin’s interpretive metasemantics. The first, which I called traditionalism, was conventionalist metasemantics (or, as Dworkin puts it, the view that a concept is criterial). The second, realism, was an alternative to conventionalist metasemantics that offers the possibility of genuine disagreement over the content of a concept. Realism is most plausibly applied to natural-kind concepts (like the concepts of tiger or water). Language-users’ convergent beliefs and attitudes identify paradigm samples falling under the natural-kind concept, but the actual nature of the paradigm samples—something about which language-users may have diverging or even no beliefs—determines the concept’s content. If realist metasemantics applies to the concept of water, we can say that we share the same concept with people in the sixteenth century, even though the criteria they associated with the concept (being an odourless, colourless, potable liquid) are different from our criteria (having the structure H2O). 51 Ibid at 1909–1915; 1918 (arguing that ‘it took a good deal of interpretation on my part to draw an interpretive theory of meaning out of Dworkin’s work’). 52 Ibid at 1904. 46 Oxford Journal of Legal Studies VOL. 28 at C olege of W iliam nd M ry on N ovem er 9, 2010 ojls.oxjournals.org D ow nladed fom The concept of water had the same content (namely H2O) for them, as it does for us, because our common paradigm samples of water actually have the structure H2O. Realists feel that their metasemantics is superior to conventionalism, because conventionalism leaves no place for conceptual disagreement in science: Einsteinians and Newtonians cannot meaningfully disagree about the nature of mass for, if they do disagree, they are simply employing different concepts of mass and so talking past one another. Given that realist metasemantics can account for meaningful disagreement about the content of a concept, I asked whether Dworkin might be a realist with respect to the concept of law. But I quickly rejected this option: One might expect [Dworkin] to conclude, in realist fashion, that . . . reference is fixed by the underlying essential structure of the stuff that people were pointing to when they first used the word. But Dworkin avoids such an approach, apparently because he worries about the intelligibility of talk about metaphysically real entities with underlying structures that could fix this reference. At the time, I didn’t have that much to go on in coming to this conclusion, but corroborating evidence has now appeared in Justice in Robes, where Dworkin refuses to apply realism to interpretive concepts for the very reason I identified in my article: Do these concepts describe, if not natural kinds, at least political kinds that like natural kinds can be thought to have a basic ingrained physical structure or essence? Or at least some structure that is open to discovery by some wholly scientific, descriptive, non-normative process? Can philosophers hope to discover what equality or legality really is by something like a DNA or chemical analysis? No. That is nonsense. What interpretive concepts are about has no natural structure that could be scientifically investigated: ‘[P]hilosophical analysis of political concepts cannot be shown to be descriptive on the model of scientific investigation into natural kinds. Liberty has no DNA’. And ‘neither does law’. It is worth noting that in Justice in Robes Dworkin offers a second reason to refuse to apply realist metasemantics to the concept of law. Realism, he argues, is still conventionalist in the sense that it depends upon agreement about paradigm samples. And certain concepts can lack even that level of agreement: People can share [an interpretive] concept even when they disagree dramatically about its instances. So a useful theory of an interpretive concept . . . cannot . . . simply excavate the deep structure of what people mainly agree are instances. 53 Ibid at 1909–1910. 54 Dworkin, above n 5 at 152. 55 Ibid at 153. 56 Ibid at 166. 57 Ibid at 11–12. Unfortunately, Dworkin cannot seem to make up his mind on this issue, for elsewhere in Justice in Robes he suggests that ‘[t]he purpose of my ‘‘semantic sting’’ argument was to make [the following] point: sharing a concept does not necessarily mean sharing criteria for its application, but might instead mean sharing paradigms as the basis for interpretive claims’. Dworkin, above n 5 at 218. SPRING 2008 Does Dworkin Commit Dworkin’s Fallacy? 47 at C olege of W iliam nd M ry on N ovem er 9, 2010 ojls.oxjournals.org D ow nladed fom I then moved on to describe a third metasemantic theory, called ‘interpretive’, which I attributed to Dworkin in the rest of the article. Under this theory, the linguistic practice of using a concept has the capacity to reform itself over time. As it reforms itself, it reveals content that was always there binding participants. But, unlike in realism, this reformation does not occur by reference to something, like the structure of a paradigm sample, that is external to the practice. It occurs immanently, that is, by drawing upon the reserves of the linguistic practice itself. As I put it in my article: A word will have constancy and determinacy of meaning, despite changes and disagreement concerning the criteria for its use, because the practice of using the word evolves through critical reflection, which reforms it in light of its prereflective commitments. The meaning of the word remains the same despite changes in the practice of its use because critical reflection manifests limitations that were, in a sense, always binding the participants. Even though Dworkin and Hart disagree about the content of the concept of law, there is—and always was—an answer to the question of the content of that concept. This answer is not determined by the currently shared beliefs of language-users (for there are no such shared beliefs) nor by reference to the DNA of law (for law has no DNA), but by reference to the critical unfolding of the practice of using the concept of law itself. I admit that this metasemantic theory is vague. It also may be completely indefensible. But at the very least it is a metasemantic theory, for it—unlike Dworkin’s view that the concept of law is interpretive—does provide an answer to the question of why the concept of law has the content that it does. It is 58 At one point in Justice in Robes Dworkin concludes, bizarrely, that the metasemantic theory I attribute to him is realist: [Green] discusses at some length a philosophical dispute between what he calls a ‘traditionalist’ and a ‘realist’ view of the concepts of gold, water and tiger. The former view treats these concepts as criterial, and the latter insists that they are not criterial but function rather as natural kind concepts. Green apparently assumes that this is a dispute about how all concepts should be understood. So he says that I make ‘arguments about meaning similar to those of the realists in order to attack the jurisprudential position that law is exhausted by agreement or convention’. That is not correct, and the misunderstanding is important. I did reject a criterial account of the doctrinal concept of law. But I did not adopt a natural kind semantics for that concept; in Chapter 6 of this book I explicitly rejected a natural kind interpretation of any of the concepts of law. Instead I said that the doctrinal concept, like other important political concepts, is interpretive. Ibid at 227. Dworkin comes to this conclusion about my argument on the basis of my single observation that Dworkin ‘employs arguments about meaning similar to those of the realists’. Green, above n 1 at 1908. But Dworkin does employ arguments similar to the realists, since he insists that questions about the content of the concept of law can have right answers even when there is disagreement about what that content is. Indeed, in Justice in Robes, Dworkin himself speaks of ‘instructive similarities’ between natural-kind concepts and interpretive concepts like the concept of law, because the content of these concepts is not beholden to people’s current attitudes. Dworkin, above n 5 at 154. See also Dennis M. Patterson, ‘Dworkin on the Semantics of Legal and Political Concepts’ (2006) 26 OJLS 545–557. To say that Dworkin’s metasemantics is similar to realism in this respect is not to say that he was a realist. Indeed, after introducing the possibility that Dworkin’s metasemantic theory is realist, I immediately and clearly reject the idea, Green, above n 1 at 1909–1910, and attribute an interpretive theory of meaning to him repeatedly in the rest of the article. Ibid at 1914, 1915, 1920, 1921, 1924. 59 Ibid at 1913–1914. 48 Oxford Journal of Legal Studies VOL. 28 at C olege of W iliam nd M ry on N ovem er 9, 2010 ojls.oxjournals.org D ow nladed fom also an alternative to metasemantic conventionalism, for it does not associate the content of the concept with language-users’ current criteria for the concept’s use. 5. Is Interpretive Metasemantics Nominalist? But in his response to my article, Dworkin rejects this interpretive metasemantics, arguing that it is a form of ‘extreme nominalism’ similar to metasemantic conventionalism. One reason is an analogy I drew in my article between interpretive metasemantics and a method for justifying rules of logic advocated by the philosopher Nelson Goodman in his seminal paper ‘The New Riddle of Induction’. Here is what Dworkin says about my analogy: Green . . . suggests that perhaps my account of interpretive concepts is . . . ‘similar’ to Nelson Goodman’s starkly nominalist theory of logic. Goodman held that our shared sense of the validity of certain rules of deductive inference is the upshot not of our perceiving Platonic forms of valid inference but of our together achieving an equilibrium between the inferences we are disposed to accept and the rules of inference we are disposed to embrace. Goodman’s account supposes that the equilibrium we have achieved after a (mythic) process of critical adjustment is both wholly contingent – ‘we’ might have settled on a very different equilibrium and then had a very different logic – and wholly a social construction: rules of logic could not have the force they do for ‘us’ unless we all unquestionably now accepted that if A, and if A then B, then B. In that way Goodman’s nominalism about deductive and inductive logic is like what Green calls traditionalism. The meaning of terms is settled by a uniform practice achieved after some period of semantic negotiation. Nothing could be more remote from my account of interpretive concepts than this extreme nominalism. But the words ‘nominalism’ or ‘nominalist’ never occur in my article and, far from suggesting that the content of a concept is contingent under interpretive metasemantics, I insisted that under this approach it is ‘determinate and stable’. There are right answers to questions of content, I said, because ‘critical reflection manifests limitations that were, in a sense, always binding the participants’. My use of Goodman does not undermine this interpretation, for Goodman too never mentions the words ‘nominalism’ or ‘nominalist’ in his article, nor does he suggests that logic is a contingent social construction. Indeed, he insists that ‘[t]he validity of a deduction depends not upon conformity to any arbitrary 60 Dworkin, above n 5 at 289. 61 Nelson Goodman, ‘The New Riddle of Induction’ in Fact, Fiction, and Forecast (Indianapolis, Indiana, 3rd edn, 1979) 59. 62 Dworkin, above n 5 at 289. 63 Green, above n 1 at 1915; see also ibid at 1916. 64 Ibid at 1914. SPRING 2008 Does Dworkin Commit Dworkin’s Fallacy? 49 at C olege of W iliam nd M ry on N ovem er 9, 2010 ojls.oxjournals.org D ow nladed fom rules we may contrive, but upon conformity to valid rules’. To be valid the rules must be ‘justified’. What he describes in his article is the method by means of which such justification proceeds. So where do Dworkin’s allegations of nominalism come from? Apparently from the fact that Goodman expresses a commitment to nominalism in some of his other writings. Dworkin’s reading is particularly bizarre, because it makes John Rawls a nominalist. As I noted in my article, Rawls cited Goodman’s article as the model for the method of reflective equilibrium that he used in A Theory of Justice. The problem faced by Rawls is similar to Dworkin’s: How can we arrive at a theory of justice (that is, an account of the content of our concept of justice), given that there is such disagreement over what this content is? As I put it in my article, Rawls uses Goodman’s method to ‘navigate[ ] a course between the Scylla of a metaphysical realism about justice and the Charybdis of a traditional approach, under which any change in our criteria for using the word ‘‘just’’ becomes a trivial decision to give the word ‘‘just’’ a different meaning’. If Dworkin is right that Goodman’s method is nominalist, then Rawls sailed right into Charybdis. But far from rejecting Rawls’s method of reflective equilibrium as trivially redefining the word ‘just’, Dworkin speaks favourably of the method and indeed draws analogies between it and his own. To be sure, he also draws distinctions, but only because the scope of his project is different from Rawls’s, not because Rawls’s method treats the concept of justice as criterial. But setting Goodman aside, doesn’t Dworkin still have a point? Since the process of immanently reforming the linguistic practice of using the concept of law is contingent (after all, certain participants in the reformation of the practice, like Dworkin, might never have been born), isn’t interpretive metasemantics really conventionalist, because the content of the concept is whatever results from this contingent process? To be sure, it does not depend upon the criteria people currently associate with the concept. But does it really help to say that it depends upon the criteria people eventually will associate with it, after this process of reflection? To avoid conventionalism, mustn’t the content of the concept depend upon something unrelated to our practice of using it? 65 Goodman, above n 61 at 63. 66 Ibid. 67 Nelson Goodman, The Structure of Appearance (Cambridge, Massachusetts: Harvard University Press, 1951). 68 Green, above n 1 at 1912. 69 John Rawls, A Theory of Justice (Cambridge, Massachusetts: Harvard University Press, rev edn, 1999) 17–18. 70 Green, above n 1 at 1912. 71 Dworkin, above n 2 at 424 n 17; Dworkin, above n 5 at 161. 50 Oxford Journal of Legal Studies VOL. 28 at C olege of W iliam nd M ry on N ovem er 9, 2010 ojls.oxjournals.org D ow nladed fom That is an interesting argument, and I spent some time worrying about it in my article. But it is one that Dworkin himself has rejected as ‘Archimedean’: Philosophers look down, from outside and above, on morality, politics, law, science and art. They distinguish the first-order discourse of the practice they study – the discourse of non-philosophers reflecting and arguing about what is right or wrong, legal or illegal, true or false, beautiful or mundane – from their own second-order platform of ‘meta’ discourse, in which first-order concepts are defined and explored, and first-order claims are classified and assigned to philosophical categories. I have called this view of philosophy ‘Archimedean’, and this is Archimedeanism’s golden age. Indeed, in spelling out interpretive metasemantics, I was trying to do justice to Dworkin’s rejection of Archimedeanism. Dworkin’s claim that interpretive metasemantics is nominalist worries me, for it suggests that, in his attempt to avoid metasemantic conventionalism, he has sailed into the Scylla of Archimedeanism and realism. Something like law’s DNA—not a process of immanent criticism—will answer the question of whether he or Hart got the content of the concept of law right. What is worse, in another comment in Justice in Robes, Dworkin appears to sail into the Charybdis of metasemantic conventionalism. At one point Dworkin gives us a glimpse of what it would take for the concept of law to have the content that he assigns to it. He says that interpretive concepts ‘require that people share a practice: they must converge in actually treating the concept as interpretive’. It appears that Dworkin thinks currently accepted criteria for using the concept of law determine its content. This comment is tossed off quickly without any elaboration. But if Dworkin’s view is that people share an interpretive concept of law only if they agree that the concept of law has such content, he suffers from the semantic sting: After all, Hart does not agree with him that the concept of law is interpretive, so Dworkin and Hart, it seems, are talking past one another. 72 Green, above n 1 at 1914 n 50. 73 Dworkin, above n 5 at 141; see also Ronald Dworkin, ‘Objectivity and Truth: You’d Better Believe It’, 25 Philosophy & Public Affairs 87–139 at 88 (1996) (describing as ‘archimedean’ a theoretical approach that ‘purport[s] to stand outside a whole body of belief, and to judge it as a whole from premises or attitudes that owe nothing to it’). 74 Green, above n 1 at 1910, 1913. 75 See Patterson, above n 58. 76 Dworkin, above n 5 at 11. 77 In Law’s Empire, Dworkin also suggests that language-users’ current attitudes are what determine the content of their concepts. He argues, for example, that the concept of law is interpretive because ‘[j]udges normally recognize a duty to continue rather than discard the practice they have joined. So they develop, in response to their own convictions and instincts, working theories about the best interpretation of their responsibilities under that practice’. Dworkin, above n 2 at 87. SPRING 2008 Does Dworkin Commit Dworkin’s Fallacy? 51 at C olege of W iliam nd M ry on N ovem er 9, 2010 ojls.oxjournals.org D ow nladed fom 6. Dworkin’s Misderivation of His Interpretive Theory of Law from Interpretive Metasemantics Despite Dworkin’s manifest confusion on metasemantic issues—indeed, his inability to be clear on just what a metasemantic theory is—I nevertheless believe the evidence is in favour of his holding something like the interpretive metasemantics I presented in my article. I also believe that he misderives his interpretive theory of law from this metasemantics. As we have seen, when Dworkin offers his account of interpretive concepts, he simply assigns content to concepts, without providing a genuinely metasemantic theory of why they have this content. Interpretive concepts are moral, which leaves open the question of why they have moral content rather than the non-moral content that others might attribute to them. Furthermore, they involve the moral interpretation of certain social practices, which leaves open the question of why they are about social practices rather than something else. Nevertheless, Dworkin sometimes describes interpretive concepts in genuinely metasemantic terms as ‘essentially contested’. Consider the following passage from Justice in Robes: Sharing an interpretive concept does not require any underlying agreement or convergence on either criteria or instances. Liberals and conservatives share the concept of justice but they do not agree either about the criteria for judgments about justice or about which institutions are just and which unjust. They share the concept because they participate in a social practice of judging acts and institutions just and unjust and because each has opinions, articulate or inarticulate, about what the most basic assumptions of that practice, its point and purpose, should be taken to be. They draw from those assumptions more concrete opinions about the right way to continue the practice on particular occasions: the right judgments to make and the right behavior in response to those judgments. This is a metasemantic position, for it concerns how concepts get their content; it does not simply assign content to concepts. To be sure, in the passage Dworkin speaks of a practice being interpreted. But this practice is the ‘social practice of judging acts and institutions just and unjust’. In other words, he is talking about the linguistic practice of using the concept of justice and offering a metasemantic theory of how the content of the concept can be revealed to the linguistic practice’s participants. Indeed what Dworkin seems to be articulating here is the interpretive metasemantics that I attributed to him in my article. 78 Dworkin, above n 5 at 221. 79 Ibid at 224. See also ibid at 221. 52 Oxford Journal of Legal Studies VOL. 28 at C olege of W iliam nd M ry on N ovem er 9, 2010 ojls.oxjournals.org D ow nladed fom And since this is a genuinely metasemantic theory, it is compatible with concepts having any type of content. After all, there is such a linguistic practice for every concept. Consider, for example, the concept of a convention. There exists the ‘social practice of judging acts or institutions’ to be conventional or not conventional. And everyone ‘has opinions, articulate or inarticulate, about what the most basic assumptions of that practice, its point and purpose, should be taken to be’. In other words, we all have views about what should or should not be called a ‘convention’. Furthermore, people ‘draw from those assumptions more concrete opinions about the right way to continue the practice on particular occasions’. We draw upon these assumptions when determining, in a case of disagreement, what should be called a ‘convention’. The concept of a convention can, in short, be an interpretive concept in this metasemantic sense. But it would be absurd to draw from this an interpretive theory of conventions—that is, a theory in which conventions themselves are not determined by agreement among a practice’s participants. The concept of convention is about agreement among these participants—even though the fact that it is about such agreement is not due to agreement among those who use the concept of a convention. For the same reason, the concept of law as identified in Hart’s conventionalist theory of law is compatible with interpretive metasemantics. Because interpretive metasemantics is compatible with the concept of law having the content identified by a conventionalist theory of law, it does not on its own support interpretive theories of law. It is in conjoining interpretive metasemantics with his interpretive theory of law—as if the latter had anything to do with the former—that Dworkin commits Dworkin’s fallacy. And he commits his fallacy for the same reason he did when interpreting Hart, because he conflates linguistic and legal practices. It is one thing to say that the content of the concept of law is revealed through reflection on the linguistic practice of using that concept. It is quite a different thing to say—as Dworkin does—that the law of a jurisdiction is revealed through the moral interpretation of the jurisdiction’s legal practices. After all, it may turn out that reflection on our linguistic practice of using the concept of law shows that the law of a jurisdiction is identified though Hart’s conventionalist theory of law—in which the moral interpretation of the legal practices of jurisdictions is irrelevant. Dworkin treats the plausibility of his interpretive metasemantics as a reason to accept his interpretive theory of law. In so doing, he commits Dworkin’s fallacy. The point at which this fallacy appears in his writings is when he moves from his rejection of conventionalist metasemantics to his advocacy of an interpretive theory of law. One place this occurs is on page 87 of Law’s Empire. Having criticized conventionalist metasemantics for falling prey to the semantic sting, Dworkin says the following: ‘We have now drawn the semantic sting and no longer need the caricature of legal practice offered in semantic SPRING 2008 Does Dworkin Commit Dworkin’s Fallacy? 53 at C olege of W iliam nd M ry on N ovem er 9, 2010 ojls.oxjournals.org D ow nladed fom [conventionalism]. We can see more clearly now and this is what we see. Law is an interpretive concept . . .’ He then proceeds to spell out his interpretive theory of law, in which the law is what is identified through the moral interpretation of legal practices. By offering the reader the false choice of a conventionalist metasemantics or an interpretive theory of law, Dworkin implies that an interpretive theory of law follows from the decision to adopt a non-conventionalist—that is, interpretive— metasemantics. And, as we have seen, that is a mistake. Dworkin makes the same mistake on page 12 of Justice in Robes. Prior to that point, Dworkin has spelt out metasemantic conventionalism and a realist metasemantics for natural kind terms, which, as we have seen, he also believes to be fundamentally conventionalist. Once again, he criticizes the applicability of these metasemantic theories to the concept of law on the ground that people disagree too much concerning its content. If either of these metasemantic theories applied, these people would be talking past one another. Dworkin then offers an allegedly competing metasemantic theory, in which the concept of law is understood, not as criterial, but as interpretive: In my view the doctrinal concept of law functions as an interpretive concept, at least in complex political communities. We share that concept as actors in complex political practices that require us to interpret these practices in order to decide how best to continue them, and we use the doctrinal concept of law to state our conclusions. We elaborate the concept by assigning value and purpose to the practice, and we form views about the truth conditions of the particular claims that people make within the practice in the light of the purposes and values that we assign. At first glance, this could indeed be a competing metasemantic theory, for Dworkin might be understood as talking about how language-users elaborate the content of the concept of law by interpreting the linguistic practice of applying the concept. The process of elaboration would determine the concept’s content (and thus decide whether Hart’s or Dworkin’s theory of law is right). This is very similar to the interpretive metasemantics that I attributed to Dworkin in my article. But this can’t be what Dworkin means, for such a metasemantic theory would not take a stand concerning Hart’s or Dworkin’s account of the content of the concept of law. It would merely describe what makes Hart or Dworkin right, whoever is right. But the theory Dworkin offers is clearly a rejection of Hart’s theory of law. The practices to which Dworkin refers are not the linguistic practices of using the concept of law, but the legal practices (or, as Dworkin puts it, the ‘complex political practices’) of a jurisdiction. And the interpretation he speaks of is a moral interpretation of these practices. Rather than presenting a metasemantic theory, Dworkin has simply offered his views 80 Dworkin, above n 2 at 87. 81 Dworkin, above n 5 at 12. 54 Oxford Journal of Legal Studies VOL. 28 at C olege of W iliam nd M ry on N ovem er 9, 2010 ojls.oxjournals.org D ow nladed fom on the content of the concept of law—that is, his interpretive theory of law. Once again, by offering the reader the false choice of conventionalist metasemantics or an interpretive theory of law, Dworkin implies in Justice in Robes, as he did in Law’s Empire, that a non-conventionalist (that is, interpretive) metasemantics entails, or is equivalent to, an interpretive theory of law. That is Dworkin’s fallacy. The fallacy arises wherever Dworkin speaks of the view that the concept of law is interpretive as if it were a metasemantic theory in competition with metasemantic conventionalism. One place where he makes this mistake is when arguing, in response to my article, that he does not commit Dworkin’s fallacy. Dworkin’s argument that he does not commit Dworkin’s fallacy is itself an example of the fallacy. The fallacy is rightly named. SPRING 2008 Does Dworkin Commit Dworkin’s Fallacy? 55 at C olege of W iliam nd M ry on N ovem er 9, 2010 ojls.oxjournals.org D ow nladed fom