1. R. Dworkin, Law’s Empire (Cambridge, MA: Belknap Press, 1986) chs. 2, 3 [hereinafter Law’s Empire.]

2. S. Fish, Is there a text in this class? (Cambridge, MA: Harvard University Press, 1980); ‘Working on the Chain Gang: Interpretation in the Law and in Literary Criticism,’ in W.J. Mitchell, ed., The Politics of Interpretation (Chicago, IL: University of Chicago Press, 1983) 271.

3. As a matter of fact, this statement is not quite accurate. Certain forms of interpretation are not evaluative, at least not in any substantial manner. But for the sake of the argument I will not insist on this point. After all, standard and familiar forms of legal interpretation do involve evaluative considerations.

4. See Fish, supra note 2.

5. Intention-based semantics, drawing on Grice’s analysis of communication intentions, has never been thought to form an adequate account of ‘meaning’. See, for example, S. Schiffer, Remnants of Meaning (Cambridge, MA: MIT Press, 1989).

6. See, for example, J. Searle, ‘Literal Meaning’ (1978) 13 Erkenntnis 207.

7. See H. Putnam, ‘Sense, Nonsense, and the Senses: An Inquiry into the Powers of the Human Mind’ (1994) 91 J. of Phil. 445 at 458.

8. This last insight is not particularly Wittgensteinian. See, for example, D. Davidson, Inquiries into Truth and Interpretation (Oxford: Clarendon Press, 1984) at 153.

9. It may be objected that my use of the term ‘interpretation’ is too restrictive. Some philosophers simply use it to describe the ‘mental process’ that takes place in our minds when we grasp the meaning of an expression (or, perhaps, when we grasp anything). According to this stipulative definition, it is of course true by definition that every understanding of the meaning of an expression involves interpretation. There is nothing wrong with stipulative definitions. On this broad definition of ‘interpretation’, however, Dworkin’s second premise of the argument loses its grip; it is no longer true that every interpretation, understood as such a mental process, necessarily involves evaluative considerations.

10. L. Wittgenstein, Philosophical Investigations. trans. G.E.M. Anscombe, 2d ed. (Oxford: Blackwell, 1958).

11. Ibid. at # 87

12. Ibid. at # 198. See also # 201: ‘What this shews is that there is a way of grasping a rule which is not an Interpretation. but which is exhibited in what we call ‘obeying the rule’ and ‘going against it’ in actual cases.’ I have offered an explanation of this Wittgensteinian argument in my Interpretation and Legal Theory (Oxford: Clarendon Press, 1992) ch. 7.

13. The problem of trans-cultural translation does pose a difficulty for legal theory for it calls into question the possibility of a general, universal, account of the nature of law. I agree with Raz, however, that the problem is generally exaggerated. See Raz, Reason, Value, and Morality (Oxford University Press, forthcoming).

14. See, for example, Raz’s excellent account of ‘directed powers’ in the law, in Ethics in the Public Domain (New York: Oxford University Press, 1994) ch. 10.

15. I do tend to agree with Coleman and Leiter, however, that lawyers’ general ability to predict legal outcomes is much better than usually assumed by theorists. We can often predict quite accurately what the courts will do in fact, due to our ability to rely on general information about politics, psychology, and similar considerations. Folk-psychology has its limits, of course, but it often works well enough. See J.L. Coleman & B. Leiter, ‘Determinacy, Objectivity, and Authority’ in A. Marmor, ed., Law and interpretation (Oxford: Clarendon Press, 1995) 203.

16. R. Dworkin, Taking Rights Seriously (London: Duckworth, 1977) chs. 2–4 [hereinafter, Taking Rights Seriously].

17. Dworkin has also been held to maintain that legal principles, as opposed to rules, do not have a canonical formulation. This is probably wrong even as a rough generalization. There is nothing in the logic of principles to entail that they cannot have a canonical formulation. Principles may well acquire a canonical formulation, either by judicial custom or by the enactment of a legislative body. Most standard examples of legal principles do have, in fact, a canonical formulation. Many principles are embodied in constitutional documents, and many others have gained canonical formulation in important precedents (e.g., the famous Learned Hand formulation of the principle of negligence liability). This mistake about canonical formulations should not be confused, however, with the separate question concerning the source of validity of principles, which I discuss in the text infra.

18. Well, this is actually not quite accurate. It is possible that even morality is partly, though in a very limited way, dependent on social conventions.

19. Wil Waluchow tells me that it is not so accurate to speak of Dworkinian principles as valid at all, since Dworkin sees validity and weight as logically incompatible properties. This, of course, hinges on Dworkin’s claim that legal principles cannot satisfy the criteria of validity embodied in the rules of recognition. I agree that this is how Dworkin sees the point, but I doubt that we have any good reasons to follow such a stipulative definition. We will see later in the text that Dworkin has very good reasons to distinguish between moral principles and legal principles, that is, principles which form part of the law. Forming part of the law is what I call legal validity here. I don’t think that Dworkin can object to that.

20. Law’s Empire, supra note 1 at 225.

21. J. Raz, ‘Legal Principles and the Limits of Law’ in M. Cohen, ed., Ronald Dworkin and Contemporary Jurisprudence (London: Duckworth, 1984) 73.

22. There is, of course, a distinction between generality and vagueness of the norm-act’s definition. Not every norm act defined in general terms is necessarily vague, and vice versa. Often the definition of a norm act, however, is both general and vague (e.g., “act with reasonable care”). For a more comprehensive account of these distinctions, see F. Schauer, ‘Prescriptions in Three Dimensions’ (1997) 82 Iowa L. Rev. 911.

23. Taking Rights Seriously, supra note 16 at 72–73.

24. Here is a typical case of conflict of rules. It sometimes happens that one action in tort constitutes or supports two different torts, so that one of the torts comprises a defense denied by the other. In such cases, the question which tort ought to prevail must be decided by a straightforward comparison of the relative reasons supported by each tort rule. How could such conflicts be resolved if we assume that rules have no dimension of weight?

25. L. Alexander & K. Kress, ‘Against Legal Principles’ in A. Marmor, ed., (Oxford: Clarendon Press, 1995) 279. A symposium was held on this article, see (1997) 82 Iowa L. Rev.

26. This conception of legal principles is often referred to as the ‘standard-rule’ distinction.

27. R. Dworkin, “A Reply by Ronald Dworkin” in M. Cohen, ed., supra note 21 at 263.

28. Law’s Empire, supra note 1 at 66–67.

29. In a recent article published in the Iowa Law Review colloquium on legal principles (see supra note 25 at 787–819), Stephen Perry has suggested an alternative account of legal principles which may seem to avoid some of these objections to Dworkin’s original thesis. I cannot go into the details of Perry’s arguments here. Suffice it to say, however, that the role Perry assigns to legal principles does not seem to be at odds with the essentials of the separation thesis. What both his arguments and detailed examples seem to show is only the way in which moral considerations are combined with constraints of fit, deference to precedents, etc., in the formation of acceptable legal solutions to hard cases. No legal positivist is bound to object to such an account.