H.L.A. Hart's The Concept of Law is the most important and influential book in the legal positivist tradition. Though its importance is undisputed, there is a good deal less consensus regarding its core commitments, both methodological and substantive. With the exception of an occasional essay, Hart neither further developed nor revised his position beyond the argument of the book. The burden of shaping the prevailing understanding of his views, therefore, has fallen to others: notably, Joseph Raz among positivists, and Ronald Dworkin among positivism's critics. Dworkin, in particular, has framed, then reframed, the conventional understanding, not only of Hart's positivism, but of the terms of the debate between positivists and him. While standing on the sidelines, Hart witnessed the unfolding of not only a lively debate between positivists and Dworkin, but an equally intense one among positivists as to positivism's (and his) core claims. The most important debate has been between so-called inclusive and exclusive positivists: a debate as much about Hart's legacy as about the proper interpretation of legal positivism.

To understand H.L.A. Hart's general theory of law, it is helpful to distinguish between substantive and methodological legal positivism. Substantive legal positivism is the view that there is no necessary connection between morality and the content of law. Methodological legal positivism is the view that legal theory can and should offer a normatively neutral description of a particular social phenomenon, namely law. Methodological positivism holds, we might say, not that there is no necessary connection between morality and law, but rather that there is no connection, necessary or otherwise, between morality and legal theory. The respective claims of substantive and methodological positivism are, at least on the surface, logically independent. Hobbes and Bentham employed normative methodologies to defend versions of substantive positivism, and in modern times Michael Moore has developed what can be regarded as a variant of methodological positivism to defend a theory of natural law.

It is hard to think of a more banal statement one could make about the law than to say that it necessarily claims legal authority to govern conduct. What, after all, is a legal institution if not an entity that purports to have the legal power to create rules, confer rights, and impose obligations? Whether legal institutions necessarily claim the moral authority to exercise their legal powers is another question entirely. Some legal theorists have thought that they do—others have not been so sure. But no one has ever denied (how could they?) that the law holds itself out as having the legal authority to tell us what we may or may not do.

There are two questions I would like to address in this article. The first and main question is whether there are rules of recognition, along the lines suggested by H.L.A. Hart. The second question concerns the age-old issue of the autonomy of law. One of the main purposes of this article is to show how these two issues are closely related. The concept of a social convention is the thread holding these two points tightly knit in one coil. Basically, I will argue that a novel account of social conventions can be employed to reestablish Hart's thesis about the rules of recognition, and that this same account shows why, and to what extent, law is partly an autonomous practice.

The American Legal Realists, as I read them, are tacit legal positivists: they presuppose views about the criteria of legality that have affinities with positivist accounts of law in the sense that they employ primarily pedigree tests of legal validity. Ever since Ronald Dworkin's well-known critique of H.L.A. Hart's positivism a generation ago, however, it has been hotly contested whether there is anything about positivism as a legal theory that requires that tests of legal validity be pedigree tests. So-called Soft or Inclusive versions of positivism are willing to relax the restrictions on the content of a Rule of Recognition to admit non-pedigree criteria of legal validity; Hard or Exclusive versons of positivism deny that such a move is compatible with the central commitments of positivism. Hard Positivism, of which Joseph Raz has been the leading proponent, thus competes with various Soft Positivisms, defended by, among others, Coleman, Lyons, Soper, Waluchow, and now, explicitly, Hart himself in the “Postscript.” If the Realists are positivists, as I claim, then it cannot be the case that Soft Positivism is a genuinely positivistic doctrine. But there is more at stake here than just labels. Realist arguments for the indeterminacy of law—arguments central to the whole Realist enterprise—depend crucially on their tacit Hard Positivism. If, in fact, positivism has a more relaxed view of the criteria of legality than Hard Positivism supposes, then Realist arguments depend on unsound tacit premises about legal validity. What is at stake, then, is not whether Realists should be called (tacit) “Positivists” or merely (tacit) “Hard Positivists,” but whether their underlying view of the criteria of legality is sound. It can only be so if the best arguments favor Hard Positivism.