This book is concerned with the conflict between the goals of justice and economic efficiency in the allocation of risk, especially risk pertaining to safety. The book approaches the subject from the premise that the market is central to liberal political, moral, and legal theory. The first part of the book rejects traditional rational choice liberalism in favor of the view that the market operates as a rational way of fostering stable relationships and institutions within communities of individuals with broadly divergent conceptions of the good. However, markets are needed most where they are most difficult to create and sustain, and one way to understand contract law in liberal legal theory, according to this book, is as an institution designed to reduce uncertainty and thereby make markets possible.

In emphasizing the importance of the separability thesis, legal philosophers have inadequately appreciated other philosophically important ways in which law and morality are or might be connected with one another. In this article, I argue that the separability thesis cannot shoulder the philosophical burdens that it has been asked to bear. I then turn to two issues of greater importance to jurisprudence. These are 'the moral semantics of law' and 'the normativity of theory construction in jurisprudence'. The moral semantics claim is that legal content is best understood as moral directives about what is to be done and who is to decide what is to be done. The problem is that legal positivists typically hold that only social facts contribute to the content of law, and it is hard to see how a positivist can hold both the social-facts claim and the moral-semantics claim. I argue that not only are the two claims consistent with one another, but that legal positivists must hold some version of the moral semantics claim if they are to make sense of the claim that legal reasons purport to be content-independent moral reasons for acting. In Section 3 of the article, I take up the question of whether theory construction in jurisprudence is normative or descriptive. This is hard to do in part because so little attention has been paid to correctly formulating the issue. I suggest a demanding test for descriptivism; namely, that an adequate analysis of law can be provided entirely in terms of its formal features. I then defend this claim against three arguments designed to show because governance by law is necessarily desirable or valuable that, we cannot characterize law without making reference to those values or to other material features of law. This constitutes a limited but powerful defence of descriptive jurisprudence. † This is such an extensively revised version of the Hart Lecture that I presented at Oxford in May, 2006 that I hesitate to call it my 'Hart Lecture'. In that lecture, I explored the question whether the fact of theoretical disagreement in law requires that judges have theories of law as a silent partner in adjudication. I am sure there is a path that took me from those remarks to this article, but I cannot imagine that it would be worth the effort to reconstruct it. In any case, along the way I benefited from numerous conversations and from comments on previous drafts of both this article and its distant relatives. I am particularly grateful to Matthew Smith,

No abstract

This chapter presents and defends a version of legal positivism, relying on two distinctions: one between epistemic and ontological conceptions of the rule of recognition, and the other between negative and positive positivism. The chapter argues that positivism is committed to the rule of recognition as a semantic or ontological rule only, and that its essential positive claim is that the authority of law everywhere is a matter of social convention. These points, taken together, make legal positivism both interesting and defensible. Because negative positivism is essentially a negative thesis, it cannot be undermined by counter-examples, any one of which would show only that, in some community or other, morality is a condition of legality at least for some norms.

No abstract