This article examines a thesis of interest to social epistemology and some articulations of First Amendment legal theory: that a free market in speech is an optimal institution for promoting true belief. Under our interpretation, the market-for-speech thesis claims that more total truth possession will be achieved if speech is regulated only by free market mechanisms; that is, both government regulation and private sector nonmarket regulation are held to have information-fostering properties that are inferior to the free market. After discussing possible counterexamples to the thesis, the article explores the actual implications of economic theory for the emergence of truth in a free market for speech. When confusions are removed about what is maximized by perfectly competitive markets, and when adequate attention is paid to market imperfections, the failure of the market-for-speech thesis becomes clear. The article closes by comparing the properties of a free market in speech with an adversarial system of discourse.

Although much legal scholarship discusses the meaning of the religion clauses of the U.S. Constitution, very few articles analyze the ways in which state regulation affects actors' incentives to engage in religious behavior. Yet the question of how a law influences religious behavior is important for determining whether various laws are desirable, and whether they violate constitutional constraints. This article draws on recent economic models of religious organization to analyze the ways in which laws affect the behavior of religious groups. Religious groups produce collective goods for their members, and the effect of laws can be analyzed by examining how they modify the payoffs members receive for cooperating or free riding. The article examines the use of laws to establish religious groups, to subsidize them with cash or tax benefits, to provide accommodations for them, to provide symbolic support for them, to provide secular substitutes for the collective goods they produce, and to regulate disputes between members. The article also briefly discusses the constitutional implications of the analysis.

Herbert Simon once observed that watching an ant make its way across the uneven surface of a beach, one can easily be impressed—too impressed—with the foresight and complexity of the ant's internal map of the beach. Simon went on to point out that such an attribution of complexity to the ant makes a serious mistake. Most of the complexity is not in the ant but in the beach. The ant is just complex enough to use the features of the beach to find its way.

In his interesting comment on my recent article, “The Appeals Process as a Means of Error Correction,” Edward Schwartz makes two criticisms of my analysis. The criticisms have essentially to do with my assumption that an appeals court judge will base his or her decisions only on what happened at trial, and not on any inference that can be drawn from the fact that an appeal was brought. Before explaining why I do not find Schwartz's criticisms problematic, it will be helpful for me to restate the main features of the model that I examined in the article.