Using as data some appellate courts'reviews of projects to maintain moral order in the city, this article shows that an analysis of legal knowledge production that is (a) dynamic and that (b) refuses to treat people and texts as totally different entities, one studied by social scientists and the other studied by lawyers, can tell us much about such familiar but seldom theorized legal manouvres as judicial review and constitutional challenges. Standing back somewhat from the content of the various claims (about urban vices, in our case studies), I focus instead on the dynamics of knowledge - the ways in which knowledge claims circulate and get transformed as they proceed through various legal, and para-legal, stages. Choosing to analyze the dynamics rather than the content of knowledge is inspired by Actor Network Theory, and Bruno Latour's work in particular: but it also reflects the fact that judicial review also privileges process and form and tends to avoid making judgements about the content of impugned laws or ordinance. In addition, Actor Network Theory also insists that all components of a knowledge network, including things and texts, be treated as 'actors': this can help to take sociolegal work beyond the dichotomy of studying law in the books (texts) vs. law in action (humans).