The second lecture, ‘Droit et statistique : répresentation mathématique des lois, méthodologie de l’analyse empirique du droit’ considers the question of how far mathematical and statistical approaches can throw light on legal phenomena. It will be argued that certain structural features of legal systems can be understood using mathematical models. For example, the mathematical concept of the ‘fractal’ is quite effective at capturing the elements of hierarchical ordering, recursion and self-reference in legal reasoning. The purpose of this observation is not to suggest that in order to be valid, legal rules must comply with certain principles of mathematical order. Rather, the fractal analogy is useful a descriptive tool which may help us to make sense of the otherwise random-seeming detail of legal decision making. The data coding technique sometimes called ‘leximetrics’ takes advantage of law’s fractal nature to produce a statistical picture of the operation of legal rules. Leximetric methods have made it possible to identify patterns in the operation of labour and corporate law rules and to estimate their economic and social effects. The importance of leximetrics as an empirical approach to the study of legal systems is to be contrasted with the tendency towards ‘governance by numbers’ in administrative action and regulation. The use of metrics as a mode of governance fails to recognise the limits of statistical methods and in so doing risks undermining the rule of law and the legitimacy of public ordering.