1 With the departure of Raymond Boudon, the international sociological community has lost one of its most important figureheads. Unlike many other French sociologists, Boudon was an internationally oriented scholar in the sense that he was actively engaged in scholarly discussions originating outside of France. Boudon also played an important role internationally by supporting institutions that promoted a scientifically oriented sociology such as the European Academy of Sociology of which he was the first President.

2 Boudon’s scientific legacy spans many areas including sociology of education, social mobility, theories of action, and the methodology of the social sciences, and his work has been an important source of inspiration to analytically oriented sociologists all over the world. I believe his most important contributions to general sociology are to be found in three closely interrelated areas: 1) his emphasis on basing sociological explanations on actors rather than factors; 2) his generative and mechanism-oriented view of explanations; and 3) his broad rational-choice inspired theory of action.

3 As far as his emphasis on actor-based explanations, I can still recall the intellectual excitement I felt as a graduate student when I first read Boudon’s clear and insightful critique of Robert Hauser’s empiricist view of sociology. The context of this exchange was a review by Hauser of Boudon’s (1974) book on education and inequality. In this unusual and pioneering book Boudon used simulations to make intelligible a number of apparent paradoxes reported in the social mobility literature. Hauser, then the leading advocate of a purely statistical approach to sociological research expressed a strong disbelief in the core idea of Boudon’s approach, i.e., that an important distinction should be made between statistical and theoretical models, and that theoretical models are needed in order to explain the results of empirical analyses. Boudon argued that statistical models of the sort advocated by Hauser are useful for many purposes, but that their usefulness for explanatory purposes was considerably more restricted than assumed by Hauser. Boudon argued that explanations normally are arrived at not by the means of descriptive statistical models, but through theoretical models that show the abstract logic of the process being analyzed. In order to explain, Boudon argued, « we must go beyond the statistical relationships to explore the generative mechanism responsible for them » (1976, p. 117).

4 He developed this generative view of sociological explanations in a series of publications, but the core theme remained the same: statistical analyses do not explain the data being observed, they merely summarize it. In his contribution to Paul Lazarsfeld’s festschrift he argued: « Understanding a statistical structure means in many cases building a generating theory or model... that includes the observed empirical structure as one of its consequences. » (1979, p. 51-52).

5 With this generative view of explanations Boudon was far ahead of his time. It took some 20 years for these ideas to penetrate the discipline at large and to receive the recognition they deserved. This generative view of sociological explanations is now at the very core of analytical sociology (e.g., Hedström, 2005) as well as much of the research carried out in the agent-based modeling community (e.g., Epstein, 2007). To explain something is to build a model that shows how actors in interaction with one another can generate the outcome that is to be explained.

6 Given this focus on actors and the outcomes they generate, it is understandable that Boudon spent a considerable part of his latter career on ways in which to conceptualize and make action intelligible (e.g., Boudon, 1994, 2001, 2003). The type of action theory that gradually emerged out of this work was part of the general rational-choice tradition, but it deviated considerably from the standard rational-choice theory as it is used in economics. Boudon was particularly interested in the axiological aspects of action and he paid considerable attention to belief-formation processes. He emphasized the importance of recognizing the cognitive limitations of individuals, and that individuals often act rationally in the sense of having good reasons for doing what they do, even if these actions may not necessarily be those prescribed by expected utility theory.

7 The early Boudon showed considerable interest in formal theorizing and argued for the explanatory importance of generative models that concretely showed how groups of individuals in interaction with one another, often as an unintended consequence bring about the collective outcome to be explained. The latter Boudon was more discursively oriented and did not attempt to give his theories of action the formal structure they would have needed to have in order to be part of the type of generative models that the early Boudon advocated. An important task that lies ahead of us is to bring together these two sides of Boudon’s oeuvre.

8 Boudon will be remembered for these and numerous other important contributions to the sociological canon, but for those of us who had the good fortune to know him personally, it is the person rather that the work that we will think of. Despite his considerable academic success, he remained a warm and generous person who went out of his way to support others, and for that we will remember him dearly.