Writings on Models and Mechanisms
After I defended my PhD disseration in March 2006, I started rethink about the kind of research perspective that I adopted in my dissertation to study educational inequalities in France and Italy. Progressively I developped the sense that all I have done was a consequence of my strong desire to practice sociology by building and formalizing conceptual models.
"The Progress and Urgency of the Concept of Model in Sociology", published in French in L'Année Sociologique (57, 1, 2007, pp. 13-61), was an attempt to draw all the implications from my conviction that sociology is still highly resistent to the idea that model building is a crucial activity in science. Thus, first, I focused on a specific understanding of the concept of “model”, that is the so-called "generative models", and explained how they are related to the concepts of "mechanisms" and of a specific form of methodological individualism. Then, I attempted to explain the limitations and potentialities of “statistical models”, “computational models”, and “mathematical models” for the analysis of "generative models". Finally, I outlined a possible way of integrating these four distinct interpretations of the concept of “model” in order to create a type of sociology based on a systematic activity of model-building and testing.
Download the original article in French (2007).
In 2014 the article has been included in the selection of articles that L'Année Sociologique decided to translate in English to increase the journal's international visibility. For institutions subscribing Cairn, the 2014 English adaptation can be accessed here.
Minor writing. In 2009 I wrote a short review paper in which I investigate the use of formal models in a specific research field, namely the analysis of collective action. I attempted to show that model-based studies in this field has moved from the use of mathematic models to that of simulation-based models because of the higher flexibility of the latter, especially for the analysis of actors' heterogeneity and network embeddedness.
Read the paper (in French) -- Manzo, G. (2009) “Analyse formelle”, in O. Filieule, L. Mathieu, C. Péchu (eds.), Dictionnaire des mouvements sociaux, Paris, Presses de Science Po, 54-61.
Early 2003, during my second graduate year, I was still looking for a method suitable for going beyond the traditional quantitative approach to the analysis of social stratification and mobility. When I came across the literature on multi-agent systems in distributed artificial intelligence, I immediately had the sense that this technique had great potential for macrosociology.
"Variables, mechanisms, and simulations: can the three methods be synthesized? A critical analysis of the literature", published in French in the Revue Française de Sociologie (46, 1, 37-74) in 2005, was my first organic attempt to articulate different strands of the sociological literature with the understanding I had of agent-based simulations at that time. The article can be read as the discussion of the research program that I intended to apply in my PhD dissertation to the analysis of educational stratification in France and Italy.
In particular, the paper investigates four types of scientific contributions: 1) those that discuss the limitations of regression-based sociology; 2) those related to the notions of social action and rationality; 3) studies of the concept of generative mechanisms; 4) publications on applying simulation methods in sociology. The article is all about the attempt to show that the intersection of ideas from these four strands of literature generates a methodological project to reformulate empirical quantitative sociology. This project, I argued, can be summarized as follows: 1/ variable-centered analysis describes; 2/ mechanism modeling (where mechanisms are built in non-reductionist methodological individualism terms) explains; 3/ simulation methods activate and run (as well as test) the mechanisms assumed to generate the statistical assocations empirically observed.