SimPa – Kinship Simulations

Research project financed by the ANR (Program SysComm – Complex systems and mathematical models)

Partner institutions: Institut national d’études démographiques (INED), Centre d’analyse et de mathématique sociale (CAMS), Centre d’études des mondes africains (CEMAF)

The aim of SimPa is to arrive at an integrated model of the emergence of matrimonial circuits in kinship networks, and to develop simulation techniques allowing for a realistic modelling of these networks that can be used by the researchers (anthropologists, demographers, historians and sociologists) who study them. These techniques will be implemented as freely accessible and easy to use software. At the same time, by analysing the workings of a specific type of network, this project will contribute to the understanding of the morphogenesis of weakly acyclic networks and to the statistical theory of interdependent events. This two-fold goal – providing the social sciences with an essential analytical instrument and making significant advances in a central methodological field of network mathematics – is reflected in the composition of the project’s partners: mathematicians, statisticians, demographers and anthropologists.

The project’s immediate aim is to resolve, by means of controlled simulation, a series of problems that, until now, have remained a major stumbling block preventing any reliable interpretation of circuit censuses within kinship networks, and thus any attempt to found analysis of kinship systems on empirical marriage practices. These problems concern expected circuit distributions under realistic hypotheses, the neutralisation of corpus biases and the identification of network artefacts.

Treating these problems should result in a better understanding of the topology of weakly acyclic networks, of which kinship networks are but a special type. Indeed, kinship networks, whose composition rules would appear to be stricter, more constraining and therefore also more regular than in many other interaction networks, offer the means of resolving a number of theoretical issues regarding approximations occurring in empirical data.

The project has both a practical interest for kinship studies in that it promises to clear the way for the identification and interpretation of significant motifs within kinship networks, and a wider theoretical interest for graph theory and simulation methodology. In this respect, the project is a pioneering study: at present, there exists no integrated simulation model either operating under the constraints of unknown data or capable of identifying network artefacts, and the statistical treatment of reticular motifs (as interdependent events) is still in a nascent state.

SimPa’s contribution thus relates both to formal sciences (mathematics and statistics) and to the social sciences.