Seite - (000127) - in Knowledge and Networks

120 exist around and beyond labor markets as well. Migrants in comparatively wealthy countries attract people from the same place of origin and sometimes return there after having been overused to sweep floors and dig holes. Students can spend semes- ters pursuing their curriculum in universities of different countries before returning to their alma mater. Geographers and sociologists can ultimately examine broad residential forms of mobility of individuals and entire communities. By doing so, they see, for example, mobility in loops of neighborhoods or as life-cycle-related mobility of young adults moving together into new places, then to bigger places when they have children, then to smaller ones when the children leave home. Sociologists have also looked independently at turnover in networks of personal relations. In an increasingly rich body of literature, they have described and mod- eled relational turnover, using statistical tools designed to deepen the understanding of network dynamics (Snijders, 1996; 2005). Relational turnover is defined in this chapter as the set of changes observed in an actor’s relationships between two moments in time (e.g., the creation or addition of new relationships, the destruction or disappearance of previous relationships, and the maintenance of relationships). Dynamic models of coevolution of behavior and networks are based on analyses of this relational turnover in members’ profiles and in the composition and structure of the collective. When people close their eyes and ignore a situation marked by con- flicts of interest, is it because they have become friends with someone who tends to do the same thing and influences them in that direction? Or is it because from the outset they chose friends from among people who, like themselves, close their eyes when confronted by situations of this kind? Often, both answers are true, but each effect has a relative weight that can be measured only by observing and analyzing behavioral changes and relational turnover over time. Without such analyses of coevolution of behavior and relational turnover, intuitions about concerted igno- rance as a complex phenomenon, i.e. difficult to observe, remain poor and explana- tions of this phenomenon as a social process remain untested.2 The act of changing structural forms and relational infrastructures triggers changes in social processes downstream. All the main social phenomena, such as solidarity, exclusion and dis- crimination, social control, conflict resolution, learning, socialization, regulation, and institutionalization, have a relational dimension, are a function of relational infrastructures, and reshape structure, at least opportunity structures.3 2 In many ways work by Snijders (1996, 2005) strongly reflects the best social science epistemo- logical practice in which researchers measure, formalize, and model the coevolution of behaviour and interdependencies, of interdependencies and conflicts between actors, both individual and collective. This approach compares models against reality and measurements of reality, but also inspires new intuitions about realities too complex and difficult to observe directly. In short, mod- els, measurements, theories, and the object of analysis coevolve. 3 The list of social processes that facilitate collective action between status competitors and that can be modeled by network analysis is indefinite (i.e., there is no finite list of these processes) because no social processes exist without a relational dimension. Relational processes that characterize collective action undertaken by interdependent entrepreneurs have been the object of neostructural formalizations: integration, assimilation, cooptation, balance of power, evaluation of product qual- ity, exploitation, extraction of economic performance, discrimination, and desolidarization. These E. Lazega