Cross-cutting social circles, from Simmel to Blau

Georg Simmel devoted a chapter of his Soziologie, published in 1908, to the intersection of social circles. In older societies, he argued, belonging to one circle entailed belonging to others, with the family determining one's trade, neighbourhood and religion. Circles were nested rather than intersecting.

With the differentiation of modern societies, each person belongs to circles chosen separately, occupation, association, neighbourhood, friends, which no longer overlap. For Simmel, it is the singular combination of these memberships that makes individuality: no one else stands at exactly the same point of intersection.

The distinction has a direct consequence for the structure of relationships. In nested circles, the people one spends time with know one another and share the same memberships. In intersecting circles, the relationships of a single person may never meet, each knowing only one facet of that person's life.

This description says nothing about the size of networks. A small number of relationships may belong to very varied circles, and a large number to a single circle. What distinguishes the configurations is the arrangement of memberships, not their volume, and neither is described as superior to the other.

In 1984, in Crosscutting Social Circles, Peter Blau and Joseph Schwartz gave a macrosociological version of this idea. Their reasoning starts from homophily: if people bond mainly with others similar to themselves, the dimensions on which they resemble one another become decisive.

When differences of age, occupation, origin or religion do not line up, resembling someone on one dimension forces one to differ on another. The authors infer that relations between groups should be more frequent where these dimensions cross-cut, a hypothesis they tested against intergroup marriages in American cities.

The notion of multiplexity, which comes from British social anthropology of the 1950s, describes the other side of the phenomenon: a single relationship can combine several roles. A neighbour who is also a colleague and a sports partner connects several circles on his own.

Network surveys can measure this property by recording, for each person named, all the contexts shared. Multiplex relationships are generally described as more stable, since losing one context does not remove the others, but this stability also stems from how long they have lasted, and the two effects are difficult to separate.

Name generators collect people, rarely the circles that connect them. Reconstructing the intersections means asking, for each relationship, where and how it is situated, which makes the questionnaire heavier and explains why most surveys confine themselves to a single context of origin.

Simmel's idea is thus more often invoked than measured. It nonetheless offers a useful lens: describing a personal network means describing not only people, but the way the social worlds one passes through overlap or remain separate.