Skip to main content icon/video/no-internet

A basic element of social network theory and analysis is the dyad, or pair of actors who may or may not share a social relation with one another. A triad is a set of three actors, which also consists of three dyads. In his seminal work, Georg Simmel argued that triads are a fundamental unit of sociological analysis. He argued that three actors in a triad may allow qualitatively different social dynamics that cannot be reduced to individuals or dyads. For example, among three parties A, B, and C, party A may have a dyadic relation to C but also may have an indirect relation to C through B. Party B may then alter the strength or the nature of the relation between A and C, such as solidifying an alliance or mediating a conflict. If A and C do not interact directly, party B may broker a transaction between them and may derive power from this intermediary position.

A principal interest in the study of triads is the phenomenon of transitivity. A triad is transitive if when there is a tie from A to B and from B to a third party C, there is also a tie from A to C. If the A–B and B–C ties both exist but A is not tied to C, then the triad is intransitive. (This condition must be checked for any assignment of the labels A, B, and C to the three parties.) Transitivity is typically studied for directed relations, where a tie from A to B (denoted A→B) is directional and does not imply the same relation from B to A. For triads in directed networks, if the A→B and B→C ties both exist but A→C does not, then the triad is intransitive.

Figure 1 Illustrations of Intransitive Triads: Undirected (left) and Directed (right)

Some directed relations have a pervasive tendency toward transitivity: if A dominates B and B dominates C, then A is also likely to dominate C. Some undirected relations also tend to be transitive, as represented by the adage, “a friend of a friend is a friend.” Triads for some other relations—notably romantic ties between lovers—are overwhelmingly intransitive.

Simmel also argued that triads may serve as an analytical foundation for understanding larger social groups. Researchers have thus described networks by examining the distribution of relations at the triadic level, enumerating all qualitatively distinct triadic configurations. This “triad census” has been used to observe an overall tendency toward sociability, hierarchy, or transitivity in social networks and formed an early foundation for statistical network analysis.

Dynamics of Triads

In models of network evolution, scholars often assume a direct propensity to close triads. Dorwin Frank Cart-wright and Harary's structural balance theory offers a motivational account for triad closure in signed networks, where the relation between two actors may be positive or negative. Specifically, disagreement between positively tied friends (or agreement between negatively tied enemies) leads to psychological tension, fostering a drive among the parties to resolve this dissonance. In a friendship triad, when a friend of a friend is an enemy, then either the enemy will become a friend or one of the friends will become an enemy to resolve tension in relations. Recent work has challenged Simmel's emphasis on the triad as an irreducible lens for analyzing social dynamics, showing that triadic patterns may be by-products of social dynamics in dyads. For example, a dyadic propensity toward homophily (choosing friends who are in the same social categories) will tend to foster transitivity, even if actors have no direct propensity to close triads. In an office of engineers and lawyers, if engineers prefer other engineers as friends while lawyers prefer other lawyers, then this generates a tendency toward transitivity. In the extreme, all engineers will be friends with all engineers and all lawyers will be friends with all lawyers, and so all triads will be closed without any direct propensity to do so. Empirical research considering homophily and triad closure—recently employing exponential-family random graph models (ERGMs)—shows how entangled these dynamics are. Such interdependence of social dynamics at the dyadic and triadic levels requires a greater sophistication both in theories and in statistical models of network evolution.

...

  • Loading...
locked icon

Sign in to access this content

Get a 30 day FREE TRIAL

  • Watch videos from a variety of sources bringing classroom topics to life
  • Read modern, diverse business cases
  • Explore hundreds of books and reference titles

Sage Recommends

We found other relevant content for you on other Sage platforms.

Loading