Pre‐transitive balance mechanisms for signed networks*

Abstract
The fundamental empirical structural balance hypothesis claims that human signed networks tend towards balance over time. Balance theorists assume that there is a balance theoretic mechanism whose cumulative effect drives the evolution of signed social structures towards balance. In previous work, we used a line index of imbalance to measure the imbalance of a network through time. Consistent with balance theory, we found a steady movement towards balance in the well known “Newcomb data”. The balance mechanisms were, at best, implicit in that earlier analysis and our use of the line index of imbalance meant that we ignored triples. Here, we consider triples with the simple hypothesis that balanced triples exist at all times and, that through time, the balanced triples become more frequent while the unbalanced triples become less frequent. We examine pre‐transitive balance conditions defined in terms of the (i ?j, j ? k) ties and count the frequencies of the completion tie (i ? k) for each of the combinations of tie signs in the pre‐transitive condition. The basic structural balance theoretic hypotheses are supported ‐ but only partially. Worse, from a balance theoretic viewpoint, there are triples for which the fundamental structural balance hypothesis is contradicted. We construct three substantive arguments to account for the exceptions and end with a plea for the collection of much more appropriate data in order to disentangle multiple mechanisms.

This publication has 17 references indexed in Scilit: