The Role of Niche Signals in Self-organization in Society

Abstract

This dissertation is concerned with the emergence of social patterns. The ability of groups of humans to bring order to both the physical and abstract realms may be our species’ most distinguishing characteristic. It is dependent upon our willingness to cooperate and otherwise coordinate, yet willingness alone is not sufficient for achieving coordinated outcomes on a large-scale because the informational demands of bottom-up organizing are high. Understanding the emergence of social order then requires, in part, understanding how information flows are structured in ways that allow groups to meet the informational demands of self-organization. Of particular importance in this regard are the patterns of person-to-person interactions. In contemporary social network research these interactions are often described as the conduits through which information flows, but person-to-person interactions are also the site and source of the coordination problem needing to be solved. To resolve this tension, network interactions must be patterned in ways that allow for the free flow of information, yet social networks most often exhibit high degrees of clustering, a characteristic which can impede the free flow of information and, thus, large-scale coordination. Does this mean bottom-up processes do not drive coordination within large groups? Is resolution by fiat the only way? Many have made the argument we create and tolerate authorities for precisely this reason, but is that the only viable mechanism for the establishment of large-scale coordination? Inspired by stigmergy, a form of communication used by social insects to coordinate hive activities, this dissertation explores the value of signals occurring outside or alongside of the person-to-person interactions studied using social network analysis. Social life features an abundance of small signals—often in the form of verbal or written communication, but also physical objects and even sounds and smells—potentially freighted with meanings or embedded knowledge. Several research traditions have regarded these signals as part of the fabric of social life, but is the information these signals yield patterned in a way that can help overcome the challenges of large-scale coordination? To begin to answer whether these signals can play a role in mass coordination, this dissertation takes three distinct approaches. The first analyses coupled differential equations describing a system in which a common resource environment is structured by the ongoing actor-to-actor interactions. This system is a modification of a canonical model of molecular self-organization, the hypercycle, and succeeds in organizing vastly more complex sets of interactions than the original. This confirms the information embedded in the environment can indeed be a powerful source of information for coordination. The second paper takes this formal insight into the lab to test whether the addition of a small number of extra-network signals can enable the emergence of conventions in a large, networked group of human participants. It can, and the probability of it happening depends on the strength of the extra-network signal and the topological features of the network. The final paper uses a unique dataset and topic modeling in an attempt to track the emergence of consensus around the themes in works of fiction. While there can be movement in the direction of consensus, the path lengths of the underlying network are too long to support large-scale consensus, a finding consistent with results of the experiment. Implications of these three findings are discussed in the conclusion.