Constructivist approaches to the emergence and stability of collective identities are now widely accepted. But few of the assumptions of constructivist theory regarding repertoires of identities and their mutability in response to changing circumstances have been examined or even articulated. The article shows how different conditions of a fluid and changing environment affect the stabilization or institutionalization of an identity as dominant within a polity. We used the Agent-Based Identity-Repertoire (ABIR) model as a simulation tool and confined our attention to relatively simple identity situations. Strong evidence was found for the emergence of identity institutionalization, the existence of a "crystallization" threshold, the effectiveness of divide-and-rule strategies for the maintenance of an identity as dominant, the efficacy of a network of organic intellectuals, and hegemonic levels of institutionalization. Thresholds leading to hegemony were not observed. Preliminary results from experiments examining more complex identity situations have been corroborative.
Bibliographical noteFunding Information:
The paper from which this article was drawn was co-authored with Stacey Philbrick and presented at the annual meeting of the American Political Science Association, August 2000, Washington, DC. Software development supporting this research was done by Dr. Vladimir Dergachev. The authors gratefully acknowledge support for this research from the Carnegie Corporation, the Solomon Asch Center for Study of Ethnopolitical Conflict, and the University of Pennsylvania Research Foundation. The authors would like to thank the editors and anonymous reviewers of this journal for their comments and suggestions. The authors also wish to thank Roy Eidelson, Lilach Nir, and Joshua Pasek for their useful insights, comments, and other contributions to the research reflected in this paper. Comments should be sent to Ian S. Lustick via e-mail at ilustick@ sas.upenn.edu For a convenient exposition of various positions on the deployment of constructivist theories for analyzing identity conflicts, see the symposium edited by Kanchan Chandra (2001). Suny used this phrase in a lecture at the Solomon Asch Center for Study of Ethnopolitical Conflict, University of Pennsylvania, June 1999. The inverted commas around "decides" signal that these agents are not actually making rational choices based on comprehensive and forward-looking assessments of probabilities, options, and preferences, nor on any memory of the result of actions taken in the past. Agents respond "adaptively" in a highly bounded, simple, and automatic way. The action of each agent at each time-step is, indeed, fully and exactly determined by the few signals it processes and the components of its identity repertoire at that time-step. A precise explanation of a "time-step in the life" of an agent operating according to ABIR algorithms is provided in Appendix 1. The algorithms themselves, which govern the behavior of agents and are used to operationalize constructivist theory for the ABIR experiments reported later, are included in Appendix 2. The executable program for ABIR, capable of replicating all the experiments reported here, is downloadable at http://www.polisci.upenn.edu/abir, along with a manual explaining its use. ABIR and ABIR-inspired approaches have also been used to investigate the conditions enhancing stability and diversity in deliberative democracies; the impact of globalization on identity conflict, immigration and national identity; processes of Europeanization; interactions between economic interests and identity norms; and learning as an emergent process. For links to these papers and publications, visit: http://www.polisci.upenn.edu/abir The application of the theory in these volumes is mainly to the institutionalization of changeable state boundaries, but the theory is presented as a general model of institutionalization, without regard to the content of the beliefs or norms involved. More complex landscapes, with larger repertoires of identities per agent, and present in substantial numbers within the array as whole, were also studied (Lustick, Miodownik and Philbrick 2000). The results corroborated those obtained in work with simpler settings. For reasons of space, only the results of the simpler experiments are reported here.