Using Network Methods to Guide Archival Discovery in Transnational Political History

• In Event: Global and Transnational Sociology Section Roundtables
  In Refereed Roundtable (60 minute): 3. Knowledge and Politics II

Abstract

Historical research on transnational political movements relies on archives shaped by uneven preservation and the disproportionate survival of materials associated with prominent actors. This paper examines how network analysis can be used to guide archival discovery under such conditions. I ask two questions: How can reconstructed social networks inform where historians should look next in fragmented archival landscapes? Can network structure help identify under-documented actors and institutional sites likely to reshape the historical record of important events?

My case study is the 1919 Pan-African Congress (Paris, France), an event widely recognized as foundational to twentieth-century anti-colonial politics but whose full attendance list remains incomplete. The Congress brought together individuals whose later work shaped liberation movements across Africa and the diaspora. A more systematic reconstruction of this event would contribute directly to the study of Pan-Africanism as a transnational political formation.

Existing scholarship often relies heavily on well-preserved figures such as W. E. B. Du Bois, risking overrepresentation of archive-rich actors. I address this problem by reconstructing a directed network from biographical records, published proceedings, correspondence, and organizational affiliations of thirty-nine delegates. Edges represent documented mentions or traceable associations. I supplement this one-mode network with a two-mode person–organization projection and estimate a bipartite ERGM to assess patterns of institutional attachment.

The resulting graph is sparse but connected, with high centralization around Du Bois. Removing Du Bois reduces centralization but does not fragment the network. The two-mode analysis shows that organizations function as structural hubs, and the ERGM indicates a significant tendency for organizations to attract delegate ties.

These findings identify brokerage positions and institutional clusters that point to archives not yet incorporated into the dataset, including parliamentary, church, and association records. Network reconstruction thus provides a structured plan for archival expansion and offers a replicable framework for integrating computational modeling into historical research on transnational political formations.

Author

• Siyanda Katlo Mohutsiwa, University of Chicago