Sociotechnical Transformations of Technology Discourse in the Washington Post, 1970–2024

• In Event: Communication, Information Technologies, and Media Sociology Section Roundtables
  In Refereed Roundtable (60 minute): 15. Media Sociology

Abstract

This paper develops a longitudinal, theory-driven natural language processing (NLP) analysis of Washington Post coverage of technology (1970–2024) to examine how technology is politicized and depoliticized over time, with particular attention to the 2013 acquisition by Jeff Bezos. Drawing on science and technology studies (STS) and media sociology, the study treats news discourse not as neutral reporting on a pre-given “technology” or “politics,” but as a site where sociotechnical orders are assembled, contested, and stabilized. Building on concepts of sociotechnical systems, imaginaries, and actor-networks, the Post is approached as a node within wider configurations of ownership, platforms, algorithms, and professional norms that co-produce what counts as technological and political issues.

The analysis uses a full-text WaPo corpus with metadata (1970–2024) and constructs a technology-focused subcorpus via curated keyword filters and distributional expansion. Transformer-based topic modeling and time-sliced word embeddings identify dominant sociotechnical topics and track semantic drift in core concepts (e.g., “innovation,” “risk,” “regulation,” “justice”). Supervised classifiers, trained on hand-coded samples, distinguish governance, policy, and partisan framings and map shifts among technocratic, market, and justice-oriented rationalities. Treating the 2013 Bezos acquisition as an interruption in the actor-network, an event-study design tests for changes in how technology is articulated with electoral politics, regulation, and state power, and in how major corporate actors are positioned as political or apolitical, extending prior work on ownership, bias, and Amazon-related coverage at the .Post.

By integrating STS concepts of co-production and sociotechnical imaginaries with large-scale NLP, this project offers a theoretically informed account of how an elite newspaper helps construct what technology is, whom it serves, and how political it appears—before and after a pivotal transformation in its own sociotechnical configuration,

Author

• Annette Gailliot, University of California-Berkeley