The Evolution of Collective Memory on Wikipedia
Mapping how Wikipedia affordance use evolves differently across English and Arabic
With Prof. Brian Keegan — 2023
I designed and led a large-scale computational study to analyze how Wikipedia’s English and Arabic editions construct different narratives about the Arab Spring over time. Instead of looking only at article content, I examined how editors behave—how they add links, update information, and reference related topics—to understand how narratives are built through collective action.
Most research looks at what ends up on the page. My work investigates how information evolves: how salience, deliberation, contextualization, and consolidation emerge from editor behavior and platform structures, revealing how collective memory diverges across linguistic communities.
This was funded by the Wikimedia Foundation link to grant description
This is submitted to ICWSM ‘26 link to preprint
Research Question
How do behavioral traces—edits, link decisions, and cross-language connections—reflect differences in collective memory across linguistic communities?
Business Impact
This work demonstrates how large-scale behavioral data can reveal structural bias, information gaps, and cross-market inconsistencies. It provides a model for teams working on search relevance, content integrity, localization, and global information quality, especially on platforms serving multilingual users.
What I did
- Developed new computational measures of collective memory: salience, deliberation, contextualization, and consolidation.
- Analyzed thousands of Wikipedia revisions using Python, the Wikipedia API, and network modeling.
- Clustered wikilink evolution to identify patterns in editorial decision-making across time.
- Constructed cross-lingual knowledge networks to measure contextual alignment and information silos.
- Tracked narrative consolidation through network propagation across related articles.
Why it matters
This project shows how scalable computational methods can surface systemic differences in user behavior—informing strategies for global content quality, consistency, and equitable knowledge representation.
Where this work has been presented:
Methods
- Jones, H. Laurie, Keegan, Brian C. (2024, June 20). Temporal Clustering of Wikipedia Outlinks as Collective Memory Processes [Wiki Workshop 2024]. Virtual. https://www.youtube.com/watch?v=TrthZ-zJ8ow&feature=youtu.be
- Jones, H. Laurie, Keegan, Brian C. (2023, June 8-9). “Network perspectives on collective memory processes across the Arab Spring in Arabic and English Wikipedias“ [Wikihistories]. Virtual.https://wikihistories.net/2023/04/13/laurie-jones-network-perspectives-on-collective-memory-processes-across-the-arab-spring-in-arabic-and-english-wikipedias/
Results
- Jones, H. Laurie, Keegan, Brian C. (2024, January 11-11). Divergent Collective Memory Processes Surrounding the Arab Spring [MENA and Asia Political Methodology Conference]. NYU Abu Dhabi, Abu Dhabi, United Arab Emirates.
- Jones, H. Laurie, Keegan, Brian C. (2024, May 24-25). Divergent Collective Memory Processes Surrounding the Arab Spring [Visions in Methodology Conference]. Washington University in St. Louis, St. Louis, MO. United States