Convergence of truth through language links in historical data – a case study on Wikipedia: a semantic analysis of source texts

Abstract

Our research examines whether numerical data on historical battles, such as casualty figures or troop strengths, converge over time across multiple languages on Wikipedia. Our analysis of annual revisions, reveals no trend of convergence, in most cases the numbers rarely change, with discrepancies persisting or even increasing over time. The study also tests these patterns in underrepresented regions and simpler data categories outside of the battle context, like bridge length, while comparing Wikipedia figures and sentiment present in the articles with established reference works. Findings include persistent barriers, like editorial biases and resource disparities, that limit numerical consistency and reflect cultural and editing dynamics in global knowledge. Collectively this thesis finds barriers that limit data alignment across languages and ultimately reveals that Wikipedia’s goal of unifying knowledge globally faces unaddressed challenges.