PARIS — Wikipedia is more than a popular online encyclopedia. It is also a battlefield for intellectuals, as the volunteer editors who create the content correct each other without end until they finally reach a hypothetical consensus.
Yet, according to an international team of researchers from the Universities of Oxford and Budapest, the subjects of these fights vary widely depending on the language. The research, carried out as part of the European project ICTeCollective, quantifies the differences in the nature of controversial subjects, according to both geographical and cultural origins.
People don’t brawl over the same topics on the French Wikipedia as they do on the German, Hebrew or Czech versions. The French squabble mostly about political and ideological subjects, such as Ségolène Royal (the 2007 Socialist Party presidential candidate who lost to Nicolas Sarkozy), socialism or Russian President Vladimir Putin. The Spanish, on the other hand, relish arguments about sport, while Arab speakers quarrel over religion.
Ten different languages were selected for the study, and 57% of the 1,000 “hot” topics identified were related to politics, religion and territory. Music, books and cinema accounted for 11%, while 6% had to do with science and technology. Of the 100 most contentious subjects, the French only share two with the Germans, the English and the Spanish: homeopathy and Jesus. With the Germans only, the French share a vivid passion for psychoanalysis, Osama Bin Laden, Roger Federer and racism. And along with the Spanish, the French get worked up about Augusto Pinochet. As a matter of fact, 81% of the 100 most controversial subjects in French are exclusively “national” and not shared with English, German or Spanish.
The long article (published on arXiv.org), which will be included in a book about Wikipedia in 2014, swarms with figures and charts. The method established by the research team in 2012 consists of attributing a unique footprint to each version of an article so that they can later analyze, for instance, how many times a specific version came back online. That data would show that an editor has been fighting with another by republishing an older version of the article.
“These specialists offer an operational definition of contentious topics, independently from languages,” explains Alexandre Hocquet from the University of Lorraine who has been involved in some of Wikipedia’s editorial wars. “It is thus possible to compare, which makes our vision of Wikipedia less monolithic.”
This method had already enabled researchers to study the dynamic of controversies — in other words, slow consensus, impossible consensus or one with pauses and resumption of hostilities, etc. This time, they looked at Wikipedia articles that had been published before March 2010 in 10 different languages. It represents about five million texts with 27 million contributors.
The editorial wars that are common to different cultures could thus be quantified by their geographical location. “We were surprised to observe that topics such as religion or territories are a source of conflicts,” says Taha Yasseri, one of the authors of the articles. “We thought that more modern subjects such as science would cause more disputes.”
Janos Kertész, another co-author, says the quantitative results can be interesting to study other fields of collaboration and conflicts, which is an important subject in sociology. “We would also like to test the theoretical model of conflict analysis that we created on other cases than that of Wikipedia,” he says. And from a practical point of view, identifying problems and following them in real time can also be useful to Wikipedia administrators to improve the system’s efficiency.
All these studies are already inspiring the authors, giving them new research ideas. They think that bringing in “foreign” editors could help moderate discussions and end debates. Similarly, it would be useful to resort to discussion lists instead of quarreling, deleting new versions and publishing older ones.
Wikipedia is already a model for large-scale collaborative work, but will it also become a place of experimentation on how to resolve conflicts?