LATE last month Amanda Filipacchi, an American writer, discovered that the editors of Wikipedia, a crowdsourced online encyclopaedia, were re-categorising female American authors from "American Novelists" to to "American Women Novelists". No corresponding "American Men Novelists" subject area existed at that time. The process seemingly happened sub rosa, through the actions of several editors. After she published an article in the New York Timespointing this out, Ms Filipacchi found that her own Wikipedia entry was edited numerous times for spurious and sometimes vindictive reasons. "Wikipedia is created and edited by its users," sheobserved. But when it comes to recategorising novelists, or vetting changes to individual pages, who actually makes the decisions?
Wikipedia advertises itself as a bias-free encyclopaedia which allows any internet denizen to contribute well-sourced facts or modify existing entries. In reality, however, the site has only about 35,000 English-language and 70,000 total active editors (as every contributor is known). With few exceptions, any visitor may edit the text of an entry so long as he follows the formatting, style and editorial form. Changes typically appear immediately, but modifications or entire entries may be rejected by other editors. That in turn may lead to consensus-driven votes and lengthy discussions. A common point of contention is whether a topic or person doesn't meet Wikipedia's detailed test for "notability". Editors who register an account, and who contribute regularly and in a manner that conforms to the nature of Wikipedia, gain implicit authority. Some editors become "administrators"—about 1,400 are at the moment—able to freeze or delete entries. Administrators have a big technical stick to ensure that when "edit wars" erupt or inappropriate changes are continuously applied, they can prod or truncheon users. Users may be banned or put under strictures, while administrators themselves can have their actions overridden by any of the 41 demiurges known as "stewards", a 12-member Olympian arbitration counsel, or the site's founder and chief deity, Jimmy Wales.
Given that no one is precisely in charge of anything, who has responsibility for the accuracy or intent of any given change, such as the shift of female novelists to a sub-category? The site tracks all changes to an obsessive degree, and also maintains for each page a "talk" section in which changes are discussed ad nauseam. In another article, Ms Filipacchi documented the seven editors who relocated women authors using the record of changes for both categories' entries. In the modifications to her entry and the "talk" section, one can see the disputes and annotation of modifications to her biography. Ultimately, then, Wikipedia's ostensible fairness relies on vigilance, and editors can mark articles and be notified if any change occurs. But a million unnoticed changes can take place without any authority or agreement other than the will of the editor to make the change.
As a result of Ms Filipacchi's exposure of the category issue, a debate is now raging which an administrator will at some point bring to a conclusion. It might seem that Wikipedia's problem is that it has too many editors and is too fluid. But the opposite may in fact be the case: the site is stiffening with age. A recent academic study found that the rate of rejection of changes jumped from 6% in 2006 to 25% to 2010 for new editors who had received kudos from other users. Meanwhile, active editors for the English-language version dropped from 50,000 to 35,000 during that period and has stayed roughly steady since. This may simply be an indication of Wikipedia's maturity. But the more complete Wikipedia becomes on historical, scientific, and other settled factual matters, the fewer people there will be keeping an eye out for odd decisions or inappropriate edits to articles read by hundreds of millions of people. Who will watch the Wikipedians?