It’s often difficult to appreciate the brilliance of Wikipedia. Only eleven years old, this free, collaboratively edited and multilingual encyclopaedia is so ingrained in our everyday experience that, like so many successful cultural products, we now take it for granted. One particular charge that’s grown up with WP concerns its purported inaccuracy. Indeed, anecdotally, I’ve noticed a tendency for people to use Wikipedia synonymously with something being error-strewn. The irony of this is that, even as early as 2005, science articles in WP and Encyclopaedia Britannica were of comparable quality (although this underwent its own war of accuracy and validity).
Really, what it comes down to is people’s perception of knowledge: for a certain number of individuals, I imagine knowledge is viewed as this fixed, static entity that comes in trivia-size portions (e.g. Which artist is famously known for… Sorry, no pop trivia here). The reality is far more sobering. Yes, there are ‘facts’, but this is easy knowledge. Hard knowledge is found in the stuff that requires you to try and build upon what has already gone before; it is built on dynamic principles and is driven by consensus mechanics. Revolutions in the sciences are simply the culmination of gradual changes reaching a tipping point. In short: conflicts in knowledge are normally where the interesting stuff is taking place (irrespective of its veracity).
This brings me to a recent paper by Yasseri et al. (2012) that heads to frontlines of knowledge in Wikipedia and examines the so-called editorial wars (best coverage I’ve read is over at EMAPS). Their abstract:
In this work we study the dynamical features of editorial wars in Wikipedia (WP). Based on our previously established algorithm, we build up samples of controversial and peaceful articles and analyze the temporal characteristics of activity in these samples. On short time scales, we show that there is a clear correspondence between conflict and burstiness of activity patterns, and that memory effects play an important role in controversies. On long time scales, we identify three distinct developmental patterns for the overall behavior of the articles. We are able to distinguish cases eventually leading to consensus from those cases where a compromise is far from achievable. Finally, we analyze discussion networks and conclude that edit wars are mainly fought by a few editors only.
There’s a lot of fascinating aspects to this article. The method alone warrants a lengthy discussion. For now, though, I’m going to focus on the three distinct, long-term developmental patterns they mentioned: consensus, multi-consensus and never-ending war.
Consensus: As the name suggests, consensus-driven dynamics refer to an end point at which an article settles into a low degree of editing following two previous stages of slow growth and rapid acceleration. We can see this in the figure below:
To quickly summarise: M refers to the authors’ controversy measure and n is the number of edits. Following a special type of sigmoid function (Gompertz Function), we see how the number of edits in the Jyllands-Posten Muhammad cartoons controversy article initially starts off with a relatively low level of controversy. However, the controversy soon undergoes rapid growth, until reaching a point of saturation: here, the authors plausibly argue that articles “can become so well polished that it becomes extremely hard to pick a fight about them”.
Multi-consensus: Best described as periods of intense war followed by lulls of peace. After the initial cycle, following a similar pattern to the standard consensus, the next cycle is set off by internal (e.g. the appearance of a new editor) or external (e.g. an event taking place in the outside world) causes.
Never-ending war: Sometimes these wars are ongoing; many different editors undergo conflict at different periods, and “a steady flow of replacement armies keeps the article always far from equilibrium”. We see this clearly for the Obama and Anarchism articles where no consensus is reached (although I’d be inclined to say that Barack Obama might reach an equilibrium once he finishes his presidency):
In summary, articles with high levels of controversy are increasingly more likely to lead to never-ending wars, with a consensus only being reached if the controversy decreases during the edit history (pretty self-explanatory). Articles maintaining ongoing wars are largely fuelled by the arrival of new editors and external events. This is clearly relevant for those of us getting ready to settle down in the era of Open Science: as research moves into the realms of Open Access, peer review at all hours, and independent testing of methodologies, we should see similar dynamics emerge (they already exist to some extent). The prospect of never-ending wars in science is not too hard to imagine at the scales of both the whole discipline and in individual articles (just look at economics). In other cases, we would expect to see consensus and multi-consensus dynamics, with the most comparable example being the #arsenicgate incident.
Lastly, on a complementary note, the authors also have a useful website that includes the top-100 most controversial articles and some videos (see below for a revert map of the Wikipedia article on Israel and the apartheid analogy):
[Admittedly, these edit wars don’t quite match the visual cinematography of Apocalypse Now!]
Yasseri T, Sumi R, Rung A, Kornai A, & Kertész J (2012). Dynamics of conflicts in wikipedia. PloS one, 7 (6) PMID: 22745683