Citizendium Blog

Somebody from CZ told me to go look at a Wikipedia “essay” titled “Don’t just do whatever.”  (Archived version.)  It’s very thought-provoking.  I’m going to ruin author Matt Britt’s street cred forever by publicly agreeing with it.  But actually, I’m going to disagree with one (apparently) point, and argue for the importance of collaboration.

So here goes:

In reading some recent thoughts on Wikipedia authorship and having some discussions with folks on the Wikipedia IRC channel, I’ve come to realize that there are some flaws in the ideology of the editing process that is held in such high esteem. More specifically, the Wikipedia process as it pertains to individual articles does not, in my experience, produce high quality articles. For the purposes of this discussion, I will talk mostly about featured articles. 

Indeed, a lot of Wikipedians are deeply confused in many ways about how and why Wikipedia both works and fails to work.  It’s hard understand in all its glorious complexity.  But Britt has dismantled one of the confusions rather well.  He says that there is an “ideal editing philosophy,” or what I would call the received wisdom or the wikipolitically-correct explanation of how Wikipedia articles improve.  According to the received wisdom, a page goes from stub to brilliant featured article because more and more people add their collective wisdom and correct each other’s mistakes.  But, Britt argues, it isn’t quite that way:

The editing process we are led to believe in generally involves many editors from varied backgrounds progressively adding an article. Each one brings his or her own flavor and perspective to the table. Each one contributes something that the others may have overlooked; an explanatory sentence, a critical fact, an appropriate image, even a spelling correction. There is no argument that this is exactly the process that occurs and produces the bulk of significant-length Wikipedia articles. So what is the cause for skepticism?

Article quality is the key here. In general, the wiki process produces articles of mediocre quality at best. The quality of each article usually is roughly proportional to the number of sincere editors working on it. This is hardly surprising, but ideally we would like for every article (and realistically, every core topic) to be of featured article quality. Herein lies the fundamental problem with the process. Featured articles do not come out of the disorganized edits of a hundred different people. It is either not possible or has simply not happened that a large body of editors can create the high organization, flow, clarity, and topical relevance required for a high quality article. If the body of featured articles are taken as an example, a trend appears. The version of an article promoted to featured status is usually mostly or wholly the work of one or two editors.

I agree.  Articles are not usually brought into being by the endless “churn” of pointless grammar edits, rearrangements, additions of ancillary facts that don’t pull their weight, new sections that are subtly off-topic, etc.  What creates a good article is usually, in my experience both on Wikipedia and on the Citizendium, the driving force of at most a handful of people who actually know what’s going on.  They are like the leaders of the conversation.  They understand and care about the overall plan of the article and its style.  They are also good writers and knowledgeable the subject.  Without the leadership of such people, what might be a good conversation descends to cacophany.  The existence of so many cacophonous articles in the Wikipedia corpus led us to jettison it from CZ last January, and we haven’t looked back.  I personally am relieved not to be laboring under the burden of so much dross.

What are we to conclude from this, however?  Britt does not suggest what you might think: he doesn’t say we just get rid of collaboration and just start assigning articles to people.  But why not?  Doesn’t his argument establish that radical collaboration is a failure?

No, it doesn’t.  (I don’t mean to imply that he thinks it does, either — but it’s clear that he’s skeptical.)  The experience of Citizendium is extremely germane here, so let me refer to that.  The crafting of our approved articles was done mostly by individuals, for example, ”Biology” by Nancy Sculerati, “Life” by Anthony Sebastian, “Telephone newspaper” by Thomas H. White, and “Literature” by Russell Potter.  Furthermore, our editors have a perennial complaint, the best solution to which I personally am struggling to determine, that often, people — sometimes, relatively ignorant authors who are acting as copyeditors, but sometimes other editors — will come in and screw up “their” articles.  “Damn it,” I am sometimes told, “you’ve got to do something about people screwing up my article!”

What, after all, is the payoff to having articles open to general editing?  Why not let people assign themselves exclusive authorship of articles?  Well, I think we know the outcome of that: a community of undermotivated people sitting on half-done drafts, with neither the community nor the content growing significantly.  I’ve seen this happen time and time again as people have tried to — somehow — capture the wiki “magic” by using a wiki to host signed articles.  It’s silly to use a wiki if you aren’t going to collaborate.  It has not yet worked.  Why doesn’t it?  Why is collaboration so important?

There are at least two reasons, both important, but the second one is fundamental.  The first reason is that other people, even if they don’t actually contribute very much to the content, they correct many little flubs.  And they do frequently act as bullshit detectors.  (To borrow Hemingway’s nice phrase.  Other people are a handy backup bullshit detector when your own isn’t built-in or shock-proof.)  Furthermore, what happens sometimes is that the availability of an article to contribution by many different people means that one of them becomes dominant simply because it’s evident that person knows more about the subject than anyone else.  The collaborative process allows them to work out who ought to lead the article’s development.  In general, if you take two articles, T1 and T2, you let a single author develop T1 alone, and you let the same article develop T2 with the help of other people on a sane wiki (like CZ!), then T2’s usually going to be better.  Again, I think the Citizendium experience in developing its approved articles amply demonstrates this.

The second reason concerns motivation.  If people aren’t motivated to contribute, the project won’t happen, and so this reason is absolutely fundamental.  Collaboration, real collaboration, is totally compelling, and explains why people get “hooked” on wikis.  Developing an article in a collaborative wiki space gives you instant social visibility.  This happens in two ways.  As I’ve said, when someone edits what you have written, even if you disagree with the edit, you are rewarded, honored, by the person’s attention and by the fact that the person left most of the article intact.  That’s the visibility of co-authorship.  There’s also the visibility of speaking to an audience of which you’re a part (co-audienceship?).  It’s like speaking to a class you’ve been in for a whole school year–a different and in some ways more rewarding experience than speaking to a group of strangers you’ll never see again.  These kinds of visibility are the psychic rewards that come as part of strong collaboration.

But in a collaborative wiki space, we are motivated not only by social visibility, but by social expectations.  If someone makes an edit to an article I started, there is a small but frequently effective expectation set up that I’ll respond to the edit, either with a comment or another edit of my own.  Not responding at all is a minor foul in an enjoyable game.

These motivations explain “the wiki habit” — we get hooked precisely by being willing parts of a bottom-up collaboration among largely self-selecting people who choose what to work on and when.

I’ll be expanding upon this at a later date…