Home » Blogs » Bèrto ëd Sèra's blog

Lilliputians and Yahoos

Submitted by Bèrto ëd Sèra on Wed, 07/30/2008 - 16:28

These days there is a huge number of requests for closure of wikipedias in African Languages. The issue is controversial as obviously one can look at them from a number of conflicting points of view. From one end you can pretty much say these requests are discriminatory in nature, as they tend to ignore the potential amount of public for information published in such languages and they deliberately ignore the objective problems (low alphabetization, lack of connectivity, to name but the first that come to mind) that slow down the production rate for such projects. On the other end, there is a number of practical problems associated with the management of an empty project, as spammers do not know linguistic boundaries and must be dealt with no matter if you have users for a project or not.

I do not know what will come out of the current discussion, but I am honestly amazed by the presence of a sacred cow in the middle of the field. While lots of people are ready to discuss whether a wiki in African Language should be moved to the Incubator or not, and do discuss it with lots of detail about how much production was there under what conditions, nobody makes what would seem to be the next logical step: people are stubbornly avoiding to ask themselves whether Wikipedia is a universal model or not.

In my absolutely personal and humble opinion, it is quite evident that writing an encyclopaedia according to wikipedia standards is a 100% Euro-American behaviour. The underlying encyclopaedic model is based on that produced by the Enlightenment model and the social model of a wiki has been created by Anglo-Saxons who made a software that is perfect for their needs. Now, as it turns out, other cultures appear to be far from enthusiast about using it. Those who DO use it are mostly people who are
bilingual in English, and create a project in their languages for sake of seeing them represented in such a cliquey environment, but in the end do get their relevant info from en.wiki, as most of the planet does. Over and over again you are going to be asked by those people WHY they should take the time to translate things in their languages, when en.wiki is already there, ready to be used.

In reality, all wikis are made of very small numbers. With the incredible penetration that the internet has in the western world, I remember a total number of ~50K voters for the election of the Trustees. This would make a very small number even for a radical political group like Hamas. And yet it's all the wiki-model for sociality could gather by summoning ALL its languages together in environments that are 100% wikipedia-ready. This has nothing to do with writing an encyclopaedia, as the thousands of mediawiki based failed projects that crowd the internet can easily prove. On the opposite, the idea of an on-line encyclopaedia (and all kinds of on-line manuals) are the one and only field in which this model could manage to capture audience enough for a working project to be born.

I'm afraid that what the current situation with the African wikis is telling us is much more relevant than moving a project from one server to another. It's rather about the feasibility of imposing Euro-American media on other cultures. You can pretty much impose the European tonal system on African musicians, and they sure can learn it. Eventually you'll get results ranging from Duke Ellington to Barry White. You may simply adore them, but are they 'African' at all? The answer is no, they are not.

All cultures have their social models, and most of them considers the Euro-American social standards as weird and remote as those of the Yahoos were to their Equine Masters in Swift's works. Now when you develop a software that is broadly based upon a given model of social interaction this definitely becomes an issue. Localization is not just a matter of choosing another alphabet, it's a matter of transposing a whole train-of-thought into another and make it usable. It's NOT about making "strings" understandable, it's about making "processes" agreeable.

So I'm afraid it's time for us to quit 'educating the poor stupid savages' and simply start to humbly ask what they would find more natural to use. Because in the end it's always the market (the people, that is) who dictates the answers. And what the Africans and most Asians are saying is a big wiki-yawn. Instead of being accused to be 'lazy' (because such is the implication of most of the ongoing discussion) they should be heard. Too bad that something tells me that the average Euro-American is far from the maturity it takes for such a step.

Yet maybe this is good news, after all. It's much better to wait for a native developer to make a software that will fit his own native culture than it would be to see more and more web 2.0 marketing rubbish thrown upon them. When the Africans refuse to work with wikis they are not weak: they are strong enough to ignore the N-th colonial gadget being pushed upon them. It remains to be seen if they will have the guts to make the next step: produce a tool that better fits them and use it.

Anyway, this is but a matter of time. As globalization advances it becomes more and more of a commercial advantage for people to develop their software offshore. The number of non Euro-Americans who actually design and make everyone's IT tools can only grow. So get ready for the time in which non Euro-Americans will discuss on THEIR media whether keeping en.wiki on-line makes sense or not. Given the state of the dollar it can happen much more quickly than we would expect. And given the much larger expansion of mobile phones with respect to the Internet... maybe it's already there, and it simply happens to be located on another medium, one in which wikis do not even exist, by all practical means.

Wikis maybe relevant for western and post-communist countries (all of which have deep roots in the Enlightenment mindset), they may even become central to their everyday information flow... and it's a still a LOOOONG way from being even just known to exist on most of the planet. We, the westerners, are LOCAL, as anybody else. So who is lazy, in the end? Those Africans who do not give a damn about becoming Europeans, or those Europeans who do not take the time to respect the differences?

Funnily enough, you won't find any such issue in the meta discussion... maybe because 99% of those who are discussing happen to be... yes, you got it! They are almost all Euro-Americans, Oceania Anglo-Saxons, Japanese and post-Communists :)

Was it done on purpose? No, I don't really think so. So how did it happen? Now... try and use a bottle to gather stones. It's made for water. It will get a few stones, if they are small enough to get through the neck, but anyone in his mind will carry stones in a box, maybe on a lorry, but certainly not in a bottle. When the software you make is built to manage Enlightenment based social models what that software will gather is mostly Euro-Americans. And that's it. People can discuss as much as they please about choosing servers, but wikis are never going to be an answer for native cultures, unless those cultures quit their being "native" to become somehow "tamed".


Submitted by Bina Meusl on Wed, 07/30/2008 - 17:00.

And with this article you just showed me what is wrong with the Neapolitan wikipedia. Being German I did not consider some very basic points ... time to change things. Thank you :-)

Submitted by Bèrto ëd Sèra on Wed, 07/30/2008 - 23:31.

"We who create and edit wikis just can't apologize for things..."
who said anyone should apologize? What for? You apologize when you are guilty of something. I'm simply pointing out that data shows that there is no such thing as a universal model, and wikis are NOT an exception. Or rather, the internet at large isn't. The problem is that when you start pointing your finger towards certain governments... you are not any better than they are. They say what you do is bad, you say what they do is bad. Just another edit war...

Submitted by Don Osborn on Thu, 07/31/2008 - 12:32.

Bèrto makes some interesting points, though I prefer to think that the encyclopedic model is not so far from knowledge building in any culture that it can't find reference points. The cataloging of knowledge might be done in different ways, but it is something that one encounters inevitably as one progresses in education anywhere.
I have the impression though that the volunteer model of Wikipedia may be another cultural dimension to look at in the failure or lagging of many Wikimedia projects. Which is to say that they were set up for many languages in a "build it and they will come" mode (the phrase is from the film, Field of Dreams).
Toussaint's comments are also very interesting. The lack of attention to education in and development of Africa's first languages is arguably a problem with a lot of ramifications that have not been widely discussed. One reflection is that many Africans are literate in a second language, but not so comfortable reading and writing their first language. That's not something that can be solved overnight, of course, nor something that need be treated as an insurmountable hurdle. Wikipedia in diverse languages can be (among other things) a tool for promotion of the very literacy that it depends on - and that's a non-linear logic arguably different than the sequential expectations in Western reasoning (reductionist logic and all).

Submitted by Bèrto ëd Sèra on Thu, 07/31/2008 - 14:46.

Well, this is obviously a very large issue. I'm absolutely NOT pretending to have the one and only definite answer to the problem, but I'm also tremendously happy to bring such a subject into discussion.

What we can positively assert is that the current wikipedia model model does NOT work with most less resourced languages. This is what the few numbers we have are clearly
stating. Which part of the model (or which combination of its parts) is at failure, exactly, is largely a matter of personal impressions. I do wish the WMF cared for producing traffic numbers that would be helpful in investigating the situation, yet this seems to be unlikely to happen, at least in the short run. So we are left with our intuition and logics in order to dig out a solution.

I pretty much understand what it means to be illiterate in one's first language, as this has been my personal situation until I turned 40 y.o. and made a personal point
of being able to read and write in the language I use to think. Most of such situation are doubtless connected to political issues (none of which usually bears any relation to the culture as such, mostly you deal with fears of separatism, the need to create strong national identities where you historically have none, etc). Yet this is totally outside the field of what the WMF or any other Charity can address. We will have to take this part "as is".

My experience with my own native culture (and others strictly related to it) is giving me a few points, though:

  1. Whatever the culture, there is always a set of "notable people" in it. If and when an agreement among these people can be achieved in order to sponsor their culture this invariably has a result (I'm speaking in general, wikipedia being but one of the millions of fields in which such a lobby has impact).
  2. The need for sourcing one's assertions is a killer. This is a very understandable need for a good western encyclopedia, yet it is almost impossible to source things from within a native culture that usually has almost no existing publications that can be used for reference.
  3. It takes years to build even just a minimal team that will at least start to understand how mediawiki works and will have but a hint of the WMF internal power structure. This happens largely because all the existing material is in English and the idea that such material can be localized in minority languages is simply absurd, given the changes the original editions are constantly undergoing.
  4. People are NOT looking for a wikipedia in their languages and they do NOT find it through the usual network channels. They get to know about it from their highly specialized communication channels (which sometimes may well be internet based) but in anyway it takes years of existence to a wiki before it can be accepted as a solid reality, rather than just another experiment.
  5. It takes seconds to spoil the game because of bad personal relations within the start-up group. People tend to manifest their personal idiosyncrasies in ideological terms, rather than just saying "I hate that guy!". When such is the case, conflicting parties appear and they can go to just any length, while never saying what the underlying personal problem really is. This is true of just any wiki, but it happens to be extremely more destructive in such small human environments as wikis in less resourced languages are. Such conflicts are simply buried into the big flow in entities like en.wiki, but they can become the dominant tinge of a small project and simply kill it right away. Which brings us back to the first point: a culture is strong only if and when its chief intellectual representatives can make an harmonic team. The atmosphere of the wiki usually reflects the atmosphere of the culture at large. If there is conflict within the culture, this conflict will invariably be imported into the wiki.

So I do agree that the build it and see what happens attitude can only be bound to a failure. Nothing can enter a less resourced culture from outside, for the simple reason that these cultures are mostly used to hide and prey they won't be discovered. They either find the inner strength of raising their heads and say Hi! I'm here! by themselves, or any empty facility left there will remain just an empty facility.

My doubts about an encyclopedia being a good start for a language (any language) remain, though. I still think that given the low rate in active users wikipedia is gathering in English, it could possibly be a much better idea to start from things like a simple portal for news and discussion or even just translating Facebook and some of its games. If what you are doing is expected to capture only a very small fraction of the user base in a language, you'll better make that user base large, before considering to start publishing pedias. Otherwise it really looks like people are building their houses starting from the roof.

My take is that just any investment in video games (the kind that work on mobile phones) would have a much deeper impact on the survival of a culture than any wiki will ever have. But then again, we are expected to make games that fit in that culture, and not to translate on open source edition of SpiderMan. It's not the cultures that must fit in the IT, it's the IT that must fit within a culture.