[bad scots joke]

By now, you have probably seen the increasingly viral reddit thread about the American teen who has written almost half of the entire Scots Wikipedia despite not speaking Scots.

If not, well, that's a pretty good precis; my story for the Guardian (the writing of which is also why this newsletter is late) gets into a bit more detail. The short version is that this guy started editing the Scots wikipedia in 2013, aged 12, doing the sort of thankless minor edits die-hard Wikipedians devote their time to: adding the right category tags to pages, fixing the capitalisation of titles, and so on.

Then after a couple of months, they wrote their first article from scratch, translating a couple of sentences about Sapporo, Japan. "Sapporo is the fourth-lairgest ceety in Japan bi population, an the lairgest on the northren Japanese iland o Hokkaido." It's impressive work for a 12-year-old! The problem is that they don't actually… speak… Scots. A few months later he went back in and changed "fourth" to "fowert", but other than that, everything the 12-year-old American boy wrote in 2013 is still there on the Scots wikipedia entry for Sapporo today, largely untouched by any other editor.

Scots are split over how much this actually, you know… matters. The Redditor who brought this to the world's attention described it as an act of "cultural vandalism on an unmatched scale"; others disagree, noting that the Scots wikipedia, by virtue of its anaemic status, isn't exactly the world's greatest cultural treasure. 

There are certainly some knock-on effects here. One claimed harm comes from people having read the fake Scots on the Wikipedia and thinking the whole language is a joke; but English speakers are perfectly capable of reading real Scots and laughing at it too: English speakers suck like that. 

A more worrying problem comes from Wikipedia's position as one of the great open-source projects of our time. The Scots wikipedia is almost certainly the largest openly-licensed corpus of (supposedly) modern Scots writing in the world, and many people will have used it over the years for various projects. Want to teach a machine to speak Scots? You probably used the Wiki as a data source. Which means you've probably been hawking a machine learning system that does not, in fact, speak Scots.

But equally concerning, too, is what this says about Wikipedia. I don't think, as many will argue, that this undercuts the encyclopaedia's usefulness as a resource, or that it brings us back to the bad old days of the 00s when educators and experts would rail against Wikipedia because "anyone can edit it", when, even then, it was one of the most accurate, authoritative and well-cited sources on the net for almost any topic.

Instead, the problem this highlights is the social dynamics of the site. Wikipedia is run almost entirely by volunteers, and there is a strict hierarchy in place: if you have admin rights, you can overturn almost anything done by any user who is not an admin. You gain power on the site by being prolific, and by being first.

Crucially, being right or knowing your topic doesn't play into it. That's by design: editors are not supposed to use personal knowledge; they are supposed to cite external sources, support their claims, and be able to do so in the specific language Wikipedians expect. They are expected to do the grunt work of reverting vandalism, while formatting and structuring articles and, yes, occasionally writing new content. The best editors know how to manage the drive-by expertise of normal people who suddenly sit down and decide to write 5,000 words into the Bonobo article; the worst simply sigh and delete it all as "unsourced".

But if you're the administrator of the Scots Wikipedia, and you don't speak Scots, an awful lot can suddenly start looking like vandalism. People who accuse you of "writing English with bad spelling" are anti-Scots, not anti-you. If you are a hobbyist author with a desire to improve things, you might start by making a few small edits here and there, just like the administrator did when he was starting out. But if those edits are to correct the poor content written by the site administrator, well, expect them to be un-done. As, in fact, they were.

What's more concerning still is that Wikipedia is the good model. Wikipedia is a modern marvel, one of the seven wonders of the digital world, a pure and shining beacon of what the web could have been, and what we should one day fight for it to become. It's easy to see ways this problem could have been solved, but harder to see any that don't fundamentally work against the nature of wikipedia. Would we want the site to always hand admin rights to a credentialed expert? Would we want to ban editors who couldn't prove – prove to who? – that they could speak the language? Would we be happy with a much larger paid staff dealing with this sort of thing? Maybe – but who would pay them?

After a while, if you go down that rabbit hole, you end up not with Wikipedia at all, but with, well, Facebook.

I don't know what the answer is here. But I'm more worried by this story than, perhaps, I should be.

Fall Guys

I've got a fun story coming out this week about Fall Guys, the PS4/PC multiplayer hit, and I just wanted to use this slot in the newsletter to say, well, play Fall Guys.

If you've not seen the game in action, it's basically Takeshi's Castle meets Fortnite: 60 small fat characters running through a series of improbable obstacle courses, until one emerges the winner. It is one of the most gloriously fun and silly, games I've played in years, it's free for anyone with PlayStation Plus, it was made by a team of just 48 people mostly in London, and I love it.