Wikipedia, the Internet, and the Inequality of Passion

By now, most of the ideals that went into the creation of the early internet have been squashed by three decades of corporate lobbying, monetization, and general late-capitalist horror. I don’t want to speak for the pioneers of the web, but I think they were aiming more towards “internationalist anarcho-democratic sharing of ideas” and less “please disable your adblocker so you can watch this 30 second ad so you then can watch this video of a new phone while 13 year olds type memes and racial slurs in a comment box” or whatever the most concise description of our current hellscape is.

There have been a few suggestions as to why the modern internet sucks so bad, and is only getting worse. Zeynep Tufekci blames the ad market for incentivizing the collection of data that is pretty much only used for evil. I think she’s right, although I have my own more general theory. But I don’t think that fully explains why so many of our shared internet spaces that are meant to embody early internet ideals (free speech, an open marketplace of ideas, equal access, internationalism, etc. etc.) are likewise bad. In particular, I think Reddit and Twitter and Facebook (and Medium!) would still be shitholes even if the ad market didn’t exist at all. Perhaps if the internet were monetized differently, these sites would have already withered and died like so many Pets.coms, but the market, free or otherwise, doesn’t really have a great track record on these sorts of things.

My contention is that many of our internet spaces are constrained by a fatal inequality: the inequality of passion. This inequality infects every part of our online spaces, and makes them worse for it. It’s not the only inequality that makes online media worse (I don’t even think it’s in the top 10), but it’s overlooked, especially by the people who champion these spaces.

It’s all too easy to write about dysfunction in websites like Reddit and Youtube and Facebook, and there are so many other factors there, that it makes things somewhat muddled. Therefore, to illustrate this point, I’m going to examine a website that people (in general) like: Wikipedia.

Wikipedia’s great. It’s one of the few websites that comes even close to delivering on the promises made by the pioneers of the early internet. I can log on from anywhere on the planet and access knowledge about almost anything. It’s built on democratic principles (the encyclopedia anyone can edit!) and technocratic ones (you can work on big projects with other experts in your field!). And (crucially, I think), it’s one of very few websites where I can spend a lot of time and not feel worse, mentally, than when I started. Social media is poison for mental health and self esteem. They are Skinner boxes where attention goes in, and stuff to be mad or sad about comes out. But reading a few Wikipedia articles makes me feel like I’ve gained something, that I’m doing research and bettering myself. Even other websites that repackage Wikipedia articles into Cracked.com-style “5 Totally Cools Facts You Didn’t Know About…” clickbait don’t give me that same feeling: there seems to be something about the agency I have in clicking through “wiki wormholes” that I don’t get from other websites.

I recognize that this sense of self-improvement is an illusion. In fact, this unearned sense of expertise is probably bad for me in measurable Dunning-Kruger-related ways. A Google search and a few paragraphs of text is not enough to make me be an expert in anything, and it might be bad that it can help convince me that I am. But it’s not the worst thing that websites have done to rewire my brain.

Yet, the wiki model has a problem: the content I see is determined by who cares the most. A minuscule fraction of people care enough about the content of Wikipedia to want to edit it. A fraction of this fraction care enough to put in the time to learn how to edit Wikipedia (not just in the technical sense of how to edit articles, but also learning the argot and procedural rules that govern how edits are accepted). Since Wikipedia is so immense, this caring is a resource that has to be carefully shepherded. Yet, it is spent rather haphazardly. To make matters worse, there is very little correlation between how passion is distributed, and the importance of the information to be edited. Certainly there is very little correlation between passion and expertise.

The obvious solutions to these problems are all some flavor of anti-democratic. For instance, we could require extensive vetting of editors before allowing them to edit. Or we could employ social engineering or other gamification techniques to force people to edit pages that needed attention. Or we could employ fully automated systems to automatically compose, edit, and update articles. I hate all of these solutions, not just because they wouldn’t work, but because other big websites are all actively engaged in one or more of these flavors of apparent solutions for their own, similar problems, with potentially disastrous consequences. It points to something that seems to me to be fundamentally broken about communal web spaces, that the current internet is poorly set up and incentivized to fix.

I’m going to illustrate the different flavors that this inequality of passion can have with a series of case examples drawn from Wikipedia. None of these cases are awful in the same way that, say, Twitter verifying Nazis, or Facebook running social experiments to toy with peoples’ emotions, are, but I think they at least illustrate the scope of the problem.

The Pitcairn Dilemma

The Pitcairn Islands are a set of islands in the South Pacific that were settled by the British and Tahitian remnants of the 1789 mutiny on the HMS Bounty. There’s about 50 people living on the islands. There’s no real migration to the island, and the islands might be completely depopulated in 40 years or so. The islanders speak Pitkern, an English/Tahitian creole. There are some Tahitian loanwords and a different system of spelling, but it’s mostly intelligible to English speakers. For instance, Pitkern for “This article is a stub. You can help Wikipedia by adding to it” is “Dis artikal i’ stub. Yu ken hiiwp Wikkapedya b’ adden t’et.”

I mention this because Pitkern is one of the 288 languages with its own Wikipedia, at pih.wikipedia.org. It’s got 630 articles. The vast majority are stubs. For instance, the article for “Manajement” that I’ve clicked on in the image of the landing page reads, in its entirety, “Maenajemant esa jawb paeple cean dea.” There are periodic arguments for closing the wiki down. These arguments have so far failed to result in action for some combination of the following reasons:

1. Pitkern is a real language, even if it’s a pidgin. Scots is also mostly intelligible to English speakers, and it’s got almost 50k articles in its wiki and a very active user base.
2. Just because there are almost no native speakers, and those that remain are dying off, doesn’t mean it shouldn’t be a wiki. The Latin wiki has over 100k articles. Plus, if the language is dying, shouldn’t we try to preserve it?
3. Somebody edited something on it just a year or two ago, so it’s not completely dead, as a project.

From Wikipedia’s perspective, these are pretty compelling arguments. However, if you’re not enmeshed in the culture of Wikipedia’s policies, it’s absurd. It’s an entire encyclopedia that doesn’t function as an encyclopedia because there’s no actual content on it beyond a sentence or two about a handful of random nouns. As a corpus of a dying language it also fails because it seems that maybe one or two people made a few articles once, and then these core articles have been edited by English speakers in very minor ways, who then argue with each other in the comments.

There’s been some slight momentum in that the Pitcairn Wikipedia is now the “Norfuk & Pitkern Wikkapedya”, centering its remit to the related Norfuk pidgin of the nearby Norfolk islands (which at least has a population over a thousand, and some administrative effort to keep its language alive, although it, too, is dying). Still, as far as I know, nobody from either the Pitcairn or Norfolk islands have substantially edited this wikipedia in over a decade.

There are speakers and even poets among the Pitcairn Islanders, but they lack access, knowledge, or desire to put together an entire encyclopedia for no pay and for a small (and shrinking) audience. The democratic consensus is that something like a Pitcairn Wikipedia would be nice to have, but the truth on the ground is that nobody is willing to build one.

Our collaborative internet spaces are built around free labor, and our norms about how those spaces should be organized and run is at odds with the availability of that labor. I don’t know what the right thing to do with the Pitcairn Wikipedia is; neither, I expect, do the people who argue for its preservation or deletion when it’s periodically on the chopping block. I suspect it will continue in its current crippled status until enough of the handful of English speakers who vote for its survival stop caring about it.

Start To Double Greek

Never trust anybody that begins a sentence with “Clearly.”

Old Man Murray was a website that provided satirical, irreverent, and snarky discussions of video games in that all-to-short era before every one of the adjectives in that preceding phrase made me want to scream. One of its most notable contributions was the “Start To Crate” review system. Crates are ubiquitous in video games because they are a simple way of cluttering up a level, making it look real, and providing things for the player to do (blow up the crate, take shelter behind it, etc.), without having to do any of the extra work of, you know, designing interesting environments or objects. In other words, the first time you see a crate in a video game is a good heuristic for determining how quickly the designers ran out of ideas.

I mention this to propose my own metric of Wikipedia math articles: formula with two Greek symbols in a row in the main body of an article is where a grad student with too much free time took over. Beyond these, the article ceases to be informative communication, and is merely a quasi-ordered list of statements. If you aren’t already familiar with the subject in question, good luck. In some math topics, double Greek is unavoidable. But usually there’s some other feature, a “clearly” here, or “as an aside” there, or a “we now proceed to,” that indicates where the article transitioned to a set of hastily strung together lecture notes rather than a genuine attempt to introduce a wide audience to a concept.

Math articles on Wikipedia are sort of an odd duck. Some of them are really quite good, and have context, historical information, and plain English derivation of important facts. Others are just sort of unordered lists of random facts and properties. I suspect this is because of the inequality of passion rearing its ugly head.

There are many subjects in math that you need to have taken at least a college class or two to even understand the terminology. There are some where you need to have taken a grad class or two. There are a few where it needs to be your graduate school area of expertise to even have a prayer of knowing what’s going on. A lot of simple math, can, with the right perspective, have a lot of complex math behind it. This means that a lot of math articles get into the weeds very quickly.

Let’s take a look at the article for Pi, for example (the number π, not the Greek letter). π shows up all the time, in a lot of cool places. You learn about it in elementary or middle school, and it shows up in popular culture with some regularity. At first blush, this is a pretty solid article on π. You learn a little bit about the history of π, from the early Greeks to the modern computing era. You get to see some snapshots of a variety of different types of interesting math that can be used to approximate π, from geometry and trigonometry to calculus to infinite series to Monte Carlo methods.

As a result, once you get past the summary, the amount of math you have to know to understand sections of the article varies wildly from one section to the next. The History section is mostly in plain English, with a few tables and pictures showing how ancient cultures used to approximate π by measuring the circumference of shapes with more and more sides, and then some notes on series and things. There’s some calculus notation sprinkled throughout, but a middle schooler could ignore it, perhaps.

It’s around the Vector Calculus section that we get into trouble, where the section starts with a brief sentence or two description of what vector calc is, and then immediately starts into a proof for why there’s a π in Gauss’ law, ending in:

After that, we’re on to Cauchy’ integral formula:

Then the Gamma Function:

Etc etc. It’s almost a relief when, at the end, we’re rewarded with an In Popular Culture section where they talk about Pi day and the notorious Indiana Pi Bill. Now, the concepts that are talked about in these more complex sections are usually pretty cool. The Gamma function in particular shows up everywhere from statistics to number theory, and so it’s sort of neat that π, a number that has nothing to do in principle with either of those things, shows up in the DNA of all of these branches of math. You perhaps might stop and contemplate whether some grand interlocking design underlies the universe, or whatever you do when you see enough things that could be coincidences.

But the issue here is that, as an encyclopedia article, these sections don’t work. If I know what Gauss’ law is already, then I can think a little bit and say, “oh yeah, it makes sense that there’d be a unit sphere in there somewhere, and unit spheres rely on π for all of their interesting properties, same as circles.” But that’s a pretty big “if.” If I have no idea, then I get a lot of complexity thrown at me, but not enough tools to understand the complexity. I could try clicking on the article for Gauss’ law, but to understand that article, I need to understand flux, and its relationship with Coulomb’s law, and whoa, is that a triple integral there? And a dirac delta function? If I want to make this section accessible, either I don’t include the fact about Gauss’ law, and my article is incomplete, or I include paragraph after paragraph of context, and then my article is way too long, and barely about π in the end anyway.

There’s a mixture of three sort of flaws here:

The first flaw is the simplest to explain: most of the people who can explain complexity are the people who have studied the complexity long enough that they are partially immune to it. This is sometimes called the “Curse of Knowledge.” In my personal experience, the internet is pretty bad at providing spaces for people with different levels of knowledge. There are a few spaces for beginners, and some for experts, but a relative dearth of stuff for people in between those poles.

The second flaw is that we value completeness and universality in naïve ways. We build internet spaces for everybody, which is the same as building them for nobody. There’s that famous bit of statistics folklore about the U.S. Air Force measuring their pilots to find the average height and weight and so on, and then designing the cockpit to fit the average pilot. Since no pilot was average in every category, the planes were very difficult to fly. They were designed for an “average” person who didn’t exist. The solution, of course, was to make everything adjustable. There’s a related issue here with making internet spaces accessible (adding special accommodations to them to allow new groups to use them) versus making inclusive internet spaces (designing them, from the very beginning, to let in as many people as possible). I don’t think the modern internet has hit on very good solutions for inclusivity, in general. Would the ideal article for π not include the complex sections at all? Would I have two articles for π, one for “beginners” and one for “experts?” Would I have to “unlock” the more complex parts of the article, by showing that I can handle it? Again, I’ve seen websites try solutions with all of these flavors to manage their audiences and content, and I think all of those approaches have pretty severe drawbacks.

Lastly, complex systems often have interconnected complexity, and actually get more complex the longer you study them. The solution to this is abstraction and encapsulation, where I learn small, manageable chunks about how things work, and then use those as scaffolds to add on complexity. You learn math by doing addition and then learning about multiplication as repeated addition, etc. etc., not by being given the Peano Axioms and an empty notebook and being told to go nuts. Internet spaces, in contrast, are built around the central metaphor of the hypertext, where every page is linked to everything that it is related to. Hypertext gives us the freedom to move between related things on the internet, but at the cost of much of the scaffolding structure we need to interpret the things we find. We read books from start to finish; the internet can only be read in medias res, and refuses to gives us a glossary when we ask for one.

The P-Funk Problem

Everybody likes funk music, right? I mean, I guess you could think it was hokey or something, but it’s still a lot of fun, fuck you. George Clinton is one of the biggest names in funk, and his Parliament Funkadelic projects are the cornerstone of any history of funk. Please keep up with this mythology:

1. The world, generally, sucks. Our negative attitudes and actions are dooming mother earth.
2. Dr. Funkenstein is from space.
3. He, and others of his kind, have guided humanity through the ages.
4. Dr. Funkenstein has returned on the P-Funk Mothership to administer funk to us all, which will free our minds.

The P-Funk Mythology grew more complex and convoluted as time went on, but I think you pretty much get it. During Parliament Funkadelic concerts, the P-Funk Mothership would descend on stage, and George Clinton in his Dr. Funkenstein persona would emerge and play music. If you want to see the mothership, it’s in the African American History Museum in DC and it rules.

I’m going to tell you a story about the Wikipedia page for the P-Funk Mothership. Anyway, the article is mostly fine now, but it didn’t used to be. It was actually proposed for deletion way back in 2006, because there was a concern that it didn’t have quite enough content to exist separately from the main P-Funk Mythology page. My story isn’t about that first brush with deletion, but the second one.

You see, there’s one thing you might have missed, on the P-Funk Mothership page. It’s right at the bottom:

It’s the “Fictional spacecraft” category that nearly led to the end of this blameless page. Often, pages have categories assigned to them. This makes it easier to, for instance, to find all the lists of lists. But these categories are just like regular wikipedia pages, in that they can be edited and added to. And this means that they become somebody’s territory.

As I mentioned previously, there is way too much content on wikipedia for everybody to deal with. Usually, editors will stake out a turf of pages that they “watch.” If you’re an Aaron Burr fanatic, for instance, then you watch all the pages that are related to your main dude. That way, if somebody edits the Hamilton page to say he was shot by Joe Chill or whoever, then you get a little notification, and you can put that upstart or vandal back in their place. Other wikis or editors will watch for all edits made by new editors (to either check them for vandalism or to help them onboard with the community).

This patrolling happens at different paces in different places. For instance, I edited the Star Trek wikipedia to change “photon torpedoes” to “proton torpedoes” and that was reverted within 2 hours. Other edits take longer to be discovered. The FYAD subforum of Something Awful had a contest to see who could make the longest, funniest edits to wikipedia pages without their vandalism being discovered. Some are still up, 7 or 8 years later.

I mention all of this because there was a person whose turf was the Fictional Spacecraft category page. Now, the page itself only had a couple dozen spacecraft in it at that time, mostly from Babylon 5, so the P-Funk Mothership stuck out like a sore thumb. This person immediately put a new proposal for deletion (“prod”) on the article, which resulted in the following conversation on the P-Funk Mothership’s talk page:

If you don’t want to read all of that (and I can’t blame you), the gist of it was that the person monitoring the spacecrafts category had never heard of Parliament, and so thought the article was some joke or something, an affront to the dignity of their curated list of spaceships, and wouldn’t leave the article alone until it was “cleaned up.” This wasn’t enough to get the article deleted, but it did result in it being considerably shortened, as editors hate those “[citation needed]” superscripts. A year or two later, this resulted in the following comment:

I didn’t type this, but I wish I had.

It’s better now (mostly, or perhaps entirely, because it’s a piece of music history important enough to be in the Smithsonian), but it’s a good illustration of the twin perils of fandom and ownership.

Anybody can become an expert in Star Trek or Babylon 5 or what have you: you just have to be a fan and watch some television, memorize some trivia, and that’s enough to make headway. And, once you’re a fan, it’s very easy to come up with facts and standard of proof. E.g., “Data feeds his cat feline supplement 25 in the Star Trek: TNG episode ‘Phantoms’, but feline supplement 74 in the episode ‘Data’s Day.’” There exist entire groups that obsessively archives and debates and catalogues information about their fiction. That there are way more fans of something like Star Trek on Wikipedia than fans of funk music is naturally going to have systematic effects on how knowledge is retained and used. This is one of hundreds of reasons why diversity matters in organizations, and why it’s a problem that Wikipedia’s editor demographics are so overwhelmingly white and male.

Even beyond this inequality of representation, there is another problem with fandom and internet spaces. In particular, there seems to be a disconnect between fandom-knowledge and, you know, real world knowledge. Real information about the real world is messy. Unless you went to a Parliament/P-Funk show during about a 20ish year period, you wouldn’t have any “proof” that the P-Funk Mothership “really” (fictionally) flies. And, of course, reasonable people can disagree about things. Reality is very different from “canon.” If you’ve only got one channel and standard of verification, you’re doomed to cast too wide or too narrow of a net, whether you’re giving out blue check marks on Twitter or deciding what is or isn’t “fake news.” This “facts as canon” trope is why I think (among other reasons) something like Verrit, a website made by hangers-on of Hillary Clinton’s 2016 campaign that seeks to create repositories of “verified knowledge,” is doomed.

Likewise, lots of people like to stake out digital real estate. We build bubbles around ourselves, and web platforms make it very easy for us to do this. When people pierce these bubbles, even inadvertently, we can react with hostility. Many spaces on the internet try to support both private community-building as well as democratic open platforms. These goals are fundamentally opposed on most internet platforms. When randos show up on a twitter thread between you and your friends, or if you make a Facebook post about how much Trump sucks and a friend of a friend posts “triggered, liberal snowflake?” in it, it’s just a symptom of the overarching issue that the internet really doesn’t know how to handle the problem of circulation, and the sweet spot between amplifying voices and preserving communities.

The Silent Hill Predicament

Silent Hill is a series of horror video games where the main action is solving puzzles in foggy environments and running from inhuman monsters. There’s an in-depth lore about secret cults and towns with dark secrets and perhaps giving birth to a new deity that will cause the world to perish in flames. But, to be honest, it all gets rather convoluted. So, it’s a perfect candidate for a fan-based wiki, especially since most people haven’t played the Silent Hill pachinko machines or light gun arcade games or what have you, and so may have missed out on parts of the story.

If you checked out the Silent Hill Wikia page for the character Walter Sullivan in, oh, 2015 or so, you might have been surprised to see that, in fact, the fourth game is an elaborate allegory for the horrors of male circumcision. This might strike you as odd, since the game was made by a Japanese studio, a country where male circumcision is very rare, and most importantly, nobody in the game says or even hints towards anything about circumcision, but it was in the wiki, so it must be true.

Naturally, some well-meaning editors attempted to ask what gives. Part of the response is below:

Again, if you’re averse to words, one of the admins found male circumcision so distasteful, and its connection to the Silent Hill franchise so self-evident, that they responded with long, uninterrupted paragraphs likening the people questioning their article to Holocaust deniers. The tweet I embedded above has only part of it. If you’ve got a lot of spare time, then feel free to read the whole thing (minus the responses that the self-same admin deleted for not being “civil”). The word “sheeple” is used in earnest. The US government, Judaism, Christianity, and the news media are all accused of being Satanic agents of the Illuminati. I think you get the picture.

Eventually, long after word of this meltdown reached Twitter and parts beyond, the conversation was locked, and one of the other admins stepped in. The unhinged admin was temporarily de-modded. The anti-circumcision polemic was then removed from the article about the video game character, and life continued on mostly as normal.

I mention this story for a few reasons (besides that fact that it’s weird and, to me, funny). The first is that there is nothing particularly special about this admin. They simply cared a lot about fandom-related wikis, and had first mover advantage on the Silent Hill one. They were there first (or the right kind of active first), had the right amount of free time, and so got to be an admin. If you had made a few life choices differently, you could be inserting your own dire warnings against perceived conspiracies in, say, the Full House or Xena wikis until forcibly removed.

The second is that the ranting person was able to continue arguing as long as they wanted to, while keeping the status quo. It took direct intervention by another admin to stop them, not rhetoric. The other side tried many different tones, conciliatory and otherwise, to very little effect. If there had been fewer admins, or if the admins had had similar axes to grind, or if just a few parameters had been different, the rant would still be on that page today. Then, if you wanted to learn actual information about Silent Hill, your only options would be either to start your own, competing Silent Hill wiki (although they have the silenthill.wikia namespace, so you’d probably be stuck with real-silenthill.wikia or something), or to simply try to raise in the ranks enough to oust the admins. In short, palace intrigue would be your main way of getting things done.

There are a few lessons that I took from this, that I think apply to the wider internet:

1. Internet moderation is a thankless and difficult task that usually doesn’t even pay any money, so you have to be really passionate or really power-hungry to do it.
2. “Really passionate or really power-hungry” are not the adjectives that are best associated with “fair and even moderation.”
3. It’s very easy to get put into positions of relative power in web spaces through blind luck or just being there first.
4. “Blind luck or just being there first” are not the circumstances that are best associated with “responsible use of power.”
5. There are not very many democratic mechanisms for removing people from power in web spaces once they are put there. You frequently have to rely on cabals or back channels.
6. “Building consensus” can be at direct odds with “don’t feed the trolls.”

So much of the internet depends on moderation, from deciding who to ban or elevate from a communication platform, to deciding what ads you get to see, to even shaping what kind of content you have easy access to. It is separating the sheep from the goats, over and over again, every day, forever. In the past, this required lots of human labor. Human labor is expensive and doesn’t scale as well as computational labor, so companies tried to automate some or most of this labor. That, in many key areas, has proven to be a mistake, and now the big companies are trying to bring more humans back into the loop. Your experience on the internet is largely dependent on whether a small handful of people, many of whom are working for no or low pay, correctly identify who is a sufficiently big enough asshole in a sea of assholes. And a lot of those people may be assholes themselves.

Conclusion

I presented these vignettes not to rag on Wikipedia; as I mentioned before, I actually like Wikipedia, in general. Rather, I use them to highlight some issues with what are often seen as the ideal of web spaces: open, democratic, and interconnected platforms. To operate at their best, web spaces require passion and investment from large numbers of people. The architecture of the web is not set up to create or distribute this passion in just and useful ways. In fact, it promotes large and damaging inequalities. I’m far from the first person to point out that “the best lack all conviction, while the worst are full of passionate intensity,” but it’s important to see that this inequality damages even internet spaces that are utopian in their construction.

I think the web is going to get shittier. There are a lot reasons for this. Our encroaching post net neutrality dystopia is one. The fact that our web experience is being driven by a shrinking handful of corporations is another. But reform is only possible if we envision a better world. And I think that the status quo ante of the “free and open web” is not going to cut it.