How Wikipedia Has Responded to the George Floyd Protests

Jun 25, 2020

"There are decades where nothing happens, and then there are weeks where decades happen" is an old and likely apocryphal quote attributed to V.I. Lenin. It's been popular throughout the tumultuous and time-warping Trump presidency, but in the second quarter of 2020 the weeks have felt like eons.

The Wikipedian has written twice previously about how the encyclopedia anyone can edit has covered the coronavirus pandemic. Today I'm interested in how it has handled the George Floyd / Black Lives Matter protests, which—over the course of the past month—have grown into an international movement whose impact is being documented in the press, in the streets, and on Wikipedia in real-time.

It's a lot for even a crowdsourced encyclopedia to keep up with, and simply trying to decide what to write about was no small task. The closest thing to an overarching theme is the rapidly changing attitudes toward racial and policing disparities in what might be called the second Civil Rights movement. This post can only speak to a narrow part of that, and specifically will focus on how George Floyd and the protests in his name have been covered on Wikipedia; how articles about the numerous police killings in America are organized; how questions about diversity around the Wikimedia project are (and aren't) being addressed; and how any of the above could change in the future.

Decades more may happen in the weeks to come. For now, here's some of what's transpired:

Wikipedia Says His Name

As of this publication, approaching 1,500 Wikipedia articles mention the name George Floyd, an explosion of new content that reflects public outcry over his killing and government responses. Sadly, when you try to think of other people who became this famous immediately upon their deaths, the most prominent examples are other Black men who have met fates similar to Floyd's at the hands of police (and at least partly within view of a smartphone camera).

The other proximate comparison, for general newsworthiness, is of course COVID-19; when this blog last wrote about the topic in mid-April approximately 6,000 articles used this novel phrase, and today there are more than 24,000. The focus article in that case, now called COVID-19 pandemic, has received more traffic overall and more edits than the focus article in this situation, but it is certainly comparable.

In this case that article is about the incident and its aftermath: Killing of George Floyd. The article has been edited more than 4,200 times by more than 700 different editors since it was created almost one month ago, and it has been viewed more than 4 million times, not counting 58 additional articles in other language editions.

The biographical article about George Floyd himself has over 900 edits by more than 200 separate accounts and more than 3 million views, plus 24 additional articles in various languages. From a Wikipedia perspective, it is somewhat unusual that there is even a standalone page for him, as there is not typically a separate biographical article for victims of police violence. Eric Garner, Philando Castile, and Breonna Taylor do not have biographies separate from articles about how they died. There is one for Trayvon Martin, though, whose death was similarly racially-charged, albeit not at the hands of the police.

Then there is George Floyd protests, with more than 4,700 edits from 800+ editors, more than 1.8 million views, and 42 articles in other languages.

George Floyd mural outside Cups Foods in Minneapolis. Photo credit: Lorie Shaull

Around 80 other articles actually have Floyd's name in the title, thanks to the repeated construction of George Floyd protests in [location] across the great many places where they have occurred, both in the U.S. and around the world. Some of this owes to the secondary discussion around the many statues and monuments—of Confederates and others—torn down or formally removed from public view in recent weeks. There are too many to link, but the articles List of George Floyd protests in the United States, List of George Floyd protests outside the United States, and List of monuments and memorials removed during the George Floyd protests seek to account for them all.

Floyd's name is now tied to the reputations of many other people, places, and things:

In section headings in biographical articles about several figures in or associated with the Trump administration, including Mark Esper, Mercedes Schlapp, and Candace Owens.
In other articles about government, activist, and corporate responses to his killing, such as Donald Trump photo op at St. John's Church, Police abolition movement and Defund the police, and CrossFit.
In articles about responses and effects, including the symbolic meaning of Eight minutes and 46 seconds, Blackout Tuesday, Capitol Hill Autonomous Zone, Black Lives Matter Plaza, and the George Floyd Law Enforcement Trust and Integrity Act, introduced in Congress less than two weeks following his death.
It even intersects with older topics like When the looting starts, the shooting starts—though this article did not exist until May 29, following Donald Trump's infamous tweet.

This list only scratches the surface—here's a link to the search if you want to keep exploring. But one clear takeaway is Wikipedia's response has as little precedent as George Floyd's death has far too much.

When is a Death a Killing, a Shooting, or a Murder?

Although Killing of George Floyd is by far the most visited of these articles, it has received less traffic than a very similar article that no longer exists. Strictly speaking, they are the same article: for the first week of its existence, when the world first learned about what happened in Minneapolis on May 26, the article was titled Death of George Floyd. The name change followed a debate spanning more than 24,000 words—longer than Franz Kafka's The Metamorphosis—and there is a very good chance the page will be renamed again, eventually. As first reported by Stephen Harrison in Slate earlier this month, there is a contingent of editors that wants to move it to Murder of George Floyd. That discussion was closed after a mere newspaper article's worth of discussion, about 1,300 words.

Factors considered in the successful renaming effort included: which word was most accurate; which was most neutral; instances of "death of" and "killing of" in news reports; applicability of the site policy referred to in shorthand as WP:BLPCRIME; and, most interestingly, the precedent of similar articles about police killings of citizens. Death of Eric Garner was mentioned in about a dozen comments opposed to the change, but also by a few in support. One said: "I'm legitimately interested to know what a move for this article would mean for the Death of Eric Garner article, then, since the two cases are extremely similar."

In fact, once the change was approved—faster than in most cases, on account of the tremendous public interest—a proposal was made to rename the Garner article to Killing of Eric Garner, now citing the Floyd article as precedent.

Meanwhile, a different choice was being made about how to reposition seven other articles, beginning with Death of Breonna Taylor. But this was not connected to the Floyd or Garner cases at all; these articles were about fatal police shootings specifically, and in fact the proposal went up on May 23—three days before Floyd was killed. The discussion to change all of these from "Death of" to "Shooting of" and bringing consistency to the category of articles called People shot dead by law enforcement officers in the United States took fewer than 600 words to reach consensus.

Black Lives Matter Plaza in Washington, DC. Photo credit: author

The desire for consistency is understandable, but the more you look up similar police killings, the lack of coordination across Wikipedia articles becomes quickly evident. Among the better known cases in recent U.S. history, each is considered differently: Shooting of Michael Brown, Death of Freddie Gray, Shooting of Trayvon Martin, and the newest entry, Killing of Rayshard Brooks. While at first it seems like the use of a gun dictates the difference between "Shooting" and "Killing"—and "Death" where there is some ambiguity—the Brooks article complicates matters. As it happens, there is an ongoing debate about whether to rename that article, and it seems like it very well might happen. So far the discussion has lasted more than 11,000 words, about the length of a profile in The New Yorker.

Confusing as the above may be, Wikipedia's categories are even more of a mess. Here is a list of some largely overlapping categories that I found by clicking around for just a few minutes:

I wouldn't be surprised if some of these are consolidated at some point; in the meantime, I recommend instead looking to the Black Lives Matter template, which lists more than 70 such cases. Among them, you will find a few using the "Murder of" prefix that some would like to apply to the Floyd article. For instance: Murder of Renisha McBride, Murder of Laquan McDonald, and Murder of Jordan Davis. In all of these cases, the change followed successful prosecutions of the officers responsible. A Murder of George Floyd article will clearly have to wait, and—as of this writing, and with no arrests made—a similar article about Breonna Taylor will have to wait even longer.

Who Tells Your Story on Wikipedia?

Wikipedia can be quite adept at documenting current events, whether hurricane, pandemic, or social movement. But there is a level of understanding beyond mere documentation, and the coverage of the George Floyd protests raises questions about how well Wikipedia addresses and contextualizes topics relating to Black people, their culture, history, and wider impact.

The rest of this post will examine this from two perspectives. First, Wikipedia's demographics: how much do we know about the backgrounds of those editing, and how much does it matter? Second, Wikipedia's content: what are some ways to look at how Wikipedia handles topics relating to people of African descent? Putting them together: what, if anything, should the Wikipedia community do about any of it?

It would be wrong to say Wikipedia is not aware of diversity or systemic bias in its midst. In fact, one of Wikipedia's most famously persistent shortcomings is its low percentage of non-male editors: just 10%, even after years of organizing and repeated public acknowledgment of the problem. It's also one of the most studied questions: a search of Google Scholar yields over 100 results for the narrow phrase "gender bias in wikipedia”. The number of results for the phrase "racial bias in wikipedia”? Zero.

This problem extends to surveys conducted by the Wikimedia Foundation (WMF) itself. Most recently, a 2018 report on the diversity and "health" of Wikimedia's communities asked contributors about gender, age, education, and geography—but nothing at all about race or ethnicity. Pages written by editors to raise issues with other editors, such as Who writes Wikipedia? and Systemic bias, have little to say about demographics or the potential for racial bias. Another page about the demographics of Wikipedia editors makes no reference to race at all—likely because the surveys do not.

Curious about this omission, I asked the WMF communications team about it: in a brief email exchange, they confirmed that questions about race and ethnicity have never been included in its periodic surveys of contributors, including the 2019 report which will be released in July. As for why these questions haven't been included, they told me it was "largely due to methodological challenges: given that race/ethnicity is defined differently across countries, it can be difficult to find language and a methodology that can be applied consistently around the world." But they added that they do plan to ask these questions for the 2020 report, which will be available in the first half of 2021.

While we're on the subject, what about WMF employees? In October 2019 the Foundation released a "diversity and inclusion" report, which did ask about race and ethnicity. It showed that two-thirds of employees were white, 13% were Asian, 8% were Hispanic, and just 7% were Black. (Another 5% were of mixed race, without further specification.) The numbers are worse when you look at employees in technical roles: 78% are white, 3% are Black, but a little better when you look at executives: 58% white, 14% Black (and 14% each for Asian and Hispanic). Overall, the report states, the numbers are better "compared to last year’s diversity report, but we still have significant room to improve." (These figures cover only U.S. full-time employees, not contractors or international staff, but it's much more than a representative sample.) Nonetheless, it is entirely plausible that the Wikimedia Foundation is more diverse than the community of editors it supports. Next year we might find out.

Race / ethnicity of Wikimedia Foundation U.S. employees. Source: WMF

The best information resource, surprisingly or unsurprisingly, turns out to be an article in the main encyclopedia called Racial bias on Wikipedia. It's not a perfect article, but it is more informative than what little can be found behind the scenes.

Even as information is scarce, collaborative efforts are growing: WikiProject Black Lives Matter was created only in the first week of the Floyd protests, though it has identified just over 400 articles within scope—many fewer than I would have expected. Meanwhile, longer-running efforts like WikiProject African diaspora (with more than 10,000 articles under scope) and the Wikimedians of the Caribbean User Group are other places to collaborate specifically on Black subjects. WikiProject Countering systemic bias is another place to get involved, albeit one that is less active so far as I can tell. I suspect that it is because the topic is rather broad, and specificity matters. That is especially true for Wikipedia editors, who tend to edit on the subjects they know best.

Open Knowledge as Anti-Racism Tool

Recognizing the limits of my own knowledge on this subject, I reached out to Sherry Antoine, the executive director of AfroCROWD. The initiative was established in New York City in 2015 and sponsors or participates in at least one "edit-a-thon" or similar Wikipedia-focused event every month. In particular, I was curious how she understands the current state of Wikipedia's demographics and content from a Black perspective.

Among my first questions for her was: has there been a formal investigation of Wikipedia's content regarding Black topics? My point of comparison is WikiProject:Women in Red, a group which has measured the number of biographies of women relative to men across Wikipedia. (It's not much better than the percentage of non-male editors.) According to Antoine, there has not been one that she is aware of. Antoine is supportive of the idea, although she has a good idea what one would find: "There is little disputing there is an under-measure of information about people of African descent," she says.

The lack of a well-known benchmark—the kind of factoid that can make for a good news hook—has not stopped AfroCROWD from doing something about it. The organization has worked closely with New York's Schomburg Center for Research in Black Culture and others to identify missing biographies across professions including STEM, medicine, and the arts, as Antoine puts it, "trying to find and fill in the gaps that we may not even know about."

Here are four gaps I hadn't thought much about until researching this topic:

Activity on non-protest topics: Obviously, there is a ton of activity around the ongoing protests. But what about other Black topics? I decided to examine three articles: African-American history, Timeline of African-American history, and List of African-American firsts. I found that the first one had experienced a considerable traffic spike in late May and early June, but not the others, and none had a corresponding uptick in editing. Based on this, I assume that most of the activity is focused on current events, not the long-term coverage of Black topics, at least in the United States. Speaking generally, Antoine acknowledged the tendency for breaking news to drive editing activity. “I think and I hope that the attention that has been brought to it will continue when the news stops talking about it,” she said.

Wikimedia Commons: As Antoine pointed out to me, if you search Wikipedia’s image and media repository, Wikimedia Commons, for phrases such as “black women”, the results are dismal. If you follow the link (NSFW), yes, there is a photo of the female members of the Congressional Black Caucus. But the returned images actually contain more photos of white women, including some in various states of undress. Change it up slightly to “black woman” and the results are no better: it brings up an Egyptian figurine and women of other races, and another nude white woman, not to mention other sexualized content, before it returns any black women. Likewise, searching for a photo of a Black woman using a computer yields some relevant illustrations, also a few white women using computers, and more NSFW content. Matters improve if you find topic pages such as Black people, although it is mainly celebrity headshots, and the category African-Americans includes people like Johnny Depp and Penn Badgley, whose African ancestry is negligible to non-existent, respectively. Commons is notorious for being uncensored to the point of embarrassment, and a free-for-all to boot, but this is a bit much.

Capitalization: Should Wikipedia capitalize the word “Black” when discussing people of African descent? Capitalization has long been the practice of African-American newspapers such as The Chicago Defender, and in recent weeks has been adopted by the Associated Press, Los Angeles Times, and other mainstream publications. The Columbia Journalism Review, having opined otherwise as recently as 2015, now recommends capitalizing the word "Black" in such contexts. But Wikipedia has no formal rules on this subject. In fact, the Manual of Style’s guideline on Capital letters#People and their languages is just one sentence long, and focused on another topic. The first time I can find that the matter came up was way back in 2005, when it was reasoned that “black” and “white” are not proper nouns, and the lack of interest in capitalizing “white” would create an imbalance. The last time before the present moment it came up again that I can find was in 2018, when it was quickly shot down. A new Request for Comment was posted on June 21, and so far it has not attracted much attention, suggesting that Wikipedia will retain the status quo—with the question of what to do about “white” complicating matters as it often does. (For the purposes of this post, I have struggled with how to write it, and while The Wikipedian generally follows Wikipedia conventions, I have settled on capitalizing it for now.) In any case, Antoine is more concerned that anxieties about writing mechanics not disincline someone from contributing: “If that’s the focus, then we’re missing the point. Just write it. Get it in there. Later on we can have an argument about semantics.”

Blackout Tuesday: Wikipedia chose not to participate in Blackout Tuesday, a protest action held on June 2 where various businesses went silent, either all day long, or for 8 minutes and 46 seconds. In 2012, Wikipedia did join a blackout to protest two pieces of U.S. legislation known as SOPA and PIPA. That was very controversial in the Wikipedia community, and The Wikipedian covered it at the time. This time the opposition was much more uniform, the effort being seen as outside Wikipedia’s domain, and arguably unhelpful. As one editor put it: “We are an encyclopedia designed to provide knowledge to all for free. Going dark takes away that knowledge.”

Knowledge was very much on the mind of Spelman College professor Alexandria Lockett, as the (virtual) keynote speaker at AfroCROWD's Juneteenth edit-a-thon, which I attended last week. Lockett described Wikipedia as a "liberatory" force: the fact that anyone can edit Wikipedia is revolutionary—a point often noted since Wikipedia's founding, but even more salient in the struggle for racial equality. And yet Lockett said her students sometimes will feel they do not have the authority to make changes themselves, even with well-prepared content. "You do not have freedom," Lockett said, "if you do not feel you have the right to make knowledge."

Taken very literally, it's a chicken-and-egg problem: if Black content is missing from Wikipedia, would-be Black editors might feel less comfortable contributing. But if there are not enough Black contributors, there likely won't be enough Black content. The same would hold true for any marginalized group. This is a stylized presentation of the issue, to be sure, and in fact there are editors of color and efforts to encourage more of them, not least via AfroCROWD. But it does broadly describe the challenge involved in making Wikipedia a better resource on subjects that the straight white men who founded Wikipedia simply might not ever think to write about.

And just as white Americans and Europeans have been driven by conscience to join the George Floyd protests, so too might Wikipedia editors—whatever their ancestry—amplify efforts to increase representation of Black voices on the platform. As Antoine puts it, learning about people you might not encounter frequently "stretches your ability to understand what you might normally fear—the unknown." She adds, "It allows you to make up your mind, rather letting your immediate society make it up for you."

Any and all efforts to close information and representation gaps will ultimately strengthen the platform. The knock on Wikipedia, from some corners of the internet and academia, is that its "anyone can edit" ethos means that it ultimately lacks subject-matter authority, no matter how strong the citations. But by increasing topic and community representation, and highlighting this diversity, Wikipedia draws incrementally closer to being the type of epistemic change that many want to see.

How Wikipedia Has Responded to the George Floyd Protests

Discussion about this post