Why Codes of Conduct are a total failure in addressing misogyny
A lot has been said and written about sexism in the Free Software movement over the years, from creepy behavior towards women attending community meetings, to overly sexualized language among software development teams, to infantilizing behavior towards female contributors of projects. What probably hasn’t been talked about as much is the recent emergence of a new type of sexism, taking the form of ostracizing feminist viewpoints based on misguided attempts at inclusivity.
In this article I want to shed a light on this phenomenon, as a wake-up call to anyone who is genuine in their commitment to inclusivity and anti-discrimination principles.
But first, for those not in the know, let’s recap what Free Software is.
The GNU Project
In the earlier days of computing, the source code making up software used to be seen as something akin to scientific formulae, and as such shared freely in the spirit of openness and collaboration — principles held dearly in academic communities. But as software began to form a profitable industry, the practice of obscuring and protecting code as “intellectual property” began taking hold.
Unsurprisingly, many were unhappy with this development. Given the increasing role computers play in every aspect of our lives, it’s crucial to be able to study and understand what the software is really doing under the hood.
In 1983, Dr. Richard M. Stallman started the GNU Project, whose aim was to develop a complete computer operating system entirely under the principles of Free Software. Two years later, he founded the Free Software Foundation as a legal entity to represent the movement.
The GNU Project was a little over-ambitious with the design of a core element of the software, called the kernel. However, in 1991, Linus Torvalds would begin working on a simpler kernel himself, which came to be known as Linux, and was eventually published under the GNU General Public License, a software license written by the Free Software Foundation, representing the values of Free Software.
Although Linux is just a kernel and not a complete operating system (you can think of it as just the engine of a car), the combination of the Linux kernel and the broad suite of system components developed by the GNU Project started to be known as the “Linux operating system” (more accurately referred to as “GNU/Linux” or “GNU+Linux”) which nowadays is the leading operating system in many areas of computing, both hobbyist and professional. Industry giants such as Google, Amazon, and Meta (formerly Facebook) use GNU+Linux extensively in their technical infrastructure. Sadly, they don’t always give back to Free Software; at times they are outright hostile to it and merely leech off of its technical achievements.
Cultural Issues in Free Software
Back to the topic at hand.
One might hope that communities built around values of freedom and openness, like the Free Software scene, would be less sexist than brutally profit-oriented corporate environments. Yet women’s experiences paint a different picture. “Lad culture” seems well and alive in Free Software, as the male-dominated aspect of the world of software applies not only to industrial but also hobbyist and academic settings.
But what with all the recent developments, in which sexism and other social issues in free software began being talked about and called out more openly? Even the founder of the movement, Richard Stallman, has been “cancelled” via accusations of sexism, based on a misreading of comments he made about the Jeffrey Epstein human trafficking scandal.
Many software projects now embrace “code of conduct” (CoC) documents which explicitly prohibit bigoted behavior. One would hope that this includes the prohibition of anti-feminist hostility, and I do think that these CoC might help to curb out the more overt expressions of sexism.
Yet there are also indicators that the CoC trend is a total failure when it comes to truly respecting the breadth of anti-sexist viewpoints held by women around the world. Further down, I will demonstrate how a CoC can in fact serve to outright ostracize community members if they voice a feminist opinion that goes contrary to established orthodoxy.
The Contributor Covenant
Given that it’s hard work writing and agreeing on a code of conduct, it’s common to adopt a pre-written CoC. Unquestionably, the most well-known example is the Contributor Covenant, authored by transwoman Coraline Ada Ehmke and initially published in 2014. According to Ehmke’s website: “it’s been adopted by over 40,000 projects including Linux, Golang, JRuby, Swift, F#, Rails, and the open-source portfolios of organizations including Apple, Microsoft, Google, Salesforce, and Intel.”
That is an incredibly impressive track record. You might think then, that the Contributor Covenant must be a really well-written CoC, and as such incorporating anti-sexist ideals as well, among many other anti-discrimination principles. After all, sexism is one of the primary and major forms of oppression recognized widely in liberal democracies, perhaps on par with racism.
And yet shockingly, Ehmke has explicitly refused to even acknowledge that “sex” is a meaningful category based on which people experience discrimination and oppression. Though they would surely frame it in different wording, this amounts to denying that sexism even exists.
One of the early paragraphs of the Contributor Covenant lists a number of characteristics based on which it’s explicitly prohibited to discriminate against someone. This is meant to make things a little more concrete, so various groups can be reassured that any discrimination they may be facing in the world is recognized. For example, the list includes age, body size, disability, ethnicity, and of course gender identity. Conspicuously, it omits what should be one of the most obvious things: sex.
This omission was first brought up publicly, as far as I know, in August 2017, by Corinna Cohn on GitHub. They proposed a change to include “sex” in the aforementioned list. Following is a comment of theirs in the ensuing discussion, which succinctly sums up the issue:
I am trans, not cis, and am the originator of this change [request]. Regardless of my being trans, I still would choose to have my sex protected as a separate characteristic, and believe that it’s especially important to acknowledge the various forms of discrimination that predominantly affect those whose sex is female. There is nothing “cis-centric” in acknowledging sex-based oppression. Trans-men are susceptible to sex-based oppression regardless of their personal identity.
Corinna Cohn
The change request was ignored for almost a year, when I posted another one in June 2018, unaware of the earlier request. Ehmke seemed initially willing to accept the change, but then reverted and suggested to use the phrase “sex characteristics” instead. There was some arguing back-and-forth, which resulted in the updated document using the phrase “sex characteristics” in lieu of “sex,” and Ehmke calling my words “exclusionary and hurtful” while continuing to deny that there exists a class of humans identifiable as members of the female sex and facing oppression on that basis.
A Case Study of CoC Failure: GNU Guix
Fast forward to February 2022, in which I tried to return, as a contributor, to a Free Software project that I find extremely valuable. Developed under the GNU umbrella, it’s called GNU Guix, and while the details of it are too technical for this article, suffice to say it takes a massive leap forward in supporting Free Software ideals in actual practice in our contemporary world of ubiquitous software.
Realizing that Guix too had adopted Coraline Ada Ehmke’s Contributor Covenant, and was still on an older version that mentioned neither “sex” nor “sex characteristics,” I decided to send in a patch file that would simply add the mention of “sex” to it. I wrote to the mailing list that the original author refused to add this due to some strange views they hold, and didn’t go into detail so as not to be incendiary.
I especially did not say anything about transgender topics or gender identity when raising the issue, since I think it’s perfectly fine for a code of conduct to show equal respect to various conflicting views on a matter. My suggestion to recognize sex class discrimination was at no point an attack on the transgender perspective on sex and gender. It was only meant as a parallel addition. After all, the point of a CoC, one would hope, is to ensure a friendly work environment and not to enforce any ideology on project contributors. I see this as being similar to the issue of religious freedom: you neither want to discriminate against people of any religion, nor against atheists. This is regardless of the fact that religious people may privately believe that atheists are heretics, or that atheists may privately believe that religious people are irrational.
Maybe I should have seen it coming, but soon after I sent in the recommendation to add “sex” to Guix’s CoC, there were immediately some replies questioning the intention. They suggested that the proposed change might open the gates to “invalidating transgender identities.” Someone referenced my past debate with Ehmke and proposed it as proof that I just had an axe to grind against trans people. Of course, they paid no mind to the fact that the issue of ignoring sex in Ehmke’s Contributor Covenant had first been brought up by another trans person. Further, I’ve been told that differentiating between sex and gender is a transphobic talking point, and that humans don’t have a biological sex. At that point I didn’t argue back at the person saying these things, since it seemed like a waste of time and off-topic. I had opened the discussion to establish in the project’s CoC that sex discrimination exists, not to argue with anyone about their belief systems.
Another trans person, who had been part of the Guix community for a longer time, seemed more reasonable, so I tried to establish dialogue with them. I stressed again that the only intention of mentioning sex discrimination in the CoC was to pay equal respect to the feminist perspective that sex matters, but this didn’t help at all. Soon the person was talking about the “hurt I had caused” and speaking of suicidality. They also called in their female partner into the discussion, to “prove” that my position was not in fact a pro-women position. Ironically, I think, this was the first (and only) female person to join the discussion. The regulars of the Guix community seem to be almost exclusively men and transwomen. From what I can tell, there are more transwomen than women, though I could be wrong.
I’ve dealt with severe depression and suicidality for much of my life, so I have genuine sympathy for people suffering similar conditions. As such, I poured my heart out to this transwoman, and asked just what it was in my emails that was so triggering. That went bad. I received a lengthy response that reiterated a lot of the transgender worldview, and been told about how harmful my “TERF” talking points were. Since I consider “TERF” to be hate speech, that was a red line for me. Keeping a cool head, I responded with a final message expressing my disappointment in the way the discussion went, said that I had to conclude that the Guix community is a hostile environment to people opposing sexism, and explicitly said this would be my last message on the subject.
A few days later I checked in to see what responses there were, and it was almost unanimous hostility. The transwoman I was talking to had responded by saying that apparently, I only wanted to hurt them; the founder of the Guix project had reprimanded me, for the disruptive and distressing discussion I apparently caused; and a moderator had removed me from the list of “trusted” members who could post freely to the mailing list without manual moderator approval, insinuating that I would troll the mailing list with incendiary posts, despite my clear statement in the message he was responding to that it was my last email on the subject. Frankly, everyone’s reaction seemed sheer irrational. There was only one person who later reached out to me in private to say that they sympathized, and found the decision by the project maintainers to be a grave injustice.
For anyone distrusting my framing of events, all the messages are publicly available in the online mailing list archives of the Guix project, February and March 2022. I’ve linked to some of my own messages above, but don’t want to link to any of the responses directly, and am not naming any names, as I don’t want to turn this into a personal feud, or cause unnecessary distress to any of the trans people in the Guix community who find it traumatizing to be faced with worldviews that oppose theirs. However, that is frankly a personal problem of theirs, and cannot be used as a justification to enshrine sexism in the community rules of the project.
Are Feminists Welcome in Free Software?
Although I’ve used the Guix community as a prime example of the problem due to this particularly terrible experience, I had noticed in prior years that similar problems likely exist in other projects under the GNU umbrella. In separate occasions, I had the chance to speak to the maintainers of GNU Octave as well as GNU Guile about these issues, and they seemed completely one-sided. (It should be mentioned that Guile and Guix are closely tied projects.) There’s a good chance that these projects too wouldn’t tolerate feminist community members. And then we have to consider the perhaps hundreds of software projects out there which also adopted Ehmke’s Contributor Covenant…
None of this, of course, is meant to discourage women from joining Free Software. To the contrary, I think the more women join, the more there will be an organic shift towards less sexism. The only reason these projects can claim to be inclusive and non-bigoted while being so hostile to feminist perspectives is that there are barely any women in their communities to call out the farce.
Indeed, one of the arguments that was repeatedly used against me by members of Guix was that I was just a man trying to dominate the discourse. It’s a cheap shot, but sure, why not look into what actual feminists (women) who are also part of the Free Software movement have to say on these topics?
Perhaps one of the most prominent feminist members of the movement is Mary Kate Fain. She appeared in Free Software conferences such as LibrePlanet, and part of her talk there also appeared in the 2019 bulletin of the Free Software Foundation. The intersection of her feminist and software engineering work includes co-founding the pro-feminist Twitter alternative Spinster, founding the women-only publishing platform 4W, and helping create the pro-feminist Reddit alternative Ovarit. She also has other feminist activism to her name that isn’t related to software.
In 2018, Fain was forced out of her house and kicked out of an animal rights group she had founded. Her crime? Supporting women’s sex-based rights. In her writings, she makes more or less the same category of arguments that I brought forth to Coraline Ehmke on GitHub, and to the maintainers of the Guix project on their mailing list. I have little doubt that if she were to try to join the Guix community, she would be hounded out for her feminist views and activism, likewise for many other software projects that adopted Ehmke’s Contributor Covenant.
Another woman who I spoke to directly on the matter, who’s a bit older and a software hobbyist, said that she would not even bother trying to contribute to a project which, as she put it, “prioritizes virtue-signaling of any kind over making quality software.” She actually happens to be a contributor to GNU Emacs and Nix, two pieces of software with very close cultural and technical ties to GNU Guix. (Emacs is the preferred software development tool of most Guix developers, and Nix is literally the project from which Guix was derived; the name of Guix is a homage to Nix.)
“But Those Aren’t Real Feminists!”
I can see my detractors now claiming that these are cherry-picked individuals, and that the vast majority of women would agree to the (in my view bastardized) variant of feminist politics supported by Ehmke and the Guix community. To judge the accuracy of this line of thinking, I would suggest two methods.
First, just take a look at the number of grassroots feminist and gay rights organizations that have popped up over the last years with the explicit goal of opposing misogynistic and homophobic aspects of an absolutist interpretation of the transgender worldview. These include Woman’s Place UK, LGB Alliance, Get The L Out UK, Women’s Declaration International, For Women Scotland, Speak Up For Women, Fair Play for Women, Feminists in Struggle, ReSisters United, Reduxx, and many smaller localized groups. Joining them in their views are long established feminist individuals and organizations such as Sheila Jeffreys, Germaine Greer, Alice Schwarzer / EMMA, Women’s Liberation Front (WoLF), Vancouver Rape Relief and Women’s Shelter, Meghan Murphy / Feminist Current, Julie Bindel, and many others. Given this enormous wave of both old and new feminist and lesbian activism that has the express purpose of making sure that the sex-based analysis of women’s oppression is not ignored, it would be foolish to believe that these views only correspond to a small and insignificant minority of feminists, who are safe to ostracize and exclude while still calling oneself tolerant and inclusive.
Secondly, we can look at some public opinion polls to see what the entire population of women, regardless of whether they see themselves as feminists or not, have to say on the matter. According to a YouGov survey, when asked in a simplistic way, a majority of British women are supportive of the transgender worldview on most questions, such as whether transwomen should be allowed to use women’s changing rooms or toilets. The only exception is the question of transwomen in women’s sports, where a plurality (relative majority) disagree that they should be allowed to compete against women. At first glance, this seems damning for my claim. What is very interesting though, is how strongly the numbers shift when it’s specified that the transgender person in question might not have undergone surgery. When this is specified, a plurality of women disagree that transwomen should be allowed to use women’s changing rooms and toilets. This is a very clear indication that to most British women, a person’s sex absolutely matters.
Similar statistics don’t exist for other countries as far as I’m aware, but we can also look at what lesbians have to say about their dating preferences, based on the 2018 study: Transgender exclusion from the world of dating. According to the statistics found in the study, less than 20% of lesbians would consider dating a transwoman. Even more interestingly, about an equal number of lesbian respondents say they would consider dating a transman. (Something inconceivable from the perspective of the transgender worldview, in which transmen are men, period.) Again, we see a clear indication that a person’s sex matters a lot to most lesbians. Gender identity, not nearly so much.
At this point, I think it’s very safe to say that I’m not exaggerating in my conclusion that the Contributor Covenant creates communities that are everything but inclusive, insofar women, and in particular feminists and lesbians are concerned. As such, I would urge the maintainers of software projects and all Free Software enthusiasts to think critically about what kind of environment they want to foster. One in which only a narrow, arguably male-centric version of feminism is tolerated, and dissidents are aggressively hounded out? Or one in which people of all wakes of life, all worldviews, and all identities are equally welcome on the condition that they treat others with respect?