A few weeks ago, I posted The Computational Humanities and Toxic Masculinity? A (long) reflection   and didn’t know what reaction to expect. It would either be a shitstorm or get ignored, I assumed. It turns out, however, that the timing might have been just right, since I was contacted by the (informal) Turing Institute reading group – they wanted to discuss my blog post. I was honoured that they did and that I could participate. Here I wanted to share some new insights and reflection prompts this discussion has sparked for me.

First things first: Thanks for everyone who participated!

First of all, I was happy this discussion happened at all. That’s exactly the type of thing I had aimed for with the post, yet I was pretty surprised it actually happened. After all, you normally don’t always get you want. Maybe it’s that help will always be given at Hogwarts to those who ask for it. 😉 So anyway, I also wanted to thank all the participants for being open to this discussion. The group was made up of a wide variety of people – from early career researchers to more established staff, a good gender balance and also some members of the Computational Humanities group. I think there also was a diversity of opinions and approaches to the topic which is always important. If it had only been a group of feminists, it would have also been a good discussion but I think a very important part was to have different and especially male points of view too. Most importantly, I want to thank Katie McDonough (@khetiwe24) and Federico Nanni (@f_nanni), of course, who initiated this working group in the first place.

New insights I’ve had on the CH and feminism discussion

During the discussion, two points became more clear to me which I hadn’t fully grasped before.

Feminism vs. privilege men can’t see

First of all, I believe now that the reason men can often see feminism as aggressive is because you can’t see the privilege you have. Often men just don’t have a realistic picture of the sexist things that happen to womxn on a daily basis, so when the womxn start complaining, it can seem to the men like they are just “making a fuss” and that it would be unfair if they received “more rights” because they’re obviously already being treated really well, right? Not really but I now get the issue many men have with this a little better.

A special interest for a group who’s already in charge?

The other thing is: I was now able to pin down where my alienation with the CH group came from. It’s not with the idea of the making this group or the concept of CH in general, for sure. So I wondered why that was. One participant of the discussion pointed out that they are pro big tent DH, so they didn’t see the need for people segregating themselves away from the group. And that’s where it struck me: I think my problem with the group is that they said they felt “not accepted” by the normal DH community, basically claiming to need a “safe space” for themselves indirectly. Yet it is my feeling – and probably that of the other people who were somewhat alienated by the creation of the group (and I mean, of course I signed up for the newsletter too, so technically I’m also part of the group) – that they are quite over-represented at conferences already. They get all the attention. They get the grants. And not only the grants in general which favour CH but there also are many grants specifically targeted towards CH where DH-only people can’t even apply anymore, cutting funding for “normal” DH projects. It is easier to get accepted into an event with a CH topic than with a DH one. One of the studies cited in my last post stated that womxn had the same changes of getting into a conference but only if they chose the same topic as the men. CH clearly is a bit of a male-dominated topic because the computer science stereotype is still so clearly male all over the world. So of course it’s also easier for a man to become accepted as a valid CH person than it is for a womxn. Meaning that even if a womxn decided they needed to jump onto the CH bandwagon to survive in the field, they might not even be accepted there.

To wrap it up with the CH group: I think many people just find it alienating that a group which is clearly already dominating the field and has many more opportunities want to start an interest group. It’s like rich people wanting to start a rich people’s group to feel safe. It seems a bit ridiculous and elitist. Special interest groups are for ‘endangered’ minorities, not for the dominating 10%. This is something the group should take into account, I guess. Of course, it seems that the creation of the group was done with the purpose of discussing CH topics with likeminded peole in mind but people in the CH group should also keep in mind the power dynamics implied. Because power dynamics will automatically be implied in this topic, whether you want it or not, so they need to be addressed.

Our discussion moderator Fede also brought up  the realistic struggle of CH people at non-DH conferences: DH applications of NLP aren’t all that interesting at NLP conferences because they’re just applications of NLP, not cutting-edge new methodologies for NLP. However, these applications have their own difficulties which fit well at DH conferences. I get that. It make sense. The problem might just be alleviated if DH conference organizers were more aware of their own bias towards CH applications. Having people label their abstracts in categories (one being CH) would help organizers to make a fairer cut with regards to whom to accept and whom to reject – because reviewers will for sure be biased strongly towards CH project. If each category received a more or less equal amount of spots, conferences would have more equality with regards to the hack-yack bias.

Being tolerated is not enough: Womxn need to be actively included and encouraged

Here are a few ideas how the group could take action in empowering its female members (and other womxn in the DH):

1. offering trainings geared specifically at teaching womxn DH researchers the CH;
2. promoting female CH work specifically;
3. assuring womxn are well-represented at CH events;
4. assuring there are no man-thologies and no man-els (i.e. male-only anthologies and male-only panels – the responsibility of this not happening lies strongly with men as womxn not even being invited into the discussion can hardly influence it);
5. that the same amount of womxn teach CH as men (because teachers are role models and if there are no female role models, womxn won’t feel as welcome);
6. maybe even organizing a female interest group among the CH group where womxn have a safe space to discuss gender-related issues and have an all-female support network (which has been shown to be extremely effective in promoting womxn).
7. keeping statistics, and thus raising awareness (!) of male overrepresentation at conferences and other things – after all, statistics is what a CH group can be good for, right?

In the end, it’s about showing that the stereotype of a CH person is not the same as the stereotype of a programmer or computer scientist (which is one we can’t control and which is and will likely remain largely male-dominated for the years to come). The CH is new, it has yet to claim its space. Let’s make it inclusive! (This means we can take the same steps to actively include other minority groups as well!)

CH power dynamics in hiring situations

First of all, the DH and especially the CH will need gender policies, obviously. But gender policies exist in many places and they alone don’t necessarily make much of a difference.

An important aspect of male-biased power dynamics in hiring situations is, in my opinion, that many people who hire a DH (or CH) person are not necessarily in the DH themselves or don’t understand exactly how DH jobs work or what skills a DH person really needs. This is very common in applications which sound like someone isn’t hiring one person but rather a whole IT department (see quote below). Even if that job is a 3 year contract, oftentimes I absolutely cannot imagine how one single project is ever going to need all those skills from the job requirements.

Dear recruiters,
if you are looking for:
– Java, Python, PHP
– React, Angular
– PostgreSQL, Redis, MongoDB
– *nix system administration
– Git and CI with TDD
– Docker, Kubernetes
That’s not a Full Stack Developer.
That’s an entire IT department.

(Giulio Carrara)

But job ads like those are bound to scare womxn off who are told repeatedly throughout their lives – subconsciously or openly – that they are just less good at IT stuff than men. Asking for mastery of 3 programming languages and 7 technologies and, if possible, a Master’s degree or better PhD in Computer Science is definitely going to scare most DH womxn (who don’t have  a Computer Science degree and also, don’t need one) from applying.

If they do, the computer skills they claim to have are going to be questioned much more than in a man because men are hired on potential, womxn on previous achievements. But giving people challenging jobs is what helps them with learning new technologies. The math is simple here: Men are trusted to learn a new technology on the job, get the job and in the next application process, they’ll have experience in one more technology gained on the previous job. Womxn will hardly ever get hired to work a skill they don’t have any experience with beforehand. So unless they  spend extensive amounts of their free time polishing their Github with programming projects, womxn will end in a vicious cycle of not having experience in a technology because they didn’t get hired and then not getting the jobs because they don’t have the previous experience. Which will keep them nicely tucked away in what has been called the “feminized corners of the DH” forever. If we want to break this vicious cycle, we need to come up with something. Don’t underestimate this cycle. It plays out silently in the background but if you watch closely, you’ll find examples of it in your own surroundings, I’m sure of that. Please don’t look away!

Actions to take immediately: Let’s create guidelines for job ads and the hiring process

I think an important first step towards a fairer hiring process would be to come up with guidelines for hiring which explain in lay(wo)men’s  terms what different technologies in the DH are, what technologies are needed in which projects (and which ones are nice to have but overkill and not realistic that one person has them all at once), in which jobs it makes sense to ask for a Computer Science degree (and in which ones you’re totally scaring away DH womxn with that requirement), etc. That way, even someone  who is not in the DH themselves can tell if their requirements are reasonable.

Also, with regards to the Computational Humanities, we should offer guidelines on how you can tell whether a person has the programming skills required for a job (for example, [random number of] lines of code on Github in the language concerned; or has worked in a project where the technology was used, etc.). We should have guidelines on how to tell how to “weigh” a certificate. DH summer schools a great but many are just for complete beginners and having completed them does not necessarily point to lots of DH skills. However, there are some DH certificates which count a lot inside of our DH community. But people outside of this community will likely not be aware of that, they might not even know whether an institution you have worked for is a “prestigious institution” in terms of DH. Many excellent universities have DH programs but that doesn’t mean that their DH programme specifically is outstanding. On the other hand, smaller universities have really influential DH programmes with a lot of history in the DH but people outside of the DH have no way to know.

What I mean with all of this is that essentially, we need to make the lines clearer between what’s what. What is DH, what counts as ‘label DH’, what is CH (see P. Sahle’s article in Zfdg on the topic if you read German). And that doesn’t mean I am argueing for elitism – but we need some objective criteria or we will be ruled by bias. And as we know from studies, sadly womxn will be on the losing end of bias. “Techie, “middle ground” and “label” DH all have their benefits. Depending on the type of project, you might need a “techie” person, while in other projects a “label DH” person is more fitting. If we had better ways of labeling people (and believe me, I am not a fan of labeling people in general), this could mean fairer hiring processes. Because at the moment, many employers think “the more techie, the better” but this is not generally the case. Yes, there are projects where you need “techie” but in most, you probably don’t, so it’s quite unfair that the hiring process would be biased towards Computational Humanities rather than “normal” DH people.

Realisitic requirements in job ads are possibly more important with womxn than they are for men because it has been shown in studies that men apply for a job even when they only meet half of the requirements (maybe even less than that) while womxn probably won’t. Womxn tend to feel like imposters when they do and are unlikely to apply if there are too many “nice-to-have” technical skills included which aren’t actually needed at all to do the job well, resulting in the womxn not meeting at least 80% of the requirements, making it less likely for her to apply.

I think DH people in general would profit from such guidelines, not only womxn.  

Well, this post got unreasonably long again and I still would find more things to say on the topic. I’ll cut myself off for now and am looking forward to more discussions on the topic.

Best,

the Ninja