Gender-equity language in the workplace

What is gender-inclusive language?

Two strands of gender inclusive language seem to be salient for my current environment:

1. “Career stuff”. Are roles gendered? Does language exclude people, usually women, from roles? e.g. Actor/actress. This is interesting in the context of the project of feminism.
2. “Gender identity stuff” — does the use of some particular word impinge upon a given individual’s preference/need to enact a certain gender role? This page is for the first one.

Context: I think the project of feminism is both important and incomplete, although I personally rank language interventions as low in the triage order. I acknowledge that I am personally going to have a hard time calibrating how important language interventions are based on my personal experience, because as a man I get to ignore the negative experiences that women have here. This is a fundamental niggle of our public sphere.

First thing to get out of the way: There is a role for policing speech in the workplace. Workplaces are not free speech zones (free speech is great but not on the clock). Many speech acts have no role in the workplace: bullying, aggression, hate speech etc. I will take it as read that doing our best to ensure psychological safety is necessary, right, and helpful in a workplace. This is not about those low-hanging, commonly-accepted fruit; This is about identifying the marginal return on investment on tricky ones. How rigorously do we wish to stamp out words which seem unnecessarily gendered? Should we try to enforce a finicky pronoun or job description?

Here are two interesting takes on linguistic engineering projects for gender neutrality:

Firstly, Daniel Scholten’s Der Führerin entgegen! which sets the potential inclusivity upsides of language engineering in the context of language history, language adaptation and so on. The author is a linguist who argues that discussion about gender-neutral language conflates grammatical, biological and social gender, and he argues that this conflation does not help the main feminist project of achieving better outcomes for women. See also Follow-on video. Even if I disagree with some of his conclusion, I think the dissection of grammatical gender as opp social or biological gender is important. These concepts clearly are in contact but also clearly not the same. TODO: raid that video’s citation list.

A counter-argument is the review of Sczesny, Formanowicz, and Moser (2016). They summarise the role of gender fair language in creating inviting environments and the result is more optimistic about the usefulness. Their position is generally “it can be worthwhile to change gender language if you take lots of local context into account”.

recent research has documented that linguistic asymmetries prevent girls and women from aspiring to male-dominated roles (see Chatard et al., 2005; Gaucher et al., 2011; Stout and Dasgupta, 2011; Vervecken et al., 2013; Vervecken and Hannover, 2015) and thereby perpetuate the higher accessibility of men in these roles. Second, the use of gender-unfair language, especially of masculine generics, restricts the visibility of women and the cognitive availability of female exemplars (Stahlberg et al., 2007), which may be disadvantageous for women (e.g., in personnel selection; Stout and Dasgupta, 2011; Horvath and Sczesny, 2015). However, increasing the visibility of women with the help of novel feminine forms may also have negative consequences and may therefore be avoided, for instance, in women’s professional self-reference (Merkel et al., 2012; Formanowicz et al., 2013). Thus, the avoidance of GFL by women (e.g., avoidance of feminine job titles in grammatical gender languages), in order to protect themselves from ascriptions of incompetence or lower status, also perpetuates the reduction of gender stereotyping and social discrimination.

If I were to be persuaded that gender fair language strongly caused gender fairness I would find it more plausible that something complicated and messy like this were real. Note, however, that not al these references are slam-dunks; many are survey results which is great at detecting what people do not like, but less good at detecting what actually helps. The studies for each case seem to suffer from the usual problems of psychological research, in that they are conducted under laboratory conditions or have difficult observational inference challenges, so the evidence is not slam-dunk.

Some of the arguments we see floating around leverage a version of the Sapir-Whorf hypothesis to conflate grammatical and social gender (roughly, what you can verbalise defines what you can think) which in my classes we were trained to dismiss as debunked. That said, the examples we had were about the Sapir-Whorf hypothesis in concrete and boring settings (Can you understand income tax if your language has no word for it?), rather than in subtle social contexts. There may be a better literature on this sub-field, but I have not done the review of it.

I am generally skeptical that using gender neutral language (a.k.a. gender fair language) is strongly influential as a point of intervention to improve gender equity, not in the sense that I think it makes no difference, but in the sense that the evidence for an effect is weak, which means that we should preferentially allocate our effort budget to other things if they have better evidence. The reason that I think that is that the literature I have read on this topic has generally done a bad job at causal attribution, and a bad job at self-reported survey analysis.

However, if I feel that this problem with the literature is systemic, I really should step up and attempt to do it better. I have not done that, so you should treat my arguments with the suspicion they deserve.

Indeed, without doing a comprehensive literature review, my sense is that

1. Sczesny et al could be correct that the effects of gender-unequal language is real in perpetuating the exclusion of women from male-dominated roles.
2. Scholten could be correct that the size of this effect of this is small next to the effect of concrete economic and administrative policies to support people in a non-traditional gender moving into a field from which they are historically excluded.

They both agree that having a language with no gender markers is not itself sufficient to lead to gender equality since they both raise the example of Turkish which is an essentially genderless language, speakers of which are not automatically transformed into gender-blind egalitarians.

The argument is of course not necessarily about empirical effectiveness of changing language to cause change society. You could instead agree to change your language for a different reason, for example as socially normalized token of good will as part of a larger and more substantial suite of reforms, without hoping that the language change achieves anything by itself. Or some other reason that I cannot think of.

In this case, the argument comes down to whether the policy of linguistic engineering is worth the opportunity cost. Prima facie, changing the words we use is cheap in that talk itself is cheap and the effort for any individual speech act is low. So maybe gender fair language is a small return at small cost.

In practice, imposing such a policy has a non-trivial cost in terms of political capital, because arguing about gender-inclusive language seems highly memetically contagious. Language is salient. We all use language so we all have opinions about it (me included, clearly) but fewer of us feel agency over other policies. In practice, language change seems to be easy to mobilise people against, via the tokenism/table stakes manoeuvre. Complicated prescriptions for words can, after all be expensive.

tl;dr I am personally happy to refer to people and their roles by whatever gender terms they see fit. This seems like the least I can do. OTOH I am not going to spontaneously advocate gender fair language as a policy in all circumstances, because experience leads me to believe it is wandering into a costly argument for unclear gain.

The fact that I have spent so long digesting evidence for this issue compared to real, substantive, causal and important policy debates (parental leave, diversity quotas, diversity performance incentives, accountability and protection mechanisms against sexual harassment…) suggests to me that language norms are a memetic hazard. They might be important, they might not, but if we spend a lot of time addressing them instead of the things we think are much more likely to be important, then what are we doing?