Title: Language, Gender, and Harm

What I almost changed the title to: How Not To Do Sex Wrong

Kirby Conrod
26 min readApr 1, 2019


I am incredibly grateful to Nic Subtirelu who invited me to participate in a special panel at the Georgetown University Round Table this year on Diversifying Linguistics. I was in incredible company on the panel — an all-female panel besides me, by the way — with Minnie Quarty and Stephany Dunstan speaking on how to diversify the student body in linguistics, and Nicole Holliday, Julie Hochgesang, and Gloria Park speaking on how to diversify the faculty. Each of the panelists brought a unique perspective on the field both in terms of their research and their position within the academy; Minnie and I are both currently ABD PhD students, and we got to join junior and senior faculty in sharing our perspectives. I leave it to my fellow panelists to tell you about their talks, but I wanted to informally write up my talk and some thoughts. I did this after my talk at LSA, and I found it to be a very useful way for debriefing and putting my thoughts together after a very busy and exciting conference.

My talk is called Language, Gender, and Harm, but as you see at the top I very nearly changed the title to something much less official-sounding. This talk was really a way for me to organize and share my reactions to a recent uptick in research in linguistics looking at transgender people. I have seen many excellent talks, most of them from students (undergrad and grad) and some from student-faculty teams. I will here point out that the majority of talks I have in mind were by trans students, and none of the advising faculty are trans (that I’m aware of), which in itself is not necessarily a problem. If there is a problem, it’s in the huge asymmetry between transgender graduate students and PhD holders who have left academia on the one side, and the absolute lack of (openly) transgender faculty. Last we checked (at the LSA panel celebrating the creation of the SIG on LGBTQ+ Linguistics) Lal Zimman is the only one in North America. That asymmetry is unfortunate for a few reasons — for one thing, anecdotally I can report that trans people who left academia after earning their PhDs did so due to economic disenfranchisement and precarity (a problem all recent and near-future PhDs face, but the numbers are not on our side here to begin with). The other problem is that cisgender faculty can be extremely well-meaning and talented linguists, but in order to advise a project on trans linguistics one needs to be much further ahead of Linguistics As A Field’s collective understanding of sex and gender. I am lucky in that my own PhD advisors and committee members have been extremely supportive and have also supported me in getting extradepartmental education to round out where my gender theory and such was underdeveloped, but I have had to teach myself a lot of stuff outside of the world of Linguistics As A Field, which means that what I’m bringing to the table is not common ground when I’m going to linguistics conferences or chatting with cis linguists.

Our field suffers from a vast divide between the LavLang crowd and the LSA crowd — the former doesn’t need to talk about this stuff because they already know it and are doing other, more fun stuff, while the latter doesn’t sit through the LGBTQ sessions because they’re at the same time as a syntax session. (What are we trans syntacticians supposed to do? Time travel. Clone ourselves. Give up the gender stuff so we can go talk about relative clauses again. Give up the relative clauses because we have to talk about gender forever. The choice doesn’t feel fair, and it’s one that I know everyone on the Diversifying Linguistics panel this weekend has felt.)

At the LSA meeting in New York this January, I invoked the difficulty of reading Butler as a parallel to Chomsky’s writing — a joke that definitely got a bigger laugh than I had feared it would. But the fact is that engagement with the core of Butler’s premise of performativity is extremely elementary by even the most advanced Linguists Doing Gender Stuff. The most cutting-edge stuff that I can find in ‘mainstream’ linguistics is work like Penny Eckert’s, an absolute champion, acknowledging that gender is a complicated and not-binary social system and that sex is a categorization of bodies. She mentioned trans people at her LSA plenary, that’s a big deal! She and her cohort got the right idea, but my talk at GURT this weekend was about asking linguists to push just a tiny bit past that cutting edge, with the hope that we can catch up to our sister fields. So this is a long prelude, but here’s what I talked about.


I began my talk with the Rickford and King quote that we all really like this year:

“Language lives in a society, and so must we.”

(Rickford and King 2016)

I don’t remember whether John Rickford snuck out of the back of the room before or after he saw this slide, and I’m absolutely not going to ask. No one send him the link to this post. On this slide I also had the text: How you design your science has the power to help or harm. The theme of GURT 2019 was Linguistics and the Public Good, and my intention in invoking Prof Rickford was to set the stage for making clear that I am setting out methods that are intended to be more ethical and less harmful, both to our study populations and to people who are vulnerable to prejudice and violence as it is.

As linguists, we have a good idea that gender and sex are important to humans, and that things that are important to humans tend to be reflected in human language. I proposed that if we believe that:

· Gender and sex are a site of violence for many people

· Gender and sex can be different in single individuals

· Gender and sex are separately but collaboratively constructed through social mechanisms including speech acts / language

· And the language of an individual or group holds social significance both in what we say and how we say it,

Then the community of linguists has a responsibility to act in a way that minimizes harm in several ways:

· Linguists must stop conflating sex and gender (I mentioned that we are doing better on this one)

· Linguists must actively intervene in situations where language is used as an instrument of harm on the basis of gender and/or sex (which includes sexist language, transphobic language, harmful linguistic acts such as misgendering)

· And linguists must not discuss, analyze, or summarize gender and/or sex in a way that supports, enables, or allows harm

This is a tall order. I will first give some background on the concepts, and then outline explicit guidelines for attending to these responsibilities.

Sociolinguists are ahead of the pack on knowing about sex and gender — we have at least thirty years of established A Linguist Has Acknowledged This foundational knowledge:

The term “sex” has often been used to refer to the physiological distinction between males and females, with “gender” referring to the social and cultural elaboration of the sex difference — a process that restricts our social roles, opportunities, and expectations. Since the process begins at birth, it could be argued that “gender” is the more appropriate term to use for the category than “sex.“

(Cheshire 2002)

“We have been examining the interaction between gender and variation by correlating variables with sex rather than gender differences. This has been done because although an individual’s gender-related place in society is a multidimensional complex that can only be characterized through careful analysis, his or her sex is generally a readily observable binary variable…”

(Eckert 1989)

Cheshire and Eckert both summarize the issue quite similarly: gender is the thing that humans are doing as a social activity, an identity, a practice, a relationship; gender is the one that has lots of cool intersections with other identities and communities like socioeconomic class, race, national identity, age, etc. Sex, on the other hand, is according to these quotes just a simple categorization method for bodies. These quotes say there are two defined sexes, and they are physical and unified and binary and obvious: “a readily observed binary variable.”

I am absolutely on board with Eckert and Cheshire here on the topic of gender. I define gender as social groups which are very roughly correlated to commonalities in identity, filtered into legible categories (as both Eckert 1989 and Cheshire 2002 would say). I conceptualize gender as having (at least) two elements: gender expression (e-gender, or ‘external’ gender, if you’d like a Chomskyan metaphor I copped from Brooke Larson) and gender identity (i-gender, ‘internal’ gender). Gender identity is the way a person conceptualizes themselves in respect to the social groups they know about, while gender expression is a variety of behaviors that a person uses to show or communicate their relationship with gender (such as clothes, haircut, makeup, but also to some extent hobbies, friend groups, dating practices, etc.) I describe gender as a spectrum, not from black to white, but a spectrum in the way that visible light yields a rainbow of colors.

I also point out that gender is at least partially made out of speech acts and language more generally (cf Tannen 1990 and Butler 1990, wonderful contemporary texts. I had a very short but fun conversation with Prof Tannen after this talk.) People will use language to convey both their gender identity (inasmuch as language can communicate any internal state, but I’m not citing Chomsky or Saussure here, let me live), but they will also use language to act out their gender expression through sociolinguistic qualities such as pitch or intonation, vocal quality, discourse practices, pronoun use, and lots and LOTS of other stuff.

I stray from the linguistic literature now, to talk instead about sex. Rather than a “readily observable binary variable,” I give a much more critical definition of sex — one that is pretty established in other social sciences, and which is getting discussed in more popular literature, by the way. (Check out these Scientific American, Slate, NY Times Opinion, or Nature articles.) I will maintain the idea that gender is a mental thing and sex is a body thing, but I am here to report that the body thing is neither binary nor readily observable. Sex is the social lumping into categories of traits that sometimes cluster together, even though they are not all causally linked or even necessarily correlated. The social categories of the sexes are categories of types of bodies and lives (a vast variety of human experience). The many traits that constitute sex include chromosomal makeup, hormone levels, genital configuration at whatever life-stage, body morphology, and assignment at birth. Crucially, the categories of sex are also constructed by means of speech acts (Butler again, 1990 and lots elsewhere in her work).

HOW are bodies made out of speech acts, you ask? Don’t worry, that’s not what I mean. Instead I am pinpointing things like legal sex as very much the absolute foundation for Butlerian performativity: when you are born, a special type of wizard casts a magic spell by looking at your purple, naked infant body and then saying some special words, such as It’s A Girl.

This is an arbitrary process, and the sex-assigning wizards are fallible. Not only do doctors routinely over-interpret baby genitals so as to force intersex babies into a category of their (the doctor’s I mean) liking, but doctors absolutely will just go ahead and mess up the magic spell for declaring the sex of non-intersex babies. A tumblr post gives us an absolutely fascinating example:

[Image description: a tumblr message sent by Anonymous says, “My gf is an actual amab cis girl. They wrote male on her birth certificate by mistake.” A user responds: “holy shit tell your girlfriend congrats on the fun gender.” Another user adds, “kinda just speedran the entire transition there”]

To unpack this if you’re not familiar: “amab” is an acronym for Assigned Male At Birth. “Cis” means “someone whose sex (at birth, usually) is in alignment with their gender identity” or simply “not transgender.” And “girl” means, well, whatever you think it means. If you’re cis you possibly don’t think that “girl” means something compatible with being “assigned male at birth,” which is why a user jokes that the tumblr gf “speedran the entire transition” — by getting assigned male but already having a body that people would theoretically interpret as female, she is in some ways doing what any trans woman is doing, just faster and easier because the game is broken.

So what does this have to do with me, a linguist, you ask me. Thanks for asking! I have an aside with a short plea that I elaborate often elsewhere, and then I have a ton of variable control stuff to tell you about.

For pedagogy, especially when teaching introductory classes: you need to stop saying wrong things about pronouns in your classes. What this means is that you absolutely cannot keep saying that English third person pronouns are based upon “sex,” because as I have hopefully just shown you, that cannot possibly be true. Speakers are not investigating their referent’s birth certificates or genital situation or hormone levels before using a pronoun. Speakers are attempting to guess at a gender, with varying success. I have lots of other work I can send you on this, and I promise I will publish it soon.

You also need to be careful about how you discuss grammatical vs “natural” gender (which we really should be calling social gender, please). Grammatical gender is a confusingly-named system of noun classes that is sometimes semantically related to social gender but has a lot of arbitrariness and irregularity (as all good linguistic features do). And in more formal classes you need to be cautious about attempting to reduce “natural”/social gender to a morphosyntactic feature on a determiner unless you have read my dissertation and cited it extensively (or are citing that Sigurthsson paper where he totally, totally beat me to it. Sort of.) Features don’t just come from the aether, and it’s misleading even in very formal/theoretical contexts to pretend that they do — you should treat it much more like you treat the difference between tu and Usted, where the pronouns (and agreement with them!) are sensitive to social relationships and formal features really need to acknowledge that in some way.

For sociolinguistic research (pretty much all of it): this is the rest of the talk. Two main points:

1. If you say men and women do things at different rates but you didn’t include any trans people in your study, you have no way of knowing how much sex and gender play a role.

2. If you did include trans people in your study, you still need to build an adequate theoretical frame around your conclusions.

I will run through the possible results that you got. This is assuming that you included trans people in your study along with cis people, and you compared trans people of each binary gender to cis people of each binary gender. That’s a great way to try and pick apart the confounded variable! It means you’ve got trans men and cis men (so you can compare men who have different sex assignments) and trans women and cis women (ditto). If you don’t have all four, you can’t exactly control for both sex and gender. (If you have a bunch of nonbinary people, WOW CONGRATS, email me.)

So here’s what you might be doing with this:

I didn’t hypothesize that there would be any correlation with either sex or gender: weird, you went to a lot of trouble, then! Well, skip to the end where I talk about what you should be doing if sex and gender aren’t really your main area of study.

I hypothesized that both sex and gender would correlate with my variable: that’s going to be tricky, and you need to figure out whether one is more important than the other. To figure out how to pick that apart, you can read the rest of the possibilities and triangulate from there.

I hypothesize that gender (and not sex) is going to correlate with my variable: Okay, this is pretty common in sociolinguistics, because gender is SOCIAL and we think people are doing SOCIAL stuff, which gender would explain nicely (as communicating identity, or as part of gender expression). Here’s what you should expect/look for, then:

· Speakers have some level of control (conscious or not) over the variable, or you can manipulate it by changing the social context.

· The variable is at least somewhat related to (symbolically, indexically, whatever level of abstraction) speakers’ own conceptions of gender identity and expression*

· Transgender speakers pattern more like cisgender speakers of their same gender identity (so, for example, you would expect that trans and cis women pattern together)*

· If transgender speakers change the behavior over the course of their transition, it’s more part of their social transition than their biomedical transition*

The asterisks are stuff I’m coming back to. There’s more of them coming. The last logical option:

I have reason to think that sex (and not gender) is correlated with my variable*: Huh, okay, this means that you expect that the aspect of language you’re looking at has something to do with embodiment or body arrangement. This is absolutely something that makes sense when you’re looking at phonetic variables like /s/-fronting, pitch (range), MAYBE some other stuff to do with physical aspects of the language modality you’re looking at (so the voice in spoken languages — I’ve NEVER seen this about signed languages but if you know of something out there, email me. What would it even be? Signer height predicts the size of their signing space? I’m extremely ignorant about this and I want to know more.) Anyways, if you’re looking for variables patterning with sex and not gender, here’s what you should expect/look for:

· Speakers cannot control this variable even if they’re trying really hard, and it does not change when the social context changes*

· There does not appear to be a clear pattern with any measures of speakers’ ideas about gender

· Transgender speakers pattern more like cisgender speakers of the OPPOSITE gender identity* (so you would find trans men and cis women patterning together)*

· Transgender speakers might change over the course of biomedical transition*

All those asterisks are serious points where you must exercise caution, both in your thinking (and research design) and in your analysis and research reporting. I’ll go through them in increasing severity:

“The variable is at least somewhat related to speakers’ conceptions of gender identity and expression”

👉That means you probably need to explicitly try and measure or ask about your participants’ gender thoughts and ideas. You can use survey instruments (usually from social science) to probe this, and you can use interview or open-ended questions to probe this. If you think this might happen, then gathering data on it will save you a lot of time and confusion later on.

“If trans speakers change their behavior of the course of transition, it is related to [either social OR biomedical transition]”

👉You should ask your participants about their experiences with transition. I do not presently know of survey instruments that target this, so you may have to develop one. You should probably start with open-ended questions and interview first. Do some ethnography. Ask a trans linguist to look over your materials before you do this so we can tell you if your questions are appropriate.

“Speakers can’t control this variable even if they’re trying really hard / the variable doesn’t change when the social context changes”

👉It would be easier to know this if you actively control or test for that in the first place. Nice traditional sociolinguistic interviews absolutely do this, and you can also get creative with manipulating the social context depending on what you think your variable ‘means’ (or doesn’t mean)

“I have reason to think that sex (and not gender) is correlated with my variable”

👉Notice that this is the fundamental premise of a research question. You should not formulate a hypothesis around this without clarifying what ASPECT of “sex” might be relevant. Is it hormone levels? Body morphology? It better not be genital configuration, because as far as I know genitals and vocal / signing apparatus are not related to each other.

“Transgender speakers pattern more like cisgender speakers of the OPPOSITE gender identity”

👉You really need to have an idea of WHY that might be before you present those results. If you don’t have an idea, it means you failed to complete the step above, so go back to the drawing board. If you collected data without knowing what you were looking for, that’s not hypothesis-testing, and you can’t report conclusions based on it without explicit backup.

“(trans men and cis women pattern together)”

👉This parenthetical needed its own bullet point. If you present these results in a way that misgenders your subjects, then your subjects can and should rescind permission to use their data. You should not refer to trans men as “females” or trans women as “males” to academic or general audiences. This directly feeds in to rhetoric which fuels profound violence against trans people, and it is extremely unethical for you to further that and in doing so further marginalize the research community who was so generous in sharing their time with you. If you have done this in a past paper you should publish a redaction and explanation. If you have done this in a past paper and don’t publish a redaction and explanation I am never ever going to cite you, I am going to advise my colleagues and students not to cite you or work with you, and I am going to snub you at conferences. For REAL, this is not okay.

Okay, great. Now, to address the actual picking-apart of confounded variables that we conflate when we talk about sex. What does sex mean in sociolinguistics, anyways? The foundational premise of sociolinguistics is that there are parts of language that are inherently social and interactional, and are not ‘hard-wired’ into us at birth. So how on earth does sex fit into that?

In order to answer this question, I am going to ask you to do a mental exercise of un-categorizing bodies. Really think about the parts of bodies and embodiment that you’re interested in. Here’s how we can do that:

I think my variable might be related to sex and not gender.

👉Okay. What aspect of sex?

Well, the literature says that men and women have different vocal tract sizes.

👉Instead of relying upon the categories of bodies, why don’t you measure vocal tract size? There are more and less direct ways to do this — some body morphology measures correlate with vocal tract size, so you could ask participants their weight/height, the neckband size they wear of collared shirts? You could probably directly measure the vocal tract, but I’m not a phonetician and I don’t know how this works. I have many amazing colleagues I can direct you to, but seriously measure it instead of assuming that sex will tell you about it — sex is an abstraction away from what your actual question is, anyways.

👉Well, I need to use those categories to be able to normalize huge amounts of data so I can do Praat stuff to it. My answer to this is that either you can choose an arbitrary cutoff point (for f0, for example) and group speakers based on where they fall on either side of that cutoff point. If all you need is a way to tell a computer to group stuff so you can make your vowels fit on the same chart, then ‘sex’ is not actually part of your research question and it’s misleading to include it. If it is part of your research question, then you need to go back up to the top of this section and figure out why.

Well, I know that men and women have different hormone levels, and that affects the voice.

👉Option 1: do blood tests and find out their hormone levels. This is probably hard to get permission to do, and expensive, but maybe you can get a big grant. I have recommendations for who you should hire as a post-doc for that.

👉Option 2: use populations with known differences in hormone levels to control against the cis and trans speakers — for example, humans go through hormone changes post-menopause, or at puberty, or these kinds of things. Now you’ve got problems with conflation with age, instead, but that’s a whole other barrel of fish where social and bodily categories get conflated.

👉Option 3: transgender people undergoing hormone replacement therapy can help you control for this, but only in one direction. People who use testosterone as part of their transition have changes in their voice. Some stuff changes and some stuff doesn’t, isn’t that fascinating? I refer you to Lal Zimman’s work (start with 2012)

Well, men and women are socialized differently in childhood.

👉Oh boy, that’s a complicated topic! Real quick: look inside your heart and ask yourself if the reason you think this is because you think trans women are secretly men. If you do, you’re doing it wrong. I’m not helping you with that. Because really, this is in fact a question about gender and not sex. (It would only pattern with sex because of the arbitrary and coercive sex-assignment at birth process, which inexplicably causes adults to do a lot of gender to babies.)

👉So, okay there are legitimate, non-transphobic reasons you might think this. The critical period hypothesis totally applies to sociolinguistic behavior so that’s something you might want to test! A lot of trans people don’t know they’re trans until adulthood, so it may be that they established their linguistic practices as children and cannot change them as adults. That’s totally possible, and actually something I REALLY want to see research on. There’s an expensive way and a cheap(er) way to probe this.

👉 Expensive way: do longitudinal studies on transgender youth. IRB is going to give you hell, ethics are very difficult, takes a long time, again you probably need a huge grant for this one. Hire me as your post-doc and I will make it happen. But, expensive (both in resources and labor)

👉Cheap(er) way: consider asking your participants about their childhood. WOW, a lot cheaper! Just add some questions at the end of the interview. Two very important angles, and you must cover both angles if you’re going to look at this. (It comes back to e-gender and i-gender — children have to develop those, you know!)

☝️ Aspect 1: What kind of gender identity and expression did the participant have as a child? You can ask them who their childhood friends were, what their favorite games were, what kind of traits they valued in themselves as a child, etc.

☝️ Aspect 2: What kind of gender socialization was being imposed on them in childhood? Adults impose incredibly global gender on children who would not otherwise have strong feelings one way or another. You need to ask how gender non-conformity was policed by adults — did they get in trouble for playing with the wrong toys, wearing the wrong clothes? You need to ask them if they knew any LGBT+ people at all when they were a child, and you need to ask what opinions adults expressed in front of them. You need to ask them whether they saw their peers bullied for gender transgressions, from early childhood all the way up through the rest of their lives, really. If a child knew only one trans adult or gender-nonconforming peer, who was derided and oppressed by the whole community, then the child would internalize a severe self-policing regimen to avoid social censure and isolation. *You also need to be sensitive about how you approach these topics, because you are opening a can of PTSD worms for cis and trans adults. Gender is a site of trauma for many if not all children, and you should not be leaving participants with psychological damage from the interview.

Heavy stuff. I expected more pushback from the audience about some of this stuff, but since the individual talks in the panel were not each followed by a Q&A I think I got a reprieve. I followed this up with sort of a jokey slide, aimed (in my mind) at petulant students who resent being held to high standards of intellectual precision because it’s hard and doesn’t align with their existing prejudices. I said this:

I don’t know, I just asked them their assigned sex at birth and then there was a correlation.

👉 This tells you nothing meaningful and you cannot report these results without more investigation. What you’ve done is called a pilot study. Try publishing it as a squib! Squibs are good.

I don’t know, Kirby, I think you’re trying to be too politically correct. If there was a correlation, then there was a correlation. I’m a scientist, I know what I’m doing.

👉 I look forward to your upcoming paper on how easily-misinterpreted baby genitals have an impact on the speech habits of adults. That sounds absolutely fascinating. Perhaps you’ve discovered an important larynx-pelvis connection we didn’t know about.

👉 You are going to have a tough time dealing with cases like the Tumblr AMAB Cis Girl GF, though. Or like, any intersex people.

I mostly stopped being snarky after this slide. Thank you for laughing to ease the tension.

The next question that I promised I would eventually get to is, well, not everyone is specifically trying to study gender and sex stuff on purpose, but we still collect demographic information on it. If my study is really about something else (local identity, ethnicity, level of education, social network, etc etc) how do you still do responsible work around sex/gender?

Don’t make statements that you don’t have the data to support. This means that you can collect the data various ways. If there’s no reason to suspect that a variable (e.g. double modals? Use of like as a discourse marker?) has ANYTHING to do with sex, then do not collect information about sex (assignment or any other aspect). This will prevent you from having a weird spurious correlation and also will prevent you from making your subjects uncomfortable for no good reason. You should still use best practices when collecting gender information (Lal Zimman has slides on that; briefly, at LEAST give three options, don’t use the words “male” and “female,” don’t call the third option “other” but maybe “neither” is fine — blank write-in question is always the safest, though)

Don’t speculate if you do find a correlation with either sex or gender. If you only collected gender information, then a follow-up study may be warranted. If you did collect sex (and were vague about it) then you can report your findings in an honest way about this:

“There was a correlation between frequency of raspberry-blowing and gender (self-reported fill-in-the-blank, analyzed as 2-factor) (p<0.005). Because this study did not probe gender and sex closely, future work should target this correlation to discover explanatory factors.”

Be a good citizen of the scholarly community. This means to pay attention that you are not naively reading or citing work that has transphobic foundational assumptions. It also means that you are critical and careful about how you read and cite and assign works to your students — discuss the conflation and shortcomings of existing studies and think creatively about how future work may fill in those knowledge gaps for us. Uphold high standards for what you are willing to sign onto as a coauthor, and be vocal about these standards when you are called upon to review works. Any time a transphobic publication gets through, it means that the author AND all their coauthors AND the journal’s editors AND the peer reviewers AND anyone who sat through the conference talks failed to sufficiently put a stop to it. It is a poor reflection on our entire academic community when we allow scholarly work to circulate that may uphold transphobic rhetoric either due to bad faith or sloppy thinking. We are not letting each other get away with that stuff anymore.

Okay, that’s mostly it. So what do I want you to take away from this talk?

What do you want us to do about it?

👉Design your studies better.

Be critical and careful about controlling your variables and doing so in a way that is ethical and supportive of your theoretical claims. Know why you are doing what you’re doing.

👉 Report your research better.

Don’t report correlations if you don’t have ANY way of explaining it. If you didn’t collect data on childhood socialization or hormone exposure or vocal tract length, then don’t blame your correlations on those things. When you do report correlations, do so in a way that is scaffolded in (EXPLANATORY) theory.

👉 Don’t throw out all the foundational literature

…but in reading/citing/assigning work that doesn’t control this variable, think and discuss that gap critically and encourage your students and colleagues to do the same.

👉 Large-scale replications

…are the best way of validating existing (and important!) works to pick apart the conflated variables. I absolutely want to see a wave of work revisiting questions where differentiating sex variables from gender will add to the collective knowledge of our field.

And this got tweeted a lot so I’ll put it in its own paragraph — one extremely strong moral of this talk, which is admittedly as much an ethics talk as it is a methods talk:

If your research is potentially useful to anti-trans bad faith actors, you either designed it wrong or communicated it wrong.

As scholars we cannot anticipate every bad misreading of our work, but if our work is rigorous and careful then it should be readily apparent to any reasonable reader that the misreading is unwarranted and unsupported. If your work is making the rounds on TERF blogs in a favorable way, then you have at best done some lazy research and at worst seriously exploited and harmed your research population. As a scholarly community we must show no tolerance for that harm, and work together to eradicate the laziness that may engender it.

Works cited

Auster, C. J., & Ohm, S. C. (2000). Masculinity and femininity in contemporary American society: A reevaluation using the Bem Sex-Role Inventory. Sex roles, 43(7), 499–528.

Chambers, J. K. (1992). Linguistic correlates of gender and sex. English World-Wide, 13(2), 173–218.

Cheshire, J. (2002). 17 Sex and Gender in Variationist Research. The handbook of language variation and change, 423.

Eckert, P. (1989). The whole woman: Sex and gender differences in variation. Language variation and change, 1(3), 245–267.

Eckert, P. (2014). The Problem with Binaries: Coding for Gender and Sexuality. Language and Linguistics Compass, 8(11), 529–535.

Holt, C. L., & Ellis, J. B. (1998). Assessing the current validity of the Bem Sex-Role Inventory. Sex roles, 39(11–12), 929–941.

Labov, W. (1990). The intersection of sex and social class in the course of linguistic change. Language variation and change, 2(2), 205–254.

Pedhazur, E. J., & Tetenbaum, T. J. (1979). Bem Sex Role Inventory: A theoretical and methodological critique. Journal of Personality and Social psychology, 37(6), 996.

Zimman, L. (2012). Voices in Transition: Testosterone, Transmasculinity, and the Gendered Voice among Female-to-Male Transgender People.

Postscript: I had several amazing conversations with fellow GURT-goers following the Diversifying Linguistics panel, and I am grateful to those who came up to me to discuss my talk at a later time. I was extremely touched by some who came to tell me how relieved they were to see this work at a higher level — a lot of trans linguistics conference talks that trans linguists give are obligatorily stuck in a loop of giving the 101 level over and over again, and the people who are living a complicated life in the gender world feel like it is difficult to gain any ground when we have to repeat the basics. I am deeply empathetic with this position, grateful for your conversation and solidarity, and give you this good meme:

[Image description: a meme. On the left is a picture of philosophers debating, labeled “Talking about gender with trans people.” On the right is a a picture of an adult showing a baby a toy, labeled “Talking about gender with cis people.”]

I also had a couple extremely gratifying conversations from others who reported that this talk was a real paradigm shift for them — that I had exposed them to thinking that was entirely new to them, and that challenged a lot of their previous thought and assumptions. This is amazing and I am so grateful to those who were willing to step up and meet me at this level. Almost none of this model of sex or gender is my original proposal, and most of what I’ve been doing this year is trying to get linguists to engage with more complex theories of gender and sex that have existed for several decades. This is a really big ask — what I’m asking people to do is to step off of solid ground, and step onto an invisible plane with the trust that it will support their weight. That trust is so meaningful to me, and I hope that as we move forward together we will find that the invisible plane will hold our weight long enough that we can learn to perceive its form. Sorry about the meme up there, but in fairness the meme toddler is doing the admirable and important work of learning new ways of thinking about the world. Thank you for being willing to do that work.

Thank you again to Nic and my fellow panelists, who are absolutely amazing linguists and dinner companions. Also thank you for dinner. Thank you to the various Georgetown grad students who kept helping me find the gender neutral bathrooms and keeping me sane and grounded. Thank you also to my colleagues at the University of Washington and elsewhere whose discussion and comments were very helpful in the development of this talk and other work relating to it. The Socioling Brownbag group at UW all had very helpful feedback, and Prof Alicia Beckford Wassink is and has been an amazing mentor and advisor who helps me challenge myself, hone my thinking, and really do the hard work of being good at stuff when I don’t just automatically already know it. I am the toddler in the meme and sociolinguistics is the shapes and Alicia has taught me about a LOT of shapes. Edwin Howard and Drs Brooke Larson, Leah Velleman, and Lauren Ackerman have all been amazing supporters from afar, and discussion with them have also been instrumental in this and related work.

GURT 2019 was an amazing conference, and I want to try and write up a second blog post to talk about EVERYONE ELSE’S AMAZING STUFF, since this one is very much about my stuff and kind of a bummer besides. I’ll keep you posted.



Kirby Conrod

Dr. Conrod is a linguist and scholar sort of at large. They write about transgender stuff, the linguistics of pronouns, and ways to work with your brain.