intermediate pronoun studies: themselves and themself

Kirby Conrod
7 min readMar 26, 2022

Q: Should I say ‘themself’ or ‘themselves’ with singular they?

A: This is a question I have gotten a few times, and when I’ve answered it in live Q&A sessions, my answer has been that infuriating Linguistᵗᵐ Answer: I dunno, which one do you say? What feels more right? I can’t tell you what to do!

Which sort of leads to a follow-up: well… what do most people do? I’ve been wondering that myself! Other than this talk by Lauren Ackerman, Nick Riches, and Joel Wallenberg from 2018, I hadn’t seen much discussion of themself as a reflexive pronoun when referring to a single person. What Lauren And Friends found was that themself wasn’t peoples’ first choice — survey respondents didn’t rate it very high when asked how natural it sounded — but people with a lot of nonbinary friends did rate it higher than people without many. That’s something!

Anyways, when I moved to Philly I started hanging around my friend Byron Ahn’s place a lot, in part because he’s nice and cool and fun, and in part because he’s like an academic slightly-older-cousin who’s cool and I want to be like him. As we were hanging out we chatted about projects — he’s worked on reflexive pronouns (which end in -self/-selves in English) before, and obviously I work on singular they a ton — and what a collabo might look like. We obviously pretty much instantly decided to collab on figuring out what the heck is going on with who uses themself or themselves, and when and how.

We’ve also both been teaching linguistics (and especially syntax) for a good few years, so we’d both been (totally coincidentally! not planned!) doing informal polls in our syntax classes for a few years on themself. My observation had been, up to that point, that my students said themself sounded more natural than themselves, but only in certain contexts:

(1) Every professor evaluated themself / themselves at the end of the semester.

(2) The ideal partner takes care of themself / themselves before helping others.

(3) My friend from syntax class nominated themself / themselves for the award.

(4) Basil, who I met at a conference, considers themself / themselves an extravert.

Can you guess what my students voted, year after year? I bet you can: they said that themself was more natural than themselves in sentences like (3) and (4); they also tended to prefer themselves over themself in sentences like (1), and we got kind of mixed results for sentences like (2). So what’s going on there?

Well, Byron and I both thought that there was something going on with the grammatical properties of the antecedent — the underlined phrase in the sentences, which the reflexive pronoun referred to. Every professor, for example, is a phrase with a quantifier (every), and a lot of people get this sense that even though it’s grammatically singular, it doesn’t refer to a single person. A definite noun phrase like my friend from class, on the other hand, is very semantically singular — that’s almost certainly referring to just a single person.

One of the other things we talked about was that a lot of the research on singular they, besides not really talking about themself/themselves, really only looked at 3 types of antecedents, or sometimes 4 (like the 4 I listed above). We really wanted to see if there were finer-grained distinctions in there — for example, are quantifiers like every professor different from ones like some professor? (Both grammatically singular, again!) So we came up with seven types of antecedents:

Quantified universals => every professor, each professor

Quantified indefinites => some professor, any professor

Generic definites => the ideal professor, the typical professor

Definites where the speaker doesn’t necessarily know them: that driver in the SUV

Indefinites where the speaker might know them: a friend of mine

Definites where the speaker probably knows them: the person I talked to yesterday

Proper names: Basil, who I introduced you to yesterday

Phew, that’s a lot of types! So basically, we wanted to know how people rated themself and themselves in sentences with all those types of antecedents. Two pronouns times seven antecedents equals 14 conditions we wanted to test, which is a lot, but keep in mind we wanted to see some fine-grained distinction if at all possible.

The other thing we wanted to collect data on with this experiment was information about the people taking the survey — mainly their age and gender, because my dissertation research did find that younger people and nonbinary and trans people rate singular they with certain antecedents higher than, say, older people and cis men and cis women. And finally, we asked people questions about their feelings about two things: linguistic prescriptivism and binary gender ideology.

Linguistic prescriptivism — you might know this if you’re here from linguistics twitter — is the belief that there’s a “correct” way to use language, and that other ways of using language are “incorrect” (or improper or uneducated or irritating or unworthy, etc. etc.). A bunch of research from people like Ellis Hernandez and Evan Bradley has shown that people rate singular they as less natural when they hold these beliefs.

Binary gender ideology was a smaller scale (but one I want to expand in future experiments). Basically, we wanted to know about whether people believed that there are exactly and only two genders, and that gender is inherent and objective. We thought that people who hold those beliefs would likely also rate singular themself lower than people who didn’t.

SO, what was the experiment?? Well, we sent this whole survey out — rate some sentences! answer some questions! — and got a whopping 1,127 responses. (That’s huge for a linguistics study, we were really blown away!) Everyone rated the sentences, then did a little survey (for demographics and ideology stuff that I’ve mentioned). And here’s what we found!

antecedent and pronoun alone don’t explain stuff — but in combination they do.

Overall when we take all the results together (not splitting people out by age or anything) we found that people rated sentences slightly lower when themself and themselves referred to a specific person (with a proximal definite or proper name). Surprisingly, themselves was rated as slightly worse when referring to a specific single person.

age and prescriptivism tell us something in particular about certain combinations of antecedent and pronoun.

Age had a greater effect when we look just at proper names, and also when we looked just at themself — older people rated those as sounding less natural on average. Same deal for people who scored as more prescriptivist.

people clustered into three groups.

We used machine learning to sort people into groups, and our algorithm found that three clusters emerged:

  • Group 1: Speakers who kind of struggle with singular they no matter what, especially with a specific person — they also rated themself bad across the board. They did rate themselves pretty high with generic stuff (“everyone… themselves”) but that was pretty much the only thing they liked
  • Group 2: rated everything pretty high! They’re good with singular they, even with a specific person, and they don’t have a preference between themselves and themself
  • Group 3: rated themself pretty high, but rated themselves lower when it was about a specific person. (They were fine with “everyone … themselves” though!)

We concluded that these groups were proxies for different dialects of English that have slightly different restrictions on how singular themself/ves works for different people.

As expected, the groups had some patterns with demographics — but demographics didn’t totally determine what group someone ended up in.

For example, Groups 2 and 3 had more nonbinary people, but there were still some nonbinary people in Group 1. Likewise, Groups 2 and 3 were on average less prescriptivist than Group 1, but there were plenty of people in all three groups who represented the whole range of least to most prescriptivist.

So what does this all mean?

Here’s the big takeaways:

  • People use themself and themselves with singular they. If you’re trying to figure out which one to use — just go with your gut!
  • The way people use themself and themselves differently is slightly separate from whether people can use singular they for a specific person in general. If you’re someone who struggles with using they for a specific person, you might want to read my other posts about they/them pronouns. Don’t worry about themself vs themselves too much, in the meantime!

(Also, if you’re a linguist and you want to know more details about what we found, you can check out the slides for recent talks we’ve given on this work — we presented some of it at NELS, and some more at HSP. It’s still an ongoing project, so keep an eye out as we learn more!)

If you’d like to read more of my writing about pronouns, you might like my other posts on medium!

This work is supported by my ko-fi tips. You can also follow me on twitter.



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.