linguistics annual sleepover debrief part 2: my projects!

Okay, now that I’ve gotten the important job of telling you about spongebobussy out of the way, I’ll continue this week with some very short posts rounding up how my first Linguistic Society of America Annual Meeting (LSA for short) since 2020 has been!

What’s this sleepover you mentioned?

A very brief bit of background if you’re someone who’s not an academic linguist reading this: the Linguistic Society of America is a scholarly society of, you guessed it, linguists. I am thinking of writing up another bit in this mini-series about what exactly a “scholarly society” even is, but for now picture it as a mix of hobby group, social club, and professional organization. This year’s annual meeting is the 98th since the organization was founded, but I guess they skipped a meeting one year because next year is the centennial somehow? I don’t know.

Anyways, the annual meeting is a conference that is generally in a different somewhat-large city, almost always hosted in a hotel, usually in the first weekend in January for reasons*. There’s usually around 1,000 in-person attendees, though this year was a little smaller than usual (750ish, I heard?) and 2021 and 2022 were atypical due to, you know, pandemic. But basically, I’ve been calling this miniseries a debrief of the “linguistics annual sleepover” because in many ways that’s what an academic conference is — the presentations and chance to share our work with other linguists is ostensibly the “point” of the party, but in the same way that watching movies all night is the “point” of a middle-schooler’s sleepover. The actual point is to be together and get way sleep-deprived in each other’s company and build social ties and decide who is whose BFF, etc.

That said, I did totally give two presentations! I’m not going to completely re-tell you them here, just give a very short recap and talk about some of the really nice questions that I got after the talks. (Q&A is always my favorite part of any presentation!)

Three ways to rate themself

Do people prefer themselves or themself with singular antecedents (like the examples below)?

(1) Each officer perjured themsel__

(2) An ideal student respects themsel__

(3) Frankie admires themsel__

We did a really big experiment to find out! We got over a thousand people to rate sentences with different antecedent types and both the themself and themselves form. We found that, if you lump together all the antecedent types, themself is a little better overall than themselves, but that’s not the whole story. When we broke it apart by antecedent, people rated themselves higher with generic and quantified antecedents like (1), and rated themself higher with specific definite antecedents like proper names in (3).

We also used a machine learning algorithm to sort all our participants into three groups that showed slightly different patterns. There was a group of conservative speakers, who prefered themselves over themself, but also rated both types really low with proper names — they seemed to struggle with that use of singular they in general, regardless of the self/selves thing. Then there were two innovative groups: one group rated everything really high, suggesting that they’re fine with specific singular they and don’t have a preference between self/selves, and the second innovative group rated themself much higher than themselves, especially with specific antecedents like proper names.

We’re working on a followup experiment now to try and figure out some more of what’s going on with that first group, the conservative speakers — do they have preferences about self/selves that are being masked by their problem with singular they? We’re using ourself/ourselves to try and find out!

This talk gives us some insight into how reflexives work in English — because we found a bunch of different behaviors with different speakers, it can’t be the case (as has been previously claimed) that reflexives just automatically perfectly reflect the antecedent in number/gender. Instead, there has to be something more complicated going on. Hopefully our future work will help us learn more about what exactly that might be!

Slides are here. This work is a collab with Byron Ahn, and part of the SEPTA lab consortium.

Acceptability of neopronouns

Neopronouns like xe, fae, and ey in English are a little odd, because pronouns are supposedly a closed category — meaning that supposedly speakers don’t easily absorb new words in that category. So we have two questions: one, are (English) neopronouns actually being taken up or adopted by people; and two, are all neopronouns equally acceptable?

This is part of a larger ongoing project, so we reported the results of a pilot survey that we’re following up on — and it’s just phase one of the project. More about the larger project below!

In order to get an idea of whether people find neopronouns natural in sentence contexts, we did a ratings survey (similar to the one above!). We asked people to rate sentences with neopronouns (xe, fae, ey, and ze for the pilot) and different types of antecedents (proper names (4), noun phrases (5), and some things we expected to be ungrammatical (6–7).

(4) After Tuck recorded the podcast, xe published it right away.

(5) When the math professor finishes grading, fae always double-checks each grade.

(6) While the many clowns were running around, ze squired the audience with foam.

(7) Before the coffee table can be purchased, ey must be checked for flaws.

(As you might have guessed, (6–7) were anticipated to be bad because we think people generally treat these neopronouns as singular (so a plural like (6) is out) and animate (so (7) would theoretically be bad too.) In addition to these, we included some ‘canonical’ pronouns (he, she, they) and some ‘case error’ pronouns (him, their used as subjects). This was to give us a baseline of comparison for what people generally found fine and what was definitely ungrammatical.

Perhaps unsurprisingly, we got mixed results — people definitely rated neopronouns better than the case-errored controls, and somewhat worse than canonical pronouns. These ratings also depended a bit on age and gender (men and women patterned together, and other genders rated neopronouns higher than either men or women). What was also super interesting was the part where we asked people what they found notable about the sentences they rated — the words pronoun and neopronouns were mentioned A LOT. This suggests that neologisms in a closed class like pronouns are REALLY salient, while closed classes are usually pretty unconscious or not salient to people.

A lot of participants also explicitly commented that they find certain neopronouns easier to parse or to use based on analogy with canonical pronouns — so lots of people compared ey to they, and xe to she or he.

In our followup, we’re swapping out one of the neopronouns for another one that wasn’t included in this pilot — thon, because it doesn’t have an obvious morphophonological analogy with the three canonical singular animate third person pronouns in English. The prediction is that if analogy is an important factor for what allows people to accept neopronouns, then thon will be rated way lower than xe or ey.

We’re also still working on some more sophisticated sociolinguistic analysis of who rates what how, which is another thing we’ll be including in the followup and hopefully sharing soon. And the next phase of the neopronouns studies is going to be looking more at what happens when people actually produce neopronouns (in everyday conversation, or in certain lab conditions). So stay tuned!

Like the first talk, this project is part of the SEPTA lab consortium, and I owe a great debt to my wonderful collaborators Al Nash, Ell Rose, Max Winig, and Kyra Roepke (all of whom are linguistics undergrad students, and names to watch!!). The slides are here if you want them!

Thanks for reading! Like I mentioned, this is Part 2 of a series of separate posts — Part 1 is here if you’d like to read back.

You might also be interested in my longer previous blogpost about themself / themselves, or you might also like to read some of my previous thoughts on syntax or perhaps my methodology thoughts on gender in linguistics surveys. And you can keep up with what I’m up to on my twitter.