In August 2014, my friend Shweta Narayan (scholar of linguistics and cognitive science, and speculative fiction author and editor) published an essay on their LiveJournal, "Let's talk about category structure and oppression!" As they said to a commenter, the essay was "applying some of the 101-level concepts of my field (literally! the class at Berkeley is cog sci 101) to what I'm seeing all round." Like many, I read it and found it useful. I still frequently want to refer to it in conversations with others.

But LiveJournal is an ever-sketchier place to visit these days. So Shweta gave me permission to repost this here on my blog under CC BY-SA, to make it easier to discover and reference. Following the text of the post I'll add links to their followup posts.

Contextual note: Narayan was writing to other people in the scifi/fantasy industry, and in SFF fandom, in the context of RaceFail 2009 which had started a few years previous.

Let's talk about category structure and oppression!

Written by Shweta Narayan, first published August 20th, 2014 at Narayan's LiveJournal and licensed CC BY-SA; please note that appropriate attribution includes the people whose research Narayan cited

This has been a v long-brewing post; I've been meaning to make it, or something like it, since 2009. Many thanks to rose_lemberg, Arachne Jericho, sovay, and elsmi for helping me finally get it together in a coherent fashion. Any errors or problems are my doing, not theirs.

Editing to add: reblogging this/sharing it on any forum is totally okay!

Edting to add #2: I am not American. I am not calling Americans the default/prototype because I'm projecting, I'm doing so because they are effective and powerful imperialists who have successfully become a global default no matter how hard Europe pretends they're not. I do live in the US now. I did not when I wrote the post. (also: prototype does not mean majority & does not always correlate) - Feb 2015

---

We tend to have this idea that categories, like "bird" or "food" (or like "human" or "white", which is what this is all really about) are like solid boxes. Entities are either in them or out of them, with a clear and unchanging boundary, and everything inside is an unsorted & equal jumble, and everything outside ditto.

This notion gets strongly underscored by our cultures, so it can be hard to ... er... unpack. But the fact is, cognitive categories aren't actually like boxes. They have internal structure, and fuzzy boundaries (which people can draw in different places, and move depending on context), and these things matter hugely in how we think about and deal with oppression.

I'm going to start by talking about research on the category "bird", because there's been a lot of it (c.f. Eleanor Rosch's work in the 70s and early 80s, which kicked it off), and it's pretty neutral so it'll be easier/less triggery for people to think about the category structure.

So! The "bird" category has (somewhat culture specific) internal structure. For example, most Americans will agree that a robin is a better example of a bird than an albatross, and an albatross is a better bird than an ostrich. (And while bats are not birds, they are better birds than horses are, and horses are better birds than refrigerators are; so the gradations continue to some extent outside the category boundary).

This internal category structure has a number of cognitive effects/characteristics:

1) If you ask people to just write down as many birds as they can, they'll list the more prototypical (category-central) ones first. More peripheral members of the category do not come to mind at first.

2) In reaction time tasks where people are asked to respond yes or no depending on whether or not a presented item is a bird, people will press yes faster, with fewer errors, for prototypical birds.

3) The structure that emerges from these two experimental measures matches the structure that emerges if you just ask people to rank birds in order of which ones are the "best" birds. Once you ask people to structure their categories they have really strong, consistent, and replicable intuitions about that structure.

4) People's idea of similarity is asymmetric: they will, for example, say that albatrosses are more like robins than robins are like albatrosses.

5) People reason from the prototype to the whole category, but not the other way around. So, for example (according to experimental results), people reason that if all the robins on an island caught a disease, the ducks would catch it too; but not vice versa.

6) People's use of linguistic hedges (really, sort of, technically, etc) is based on prototypicality too. So you can say an emu is technically a bird, but you can't say a robin is technically a bird.

7) Over time, some characteristics can become more prototypical. Others can't. The US adoption of the eagle as a standard animal has made it a more prototypical bird; and the hooked beak has become a more prototypical characteristic than it used to be. But yeah, eagles can still fly. An emu is never going to be a prototypical bird.

This is all pretty innocent when it comes to birds! But there is evidence that this sort of category structure is everywhere in human cognition (e.g. people will say 4 is a better even number than 1374.) Now, robins excluding emus from the bird-category, or claiming to understand how emu-ness works because of their experience as robins, might sound like the stuff comic strips are made of; the human dynamics are less funny, and far more harmful to their targets.

So, moving domains to socially relevant categories:

1) Able neurotypical not-fat not-poor straight cis white anglophone American Christian men are considered to be prototypical humans (prototype here = privileged default). So. If you ask people to think of famous people, they will think first of famous able neurotypical not-fat not-poor straight cis white anglophone American Christian men. And their exceptions will normally fall outside this prototype in only one or two ways.

This is how a lot of casual erasure happens. (btw it's also what's happening when editors "just happened to think of" a lot of poets/writers/artists who aren't marginalized, and when poets/writers/artists "just happened to think of" prototypical characteristics to portray.)

2) If someone is not an able neurotypical not-fat not-poor straight cis white anglophone American Christian man, it will generally take people longer to categorize them as human. And the further they are from this prototype the longer it will take to make the judgment. Now, if people take that extra time, we're probably good; but do they? When they sort resumes / run job interviews, when they're trigger-happy cops, etc?

3) (horrific examples tw) Consider the structure of the category "American citizen", which often gets treated as either-or. But the prototypical citizen is white, abled, and Christian (at least). Consider who counts: who gets protected under US law. And consider whose ID gets checked, who gets stop & frisked. Whose mass incarceration and state-sanctioned murder is business as usual. Who gets called "an illegal", or told to "go back home", regardless of their actual documentation. Who gets demands for their birth certificate once elected to high office. Whose languages are considered ok if spoken in the US, whose accent if they're speaking English.

(Non-Americans, when we talk about American privilege, we need to understand that it does not apply equally to all people with US citizenship; it applies only to the people who get counted as "proper" Americans, according to this category structure & the context.)

3b) (Horrific example tw) Where you draw the category boundary can be person and culture specific. Which is okay with birds, you'll only annoy scientists if you decide an emu isn't a bird, but what about the category "human". What about the people who decide that if you're Black, or disabled, or a trans woman or all three, then you've fallen outside the human category and your murderer isn't really a murderer? The murderers who call their Black victims “it”? The settler laws about Aboriginal Australian people, that only recently categorized them as human?

3c) This also applies for categories like whiteness. Who counts as white depends on who's drawing the boundary, where, when, for what purposes. I think we do need to talk about which people's whiteness is marginal/conditional and can be revoked by category-central white people. We can't do that, however, without also talking about how people in these groups benefit from conditional/marginal whiteness, by mostly gaining white privilege while denying whiteness whenever questions of race/racism come up. I am suspicious of people who will only talk about how their whiteness is marginal when other people are talking about racism.

3d) Obviously I could go on, but consider also the category of English. Whose English counts as actual English? And within that, whose is proper English?

So yeah (3) tl;dr: This is how a lot of active casual bigotry happens.

4) An albatross is more like a robin than a robin is like an albatross; a queer WOC is more like a cishet white man than a cishet white man is like a queer WOC. Which characters in stories count as "relatable"?

Everyone is expected to relate to a cis straight white anglophone American man. We're all like them, they're just (default, category-central) people after all! But they're not like us. We're the albatrosses, here. How can the poor robins be expected to relate to us? This is why they think it's so ludicrous that they should be expected to read about marginalized characters (who are nothing like them!!) but think it's normal and fine that marginalized people should be expected to read about category-central characters.

Conversely, it's also why they think they know our experience perfectly well and can talk over us; after all, we're just like them, except in a few (stereotyped) ways. They're default people! Unlike us.

5) (Horrific example tw) While people know perfectly well that diseases will spread from category-central members of humanity to peripheral ones, they often don't realize it goes the other way too. In the 80s, a lot of people thought AIDS was a "gay disease" - it wouldn't hit straight people! (And bi/pan/polysexual people don't exist after all, c.f. the erasure caused by (1)). Sooo yeah, they didn't care, till it did start hitting a lot of straight (white) people.

6) (TERF warning.) Consider how some TERFs say, "Of course I think trans women are women! - Technically. But like, not real women."
So long as they can make that linguistic hedge in some form - so that they're not actually expected to treat trans women as fully women, as fully human - they're fine with it. This is part of how they contradict themselves so blithely without hitting cognitive dissonance.

This is one method of moving the goalposts. Our understanding of categories is fluid and context-dependent, and we shift from thinking about the prototype to the whole category and back more than we normally consciously realize, and we can use the same word, often, to refer to either; and oppressors can use that to pretend they're speaking in good faith and being "reasonable", while in fact they're changing their definitions on the fly to suit their convenience.

7) Consider whiteness again. Within a US context, some groups (e.g. "white" Jewish Americans) have become more white than they historically were, and benefit from co-signing whiteness. They are still not category-central; their inclusion may still be marginal; but by default, they are now on the inside of the category boundary. [EDITED 7 November 2023 by Narayan to add quotation marks around "white" and to note: This was more true in 2014, but we've been seeing that marginal status revoked increasingly since the rise of MAGA, and it's even worse now.] Whereas other groups (e.g. South Asian Americans) do not get to cross the line no matter how strongly they buy into whiteness, because Blackness, and therefore darkness, is an exclusionary feature. But what that means, too, is that South Asian Americans do get privileged over other groups, most notably Black Americans, and need to understand that the power dynamics of whiteness do not end at the boundary of whiteness. [EDITED 7 November 2023 by Narayan: light-skinned South Asians do try passing as white, but passing isn't the same as being.]

For more central but still not default people (both within and outside the category!), aligning with & co-signing the category-center brings clear advantages. That's not true for people who are always, definitionally, excluded.

I'm going to start my wrap-up by talking a bit about derailing (getting in before defensive-privileged-commentors do so, haha). Often it works by changing the category under discussion – forcibly redrawing the boundary, and thereby changing the center of the category & what's being talked about. Example that I see all the time: “Trans women are awesome!” gets derailed with “ALL women are awesome!” By making the category “all women”, the derailer does not merely extend the statement to more people. No, by changing the category and evoking the new category's cisnormative prototypes, they change the subject entirely – recentering themselves and pushing trans women off to the margins.

“Not All Men” works in sort of the opposite way. By creating this hypothetical subcategory of Not-All-Men and forcing attention to it, it derails discussion away from, & attempts to undermine statements about, the category as a whole.

So! When talking to other people, in fandom and outside it, we need to be aware of category-centrality as well as membership. Especially because categories like whiteness are not boxes, but rather spectrums, with a central core of “real”, unarguable members, and an uneven periphery of conditional members, who can get kicked out by the category center as convenient, but still benefit from some or all of the privilege most of the time. Understanding this helps us understand the mechanics of derailing, and the mechanics of marginalization/exclusion done by not-central members to even-more-non-central members, as well as the mechanics that central members use against us all.

Narayan also wrote a few followup LJ posts:

1. 25 August 2014: "Low spoon warning", including: "pls note that there are many big gaps in what I posted, it's only 2k words, there's only so much I can cover in 2k words, if you think there's something I left out you are almost certainly right :) The post is not meant to be a complete explanation of everything for everyone, it's just a basic explication of how conceptual category structure interacts with power structures & influences, for example, casual oppression."
2. 14 February 2015: "on white fragility"
3. later on 14 February 2015: "I am college-educated; I am not American. Which modifier did I remember to put in That Post?" including: "we are often unaware of the ways in which we match the prototype. The default is unmarked....prototypes absolutely do vary between people & communities, based on experience; and relevant categories and thus prototypes do absolutely shift with context. Take all generalizations as exactly that."

And people discussed Narayan's essay in the comments of their original post (see this thread on "most prototypical" not necessarily meaning "best"), as well as on BoingBoing, MetaFilter, and elsewhere.

Personal note from Sumana: there's an oblique monologue in Ancillary Mercy, the third Imperial Radch book by Ann Leckie, about socially constructed categories and saying things were the same but now they're different, or vice versa, that now comes to mind for me when I think about this topic.