iiinova — where your digital self keeps evolving

I should add to this the pragmatic costs to ambient eavesdroppers, which I think is the biggest problem.

You've been right about things before! I've seen it happen with my very own eyes and hit the "like" button accordingly. There's enough good takes for both of us

010

Kyle L @haikugoat.bsky.social · 23h

omg i've never been right about something before! big day for me!

110

Kyle L @haikugoat.bsky.social · 12/10/2026

Anyways, I think my original question (which I might've phrased differently) still stands: is the pragmatic risk to referrers justly proportional to the pragmatic reward for the referent?

110

Kyle L @haikugoat.bsky.social · 12/10/2026

Some of these rebuttals are good, but I don't think the first one is. Just cause something is a defeasible inference (in this case an anti-presupposition) surely doesn't mean it's socially inert or costless — often quite the opposite.

110

Kyle L @haikugoat.bsky.social · 12/10/2026

fwiw I posted this thread into Claude and the evil machine was NOT impressed by my arguments

I’d separate the observation from the conclusion. The observation holds up, but I don’t think it supports the conclusion.

What holds up: Third-party reference does create a real pragmatic wrinkle. If someone outside the common ground hears you use “it” for an absent person, they may misread you. That’s a legitimate cost to name.

Where I’d push back:

1. The dehumanizing reading is a defeasible inference, not part of what “it” means. A useful framework here is Maximize Presupposition (from Irene Heim): speakers should use the form whose presuppositions are satisfied in context and which is the most informative. He/she presuppose animacy (and gender); it is the elsewhere form with weaker requirements. So using “it” for a person normally triggers an anti-presupposition: “I’m signaling the stronger forms don’t apply,” which reads as “not a person.” But that inference depends on the common ground, and the repair is one clause: “Y uses it/its.” Once that’s established, the inference doesn’t go through. That makes it a cheap, one-time repair, closer to explaining that someone goes by a different name than a persistent double-bind.

2. The pronoun contrast carries less of the moral-consideration load than the thread assumes. English already uses “it” for persons in places (“Who’s there?” “It’s me”), and colloquial Finnish routinely uses se for people. Spoken Mandarin has a single tā for he, she, and it. Those communities encode moral consideration just fine through other means: names, verbs of cognition and emotion, the social act of the request itself.

3. The goal in 3/6 is imputed. You ask why, then rebut a goal (scrubbing animacy reflexes from language) that isn’t what most it/its users say. The reasons they give are reclamation, an other-than-gendered or nonhuman self-concept, or simply fit. None of those require purging intentional vocabulary, so the inconsistency charge doesn’t land. And for reclamation users, the form’s baggage is part of the point, and they’re accepting that cost themselves rather than imposing it unknowingly.

4. “Selfish” and “unilateral” overstate it. It’s a request about one pronoun for one person, with a small occasional cost (a metalinguistic gloss). The cost to Y of being consistently misreferred isn’t zero either, so the real question is how to weigh the two.

5. The they/them parallel is stronger than your reply to it. Your animacy-versus-gender distinction is a decent one. But early singular they for known individuals drew the same pragmatic complaint, and common ground adjusted. The animate/inanimate contrast would only be threatened if it/its usage changed the default meaning of “it” for objects, and nothing suggests that.

One framing note: “AITA,” “deeply selfish,” and “persona of someone who dehumanizes” will pull replies toward the tone rather than the argument. The sharper, more defensible version of your question is probably “does the default association of ‘it’ create a real cost for third-party reference, and who should bear it?” That’s a question people can engage with seriously.

210

Kyle L @haikugoat.bsky.social · 12/10/2026

To put in the terms of @jeffvandermeer.bsky.social's favorite intelligence agency: "There's no reward in the risk"

000

Kyle L @haikugoat.bsky.social · 12/10/2026

makes it load-bearing. But I don't think the same can be said about the animate-inanimate contrast. The difference b/w a human being and a shoe *is* in fact deeply important, politically/socially. It undergirds our notions of moral consideration and non-instrumentalization!

100

Kyle L @haikugoat.bsky.social · 12/10/2026

I assume the rebuttal is that you could (but shouldn't!) say the same thing about gender-unmarking: that asking for "they/them" pushes interlocutors to engage in pragmatically risky behavior. But the whole point is that gender *shouldn't* be socially load-bearing in the way that cisheteropatriarchy

100

Kyle L @haikugoat.bsky.social · 12/10/2026

corollary: LLMs tell us surprisingly little about machine "minds"

020

Kyle L @haikugoat.bsky.social · 12/10/2026

I'm really open to being corrected on this, and I'm sure I'm missing a lot of the nuances, but prima facie, it just seems deeply selfish to ask others to unilaterally forgo one of the main ways that people encode (the presupposition of) moral consideration for others via language. AITA?

100

Kyle L @haikugoat.bsky.social · 12/10/2026

I also think that if your goal (why exactly?) is to try to remove the reflexes of human animacy from language, you're gonna have to do a LOT more than just get rid of he/she/they. But presumably it/its people* don't remove all intentionalist or human-theoretical vocab from their own speech...

100

Kyle L @haikugoat.bsky.social · 12/10/2026

Imagine you're at a cocktail party, talking to X about Y, who's absent. Y uses it/its pronouns, but X doesn't know Y at all. In the guise of trying to respect Y, you end up coming across to X as if you're dehumanizing Y. This seems like an unethical pragmatic double-bind to knowingly put people in.

100

Kyle L @haikugoat.bsky.social · 12/10/2026

OK I have a question that would probably get me yelled at, Bluesky-style, if only I had more followers: Am I wrong to feel that folks who request to be referred to by "it/its" pronouns are basically asking everyone to take on the persona of someone who dehumanizes people? @kirbyconrod.bsky.social

200

Reposted by Kyle L

bristlyboi @lvrking-lingvist.bsky.social · 12/10/2026

between katakana and kanji I can functionally parse quite a lot of Japanese if I ignore the pesky hiragana interpolations

111

Kyle L @haikugoat.bsky.social · 12/10/2026

Hahaha, this feels both blasphemous and beautiful

010

Kyle L @haikugoat.bsky.social · 12/10/2026

I'm fascinated by the old Sinosphere scenario of Korean, Chinese, and Japanese intellectuals all successfully communicating about abstract ideas with each other, with no idea how to speak each other's languages

000

Kyle L @haikugoat.bsky.social · 12/10/2026

but which I now appreciate cause I think I get more out of texts (esp fiction and poetry obvi). But I also get kinda frustrated when my students exclusively read audiobooks. A big part of literature is the generative tension bw language as sound & language as image (enjambment being case in point)

010

Kyle L @haikugoat.bsky.social · 12/10/2026

I was def on the lookout for something metrical based on your post and the replies, but initially couldn't tell if that was simply priming me to see something not there... Anyways, I do probably subvocalize more than most when reading, which I used to feel bad about (cause I think it slows me down)

110

Kyle L @haikugoat.bsky.social · 12/10/2026

I am gonna continue staking out my idiosyncratic position that appeals to no one: LLMs are very cool precisely because human collective intelligence is very cool, and ALSO, LLMs are very uncool precisely because human political economy (in 2020s America) is very uncool.

350

Kyle L @haikugoat.bsky.social · 12/10/2026

Stridently pro-AI people are fanatics. But many stridently anti-AI people are fanatics too. And one of the ironies of the latter clique is that they're plainly audience-captured by the engagement incentives of social media! They are engaging in RL w/ an algorithmically-distorted signal lmao

1100

Reposted by Kyle L

Kyle L @haikugoat.bsky.social · 30/11/2025

I'm becoming kanji-pilled. It's actually degrading and uncivilized to have to sound out words mora by mora like a child. I'm an adult, I can ingest ABSTRACT IDEAS (as soon as I memorize all 2100 of them and their bonkers compounds and permutations)

211

Kyle L @haikugoat.bsky.social · 12/10/2026

tangentially related (though also, apparently, wrong, since I already speak English)

010

Kyle L @haikugoat.bsky.social · 12/10/2026

Wait am I going insane, or is the linked paper written in very up-and-down, easily perceptible rhythm? Or is the content of the paper just leading me to notice what was always there? Or both?

210

Kyle L @haikugoat.bsky.social · 12/10/2026

conditional and social meaning, with Trump's idiolect, or post-truth politics more generally, being an instance of the weight on social meaning being 0.99 and the weight on truth-conditional info being 0.01.

000

Kyle L @haikugoat.bsky.social · 12/10/2026

As many linguists (& increasingly some philosophers) have emphasized: information-conveyance is only one part of communication, alongside social maneuvering. @elinm.bsky.social & @rhenderson.net make this explicit (somewhat crudely but usefully) by making meaning literally a weighted sum of truth-

100

Kyle L @haikugoat.bsky.social · 12/10/2026

Well said, and could not agree more. Propaganda & social meaning are the weapons of the war, and if you're on Team Fight, you can't just bring a knife to a gun-fight

110

Kyle L @haikugoat.bsky.social · 12/10/2026

An idea I've mulling over is that there is in fact something "inside", but that thing isn't machine consciousness — it's just us! The call is coming from inside the house!

110

Kyle L @haikugoat.bsky.social · 12/10/2026

The teleporter/brain-uploader might be cool; it might be attached to a ghoulish political economy; it might be mathematically interesting. Whatever. The books are the things that make you more knowledgeable! A feeding tube is not the same thing as the nutrients that pass through it! Come on, guys

010

Kyle L @haikugoat.bsky.social · 12/10/2026

All in all, LLMs have ironically made me much more impressed by the power of human culture. Say you had a machine that could teleport you around a vast library & instantly upload any info to your brain. If you got smarter or more empathetic as a result, that would be mainly the books' doing!

110

Kyle L @haikugoat.bsky.social · 12/10/2026

This is an example of a fallacy I see all over the place: the automatic assumption that LLMs tell us something about machines, rather than something about humans. A lot of the miracles we attribute to LLMs are miracles of human collective intelligence. The LLMs just makes it possible to query that.

120

Kyle L @haikugoat.bsky.social · 12/10/2026

I'm agnostic about whether or not, say, LLMs could in principle be conscious. But, as it stands, I just don't think they tell us that much about the relationship b/w language and consciousness. More persuasive evidence, in my opinion, would be stuff from the study of animal communication...

000

Kyle L @haikugoat.bsky.social · 12/10/2026

Why did the human kissers kiss in the first place? (Probably lust or love or affection or darker emotions.) What made those kisses memorable? (They felt good, looked good, were momentous, etc) All these things involve consciousness and embodiment!

100

Kyle L @haikugoat.bsky.social · 12/10/2026

Imagine (& maybe this already exists) a robotic mouth trained, Cinema-Paradiso-style, on a massive corpus of biomechanical representations of "good kisses". It could plausibly be trained to be a v good kisser, utterly lacking qualia. This does not show that consciousness is irrelevant to kissing!!

100

Kyle L @haikugoat.bsky.social · 12/10/2026

huge role in motivating & structuring the creation of texts in the training data, the editing of those texts, the reception of those texts, the social realities of which texts get forgotten vs make it into the training data in the first place, & on & on. Consciousness seems like crucial substrate!

110

Kyle L @haikugoat.bsky.social · 12/10/2026

Been seeing the argument that LLMs show that consciousness isn't necessary for language, but this seems far too strong a conclusion to draw. All that LLMs show is that consciousness isn't necessary for statistical amalgamation of language-shaped data. It seems plausible that consciousness played a

110

Kyle L @haikugoat.bsky.social · 12/10/2026

More broadly than just politics, I think this type of libidinal rationality is key to understanding e.g. addictions, or abusive relationships. It's a mistake to say that such things aren't rational. I think almost everything humans do is, maybe trivially, rational!

010

Kyle L @haikugoat.bsky.social · 12/10/2026

I do think this type of "rationality" is systematic, structured, patterned, etc. Making money rather than power or status the default example of maximizing expected utility was a mistake

120

Kyle L @haikugoat.bsky.social · 12/10/2026

Yes. This idea has really been in the zeitgeist in Trump 2.0. It's the only way to make sense of stuff like tariffs, Elon Musk, the voting behaviors of working class conservatives. And as others point out in the replies, this is straightforwardly anticipated by Du Bois

131

Reposted by Kyle L

Tom McKay @catturd2.bsky.social · 11/10/2026

nice consensus reality you got there. shame if someone was to invest 50 trillion dollars in breaking it

0233

Kyle L @haikugoat.bsky.social · 11/10/2026

Can you expand on why you think it would be more disturbing? Do you mean because it would cut against a naturalist theory of our own minds?

110

Reposted by Kyle L

Kyle L @haikugoat.bsky.social · 24/07/2026

retweeting your own posts simply because you wish they got more attention than they did: is it all right to?

151

Reposted by Kyle L

Kyle L @haikugoat.bsky.social · 01/04/2025

Woah just occurred to me that Season 1 of Severance included a literal example of Goat Update (definite anaphora to a non-linguistically introduced goat referent: "It's not ready yet!")

221

Reposted by Kyle L

Kyle L @haikugoat.bsky.social · 22/09/2025

this is, naturally, my favorite philosophy of language paper title ever, and doubles as a soothing mantra I can repeat to myself now and then

All Update is Goat Update
Andy Egan and Timothy Sundell
Draft of June 18, 2024

131

Reposted by Kyle L

Kyle L @haikugoat.bsky.social · 03/06/2025

When I finally get my shit together and walk into the room for the very first time self-actualized and in my power, that's gonna be what formal semanticists like @schnitzafars.bsky.social call GOAT Update

211

Kyle L @haikugoat.bsky.social · 11/10/2026

every time I get back on Bluesky, I am reminded that I would like to be an Intellectual

020

Kyle L @haikugoat.bsky.social · 11/10/2026

I would never put your or anyone else's book in the evil machine!

000

Kyle L @haikugoat.bsky.social · 11/10/2026

Yes. The ideal state of being imho is "socially connected, beloved eccentric"

021

Kyle L @haikugoat.bsky.social · 11/10/2026

HDM (which I adore and read at exactly that age) requires you to imagine what it might feel like to fall in love and to challenge authority. Annihilation (and SR series more generally) asks you to imagine falling out of love and trying to become authority. (many "authority" double meanings here)

010