@teortaxesTex [same guy in a different life reading Marx] Hmm… these "lumpenproletariat" sound a lot like… dalits… What if… [picks up pen]
English
Linus Mixson
1.9K posts
Linus Mixson
@LinusMixson
Former gifted kid with ADHD, autism, low IQ, erectile dysfunction.
San Francisco, CA شامل ہوئے Mayıs 2026
178 فالونگ268 فالوورز
@teortaxesTex Can't express the unearthly delight I felt envisioning the guy painstakingly mapping divisions of labor to varnas and then smugly setting down the pen. What a bombshell!
English
disturbing thought: I believed I have a lot of Chinese followers, but they may have been merely shang people and not shi (or at least nong)
Nepali Billa 🇳🇵@nepalimeow
@teortaxesTex No , it was system follow by china where shang people from rural areas can't even enter in big city's, it was birth base, their children's don't get education in big city's schools
English
Linus Mixson ری ٹویٹ کیا
If this is true this is stunning.
Przemek Chojecki | PC@prz_chojecki
Kimi 2.7 ranked 2nd after Fable 5 and before GPT-5 xhigh We have re-run our ErdosBench smoke test on 14 problems with Kimi 2.7, Qwen 3.7 Max, Grok 4.3 and compared it with the top performers from previous runs. Kimi 2.7 is amazingly good. More below.
English
@spandrell4 @neilalexanderw1 I agree with the former point; I merely said that my schema better resembles theirs, not that it's literally the same. Your second point rests on a counterfactual and I don't think it's really worth nittering about.
English
@LinusMixson @neilalexanderw1 The median linguist definitely doesnt think Cyrillic was de novo, and to the extent they accept Hangul as one it's because they haven't looked into it and so take the Korean propaganda at face value
English
Every single living writing system besides Chinese is descended from Egyptian through Phoenician.
Linus Mixson@LinusMixson
Anyway did you guys know that the Thai alphabet is a descendant of Egyptian hieroglyphics
English
Questions of when sentences are logically-quantifying rather than distributional aside, it seems uncontroversi that "every single living writing system besides Chinese is descended from Egyptian through Phoenician" has semantics that are more like "every single man is taller than every single woman" than "men are taller than women."
And beyond that — if I said "men are taller than women" and some dimwit chimed in with "some men are shorter than some women" and provided examples of tall women, I would say "yes, that's true, but on average…" rather than doubling down and trying to argue that those women actually are shorter than every man alive.
English
@LinusMixson There's two independent living branches, Egyptian and Chinese, and two dead, Mesoamerican and Mesopotamian (oddily).
Saying anything else is rather pointless nitpicking, like forcing people to say "on average" when stating men are taller than women.
English
I could be convinced that Cyrillic is de novo. I am, in fact, now leaning in that direction, although for the reasons I describe towards the bottom of the attached post I think it's somewhat frivolous exercise to try to resolve that one.
I'm confused as to why you're suggesting that I'm creating a new taxonomy when the taxonomy that you propose is accepted by essentially no one (see: your consensus-rejecting take on hangul), and mine is far more in line with the taxonomy held by the median linguist. One of us is proposing a new system, and I really don't think it's me.
Linus Mixson@LinusMixson
Descent clearly *is* a thing in orthography, just as it is in spoken language. Changes occurring gradually, that do not break legibility from the parent (or that occur via a stochastic process) constitute mutative changes almost exactly analogous to those occurring in the process of biological descent. I feel like you're pushing a metaphysical point about "descent" here that violates the sane sense of the term and doesn't really hold water. The issue isn't that hangul involved explicit design; like you said, orthography is full of explicit changes over time that don't break heredity. It's that hangul constitutes a *sharp break* from its antecedents. A reader of 'Phrags-pa would be baffled by hangul, and there's no trail of intermediate scripts that would allow them to reach hangul by tracing the gradient between them. The only way for a 'Phrags-pa reader to understand hangul is by explicit reeducation. If you projected hangul and 'Phrags-pa into some sort of reasonably-comprehensive orthographic feature space, they would be closer than, say, hangul and Latin, but would they form a cluster? No. Hangul would be an island on its own, isolated, farther away from the closest point in the Brahmi landmass than any script in it is from its nearest neighbor. Katakana, for what it's worth, along with all Japanese scripts, is an oddity: the letterforms simply are hanzi letterforms (with modifications made for convenience in the case of the kana), while kanji, obviously, preserves the meaning behind the Chinese originals. But kanji does involve a wholesale replacement (in many cases) of the Middle Chinese pronunciation with the native Japanese semantic equivalent of the corresponding word. This is a hard break and does throw into question the idea of heredity. But, on the other hand, the semantics of the glyph are preserved. So is kanji (and transitively the kana) a descendant of hanzi? Who knows? And, frankly, who cares? I'm personally not of the opinion that every odd intermediate case needs to be resolved any more than I'm worried that the sex binary is "troubled" by the occasional intersex pathology. These are observations about consistent patterns of function that we make concerning the natural world. Aberrations are to be expected.
English
Some scripts did change somewhat gradually in an organic way, sure, but plenty didn't, e.g. Cyrillic was St. Cyril explicitly adapting Greek to Old Slavonic. Some breaks are pretty big too, we just don't know how they happened. Did Brahmi evolve gradually from Syriac? Greek from Phoenician? Maybe, maybe not.
Again by the traditional taxonomy it's basically impossible to create a script de novo after the Neolithic as script creators will in all likelihood be literate themselves and use existing scripts as inspiration.
But that's just how this works. If you want to create a new taxonomy you're very entitled too, and you might get some takers, but it's just not very interesting as a historical statement to e.g. define Hangul as de novo but Katakana as derived.
English
@spandrell4 You said that Cherokee is Latin, not that it doesn't count. I backed down on 'Phrags-pa when it became apparent that I was mistaken; I fail to see the challenge in simply saying *almost* every living writing system, which is true and beyond all possible dispute.
English
@LinusMixson Because I disagree about Hangul and don't think Cherokee counts.
English
@teortaxesTex I know people who graduated with degrees in "localization"… in 2025. Hard to imagine a rougher start to your adult life. Good move; sad we won't see something similar in the US.
English
A perfectly Xi-like solution to "youth unemployment" lmao
next: cultivating the wastelands (hopefully with robots)
Whale Insider@WhaleInsider
JUST IN: 🇨🇳 China eliminates 12,000 ‘obsolete’ university degrees in push to prepare for the AI era.
English
@neilalexanderw1 @spandrell4 I elsewhere drew a distinction between independent invention and non-descent that I would stick to here, as well, although this is an interesting variation in conception. But this is becoming a semantic quibble that would likely waste both of our times.
English
@LinusMixson @spandrell4 Writing systems are not species; they’re technologies. Cherokee absolutely descends from Latin script as a technology, whereas Mayan script does not. A case that interests me is Luwian hieroglyphs: they’re officially inspired by surrounding systems, but they look so different.
English
@spandrell4 Almost every! That's all I wanted. Now we're on the same page. I don't know why you couldn't have just said that in the first place.
English
Yeah it was also made very recently which is kinda cheating. Deseret has a better claim at being de novo but it's just some guy, it has zero historical value so who cares?
Do I really have to mention freaking Deseret every time I make the interesting point that almost every script on earth descends from Egyptian?
English
Descent clearly *is* a thing in orthography, just as it is in spoken language. Changes occurring gradually, that do not break legibility from the parent (or that occur via a stochastic process) constitute mutative changes almost exactly analogous to those occurring in the process of biological descent. I feel like you're pushing a metaphysical point about "descent" here that violates the sane sense of the term and doesn't really hold water.
The issue isn't that hangul involved explicit design; like you said, orthography is full of explicit changes over time that don't break heredity. It's that hangul constitutes a *sharp break* from its antecedents. A reader of 'Phrags-pa would be baffled by hangul, and there's no trail of intermediate scripts that would allow them to reach hangul by tracing the gradient between them. The only way for a 'Phrags-pa reader to understand hangul is by explicit reeducation.
If you projected hangul and 'Phrags-pa into some sort of reasonably-comprehensive orthographic feature space, they would be closer than, say, hangul and Latin, but would they form a cluster? No. Hangul would be an island on its own, isolated, farther away from the closest point in the Brahmi landmass than any script in it is from its nearest neighbor.
Katakana, for what it's worth, along with all Japanese scripts, is an oddity: the letterforms simply are hanzi letterforms (with modifications made for convenience in the case of the kana), while kanji, obviously, preserves the meaning behind the Chinese originals. But kanji does involve a wholesale replacement (in many cases) of the Middle Chinese pronunciation with the native Japanese semantic equivalent of the corresponding word. This is a hard break and does throw into question the idea of heredity. But, on the other hand, the semantics of the glyph are preserved.
So is kanji (and transitively the kana) a descendant of hanzi? Who knows? And, frankly, who cares? I'm personally not of the opinion that every odd intermediate case needs to be resolved any more than I'm worried that the sex binary is "troubled" by the occasional intersex pathology. These are observations about consistent patterns of function that we make concerning the natural world. Aberrations are to be expected.
English
@LinusMixson @neilalexanderw1 "Descent" is not a thing here. There's always some guy involved doing the changes. It just happens we often forgot about him.
Sure Hangul involved more explicit design than adapting Latin from Greek but there's no discrete line here. Is the creation of Katakana that different?
English
Cyrillic borrows massively from Latin orthography and has far fewer novel letterforms. It *could be* construed as de novo; there's a sane argument to be made there, but it's barely worth discussing because the overlap is so huge. Cherokee, conversely, has ~no orthographic correspondence to Latin. It's not even an alphabet.
English
@LinusMixson Again by that logic Cyrillic is de novo. You can disagree about the whole discipline but not how this works
English
@neilalexanderw1 @spandrell4 But neither of those is genealogical descent. They're horizontal transfer at best.
English
@spandrell4 @LinusMixson I agree: Cherokee is a clear case of diffusion, though it’s not structural borrowing; Korean is an even mix of structural borrowing (alphabetic symbols) & Chinese syllabic-script-presentation influence.
English
Will never forgive myself for not going to see Yung Lean when he played literally one block from my apartment.
English
Correction: Despite my earlier assertions to the contrary, Phags-pa seems to be sufficiently derivative of Tibetan to be arguably not de novo. But this is somewhat irrelevant because hangul is pretty clearly not a descendant of 'Phags-pa in any proper sense, incorporating at most a handful of 'Phags-pa forms and plausibly being influenced by the earlier script in its overall conception. There seem to be Reddit theories contending otherwise, of course, but nothing credible.
English
'Phags-pa is de novo, so whether hangul is a derivative or not is irrelevant. Cherokee letterforms are (in many cases) Latin-derived but their reading is not, and bears no relation to the Latin characters that inspired them. And scripts clearly do evolve; your argument here is like arguing that animals don't "evolve" because each new animal was given birth to by its mother. This is really a stunningly callow and blank comment start to finish.
English
It's a slop indicium, but more importantly, layers, like all complexity, are a flaw & not a feature. The reason to have them is to provide transparent supports for a more elegant interface, not to be an advertising point.
English
This is going to come across as a bit blasé, but whenever you hear someone describing a software system like it's a cake you're justified in immediately dismissing it as some bullshit.
Simplifying AI@simplifyinAI
Tencent just open-sourced Hy-Memory. A memory plugin that gives Al agents real long-term memory using a 6-layer framework with dual reasoning. → System1: fast pattern matching for instant recall → System2: deep reasoning for complex memory retrieval → 35% reduction in token usage → 70% less memory bloat over time Most agents forget everything between sessions. This fixes that. Works for long-running collaborative Al agents that need persistent context. 100% Open Source.
English