The Wind, a Pole, and the Dragon
Posted3 months agoActive3 months ago
entropicthoughts.comTechstory
calmmixed
Debate
70/100
Machine TranslationLanguage BarriersError Debugging
Key topics
Machine Translation
Language Barriers
Error Debugging
The post discusses a bizarre error message that appears to be the result of a mistranslation, sparking a lively discussion about language barriers and machine translation.
Snapshot generated from the HN discussion
Discussion Activity
Very active discussionFirst comment
1h
Peak period
25
0-6h
Avg / period
11
Comment distribution99 data points
Loading chart...
Based on 99 loaded comments
Key moments
- 01Story posted
Sep 25, 2025 at 6:39 AM EDT
3 months ago
Step 01 - 02First comment
Sep 25, 2025 at 8:08 AM EDT
1h after posting
Step 02 - 03Peak activity
25 comments in 0-6h
Hottest window of the conversation
Step 03 - 04Latest activity
Sep 29, 2025 at 10:44 AM EDT
3 months ago
Step 04
Generating AI Summary...
Analyzing up to 500 comments to identify key contributors and discussion patterns
ID: 45371309Type: storyLast synced: 11/20/2025, 6:30:43 PM
Want the full context?
Jump to the original sources
Read the primary article or dive into the live Hacker News thread when you're ready.
Here's one I was thinking of the other day. In German, to breathe is just a verb: atmen. But in Danish, it's an idiom: at trække vejret. If you just found the straight dictionary definition of "at trække" you would find "to pull". "Vejret" is the weather. So if you didn't know that bringing the two together meant to breathe, you would end up being confused.
It gets worse if the idiomatic phrase has grammatical significance. "I used to eat meat". If you only just learned a bit of English, you would wonder what "to use" in past tense meant that had to do with the rest of the sentence. Or perhaps you would theorize that a word was left out (eg using a fork to eat meat). But you'd be completely wrong, since what it actually means is that I stopped eating meat, though I had done so for a period in the past.
My guess is that Japanese is far enough away from English that early translating software couldn't figure out these kinds of things. By contrast, I've never had modern LLMs write anything that didn't seem native, presumably because they are complicated enough to absorb the knowledge from the training data.
Literal words don't mean anything specifically unless they have a context, which is idiosyncratic.
LLMs never are able to resolve the double whammy of lexemes and the conduit metaphor paradox. That's how they need constant specific supervision. They're babblers that have no idea what they're saying.
Coders can't debate this, this is the inherent problem to language that generative and NLP waves away in a sleight of hand trick. Study the initial conditions to language folks: language doesn't really mean anything.
To make that simpler, each statement in any existing language is about replacing language.
Not only your two posts above read like word-salad, I went through your comments history and almost every comment of yours reads like this (except maybe a couple).
If so, could you maybe stop your experiment? I wasted time trying to parse your comment as if it was generated by a human.
To accuse a human of being a bot is really poor manners.
Far from word salads, they are based in deep theories from empirical demonstrations that words are our most fundamental illusions. If you want the deep research we use internally (about 100 citations) start here.
https://docs.google.com/document/d/1cXtU97SCjxaHCrf8UVeQGYaj...
I suggest that CS has little ability to debates the scientific reality of language if posters are going to complain scientific statements that are defined and defendable are word salads.
That's a warning CS hasn't done its homework.
What's your startup? What are your publications? That Google doc link is more word-salad and random quotes and citations.
What on earth is a "start up that's about replacing the symbolic"?
> If you understand the nature of language fully, then its ultimate function is to refute itself. In other words, all our use of arbitrary signals leads eventually to specific signals. To make that simpler, each statement in any existing language is about replacing language.
This is word salad, for example. You're not trying to "make it simpler", you're trying to obfuscate (or it's just random).
> That's a warning CS hasn't done its homework.
What's a "CS"?
I'm fine with each sign being arbitrarily chosen at some point. People can even make that sound kind of cool with the Helen Keller "water" thing. But if the whole conversation is meaningless, that's a snoozer. I'd be happier staring at the clouds.
“There is nothing red about the word red, and the word big is itself rather small.” (Cuskley, Simner, & Kirby)
https://www.sciencedirect.com/science/article/abs/pii/S00033...
The idea that "language has no meaning" highlights the arbitrariness of language, meaning there's no natural, necessary connection between a word's form and its concept. The relationship between a word and what it signifies is a result of social convention and agreement within a linguistic community. For example, the sound of the word "dog" does not intrinsically resemble the animal itself; rather, we've all agreed to associate those sounds with the canine creature.
A start-up that replaces the symbolic engages with specifics in signals in the record (animals, entoptics, onomatopoeia, calendrical signs, action-syntax externals, cinematic action-glyphs) and figures out what representations veer referential and offers concatenation.
The idea that language refutes itself is as old as the pre-socratics. I won't go into the detailed history here, but there are 1000s of reference for this statement. I'll use Cassirer's, which is pretty succinct.
..at some point a direct contact must occur between knowledge and reality. If we succeed in freeing ourselves from all these interpretations – if we above all succeed in removing the veil of words, which conceals the true essence of things, then at one stroke we shall find ourselves face to face with the original perceptions..
Ernst Cassirer The Philosophy of Symbolic Forms
Freeing ourselves from the veil of words is theoretically equivalent to language's ultimate function is to refute itself.
CS is computer science.
btw it's clear the drive to encapsulate in plain English was and is the achilles heel to CS, coders are forced to finger point using characterizations of "word salad" at scientifically complex theories and analytic ideas.
The critique you're dismissing is from an English professor who teaches argument, not "engineering." So your writing is in fact being criticized by both "CS" and "English" (as you might put it). I'm not implying that taking license with punctuation, vocabulary, or grammar is necessarily bad -- I studied writing under two poet laureates, so I'm no stranger to unusual syntax -- but what are you writing on an Internet forum for if not to be understood? Your arguments against writing more plainly are quite familiar to me; engineers make similar arguments when told they need improve their documentation or user training materials.
Reducing everything you don't understand to "must be a bot" seems uncharitable.
That seems a non sequitur.
I didn't say I didn't understand this person, I said their comments are word salad and full of logical disconnects. You'll see, for example, the first reply bears almost no relation to TFA nor to the comment they were replying to. Then, when pressed, they responded with obfuscation. They reply with variations of "language refutes itself" and links to google docs with a salad of links that bear no relevance to the topic at hand. They claim to be researchers in a startup but when pressed, they backpedal into anonymity. That's... not a good sign.
There's plenty I read that I don't understand, and I don't assume it was written by an LLM. Most of it predates LLMs!
But surely you sometimes read stuff you suspect was written by an LLM?
PS: the commenter directly used AI (Gemini) in this response, with no semblance of relevance to the comment they were replying to (except responding to keywords, exactly what an LLM would do): https://news.ycombinator.com/item?id=45374270
If you've read something like Ricoeur's The Rule Of Metaphor, their initial comment makes more sense I think.
To be 100% honest, what drew my attention to the initial comment is that I thought it was joke about the TFA: responding to an anecdote about cryptic machine (mis)translation with some LLM-generated mumbo jumbo! Alas, it seems not to be the case.
[1] though, I suppose, if one is trying to make the point that "all language refutes itself"...
The reply deals directly with both the article in question and the post heading this thread. Neither get at the cause of the confused translation. My post provides a direct channel past idiomatic expression into the source pf all language mistranslation, that words are arbitrary, and particularly what sabotages exchanges between Western agentic language and Eastern nonagentic languages (Chinese, Korean, Japanese - each with their own peculiar forms of non-agency.
If you don't understand a statement, then ask questions. If you think it's obscure, then detailed questions. If you understand it, then probably very bad protocol to accuse anyone of being a bot. In any case, this is a protocol network, each exchange is negotiable. If you want to participate, do it in good faith and keep the outlook rosy, avoid characterization: you are not a mindreader. One thing is clear here, there's a very big divide in here between the intellectually curious, and the intuited pretenders who seem to have only a background in pseudoscience and folk science/psychology who are posing as scientific thinkers. Some of you may have extensive math backgrounds, but this is not enough to parse theory and demonstration in the linguistic and neuroscientific fields. Keep your minds open.
If you need some background that goes into the statement, I'd do some research, here's a section of the citations from Nisbett's Geography of Thought that delineate how distinct Eastern and Western perception and language are:
Gentner, D. (1982) Why are nouns learned before verbs: Linguistic relativity versus natural partitioning. In S. A. Kuczaj (Ed.), Language Development: Vol. 2 Language thought and culture. Hillsdale, NJ: Erlbaum Gentner, D. (1981). "Some interesting differences between nouns and verbs." Cognition and Brain Theory 4, 161-178. Imai, M., and Gentner, D. (1994). "A cross-linguistic study of early word meaning: Universal ontology and linguistic influence." Cognition 62, 169-200. Ji, L., Peng, K., and Nisbett, R. E. (2000). "Culture, control, and perception of relationships in the environment." Journal of Personality and Social Psychobgy 78, 9 Masuda, T., and Nisbett, R. E. (2001). "Attending holistically vs. analytically: Comparing the context sensitivity of Japanese and Americans.” Nisbett, R. E., Peng, K., Choi, I., and Norenzayan, A. (2001). "Culture and systems of thought: Holistic vs. analytic cognition.” Norenzayan, A., Smith, E. E., Kim, B. J., and Nisbett, R. E. (in press). "Cultural preferences for formal versus intuitive reasoning.” Norenzayan, A., and Kim, B. J. (2002). A cross-cultural comparison of regulatory focus and its effect on the logical consistency of beliefs. Unpublished manuscript, University of British Columbia, Vancouver, B.C. Norenzayan, A., Choi, I., and Nisbett, R. E. (2002). "Cultural similarities and differences in social inference: Evidence from behavioral predictions and lay theories of behavior." Personality and Social Psychology Bulletin 28, 109-120. Peng, K., and Nisbett, R. E. (1999). "Culture, dialectics, and reasoning about contradiction.”
"Tell me what you think about translation, and I will tell you who you are."
-- Heidegger, Hölderlin's Hymn "The Ister"
"All mistranslations are good."
-- Deleuze, Dialogues II
I found that difficult to parse. Moreover, it's a strong claim. Can you provide some evidence as to why it's true? I can find some trivial counter-examples. For instance, which part of "I like oranges." evolves English as a language?
“We refute (based on empirical evidence) claims that humans use linguistic representations to think.” Ev Fedorenko Language Lab MIT 2024
.at some point a direct contact must occur between knowledge and reality. If we succeed in freeing ourselves from all these interpretations – if we above all succeed in removing the veil of words, which conceals the true essence of things, then at one stroke we shall find ourselves face to face with the original perceptions.. Ernst Cassirer The Philosophy of Symbolic Forms
https://en.wikipedia.org/wiki/Argument_from_authority
Besides, I was largely agreeing and building on what you said (with a quibble over one word choice), so I'm not really sure what you're objecting to.
if you understand logic, you may grasp logic was not required to reach any analytic conclusions in multiple contradictions, scientific discourse does just fine without it and probably needs to jettison logic en route to new formats.
The real problem is where does the post-hoc nature of words interact with the world of actions? They're highly separable. In one hand we have conscious will created by words, on the other is the wordless world of memory integration, consolidation, cognitive mapping, planning. All thought is wordless, we know this as of 2016 in aphasia studies, the question is if the words are arbitrary, what we doing using them in place of thought? Are we dumb, insane, or just wiling to live in slumber using them to go nowhere? If you look around carefully, nowhere is where we're going, and fast.
Thoughts aren't about things, they are things.
“We refute (based on empirical evidence) claims that humans use linguistic representations to think.” Ev Fedorenko Language Lab MIT 2024
https://pmc.ncbi.nlm.nih.gov/articles/PMC7415918/
btw the levels of folk science inhabiting engineering is tsunami. the whole field needs to be taught how to deal with the contradictions in scientific practice and discourse to get out of the vortex you've built for our species.
Probably you're not really a linguist, you're in a sub-field like NLP, generative or construction, which is only linguistics coded for engineers.
This is amateur hour in HN.
- It's great that you mention Stitch, because what I'm trying to tell you is directly related to his theory as presented in e.g. Mindreading. What he terms "desire detection mechanisms" are what I'm referring to as our ability to discern intent, which is how you're able to make sense of other people's actions (whether verbal or non-verbal).
It's either a troll or someone with problems. Since the username is "mallowdram" (melodram(a)) I'm leaning towards prankster.
To "use" can mean (though in most contexts this meaning is obsolete) to do something regularly or habitually. So "I use to do X" means "I am in the habit of doing X", and "I used to do X" means "I was in the habit of doing X". The implication that you don't do it any more is a Gricean thing -- if you were still doing it you'd say "I use to do X" rather than "I used to do X".
Nowadays no one uses "use to" in the present tense and no one is thinking of the above when they use "used to" in the past tense. But that's where it comes from.
[EDITED to add:] I see that this has got some downvotes, but I'm having trouble figuring out why. If you read this and didn't like it, I'd be very grateful if you let me know what problem(s) you see. Thanks!
(Of course maybe it's just random drive-by haters, but more likely there's something it would be useful for me to know. Maybe something I wrote is wrong, in which case I want to know. Maybe to some readers it looks like I'm insulting the person I was replying to or something, which wasn't at all my intention and if that's going on then I should probably clarify or apologize or something. I dunno.)
Sure thing, I'll try. Your comment does not contribute to the discussion. It appears that you completely missed the point the comment you replied to made. The point was that “used to” has no inherent meaning and would seem odd to someone who is very new to English, and your response was just to explain what it means. You argue that it's not actually that odd but then basically prove how odd it is by highlighting how much obscure/obsolete knowledge is required to fully explain it.
[1] Which may not be much; I will in no way be offended if you don't care.
I didn't miss that the person I was replying to said that "used to" is odd and confusing. I wasn't arguing that "used to" isn't odd (still less that it isn't confusing for novice English-speakers -- it certainly is).
I was arguing that it's a different kind of odd from e.g. "at trække vejret" in Danish, and (to me, but evidently not to you!) I think it's an interestingly different kind of odd, which is why I thought it was worth pointing out.
I wasn't attempting to "explain what it means", which obviously the person I was replying to already knows. I was attempting to explain why it means what it does.
(In particular, it isn't true that "used to" has no inherent meaning[2]. It really is a past tense of "use", and while the specific meaning it's a past tense of is largely dead you can still, if you squint at it, see how it's of a piece with the other meanings of "use".)
[2] Except in so far as no word has inherent meaning.
Evidently, none of that came across the way I intended (or, perhaps, it came across fine but I misjudged how interested anyone else might be in the history): I should probably either have been more explicit or not bothered at all :-).
I think maybe the tone of your comment has a bit of "well actually" that bothers some people. It's the difference between sharing a fun fact versus policing correctness. The interesting point is that "used to" is weird, but not for the reason that the parent comment assumed. That's a genuinely fun fact that I enjoyed learning.
There's an interesting rabbit hole to go down for English words from Latin "spirare" - to breathe, and how it connects air and breath of life, spirituality. We get respiration (re-spire, re-breathing), inspire (blowing on something, figuratively breathing life/fire into it), expire (breathing out the last breath of life), spirit (a vital substance of life), conspire (to breathe together, speak similar thoughts).
https://www.etymonline.com/word/respire
https://www.etymonline.com/word/inspire
https://www.etymonline.com/word/spirit
etc. Words that were almost literal Latin that are now more figurative English based on a sort of French / medieval version of the Latin version. [edit: just remembering that the rather Germanic and plain "Breathe on me breath of God" used to give me the image of someone going "huuuhhh" on their spectacles before polishing them. The next bit is "till all this earthly part of me Glows with thy fire divine" so it is supposed to be exciting and soul-sparking. I only feel the poetic imagery of "fill me with life, blow on the embers of my soul, oh life giving wind-deity", through Romance *spire* words not Anglish *breath* words. "inspire my spirit(2) through respiration, you spiritual spirit(1)"]
If you want to go down the Proto-Indo-European rabbit hole, the Sanskrit word for soul is atman.
In Spanish (from Spain), someone who "eats both meat and fish" is an expression for someone who is bisexual.
So my take would be asking if there are JSP settings(defaults,limits, flags) that interact With the runtime.
My 2c.
For the record the original question is much more enjoyable :-D
Going from フラグ to 龍旂 is no less weird than going from English "flag" to "dragon."
风 and 极 are Simplified Chinese. The corresponding Japanese characters are 風 and 極 and they don't connect to defaults or limits in software either. Now it's of course possible that 松本武 is not Matsumoto Takeshi (松本•武) but instead Song Benwu (松•本武) and the original message was in Chinese, but that doesn't explain the curious word choice either.
Really? What makes one style or the other easier for you to write?
Note that you didn't actually succeed at using simplified characters; 风 is simplified, but 龍 isn't.
> And Japan still uses trad Chinese characters as one of its writing systems.
No, that's just false. They use their own system, which involves some characters that match traditional Chinese, some characters that match simplified Chinese, and some characters that are specific to Japanese.
(edit: never mind, I see the original is linked there already. Please apologise for your stupidity!)
Note that in the original thread, there was someone who requested (in Japanese) to repeat the question in Japanese, and was ignored.
Trying to reverse engineer the translation errors when you know zero Japanese is absurd.
I'm only semi-fluent in Japanese, but none of it makes sense to me. "Runtime" (in the computing-related sense) in Japanese is 実行時, or one might use the English word tansliterated to Kana, ランタイム, but there is absolutely no connection from either of those to goats.
^ Hello, Mr. Matsumoto. This is Nate. Google Translate is incompetent and vulgar. (LOL) Please email me directly. You can write in Japanese. I intend to help you. See you,
It appears that the user is dealing with an error that occurs intermittently with shibboleth, an SSO interface. If some daemon is polling some kind of thing in a way that triggers an error, it would probably seem to happen at random.
Why? Those aren't similar concepts. There's no connection between them.
More wild speculation for the curious :-)
Whenever I open it, it gives me a good laugh.
> Minecraft installer keeps crashing, I'm using a Macbook pro 2014. (...)
Based on that, if I had to guess "wind" is actually a coercion of the katakana ウィンドウ into ウインド based on simple pattern-matching, then translated into "wind". I'm sure you can guess where this is going... that's right, ウィンドウ is "Window/Windows" (for an example of how it is used as tech terminology, see https://ja.wikipedia.org/wiki/%E3%82%A6%E3%82%A3%E3%83%B3%E3...). Typing ウィンドウ into google search also autocompletes with ウィンドウ 10 , ウィンドウ 11 (this should be easy enough to guess).
Then the whole thing can be read as
> When I try to install the runtime it throws an error. Does this happen a lot on Windows [Version]? I tried two, three times and it throws an error
[Error]
> That's not the exact error but it's about right. Is the full error in the runtime log? [Something]? Is this a problem with the JSP error handler on Windows [Version] when you install the runtime? Or maybe I just got something wrong with the runtime?
Otherwise I subscribe most to @brazzy 's point about this possibly just being a joke or a prank with multiple layers of machine translation. Still, this doesn't exactly seem like a big newsgroup so I don't see why someone would go to the effort.
Chat thinks it’s a kana segmentation error in a list of possible hypothesized problems/causes that get mis-rendered as literal nouns, such as “XML風, 設定方法, 処理の流れ / 処理流” such that roughly “XML style” -> “-style” -> “wind”, “setting method” -> “-method” -> “pole”, “processing flow” -> “-flow” -> “dragon”
(Note that XMl, setting, and processing are just examples to illustrate.)
The LLM assures me that “Japanese writers often end diagnostic questions with a compact list of possible causes (A、B、Cでしょうか?).”
Its final verdict: “very likely the English “wind, pole, dragon” = MT literal translations of a compact Japanese list such as 〜風、〜法(方法)、〜流 (-style, -method, -flow). The pattern, repetition, and the fact they appear at question ends all support this strongly.”
I watched it for the first time after somebody referenced it, as I did just now, as an example of this kind of problem. Despite my knowing the point of the plot beforehand, I found the episode was still interesting.
I wish I could mention this episode here for language enthusiasts to enjoy without revealing the main plot point (that idioms in languages are hard to translate). Shaka, when the walls fell. But I think the very act of mentioning it in a thread on this topic does so unavoidably. Temba, at rest.
Yeah I know it's completely different, but HNers will enjoy discovering this rabbit hole so let me lead onwards: Time Team reconstructed one https://www.youtube.com/watch?v=zNYcVuO-PoE :D
Oh and someone downvoted that. I love this site and all you wonderful people.
4 more comments available on Hacker News