to be stahn this could be a flawfst gwuj for pleedh mwaw up for a shrermp and such, ha weird words to get the skawrpsk used to keeng in a flow when they mahrstst
Some of them are outright English, and some of them are just... not Englishally orthographic. Like "pleedh" and "mahrstst", yeah you don't see anything like those consonant groups in English :P
I like how the Glish word for 'having' is literally the Norwegian word 'ha'. What I love about Norwegian is how simple it is and there are far more monosyllabic commonly used words than in English or any other European language I know. It's what feels like a 'clean' language to me
I love the two hand/cursor thing you came up with. so creative. gives you a personality and face without showing your actual face. the animations are so high quality. you did a really good job.
Nooo you can’t stop with that. We just need to bias the machine to start syllables reminiscent of the starting syllables. That and use predictable extra letters for allomorphs.
This happens naturally. Most recent example is "riz" for "charisma". The algorithm is a little naïve and ignores how some words relate to one another. Hence, the weird mapping for "thirty" when "three" and "third" are presumably untouched due to being monosyllabic already. "thirty" should probably be categorized as an allomorph.
I think riz could be considered a diff word, the connotations for riz is more romantic like trying to seduce someone, while charisma is an adjective thats more passive
@@arrakistoxic1765 It shouldn’t be, it’s literally charisma shortened, ChaRIZZma. Also how often do you actually use the word charisma for platonic relationships?Regardless, you unintentionally highlighted another example, from different to diff.
@@arrakistoxic1765It could be considered a different word sure, but I think it’s a good thing to consider the etymology of the word. Its origin is the word charisma, just shortened
I think one very important thing you missed is composite words, like hotdog, outside, inside, bathroom, sunflower, cowboy, etc. There's probably a lot of them that are "taking space" innecesarily. For example, watermelon gets translated to wult, but water is twawstst and melon is flem.
Things like inside outside can be simplified through context as 'ins' and 'outs' ... see my comment above. I think to make this work, there has to be a stupid understanding of contextual identifiers within the sentence.
Nice work. This actually reminds me of a lot of english creole languages. Maybe if we just deliberately shortened a lot of common english words to already used colloquialisms or slang (like "about" to "bout" in verbally speech, and "already" to "alr" like we do in text), we'd be able to tackle a lot of it already.
Congrats you just showed why Chinese being mono-syllabic makes it so efficient, but also why tones are essential. Tones allow you to reuse the same syllable sounds but still have different meanings. The result is a language with an incredibly high "meaning" per syllable ratio. That's why with Chinese proverbs, you have the ability to densely pack entire philosophical concepts and lessons into just 4 words.
Yes! I am learning Chinese (mostly just for fun, but I did take a class in college), and when he mentioned the homophone problem, I was just like "here is where tones come in" lol
And don’t forget there’s also just a ludicrous amount of simultaneously same-sound same-tone words, so even more than tones context may be the only way to tell when speaking. The words for “sixteen” and “pomegranate” (shíliù, shíliu) will practically sound identical. There are so many more homophones in Chinese than English
I mean any language can do short proverbs. They still have to be explained or understood already for the few words to actually be interpreted correctly. The culture keeps the meaning alive, but if you were disconnected from context, the proverb would sound like gibberish.
@@yolt9786buddy pomegranate and 16 do not sound the same. liu in pomegranate on its own turns into liú and is then distinguishable from 16. Also, context says that 我吃十六 is probably _not_ what you just said.
in the example paragraph at 10:34, there are already some words in the original that could be swapped out for shorter ones. the word "only" could be substituted for "one". "the _one_ problem left" is already something valid in english, even if not as instinctive as saying "only", and sometimes saying "the ONE thing" can have different meanings. so maybe part of the length of english comes down to word choice. but then theres expressions and stuff that could just be shortened to other phrases, and even if they were to sound awkward at first, if the goal is efficiency and speed then it would probably work in time. i know glish was made with the entire point of it being "english but short" but making a conlang out of english with shorter wording could work too. i love how english, such a short language already, can be made even shorter, like! portuguese WISHES it were this short already 😂 and im sure the finnish watching this are all like, *_you want even LESS syllables?!_* and after all this, antidisestablishmentarianism isnt even translated into glish with the translator in the desciption!
To make the Glish words more similar to their English equivalents, you could give the generated Glish word a similarity score. The easiest similarity function here is probably Levenshtein distance. You might want to play with the weights for the operations to discourage transpositions and subsitutions. Perhaps even boost the score for deletions! This means you will have to generate many more Glish candidates, but the ones selected will be of higher quality.
My man just came out of the shadows and made a perfect youtube video for like, no reason. Like this isn't his job. He just made a creative, well researched, educational, well animated, easy to understand video for the fun of it. Holy shit. I'm blown away. ❤ much love
@@Pain.- Probably cos you're not used to them, just like most people won't be used to abhorrent vocab usage like "weird af bro'". I hope you at least think English sounds abhorrent too cos phonetically it's an inconsistent mess
@@sktzn6829 no, I'm german, Im very used to the german language and I think it sounds horrendous Edit: English is shit too, general American accent is mostly fine, but those heavy ami accents, where you cant even understand anything are shit, british is shit, canadian is shit, Australian is shit and so on..
I remember my mixed emotions during those clapping exercises, in elementary school. It was fascinating and curious, but also frustrating and confusing. They usually had me feeling angry or insecure, by the end. Sometimes a caramel vs _“carmel”_ or chocolate vs _choclate”_ thing. Sometimes a “How are words like *scraped, bridge, truth,* and *desks* one syllable?” thing.
Before you count the vowels, you gotta re-spell the word to be like how you pronounce it. Yes, "scraped" and "bridge" both have more than one vowel. But we actually pronounce them "scrayp'd" and "bridj", which both only have one vowel.
I think there's probably a step after setting up the directed graph, where you run a loss optimization function across some common corpus, weighting words by frequency and aiming for glish versions involving the least change across the whole corpus, not just per-word. I also think it would also be important that some rules, like plurality, are applied somewhat uniformly. Of course, we don't do that reliably in English already, but sometimes that's a result of sound combinations not working. Like, 'changes' is only two syllables because 'zj-s' is difficult. But assuming pronunciation similarity is less important than rule following, 'change' could be 'chang' and 'changes' could be 'changs'.
I did a similar project without the strict one syllable rule, I called it Minima, the goal was to make english more logical - sometimes borrowing words from other european languages - it only has 16 letters. I also made a translator for it!
So glad you didn't mention how we could see it in action. Even if you respond now with a youtube link I'm unlikely to ever see this conversation again. Nice idea though.
I will share/let you know when I get around to making a video or something of it as it was just a personal project! Happy to if anyone's interested though@@nam4032
Just a thought for if you made a second part to this video, all numbers from 0-9 are monosyllabic except for seven. As a math guy, I personally get really excited when we can represent any number in one syllable, and calling seven Sven is amazing to me. Then instead of 35 being meedhd-five, it could just be three-five. or if the number was 777, it could just be said as Sven-Sven-Sven. Or we could call it Bjorn
Great video. Didn't know I'd find a linguistics experiment that captivating. Recently I've been playing Cyberpunk 2077, and they invented some practical monosyllabic slang words that are quite recognizable: "sitch" (situation), "preem" (premium, great, awesome), "klep" (stealing, stemming from Greek "kleptein", I presume).
A small thought for Glish 2.0: it should be able to account for things like "totally" = "totes", which is more in the spirit of the project, I believe. Of course, "legitimate" = "le • git" in this convention... maybe allowing two syllables [occasionally] would be a useful compromise, as well as permitting homophones to be distinguishable in context ("comp" is "compromise" here, "computer" there, "compress" elsewhere). Just spit balling.
Totes is already a word though. And the problem with allowing homophones is that typically they are just pronounced the same but spelt differently (new vs knew). When they are all spelt the same it becomes illegible. “I comp on comp comp because comp are comp to comp. = I compromised on computer compression because computers are comparable to compost.
@@jmoney4695The thing is that while "totes" is indeed a word, practically no one uses it (with its proper meaning) compared to using it as a substitute for "totally", or just saying "totally" in general. Way more common word.
@@hahasamian8010 that is just an example - i am sure there are many other comparable examples. The problem is that if you start allowing arbitrary shortenings (totally into totes), it opens the metaphorical Pandora’s box. Keeping it in a more systematic way is the only way to ensure it is somewhat understandable. Furthermore, “totes” is slang - and slang is not consistent across regions. Therefore, the number of idiosyncrasies that would be introduced to allow for certain, arbitrary slangs to remain would make it an overall much more complex system.
Glad I stumbled across this video. As a computer science student, the amount of effort put into the programming, explanations and animations is phenomenal. Definitely deserves the support of the algorithm.
Really cool from a programming perspective, but the linguistics in this is so oversimplified that it at times borders on incorrect. As a programmer and linguist, I am conflicted.
It looks like he made separate words for singular and plural. A better approach would have been to use a word like “many” before any plural, hence cutting the number of new words in half. For past tense, using a word like “past” before the verb. Making new words for each variant means Glish becomes a weird language real quick.
@@joostine3720 Good question! I will reiterate, before I start, the key words in my original comment are "borders" and "oversimplified". One of the most-obvious things, which even non-linguists could have noticed during the course of the video, is that [s] is higher on the sonority hierarchy ("more sonorant") than [t]; indeed, this is a very common and well-known exception to the rule-of-thumb that is the Sonority Sequencing Principle in many Indo-European languages. Yet, the presenter makes no mention of this (instead (as I recall) presenting the SSP as a kind of universal law of language (It's more of a universal guideline, and many (most? Hmm, this is a good question...) languages have explicit exceptions to it.). No mention of there even being such exceptions was made in the video (at least that I heard), and worse: 3:06 shows [s] as being less sonorant than [p]! This is plainly factually incorrect, but I assume the author just wanted a prettier squiggle, and decided to say "chàbùduō" and draw it incorrectly anyway, either to avoid having to talk about there being exceptions, or because the author didn't know that exceptions existed. Another thing that turned me off was that the author, as I recall, reduced the sonority hierarchy into a matter of how loud phones were which is... not correct, and beyond mere oversimplification. It's actually kinda difficult to explain sonority to a layperson. One way you can think of it in terms of distinctive features, which are the characteristics of a phone. Some features make a phone more sonorant than others. The ¿best? way to think of sonority is *probably* in terms of acoustics. Unfortunately, I'm not an acoustic phonetician (though I'd like to dive deep into it someday); but if I had to try to *hazard* an acoustic definition of sonority, I'd perhaps say that the more well-defined and steady the formants are, the more sonorant a sound is. If there's an acoustic phonetician out there, *please* correct me if I'm missing something with this definition. The loudness that the author said was the defining characteristic of sonority is at worst more of a side-effect of sonority, and at best just one small part of the puzzle. If you want to quickly grok the sonority hierarchy, you can essentially do so by going row-by-row in the IPA chart; the rows are the manners of articulation, and they're mostly ordered by sonority (though this is not true for some rows, such as the laterals, which aren't more sonorant than their unlateralized counterparts). An additional, though very minor point, is that the author exclusively uses "Sonority Sequencing Principle" in places where he meant "sonority hierarchy" or just "sonority" This is really not at all a *real* problem, since people can figure things out from context, or just reduce everything to the word "sonority"; but I bring this up because it's one of many tell-tale signs that the author is inexperienced with the subject matter. Which, I want to stress that that is fine; we're allowed to go outside our fields of expertise (Good heavens, imagine if we couldn't!). But what was covered was lackluster, akin to being tutored by someone who is still, themselves, learning the material they are trying to tutor you on. The author also had this idea that fewer syllables means faster communication, but per my understanding, this isn't true. I don't have a study off-hand to point to, but my recollection is that the rate of information transfer during human speech is consistent regardless of how syllable-laden the language is; that is to say: languages with more syllables are simply articulated faster than languages with fewer syllables. As an English speaker, you may have experienced this phenomenon when hearing Spanish spoken: it *sounds* really fast, because each syllabe really is being pronounced faster; but Spanish words have on average so many syllables that they aren't actually communicating more-rapidly than you with your less-syllabic English. The language faculties of the brain can only handle so much information at once, regardless of how quickly your mouth is able to move. In any case: these were the main things I remember having noticed when I watched it a couple days ago. Please don't read this as a total condemnation of the author, because it isn't; I'm just answering your question about why I felt the linguistical side to this video was so lackluster. And hence, why I was so underwhelmed by it, despite finding the software part quite cool.
Content is great and the way you present it is very engaging, made me curious about a language that doesn't even exist! And your voice is very soothing. Thanks for the interesting video and please keep making more, if possible.
In a weird way, I find it neat how this highlights the importance of root words and multisyllabic words in general. A word that's based one or multiple others has an easily identifiable meaning, even if someone's never heard that specific word before. It eliminates the need to memorize unique sounds for _every single_ word, and instead allows things to build on each other, and - in turn - build on a person's prior knowledge of the language.
No i mean, like, they only have 7K subs It realistically is my fault for not checking the channel to see their previous videos (of which there is ONE) dats my bad@@YamamotoTV2021
There's a bug in your application. If you enter a word such as "multi-directional", the second word doesn't get translated at all, compared to when it was separated with spaces. I suggest you to use a tokeniser to pass on non-word characters unchanged and capture groups of word characters, instead of whatever you did to capture only the first half of a hyphenised compound word ;)
Great catch; that's sure to improve the quality of the output! In my opinion, I don't think any "fancy"/actual implementations of a tokenizer (lexical/probabilistic) would be needed for this specific issue, mainly because a simple regular expression pattern would be able to handle most of the cases we'd want to (e.g., "thirty-five"), such as with a first basic pass: /\w+/g or /(\w+)/g for the capture groups, as you mentioned. (Use whatever modifiers you want, like m, etc.) This pattern doesn't handle numeric digits, but I considered digits regardless since they're out of the scope anyway. If we cared more for semantics or understanding, moving towards more complex natural language processing techniques would surely help to improve the sound/tone/flow/etc. Technically, it would offer much-improved consistency, but that’s a much more substantial change for a more complex, different problem. I'm not overly familiar with linguistics, though, so I might be missing some more significant pieces from my ignorance or lack of experience, where my understanding is rougher and causes me to make more assumptions. Edit: I just saw you put in a PR! I appreciate your diligence!
This video is an incredible insight into language and how it works fundamentally. I'd like to see more come from this, it seems like a solid foundation for a project!
Well, actually, nothing is bad to have some two-syllables or even more-syllables words. They introduce some variations in the prosodical patterns making the language speaking more of a pleasant occupation.
one problem: many words in english have multiple roots like pterodactyl, it would make less sense to invent a new syllable than to make the one word into two (calling them wing-fings) i think the english language already has enough monosyllabic words to do this, there's this game called poetry for neanderthals where you have to describe a word or concept on a card using 1 syllable words, when you practice it you can get very quick at it
Did you know that helicopter is not made up of Heli and Copter, but actually helico (like helix) and pter (like pterodactyl)? I found this out recently and thought it was fascinating.
Yeah, even so-called monosyllabic languages like 漢語 and tiếng Việt are filled to the brim with multisyllabic words like 自己 and hạnh phúc - they just happen to have 1:1 morpheme to syllable ratio. I know it somewhat defeats the point to make words like these, but it avoids consonant clusters that are not only hard to pronounce for english speakers but also likely to evolve into multisyllabic words in the future, and besides, you already let thirty-five slide, right?
Such a unique presentation style! The hands give everyone something to look at in the same way pointers work on slides, but their versatility in doing other gestures is super cool.
Good sir, I completely nerded out to this video. It was like hearing about Toki pona for the first time. What a neat idea! I’ll definitely be looking at Glish some more.
i don't know if you'll ever see this but this is one of the most creative videos ive seen in such a long time on youtube, i mean the animation, the presentation of your hands through those cute hands, the topic, the delivery, the writing, all of it is FLAWLESS and im dumbfounded how you only have 10k subscribers when you should be closing in on 8 figures with ths level of content.
well, if you check the videos this channel posted you will see there are a total of 2 videos, this one and another one posted 2 years ago, if he continues posting videos like this one at a reasonable rate I believe he would have much more subscribers.
Amazing video, I love that this has some literal parallels to stenography. Seeing the graph you made to assign monosyllabic words was kind of cool because it is LITERALLY the steno keyboard layout. Syllables add up.
The sonority principle is not the only thing. It is too general to be descriptive of English. Some languages disallow certain syllable structures. Like a lot of asian languages follow strict CV (consonant-vowel) structure. So there is always a vowel after a consonant. On the other hand here is the Czech word for wolf: "vlk". Yes it doesn't have vowels, but the "L" kinda functions like one. English also has theese things. For example in old English the sound "g" as in "good" began to shift to the sound "g" (as in "gene") before front vowels, like the one in "gene" or "green". This btw is the basis of the whole gif/gif debate.
@@kmr_tl4509 No I believe those all follow the principle correctly. pesps isn't really correct due to the sps part at the end since s should have greater sonority than p. However, english typically allows s to occur after or before voiceless plosives like p even if it breaks the sonority sequencing principle so I would say it's not really wrong in english.
The attention to detail in this video is crazy, this man even animated the typing hands towards the beginning to match the real keystrokes. Great job 👍
I love that "mapping" is "mip". MIP mapping is a common technique in computer graphics to avoid aliasing. It means _multum in parvo_, or "much in little".
This video has made me realize that the English language is like a completely trashed and barely functional cruise ship with a metric ton of duct tape keeping it afloat.
I think it would help a lot to make it much more intelligible if you at least tried to force letters to have the same phoneme order. I know your directed graph is supposed to somewhat do this, but it allows consonants that are in the first syllable to be the coda of the Glish word. If you had your directed graph instead take 2 lists of IPA symbols (one for the onset of the syllables and one for the codas, it might make some easier to understand words. For example: Problem -> Rahrmp. While all these sounds appear in the word. The P in "Problem" going to the end makes it arguably harder to distinguish than if it was left out entirely, Similar for the R that appears there too. Despite them being phonemes that are in the original word. If instead you initially restricted the /p/, /r/, and maybe /b/ to be available in the onset and maybe the /b/, /l/ and /m/ to be available in the coda, it might generate some easier to read mappings. Along with following your strategy of relaxing these restrictions as the generator fails. The problem I see with the generated words is that while they may have the same set of sounds, they are so jumbled up (and missing some) that it makes it impossible to try to guess. So anything that could somehow restrict the order of the sounds to be more like the original would help a lot for readability.
very true! That would likely help out. I actually have a TODO for that in my code that I never got to 😅 Though additional restrictions like this do make it harder to generate as it runs out of valid syllables a bit faster, so it may weirdly lead to more words having strange mappings. Though hopefully that would skew towards uncommon words, and it's probably still worth doing.
@@paralogical-devto help with that problem, maybe it could be added as another stage before the existing "try to form a syllable from existing sounds" one, so that (in theory) it never makes a worse mapping than the existing code does
it’d also help to generate words in multiple stages. the first pass would generate a set of potential words that could all be considered “good” mappings. the second pass would be about deciding which words get to reserve each syllable. this would solve cases where a word with 2 good mappings picks one that collides with many other words down the road when its other choice didn’t have any collisions at all.
I enjoy that this is a light hearted topic, and the video is presented as being some what silly. All while being a lowkey introduction and overview of some of the most powerful ontological/linguistic tools there are in computer science, and even some of the more complicated computer science concepts like graph theory. This is the type of video that keeps me coming back to youtube. Subscribed.
Here I am, thinking I just discovered a new awesome youtube channel and I'm about to binge hours of funny educational content. Now imagine my despair realizing there are only 2 videos. Please please please make more, this was awesome!
I just wanted to say how amazing this video is! I love the hand cursors and the animations are so smooth. Your explanations are amazing! Thank you for making this and you deserve more support from the UA-cam algorithm. ❤
My very own splawm preft. This ended up sounding like the kind of random gibberish you'd expect to hear in an improvised Rick and Morty skit. And I kind of love it.
When I was a kid, the clapping method for syllables always confused me. It took an interest in linguistics to finally realize the vowels were what you were really counting. This is made even more obvious in Japanese. In Japanese, people often count each hiragana/katakana character as a single syllable. Japanese has no lone consonant sounds except for "nn." This is also why their accents do stuff like "MaKuDoNaRuDo" and "hotto doggu." Notice how they inject vowels into everything even when there are none.
no, their english sounds like that because of their phonotactics. in japanese, the largest syllable possible is a consonant plus a "y" sound, plus a vowel, plus an ending n.
@@cordeaux the *timing* is said to follow mora but syllabic analysis still works and especially useful if you want to compare the number of syllables between languages
Sort of! Japanese people don't count syllables like we do in English (they count mora which is the minimum sound unit) but syllables are a universal principle of language so they still have them. The reason they inject vowels in places is because Japanese has different rules than English for building a syllable. In English we are allowed to have complex consonant clusters or 3 or even 4 consonants together e.g STROBES (s-t-r-oh-b-z) or TWELFTHS (t-w-e-l-f-th-s). Japanese syllables on the other hand are not allowed consonant clusters and have to conform to a (consonant)-vowel-(N) structure. So STROBES would become su-to-roo-bu-zu. Notice that they retain the long vowel length of the English diphthong and that the final consonant is a Z not an S. So they are basing the transliteration on the pronunciation not the English spelling. When spoken however some of the vowels can become muted to the point that they sound like they are omitted. Su-to-roo-bu-zu could be spoken as sto-roob-zu which is 3 syllables.
@@pb7199 I understand though I could argue against syllables being universal to every language - the need for a syllable to have a nucleus kind of disallows some constructions made in Nuxalk (and other languages) with the existence of the phrase clhp'xwlhtlhplhhskwts' [xɬpʼχʷɬtʰɬpʰɬːskʷʰt͡sʼ], without a single sonorant (though, even though it is definitely not conceptualised in syllables, syllables with voiceless vowels as nuclei - as in the u in japanese tsuki, suki, desu, and other words - do exist and could then explain this as being syllables with fricative nuclei - as in some sino-tibetan languages like Yi/nuoso - but voiceless)). But, I can find one example of a word in nuxalk which does not support this - qʼtʰ - which does not have a true syllable nucleus. My understanding of syllables is that they are not innate but a human way of understanding speech - just as word boundaries are not innate to the speech content we produce. If voiceless fricatives can constitute the nucleus of a syllable, there’s no reason why a word like strengths couldnt be analysed as bisyllabic streng kths besides from preconceived notions of phonology and syllables in english (the same goes for some pronunciations of perpetually, where the “per” is pronounced solely as an aspirated voiceless bilabial plosive - does the syllable nucleus become the glottal fricative or technically a voiceless schwa (even though english speakers might not perceive it that way) or is there now a syllable without a nucleus in that word?
I don't know how YT started playing this video, I don't have automatic play on and I don't follow you, I was on a different tab, but your presentation style hooked me in. You have some real skills. I loved this video.
You should see a conlang called Xdi8 or Shidinn, a conlang based on Chinese but made many characters multisyllable in order to remove all homophones from Chinese. It has also a few translators on the internet. It's a «mirrored» version of your idea, and in many ways very similar (or right opposite) to your conlang idea!
4:52 apparently most languages just don't have spelling bees because unlike english, their spelling actually makes sense lmao so it's more or less trivial
what a cool project! I kinda want to see a variant where you allow 2 syllable words. It won't be as efficient but would make mappings more recognizable. You can call it two-glish lol.
Delightful. For glish+ try a morpheme matching step and a standardized ending set. Gish++ contractions, Gish+++ maximize information per syllable. Btw you are in good company Shanon has a paper on this and so did Chomsky on info of language.
You built an anticipation through this video by focusing on the requirements and goals for 2/3 of it before you introduced that a computer was going to do the work. This did a great job at breaking down the problem and making it very understandable. Saying that a computer will solve the problem with code earlier would’ve somehow reduced the quality of the video in a way I don’t know if I could describe well. I am a programmer too for context.
Learning Japanese turned my understanding of syllables on its head. Consonant contractions are considered multiple syllables in disguise, and the end consonant would also count a syllable, so GLISH would be 3 syllables, GU LI SHI. But now I understand why it's so hard to tell when one word starts and ends in Japanese. They don't really use the sonority triangles so it all just runs together...
The sonority sequencing principle applies to Japanese just as well. One very important distinction is that between morae (singular: mora) and syllables. Gurishi would be a three-mora word, but could be pronounced with one, two, or three syllables. Two other examples: "Hakken" and "Nippon" - Each 4 morae, but 2 syllables This is also where the idea of a haiku poem becomes very different in English compared to Japanese. In Japanese haiku, you count the morae.
Japanese music lyrics can also just decide how many syllables a word needs to be to sound good. "Konnichiwa" can be 4 syllables or stretched out to 5 like "Ko-n-ii-chi-wa" if it would fit better. You can't really do this in English and make it sound good. It's probably part of the reason music translated from Japanese sounds so off to me. Maybe it sounds that way to a native JP speaker too, but I'm not good enough to tell.
This video is soooo good it's by now the most helpful and inspiring linguistic video about English phonology I have ever seen. It answered the question in my head that even myself can't figure out what I'm actually asking. Absolutely creative!!! This vid is soooo good it's by now the most fluh and spirnk skwuung vid bowft glish soulnt I have vevdz seen. It snerrd the usk in my head that nawrnv fleh can't frig out what I'm wul angsk. sutsts kriv!!!
The primary problem with aligning orthography with phonology is that you project accents onto written text, i.e. two words that mean the same thing are not spelled the same if one is written by an American and another by an Australian. There have been several serious attempts at this in the past, both with IPA and, specifically for English, Shavian and Quickscript. They look cool, but don't catch on for this reason; it becomes really hard to read at speed if you have to know a different sequence of characters for different accents. Imo, I actually like the other extreme: ideographic scripts, precisely for this reason.
@@saoirsedeltufo7436 It simply means unification of spoken language so there is no difference in pronunciation between regions where it's spoken. Then using alphabet that is phonetically correct won't cause issues and will set these changes in stone.
This reminds me so much of one of those kid languages taken to the extreme and fledged out. I can't be the only one that had a code language with some of my friends in school. Soon, if not now, kids will have AI assist them in fully developing it. Never occurred to me that AI may have an impact on language far quicker than traditional shift.
This was a beautiful experiment and as a former linguistics major I thought it was really fun! I think one thing you might want to consider is how to treat grammar. Bubble, bubbles & bubbled are not really 3 separate words, in practice we understand it is the same word just conjugated for different tenses. This is because humans do not rote learn every individual word and its separate forms, we learn root words and grammatical patterns. There's no problem with treating them as unique words but unless the shortened syllables also have a consistent grammatical pattern, doing so would in effect remove certain grammatical functions from Glish that exist in English. Perhaps you could try a different approach where only the root of each word would become a single syllable and grammatical particles are tacked onto the end like in Korean. One other thing to consider is whether or not you want to remain within the phonotactic constraints of English. Many of the generated syllables in your examples do not follow English constraints, like "meedhd" for "thirty", or "blegd" for "bubbled". If you remain within English constraints your directed graphs would be a lot smaller but so would your pool of legal syllables.
Conjugation strictly refers to verbs. Only bubbled is a conjugation of the verb "to bubble" (simple past or past participle for the perfect). If it can also be used as an adjective, in which case it is a derivation of bubble, but still considered a separate word even while containing the free morpheme bubble. Only "bubble" and "bubbles" are the same word as they are both a noun inflected for number.
I would say "blegd" is a perfectly reasonable English syllable. We have the one-syllable words "blogged" and "begged", just smoosh them together and you have "blegged".
I feel like this could almost be used in creating another "What English Sounds Like To Non-English Speakers" part
to be stahn this could be a flawfst gwuj for pleedh mwaw up for a shrermp and such, ha weird words to get the skawrpsk used to keeng in a flow when they mahrstst
no some of the consonant pairs look down right slavic
@@eksplosiveknight8891 except they exist in english?
No because many of the words are the same.
Some of them are outright English, and some of them are just... not Englishally orthographic. Like "pleedh" and "mahrstst", yeah you don't see anything like those consonant groups in English :P
Glish and Pig-Latin seem to be on opposite ends of the syllable spectrum
And, now.... We need to invent Glish-Latin. Thanks for that!
@@carstenjorgensen2607 Lom ips dol sit amt. Conc ang lit, sed do eid temp int ut lab et dor mag qua.
Pubig-Lubatubin ubis nubot qubite ubas fubar ubin thubat dubirubectubion ubas Ubububy Dubububy.
@@carstenjorgensen2607 Tin
Yeah, until you look at Japanese.
I love how the finished product sounds like he reverses engineered modern English into old English or a pre english Nordic dialect. love it
I like how the Glish word for 'having' is literally the Norwegian word 'ha'. What I love about Norwegian is how simple it is and there are far more monosyllabic commonly used words than in English or any other European language I know. It's what feels like a 'clean' language to me
sounds like German bro
To me too @@shatteredvidrio
@@shatteredvidrio nur das Deutsch ungefähr fünfhundert mal so viele Silben pro Satz hat
This is by far the best explanation of what syllables are and how our brains interpret them that I've seen
I love the two hand/cursor thing you came up with. so creative. gives you a personality and face without showing your actual face.
the animations are so high quality.
you did a really good job.
It reminds me of the Wii Menu
YES i love it!!
yess it's so unique!
Nooo you can’t stop with that.
We just need to bias the machine to start syllables reminiscent of the starting syllables.
That and use predictable extra letters for allomorphs.
I'd love to introduce you to the concept of the Vtuber XD
This happens naturally. Most recent example is "riz" for "charisma". The algorithm is a little naïve and ignores how some words relate to one another. Hence, the weird mapping for "thirty" when "three" and "third" are presumably untouched due to being monosyllabic already. "thirty" should probably be categorized as an allomorph.
Thirty is related to Twenty, Forty, Fifty, Sixty, Seventy, Eighty, Ninety. You can't escape the numbering structure .
I think riz could be considered a diff word, the connotations for riz is more romantic like trying to seduce someone, while charisma is an adjective thats more passive
@@arrakistoxic1765 It shouldn’t be, it’s literally charisma shortened, ChaRIZZma. Also how often do you actually use the word charisma for platonic relationships?Regardless, you unintentionally highlighted another example, from different to diff.
@@arrakistoxic1765It could be considered a different word sure, but I think it’s a good thing to consider the etymology of the word. Its origin is the word charisma, just shortened
@@tylerdavis3 Go complain to language department if such a thing exists, "sHoULdnT bE" 🤡🤡🤡🤡🤡
I think one very important thing you missed is composite words, like hotdog, outside, inside, bathroom, sunflower, cowboy, etc. There's probably a lot of them that are "taking space" innecesarily. For example, watermelon gets translated to wult, but water is twawstst and melon is flem.
wATER IS WHAT
Hmm, I feel like a better mapping is water -> wult, melon stays as flem, then watermelon is wultflem.
@@jacob-shafferyeah then its no longer monosyllabic which was his goal. But I can see you could make an exception for these composite words.
@@bryce4395Twawstst! 👍🏻
(Coughs up a whole-ass snake)
Things like inside outside can be simplified through context as 'ins' and 'outs' ... see my comment above. I think to make this work, there has to be a stupid understanding of contextual identifiers within the sentence.
Nice work. This actually reminds me of a lot of english creole languages. Maybe if we just deliberately shortened a lot of common english words to already used colloquialisms or slang (like "about" to "bout" in verbally speech, and "already" to "alr" like we do in text), we'd be able to tackle a lot of it already.
Never seen someone say alr. No way thats a thing, its so goofy
alr
LMAOO
Nah its a thing @@Pain.-
Congrats you just showed why Chinese being mono-syllabic makes it so efficient, but also why tones are essential. Tones allow you to reuse the same syllable sounds but still have different meanings. The result is a language with an incredibly high "meaning" per syllable ratio. That's why with Chinese proverbs, you have the ability to densely pack entire philosophical concepts and lessons into just 4 words.
Yes! I am learning Chinese (mostly just for fun, but I did take a class in college), and when he mentioned the homophone problem, I was just like "here is where tones come in" lol
And don’t forget there’s also just a ludicrous amount of simultaneously same-sound same-tone words, so even more than tones context may be the only way to tell when speaking. The words for “sixteen” and “pomegranate” (shíliù, shíliu) will practically sound identical. There are so many more homophones in Chinese than English
I mean any language can do short proverbs. They still have to be explained or understood already for the few words to actually be interpreted correctly. The culture keeps the meaning alive, but if you were disconnected from context, the proverb would sound like gibberish.
@@yolt9786buddy pomegranate and 16 do not sound the same. liu in pomegranate on its own turns into liú and is then distinguishable from 16. Also, context says that 我吃十六 is probably _not_ what you just said.
It also makes it horrible to learn as someone with english as their first language.
Man scientifically invented slang
in the example paragraph at 10:34, there are already some words in the original that could be swapped out for shorter ones. the word "only" could be substituted for "one". "the _one_ problem left" is already something valid in english, even if not as instinctive as saying "only", and sometimes saying "the ONE thing" can have different meanings. so maybe part of the length of english comes down to word choice.
but then theres expressions and stuff that could just be shortened to other phrases, and even if they were to sound awkward at first, if the goal is efficiency and speed then it would probably work in time. i know glish was made with the entire point of it being "english but short" but making a conlang out of english with shorter wording could work too.
i love how english, such a short language already, can be made even shorter, like! portuguese WISHES it were this short already 😂 and im sure the finnish watching this are all like, *_you want even LESS syllables?!_* and after all this, antidisestablishmentarianism isnt even translated into glish with the translator in the desciption!
To make the Glish words more similar to their English equivalents, you could give the generated Glish word a similarity score. The easiest similarity function here is probably Levenshtein distance. You might want to play with the weights for the operations to discourage transpositions and subsitutions. Perhaps even boost the score for deletions! This means you will have to generate many more Glish candidates, but the ones selected will be of higher quality.
My man just came out of the shadows and made a perfect youtube video for like, no reason. Like this isn't his job. He just made a creative, well researched, educational, well animated, easy to understand video for the fun of it. Holy shit. I'm blown away. ❤ much love
He’s the best
this isnt his job, yet
Thanks for speenk my mind. I can't deal with these sibz.
It's not well-researched. He incorrectly defines allomorphs.
@@BecauseICantEdit well it can still be well researched and have mistakes! it’s more about the work put into it- mistakes can always happen
That "berzdzdz" right at the end. Brilliant. You tried to make Glish, but accidentally made Polish instead.
As a Pole, I think more languages need to create phonemes with s, z, and ch sounds. It's strangely satisfying to speak those combos.
@@Hendrixski your language sounds like a mosquito trying to speak Russian
@@Hendrixskino its not😭 and it sounds weird af bro.
But then again, I think most languages sound absolutely abhorrent
@@Pain.- Probably cos you're not used to them, just like most people won't be used to abhorrent vocab usage like "weird af bro'". I hope you at least think English sounds abhorrent too cos phonetically it's an inconsistent mess
@@sktzn6829 no, I'm german, Im very used to the german language and I think it sounds horrendous
Edit: English is shit too, general American accent is mostly fine, but those heavy ami accents, where you cant even understand anything are shit, british is shit, canadian is shit, Australian is shit and so on..
I remember my mixed emotions during those clapping exercises, in elementary school. It was fascinating and curious, but also frustrating and confusing. They usually had me feeling angry or insecure, by the end. Sometimes a caramel vs _“carmel”_ or chocolate vs _choclate”_ thing. Sometimes a “How are words like *scraped, bridge, truth,* and *desks* one syllable?” thing.
Before you count the vowels, you gotta re-spell the word to be like how you pronounce it.
Yes, "scraped" and "bridge" both have more than one vowel. But we actually pronounce them "scrayp'd" and "bridj", which both only have one vowel.
Thank goodness my first language is Portuguese so i never had to face this problem
I grew up in the Netherlands and splitting words up into syllables is actually so much harder in dutch
I see why you were confused about “scraped” and “desks”; but why “truth”?
@@betin731They don’t have more than one *vowel*; they have more than one letter *usually* making a vowel sound. Linguistically, it’s just one vowel
I think there's probably a step after setting up the directed graph, where you run a loss optimization function across some common corpus, weighting words by frequency and aiming for glish versions involving the least change across the whole corpus, not just per-word.
I also think it would also be important that some rules, like plurality, are applied somewhat uniformly. Of course, we don't do that reliably in English already, but sometimes that's a result of sound combinations not working. Like, 'changes' is only two syllables because 'zj-s' is difficult. But assuming pronunciation similarity is less important than rule following, 'change' could be 'chang' and 'changes' could be 'changs'.
I did a similar project without the strict one syllable rule, I called it Minima, the goal was to make english more logical - sometimes borrowing words from other european languages - it only has 16 letters. I also made a translator for it!
I’m curious to see it
So glad you didn't mention how we could see it in action. Even if you respond now with a youtube link I'm unlikely to ever see this conversation again. Nice idea though.
sounds really interesting, could you send a link to it please?
I will share/let you know when I get around to making a video or something of it as it was just a personal project! Happy to if anyone's interested though@@nam4032
@@foobars3816That's rude
Just a thought for if you made a second part to this video, all numbers from 0-9 are monosyllabic except for seven. As a math guy, I personally get really excited when we can represent any number in one syllable, and calling seven Sven is amazing to me. Then instead of 35 being meedhd-five, it could just be three-five. or if the number was 777, it could just be said as Sven-Sven-Sven. Or we could call it Bjorn
Seven should be changed to either Sev, Siev, or Zieb, anything else is just wrong
Sieben
What about zero? Zer?
But then "one trillion" would be called "one-zer-zer-zer-zer-zer-zer-zer-zer-zer-zer-zer-zer".
@@chr13whatever happened to nil?
@@chr13 null.
Isn't that just Dutch
yea but worse
Great video. Didn't know I'd find a linguistics experiment that captivating. Recently I've been playing Cyberpunk 2077, and they invented some practical monosyllabic slang words that are quite recognizable: "sitch" (situation), "preem" (premium, great, awesome), "klep" (stealing, stemming from Greek "kleptein", I presume).
Pretty sure Kim Possible invented "sitch"
@@tonywebert8326 The game is based on a tabletop RPG from the 90s (and books, I believe), though I don't know when the words were actually invented.
I've actually heard sitch long before Cyberpunk 2077. It's a very 2000s slang term to me.
sitch is definitely old, but the other ones are new for me!
klep is from kleptomaniac, and calling someone a klepto is already sorta a thing
A small thought for Glish 2.0: it should be able to account for things like "totally" = "totes", which is more in the spirit of the project, I believe. Of course, "legitimate" = "le • git" in this convention... maybe allowing two syllables [occasionally] would be a useful compromise, as well as permitting homophones to be distinguishable in context ("comp" is "compromise" here, "computer" there, "compress" elsewhere). Just spit balling.
Totes is already a word though. And the problem with allowing homophones is that typically they are just pronounced the same but spelt differently (new vs knew). When they are all spelt the same it becomes illegible. “I comp on comp comp because comp are comp to comp. = I compromised on computer compression because computers are comparable to compost.
manatee vs pizza yousyn’d
@@jmoney4695The thing is that while "totes" is indeed a word, practically no one uses it (with its proper meaning) compared to using it as a substitute for "totally", or just saying "totally" in general. Way more common word.
@@hahasamian8010 that is just an example - i am sure there are many other comparable examples. The problem is that if you start allowing arbitrary shortenings (totally into totes), it opens the metaphorical Pandora’s box. Keeping it in a more systematic way is the only way to ensure it is somewhat understandable. Furthermore, “totes” is slang - and slang is not consistent across regions. Therefore, the number of idiosyncrasies that would be introduced to allow for certain, arbitrary slangs to remain would make it an overall much more complex system.
@@jmoney4695 Glish already has enough words messed up that it needs to be learned, this sort of change is just an optimization
Glad I stumbled across this video. As a computer science student, the amount of effort put into the programming, explanations and animations is phenomenal. Definitely deserves the support of the algorithm.
Really cool from a programming perspective, but the linguistics in this is so oversimplified that it at times borders on incorrect. As a programmer and linguist, I am conflicted.
It looks like he made separate words for singular and plural. A better approach would have been to use a word like “many” before any plural, hence cutting the number of new words in half. For past tense, using a word like “past” before the verb. Making new words for each variant means Glish becomes a weird language real quick.
@@Swenthorian what parts border on incorrect?
Yeah, fun :) 6:40 looks like a Markov chain to me.
@@joostine3720 Good question! I will reiterate, before I start, the key words in my original comment are "borders" and "oversimplified".
One of the most-obvious things, which even non-linguists could have noticed during the course of the video, is that [s] is higher on the sonority hierarchy ("more sonorant") than [t]; indeed, this is a very common and well-known exception to the rule-of-thumb that is the Sonority Sequencing Principle in many Indo-European languages. Yet, the presenter makes no mention of this (instead (as I recall) presenting the SSP as a kind of universal law of language (It's more of a universal guideline, and many (most? Hmm, this is a good question...) languages have explicit exceptions to it.). No mention of there even being such exceptions was made in the video (at least that I heard), and worse: 3:06 shows [s] as being less sonorant than [p]! This is plainly factually incorrect, but I assume the author just wanted a prettier squiggle, and decided to say "chàbùduō" and draw it incorrectly anyway, either to avoid having to talk about there being exceptions, or because the author didn't know that exceptions existed.
Another thing that turned me off was that the author, as I recall, reduced the sonority hierarchy into a matter of how loud phones were which is... not correct, and beyond mere oversimplification. It's actually kinda difficult to explain sonority to a layperson. One way you can think of it in terms of distinctive features, which are the characteristics of a phone. Some features make a phone more sonorant than others. The ¿best? way to think of sonority is *probably* in terms of acoustics. Unfortunately, I'm not an acoustic phonetician (though I'd like to dive deep into it someday); but if I had to try to *hazard* an acoustic definition of sonority, I'd perhaps say that the more well-defined and steady the formants are, the more sonorant a sound is. If there's an acoustic phonetician out there, *please* correct me if I'm missing something with this definition. The loudness that the author said was the defining characteristic of sonority is at worst more of a side-effect of sonority, and at best just one small part of the puzzle.
If you want to quickly grok the sonority hierarchy, you can essentially do so by going row-by-row in the IPA chart; the rows are the manners of articulation, and they're mostly ordered by sonority (though this is not true for some rows, such as the laterals, which aren't more sonorant than their unlateralized counterparts).
An additional, though very minor point, is that the author exclusively uses "Sonority Sequencing Principle" in places where he meant "sonority hierarchy" or just "sonority" This is really not at all a *real* problem, since people can figure things out from context, or just reduce everything to the word "sonority"; but I bring this up because it's one of many tell-tale signs that the author is inexperienced with the subject matter. Which, I want to stress that that is fine; we're allowed to go outside our fields of expertise (Good heavens, imagine if we couldn't!). But what was covered was lackluster, akin to being tutored by someone who is still, themselves, learning the material they are trying to tutor you on.
The author also had this idea that fewer syllables means faster communication, but per my understanding, this isn't true. I don't have a study off-hand to point to, but my recollection is that the rate of information transfer during human speech is consistent regardless of how syllable-laden the language is; that is to say: languages with more syllables are simply articulated faster than languages with fewer syllables. As an English speaker, you may have experienced this phenomenon when hearing Spanish spoken: it *sounds* really fast, because each syllabe really is being pronounced faster; but Spanish words have on average so many syllables that they aren't actually communicating more-rapidly than you with your less-syllabic English. The language faculties of the brain can only handle so much information at once, regardless of how quickly your mouth is able to move.
In any case: these were the main things I remember having noticed when I watched it a couple days ago.
Please don't read this as a total condemnation of the author, because it isn't; I'm just answering your question about why I felt the linguistical side to this video was so lackluster. And hence, why I was so underwhelmed by it, despite finding the software part quite cool.
This is such a creative and fun to watch art style. Keep up the good work !!!
Content is great and the way you present it is very engaging, made me curious about a language that doesn't even exist! And your voice is very soothing. Thanks for the interesting video and please keep making more, if possible.
In a weird way, I find it neat how this highlights the importance of root words and multisyllabic words in general. A word that's based one or multiple others has an easily identifiable meaning, even if someone's never heard that specific word before. It eliminates the need to memorize unique sounds for _every single_ word, and instead allows things to build on each other, and - in turn - build on a person's prior knowledge of the language.
conceptualization and integration instead of deconstruction?
aristotle instead of plato?
in our modern culture?
Exactly. Imagine what a nightmare med school would be if every word for every bone, muscle, and organ was totally unique
This is so well animated and it's so cool? How the hell has this not caught on yet
It's only one day old.
No i mean, like, they only have 7K subs
It realistically is my fault for not checking the channel to see their previous videos (of which there is ONE)
dats my bad@@YamamotoTV2021
this has been out for 20 hours my guy itll catch on
YEAH I REALISED OK@@dant3838
The video is less than one day old if that’s what you’re referring to
This is the most perfect educational(?) video I have ever seen. Also the animation is really appealing and almost took up my entire focus lol
Love your videos!! I was sad to find that there are only 2 on your channel. Hope you keep making these, they really are very interesting.
There's a bug in your application. If you enter a word such as "multi-directional", the second word doesn't get translated at all, compared to when it was separated with spaces. I suggest you to use a tokeniser to pass on non-word characters unchanged and capture groups of word characters, instead of whatever you did to capture only the first half of a hyphenised compound word ;)
Great catch; that's sure to improve the quality of the output!
In my opinion, I don't think any "fancy"/actual implementations of a tokenizer (lexical/probabilistic) would be needed for this specific issue, mainly because a simple regular expression pattern would be able to handle most of the cases we'd want to (e.g., "thirty-five"), such as with a first basic pass: /\w+/g or /(\w+)/g for the capture groups, as you mentioned. (Use whatever modifiers you want, like m, etc.)
This pattern doesn't handle numeric digits, but I considered digits regardless since they're out of the scope anyway.
If we cared more for semantics or understanding, moving towards more complex natural language processing techniques would surely help to improve the sound/tone/flow/etc. Technically, it would offer much-improved consistency, but that’s a much more substantial change for a more complex, different problem. I'm not overly familiar with linguistics, though, so I might be missing some more significant pieces from my ignorance or lack of experience, where my understanding is rougher and causes me to make more assumptions.
Edit: I just saw you put in a PR! I appreciate your diligence!
This video is an incredible insight into language and how it works fundamentally. I'd like to see more come from this, it seems like a solid foundation for a project!
I know right, I was in awe after watching this
Well... kind of. More precisely, into how a single, rather simple (and not universally applicable) concept works xD
antidisestablishmentarianism
Well, actually, nothing is bad to have some two-syllables or even more-syllables words. They introduce some variations in the prosodical patterns making the language speaking more of a pleasant occupation.
one problem: many words in english have multiple roots like pterodactyl, it would make less sense to invent a new syllable than to make the one word into two (calling them wing-fings)
i think the english language already has enough monosyllabic words to do this, there's this game called poetry for neanderthals where you have to describe a word or concept on a card using 1 syllable words, when you practice it you can get very quick at it
Absolutely in love with wing-fings
Wing-fings, it wings and fings.
Did you know that helicopter is not made up of Heli and Copter, but actually helico (like helix) and pter (like pterodactyl)? I found this out recently and thought it was fascinating.
So... Twist-Wing?
Yeah, even so-called monosyllabic languages like 漢語 and tiếng Việt are filled to the brim with multisyllabic words like 自己 and hạnh phúc - they just happen to have 1:1 morpheme to syllable ratio. I know it somewhat defeats the point to make words like these, but it avoids consonant clusters that are not only hard to pronounce for english speakers but also likely to evolve into multisyllabic words in the future, and besides, you already let thirty-five slide, right?
Such a unique presentation style! The hands give everyone something to look at in the same way pointers work on slides, but their versatility in doing other gestures is super cool.
Good sir, I completely nerded out to this video. It was like hearing about Toki pona for the first time. What a neat idea! I’ll definitely be looking at Glish some more.
This is genuinely really fascinating!
i don't know if you'll ever see this but this is one of the most creative videos ive seen in such a long time on youtube, i mean the animation, the presentation of your hands through those cute hands, the topic, the delivery, the writing, all of it is FLAWLESS and im dumbfounded how you only have 10k subscribers when you should be closing in on 8 figures with ths level of content.
well, if you check the videos this channel posted you will see there are a total of 2 videos, this one and another one posted 2 years ago, if he continues posting videos like this one at a reasonable rate I believe he would have much more subscribers.
Bro is the irl newspeak engineer 💀
German chilling with the Rindfleischettiketierungsüberwachungsaufgabenübertragungsgesetz
the dislikes are from english teachers
Amazing video, I love that this has some literal parallels to stenography. Seeing the graph you made to assign monosyllabic words was kind of cool because it is LITERALLY the steno keyboard layout. Syllables add up.
The sonority principle is not the only thing. It is too general to be descriptive of English.
Some languages disallow certain syllable structures. Like a lot of asian languages follow strict CV (consonant-vowel) structure. So there is always a vowel after a consonant.
On the other hand here is the Czech word for wolf: "vlk". Yes it doesn't have vowels, but the "L" kinda functions like one.
English also has theese things. For example in old English the sound "g" as in "good" began to shift to the sound "g" (as in "gene") before front vowels, like the one in "gene" or "green".
This btw is the basis of the whole gif/gif debate.
Aaaand you didn't use the principle
@@norude If I'm getting the principle right, then the odd words are rahrmp, pesps, and slirnjd, right?
@@kmr_tl4509 google English phonotactics
@@kmr_tl4509 No I believe those all follow the principle correctly. pesps isn't really correct due to the sps part at the end since s should have greater sonority than p. However, english typically allows s to occur after or before voiceless plosives like p even if it breaks the sonority sequencing principle so I would say it's not really wrong in english.
Note that "English" does not always have a /g/; for many speakers, it's just the velar nasal.
1:09 the hand cursor was a suuuuuper awesome anchor. brilliant. ingenuine (:
Your style is super unique, I love it!
The attention to detail in this video is crazy, this man even animated the typing hands towards the beginning to match the real keystrokes. Great job 👍
I love that "mapping" is "mip". MIP mapping is a common technique in computer graphics to avoid aliasing. It means _multum in parvo_, or "much in little".
This is a gold mine of search terms and information about phonetics. I wish I had this video a year ago.
This video has made me realize that the English language is like a completely trashed and barely functional cruise ship with a metric ton of duct tape keeping it afloat.
10:52 dude just incarnated as the bastard child of Shakespeare and Dr. Seuss
I think it would help a lot to make it much more intelligible if you at least tried to force letters to have the same phoneme order. I know your directed graph is supposed to somewhat do this, but it allows consonants that are in the first syllable to be the coda of the Glish word. If you had your directed graph instead take 2 lists of IPA symbols (one for the onset of the syllables and one for the codas, it might make some easier to understand words.
For example: Problem -> Rahrmp. While all these sounds appear in the word. The P in "Problem" going to the end makes it arguably harder to distinguish than if it was left out entirely, Similar for the R that appears there too. Despite them being phonemes that are in the original word. If instead you initially restricted the /p/, /r/, and maybe /b/ to be available in the onset and maybe the /b/, /l/ and /m/ to be available in the coda, it might generate some easier to read mappings. Along with following your strategy of relaxing these restrictions as the generator fails.
The problem I see with the generated words is that while they may have the same set of sounds, they are so jumbled up (and missing some) that it makes it impossible to try to guess. So anything that could somehow restrict the order of the sounds to be more like the original would help a lot for readability.
very true! That would likely help out. I actually have a TODO for that in my code that I never got to 😅 Though additional restrictions like this do make it harder to generate as it runs out of valid syllables a bit faster, so it may weirdly lead to more words having strange mappings. Though hopefully that would skew towards uncommon words, and it's probably still worth doing.
@@paralogical-dev of course, it’s easier said than done. Really enjoyed the video though, it was a fun idea!
@@paralogical-devto help with that problem, maybe it could be added as another stage before the existing "try to form a syllable from existing sounds" one, so that (in theory) it never makes a worse mapping than the existing code does
@@paralogical-dev i made this:
ABEFGHIKLMNOPRSTUVWY
i calling it the park alphabet
it’d also help to generate words in multiple stages. the first pass would generate a set of potential words that could all be considered “good” mappings. the second pass would be about deciding which words get to reserve each syllable. this would solve cases where a word with 2 good mappings picks one that collides with many other words down the road when its other choice didn’t have any collisions at all.
I love the animation style!
Your animations look really cool and unique!
I enjoy that this is a light hearted topic, and the video is presented as being some what silly. All while being a lowkey introduction and overview of some of the most powerful ontological/linguistic tools there are in computer science, and even some of the more complicated computer science concepts like graph theory. This is the type of video that keeps me coming back to youtube. Subscribed.
Here I am, thinking I just discovered a new awesome youtube channel and I'm about to binge hours of funny educational content. Now imagine my despair realizing there are only 2 videos. Please please please make more, this was awesome!
I just had this same moment lol
This video style with the cursor hand is so absolutely charming
this channel is an enigma wtf
I love your unique style of presentation. Very inspiring and fun to watch!
This is so cool!! Amazing animation btw. Love the original art. Also you’re quite funny
i got a grammarly ad before this 💀
"meedhd" has major Gaelic energy.
I just wanted to say how amazing this video is!
I love the hand cursors and the animations are so smooth.
Your explanations are amazing!
Thank you for making this and you deserve more support from the UA-cam algorithm.
❤
This is basically how anyone under the age of 30 already communicates
Nice vid dude, it was sleyez kming. Keep up the good work!
P.s. dayps dwitsts sound nahrj flawlm stunt heyeh.
My very own splawm preft.
This ended up sounding like the kind of random gibberish you'd expect to hear in an improvised Rick and Morty skit. And I kind of love it.
The art style is really unique and fluid!
(The art style is lih neek and floodhd!)
When I was a kid, the clapping method for syllables always confused me. It took an interest in linguistics to finally realize the vowels were what you were really counting.
This is made even more obvious in Japanese. In Japanese, people often count each hiragana/katakana character as a single syllable. Japanese has no lone consonant sounds except for "nn." This is also why their accents do stuff like "MaKuDoNaRuDo" and "hotto doggu." Notice how they inject vowels into everything even when there are none.
no, their english sounds like that because of their phonotactics. in japanese, the largest syllable possible is a consonant plus a "y" sound, plus a vowel, plus an ending n.
japanese doesn't have syllables... it's useless to analyse it with syllables when it is based around morae
@@cordeaux the *timing* is said to follow mora but syllabic analysis still works and especially useful if you want to compare the number of syllables between languages
Sort of! Japanese people don't count syllables like we do in English (they count mora which is the minimum sound unit) but syllables are a universal principle of language so they still have them. The reason they inject vowels in places is because Japanese has different rules than English for building a syllable. In English we are allowed to have complex consonant clusters or 3 or even 4 consonants together e.g STROBES (s-t-r-oh-b-z) or TWELFTHS (t-w-e-l-f-th-s). Japanese syllables on the other hand are not allowed consonant clusters and have to conform to a (consonant)-vowel-(N) structure. So STROBES would become su-to-roo-bu-zu. Notice that they retain the long vowel length of the English diphthong and that the final consonant is a Z not an S. So they are basing the transliteration on the pronunciation not the English spelling. When spoken however some of the vowels can become muted to the point that they sound like they are omitted. Su-to-roo-bu-zu could be spoken as sto-roob-zu which is 3 syllables.
@@pb7199 I understand though I could argue against syllables being universal to every language - the need for a syllable to have a nucleus kind of disallows some constructions made in Nuxalk (and other languages) with the existence of the phrase clhp'xwlhtlhplhhskwts'
[xɬpʼχʷɬtʰɬpʰɬːskʷʰt͡sʼ], without a single sonorant (though, even though it is definitely not conceptualised in syllables, syllables with voiceless vowels as nuclei - as in the u in japanese tsuki, suki, desu, and other words - do exist and could then explain this as being syllables with fricative nuclei - as in some sino-tibetan languages like Yi/nuoso - but voiceless)). But, I can find one example of a word in nuxalk which does not support this - qʼtʰ - which does not have a true syllable nucleus. My understanding of syllables is that they are not innate but a human way of understanding speech - just as word boundaries are not innate to the speech content we produce. If voiceless fricatives can constitute the nucleus of a syllable, there’s no reason why a word like strengths couldnt be analysed as bisyllabic streng kths besides from preconceived notions of phonology and syllables in english (the same goes for some pronunciations of perpetually, where the “per” is pronounced solely as an aspirated voiceless bilabial plosive - does the syllable nucleus become the glottal fricative or technically a voiceless schwa (even though english speakers might not perceive it that way) or is there now a syllable without a nucleus in that word?
it felt like a stroke reading the end result
Great vid, you have done a stal job my friend.
0:28 I'm very disappointed that the Vsauce music didn't start playing.
Very surprised to see this channel have only two videos. Definetly subscribed. Looking forward for more!
Out here making newspeak without the dystopian edge 😆
You,
Have successfully predicted the future of Internet slang
3:52 is my favourite part of this video.
Duckduckduckduckduck
Actually, with duck, it’s pronounced
Duckduckduck duck duckduck dduckk-
The order of vs and o is as follows
Duck duck duck
8:21 I like shnek. It's a good word.
I don't know how YT started playing this video, I don't have automatic play on and I don't follow you, I was on a different tab, but your presentation style hooked me in. You have some real skills. I loved this video.
This is some great production value wow
This is amazing. Great job and I love the idea of the talking hands; So expressive
You have the coolest and most unique video style man, those hands are such an interesting addition
You should see a conlang called Xdi8 or Shidinn, a conlang based on Chinese but made many characters multisyllable in order to remove all homophones from Chinese. It has also a few translators on the internet. It's a «mirrored» version of your idea, and in many ways very similar (or right opposite) to your conlang idea!
"resembles" -> "berzdzdz" is one of the funniest things
4:52 apparently most languages just don't have spelling bees because unlike english, their spelling actually makes sense lmao so it's more or less trivial
fun fact I learned thanks 👏
It was always fun watching movies based on english culture and having characters fail at spelling bees when the movie itself is dubbed to spanish
what a cool project! I kinda want to see a variant where you allow 2 syllable words. It won't be as efficient but would make mappings more recognizable. You can call it two-glish lol.
This is what they should use for Sims 5
"I love it here. Everyone's so numbed by the sun that if you use a three-syllable word, they think you're a professor."
- Sun Tzu, The Art of War
You could certainly tutor game developers into creating more pronounceable procedural names.
Delightful. For glish+ try a morpheme matching step and a standardized ending set. Gish++ contractions, Gish+++ maximize information per syllable. Btw you are in good company Shanon has a paper on this and so did Chomsky on info of language.
This feels like a linguistics major's class project but I'm so here for it
I really like your editing style
You built an anticipation through this video by focusing on the requirements and goals for 2/3 of it before you introduced that a computer was going to do the work. This did a great job at breaking down the problem and making it very understandable. Saying that a computer will solve the problem with code earlier would’ve somehow reduced the quality of the video in a way I don’t know if I could describe well. I am a programmer too for context.
Learning Japanese turned my understanding of syllables on its head. Consonant contractions are considered multiple syllables in disguise, and the end consonant would also count a syllable, so GLISH would be 3 syllables, GU LI SHI. But now I understand why it's so hard to tell when one word starts and ends in Japanese. They don't really use the sonority triangles so it all just runs together...
SHU
The sonority sequencing principle applies to Japanese just as well.
One very important distinction is that between morae (singular: mora) and syllables. Gurishi would be a three-mora word, but could be pronounced with one, two, or three syllables.
Two other examples: "Hakken" and "Nippon" - Each 4 morae, but 2 syllables
This is also where the idea of a haiku poem becomes very different in English compared to Japanese. In Japanese haiku, you count the morae.
Dutch is similar in that a sentence sounds like one long word
Japanese music lyrics can also just decide how many syllables a word needs to be to sound good. "Konnichiwa" can be 4 syllables or stretched out to 5 like "Ko-n-ii-chi-wa" if it would fit better. You can't really do this in English and make it sound good. It's probably part of the reason music translated from Japanese sounds so off to me. Maybe it sounds that way to a native JP speaker too, but I'm not good enough to tell.
@@ExSheriffFattyBoySkinnyArmsDat klopt.
right after this video ended, i got an ad where a guy was speaking in an accent so thick i legitimately thought he was speaking glish for a sec
This video is soooo good it's by now the most helpful and inspiring linguistic video about English phonology I have ever seen. It answered the question in my head that even myself can't figure out what I'm actually asking. Absolutely creative!!!
This vid is soooo good it's by now the most fluh and spirnk skwuung vid bowft glish soulnt I have vevdz seen. It snerrd the usk in my head that nawrnv fleh can't frig out what I'm wul angsk. sutsts kriv!!!
The primary problem with aligning orthography with phonology is that you project accents onto written text, i.e. two words that mean the same thing are not spelled the same if one is written by an American and another by an Australian.
There have been several serious attempts at this in the past, both with IPA and, specifically for English, Shavian and Quickscript. They look cool, but don't catch on for this reason; it becomes really hard to read at speed if you have to know a different sequence of characters for different accents.
Imo, I actually like the other extreme: ideographic scripts, precisely for this reason.
Just get rid of accents then.
@@kyurenm5334 what does that even mean
He did say he was only looking at North American English
@@saoirsedeltufo7436 It simply means unification of spoken language so there is no difference in pronunciation between regions where it's spoken. Then using alphabet that is phonetically correct won't cause issues and will set these changes in stone.
@@kyurenm5334 and how exactly do you propose achieving that?
This reminds me so much of one of those kid languages taken to the extreme and fledged out. I can't be the only one that had a code language with some of my friends in school. Soon, if not now, kids will have AI assist them in fully developing it. Never occurred to me that AI may have an impact on language far quicker than traditional shift.
only 18k subs? that's criminal! this is the best video i've seen in a while!
This is a great way to generate character names for my goblins.
This was a beautiful experiment and as a former linguistics major I thought it was really fun!
I think one thing you might want to consider is how to treat grammar. Bubble, bubbles & bubbled are not really 3 separate words, in practice we understand it is the same word just conjugated for different tenses. This is because humans do not rote learn every individual word and its separate forms, we learn root words and grammatical patterns. There's no problem with treating them as unique words but unless the shortened syllables also have a consistent grammatical pattern, doing so would in effect remove certain grammatical functions from Glish that exist in English. Perhaps you could try a different approach where only the root of each word would become a single syllable and grammatical particles are tacked onto the end like in Korean.
One other thing to consider is whether or not you want to remain within the phonotactic constraints of English. Many of the generated syllables in your examples do not follow English constraints, like "meedhd" for "thirty", or "blegd" for "bubbled". If you remain within English constraints your directed graphs would be a lot smaller but so would your pool of legal syllables.
Conjugation strictly refers to verbs. Only bubbled is a conjugation of the verb "to bubble" (simple past or past participle for the perfect). If it can also be used as an adjective, in which case it is a derivation of bubble, but still considered a separate word even while containing the free morpheme bubble. Only "bubble" and "bubbles" are the same word as they are both a noun inflected for number.
I would say "blegd" is a perfectly reasonable English syllable. We have the one-syllable words "blogged" and "begged", just smoosh them together and you have "blegged".
Your voice sounds exactly like my friend's voice and it makes me smile hearing it
this is actually crazy dude, well done
Absolutely awesome video dude. Not only a talented engineer but a great animator and presenter too. This was really interesting