An online demo and a link to an informative video are here:<p><a href="https://paralogical.dev/glish/" rel="nofollow">https://paralogical.dev/glish/</a><p>Input: South Korea has said it will resume propaganda broadcasts against North Korea for the first time in six years in response to Pyongyang's campaign of sending rubbish-filled balloons across the border.<p>Glish output: South kawrshk has said it will zoordz pran byerkstz genst North kawrshk for the first time in six years in spahns to yangz kayktn of sning brish-filled noorz skruh the derr.<p>Not very practical but a lot of fun.<p>More than a century ago, “words of one syllable” was a trend in publishing:<p><i>Robinson Crusoe in words of one syllable</i><p><a href="https://archive.org/details/robinsoncrusoeindefo/page/8/mode/2up" rel="nofollow">https://archive.org/details/robinsoncrusoeindefo/page/8/mode...</a><p><i>Aesop's fables in words of one syllable</i><p><a href="https://archive.org/details/aesopsfablesinwo00aeso/page/6/mode/2up" rel="nofollow">https://archive.org/details/aesopsfablesinwo00aeso/page/6/mo...</a><p><i>History of Japan in words of one syllable</i><p><a href="https://archive.org/details/historyofjapanin00smit/page/n9/mode/2up" rel="nofollow">https://archive.org/details/historyofjapanin00smit/page/n9/m...</a>
Conclusion from the video:<p>"""<p>Meed for 30 like WTF is<p>that well I guess it technically matched<p>my requirements so I'll give it a B<p>minus and call it a a it does generally<p>reduce the number of syllables in a<p>paragraph by 30%<p>I suppose it even has some kind of pro<p>to it I was kind of hoping it would be<p>mostly intelligible for English speakers<p>without having to learn a bunch of new<p>words but that's not exactly the case<p>there's just too many words to make a<p>mapping that's immediately obvious and<p>unambiguous at least with my simple<p>method most individual word mapping seem<p>"""<p>If the goal was to be intelligible, you could probably adapt the method to be more successful. For one, I would drop the requirement to not create more homophones. Since you're hoping people will guess what the word means based on what it sounds like, you can afford to have it sound like several things and let them distinguish based on context. Besides that, according to the established rules, you could turn a single multisyllabic word into multiple monosyllabic words. Suppose you turned "better" into "more good". This is technically monosyllabic, and immediately obvious in meaning to any english speaker. You might also squeeze multiple synonyms into a single word, for example "attempt" could be turned into "try", rather than attempting to come up with a legible one syllable variant of it. Doing these things wouldn't be as faithful to the seeming goals of the project, but the result might be more interesting in terms of semi-comprehensible english.
This reminds me of efforts like Esperanto. Which I hate. There's something very technocratic about constructing/reconstructing languages which feels like the following fallacy informs them:<p>* The language has redundancy in it<p>* I notice the redundancy and don't think it's useful<p>* Let's remove it without any sense that it may have had a purpose<p>It's a contemporary of eugenics as a concept, and Esperanto (and possibly things like Glish) have the same mouth-feel to me.<p>Different languages have evolved for different reasons/needs/purposes; the main thing about English to me is its vocabulary: the precision available in English due to our million+ words is the gift (and hell) of English. Getting rid of it feels wrong, and wrong-headed.<p>That said, Glish doesn't have a 'building a eugenically purer language' manifesto, instead the explainer video vibes more like 'nerd deep dive'. On those terms, building a directed graph of possible unused single syllables to automatically build out a full dictionary qualifies a 'sweet hack' to me. But, I won't stop hating the existence of this -- the risk of it taking off is small, but the impact on the world would be terrible.
I don't know if this is buggy or if the approach is just flawed, but there are a lot of missed opportunities for clearer mappings here. Another comment showed an example where 'response' maps to 'spahns' which is a correct USian pronunciation of the second syllable. But nowhere in the corpus is the word 'sponse' used - that would have been the natural choice here, surely? Nor is 'spons' used for 'sponsor' or anything else. So it seems like the random walk of the graph is being over-used here, even though the video talks of using substrings as a default. I think focusing purely on phonetics versus edit distance is a mistake and makes this less accessible than it could be.
Fantastic explanatory video by the author: <a href="https://youtu.be/sRbcw2sGkJw?si=USBnDyShTJNAeKVj" rel="nofollow">https://youtu.be/sRbcw2sGkJw?si=USBnDyShTJNAeKVj</a><p>It’s a shame that the repo docs are so sparse in comparison.
This reminds me of a project a friend of mine worked on, where he wrote a dictionary of all the English one syllable words, but defined only in terms of other one syllable words. For example "day" would be something like "time in which the sun goes round the earth". Sadly I don't think it's posted online anywhere.
I feel like this <i>really</i> stretches the definition of "syllable".<p>For example, "interesting" becomes "snirnktstsk". That feels like at least 2 syllables, if not 3, for the same reason "pspsps" (think the sound you make when trying to beckon a cat) should be considered 3 syllables.
Vietnmese is almost all mono syllabic. However it encodes extra information in tones.<p>So you could have bây, bảy, etc all meaning totally different things.
Maybe create a Google translate language that is only the monosyllabic words in English so output would be monosyllabic but understandable. So propaganda might become lies. The demo text from the Glish explorer page:<p>Type a sentence or paragraph in here in English to see its translation.<p>This is some additional example text to demonstrate.<p>might be translated as<p>Type UK text here to see its sense.<p>This is more test text to show.
The amount of syllables in a word does not always, and in the case of this language a considerable amount of time, correspond to how long it takes to say a word nor to how difficult the word is to pronounce.<p>I really hope the language was intended as nothing more than a curiosity.
As per the translation example shown ... the word "English" translates to "glish"?<p>Seems like a flaw.<p>INPUT "I understand English better than glish."<p>OUTPUT "I sadt glish begt than glish"<p>Meaning is fully lost in translation.
Reminds me only slightly of Votgil and this entertaining video:<p><a href="https://m.youtube.com/watch?v=12bT6wGXESc" rel="nofollow">https://m.youtube.com/watch?v=12bT6wGXESc</a>
I wonder how they defined a syllable. One example from the home page I saw was "shlayr", which seems like two syllables - of course depends on pronunciation. On a quick look, it seems they used a version of CMU pronunciation dictionary split into syllables, so I supposed that appeared somewhere in there.
Anyone remember Mad Ape Den[1] where you can only use words of three letters or less?<p>[1] <a href="https://www.metafilter.com/159399/May-Ten-is-Mad-Ape-Den-Day" rel="nofollow">https://www.metafilter.com/159399/May-Ten-is-Mad-Ape-Den-Day</a>