Transcript for:
Creating a Monosyllabic Language

anti- disestablishmentarianism G yuck G English is so slow there's so many syllables all the time who thought this was a good idea okay here's the plan we make a new language based on English but we remove all the syllables well I guess we need some okay we remove all syllables except one from every word so everything is monosyllabic think about how fast we'll be able to talk without all those unnecessary syllables but what is a syllable no I'm asking does anyone know all I know is what I learned in elementary school where you clap while talking and each clap is a syllable Elementary uh five but I don't know why we know when to clap and at this point I'm a bit afraid to ask like I could clap while talking Elementary six seven who knows okay fine I'll look at up what is a syllable apparently it's all about vowel sounds was I just not paying enough attention in elementary school and everyone else already knows this uh every syllable has exactly one vowel sound and you can tack some extra consonants on top but it's not like anyone planned the English language everyone knows what a hodge podge it is so how can syllables be so neatly described well it's not an English language thing it's just a human language thing digging deeper it's all about the sority sequencing principle sounds scary but it's not each different type of sound we can make has a different inherent loudness to it keeping your speaking volume constant some sounds produce more air movement than others if we rank all the sounds we can make by how loud they are we see a trend appear vowel sounds spoken with a relatively open vocal tract are the loudest type of sounds the sority sequencing principle says that in practice human speech tends to follow a pattern of building up consonant loudness towards vow vow then back down away from them this forms little Triangles around vowels so in a word like blimp B sounds come before L come before I in our hierarchy then it goes back down to M and to P after the I vowel this isn't always true by the way but it's a general principle and humans sure love general principles Pop Quiz what's wrong with this madeup word you probably think that's dumb that can never be an English word it looks so wrong it's garbage you're an idiot but now you know why it's garbage this garbage word clearly violates the sority sequencing principle look the sority of this garbage word would look like a roller coaster instead of looking like a um Less Fun roller coaster this sequencing is what gives speech a sense of Rhythm the pattern of loudness gives a beat for speech which is why it matters for poetry and Shakespeare it helps distinguish each word from each other look all speech is just a set of triangles if it's not obvious yet a syllable is just one of these triangles and multi- syllabic words are just several triangles stap pull together so it's true that syllables are based on vowel sounds but the underlying reason is this sority sequencing the low points in the triangles are lower volume and therefore form natural breaks in words and the Cadence of speech now that we're all experts in syllables you can see why making everything monosyllabic will revolutionize the modern world look our speech pattern will be so optimal why use many word when few word do trick okay ground rules before we make everything monosyllabic one we want every word to be one syllable I guess that's the whole point two words should try to follow the sority sequencing principle so they're pronounceable three we don't want to introduce more homophones than already exist so no mapping two different words to the same syllable otherwise like what's the point every word is duck done it's all one syllable duck duck duck duck duck du duck duck duck four I'll be focusing on North American English for Simplicity though the same concept could be applied to other languages in dialects okay okay but as soon as we try to end multi- syllabic words we realize that English is a bit of a train derailment in terms of spelling like who would have thought that this is how you should spell thought if you say thought it's more like three parts th a so we're going to throw away letters and focus on sounds we can represent sounds with a special alphabet called IPA that has symbols for every different sound can make thought is spelled like this th frankly I think that's what the alphabet really should be this also helps us solve another problem when making new monosyllabic versions of Words we don't want to have two different words with the same spellings and not realize it well if we're always acting on sounds instead of letters we don't have to care about spelling cancel the spelling tests boycott spelling bees okay we're ready to do the thing for every word in English that's more than one syllable we want to assign it a new unique monic form there's many ways to make such an assignment and there's only so many single syllables we can make so we're going to have to make some decisions in order to make this monosyllabic English actually good here's the plan step one get a list of all the words in English plus a mapping of words to their IPA equivalent split into syllables step two leave all the monosyllabic words alone if it ain't broke don't fix it step three sort the list of words by their frequency of use we want to prioritize assigning good and short syllables to words that people actually use no sense assigning a nice easy syllable like ver to First similitude rather than over if we sort words by frequency and make our assignments in order one by one then all the good syllables will get used up first and the nonsense words like avuncular can get the scraps step four let's start actually assigning words and uh this is the hard part our first strategy will be to take existing syllables from the word and use it directly as long long as it's not already a word or already been assigned let's call this plan a for example changes it has two syllables chain and JZ J is not a word yet so let's just make J be the monosyllabic version of changes this is great when it works but this only works for a couple thousand words now that J has been used we can't use it for Pages or images Etc if that doesn't work our next strategy will be to generate a new syllable based on the Sounds in the word to do this we need to be able to generate new random english-like syllables we want this to follow the sority sequencing principle we talked about this gets a little complicated to Generate random syllables we'll build a data structure called a directed graph it's a fancy name for something simple some IPA letters with arrows going between them what we want is a way to tell which sounds come after which other sounds in real English words to build this we'll go through every sound in every syllable in every word in English I think it's easiest to see with an example if we were looking at the word bat we'd see that the B sound is followed by an a sound so we draw an arrow between b and a in the directed graph then we see that the a is followed by T so we draw another arrow between a and t if we see the same pair of letters again like a to T in Rat we make the at Arrow a little thicker as we continue with every other PA of sounds in every other word we end up building a web of how sounds connect to each other this web entirely encodes the sority sequencing principle in English without us actually explaining to the computer how that works there's an edge from B to R but not from R Tob now that we have this fancy graph we can randomly Traverse between sounds weighted by how often different edges are used so for example at the start we pick a random sound that we've seen at the start of a syllable then we look at all the out going arrows from that sound in our directed graph we pick one randomly but weighted by the size of that Arrow then we just keep going picking random arrows until we reach the end of the syllable this pretty magically generates syllables that kind of look like the real plausible English syllables spling kinch FS drith schneck glid the best part is you can just generate thousands and thousands of these things sping you could also use this approach to learn what letters link to others between Ables and how often which would let you generate fake multiable words but we won't bother with that here with this fancy graph at our disposal we can make better monosyllabic assignments if plan a doesn't work then plan B is to do this random syllable generation but limit our pallet of possible sounds to ones that already exist in the word for example we take all the sounds in the word color C we throw away all the other letters in our graph except for these then randomly generate a syllable as before with this pallet of sounds for example we might get CLA this works pretty well but eventually this also starts to break down so plan C is to just generate a metric ton of random syllables of any combination of letters and see if any kind of match up with the original words letters so you might get a good match like cannibalism as kibs or you might get a garbage mapping like bread box as scrud worst case scenario none of the letters line up and we just pick a random syllable and assign it to the word luckily this only has to happen on very rare words after our good strategies fail side note there's one small thing we can improve slightly there's a lot of words that are very similar to other words like bubble and Bubbles and bubbled these are sometimes called allomorphs which would also be a totally badass name for a space creature made out of a metal alloy that can morph anyway when generating random syllables we can check if the sority graph we generated is is okay with us adding any extra sounds like du Z to represent these variants when assigning syllables we can try to match up these syllables with variants first when possible so that bubbles and bubbling aren't totally different this only helps a little because it doesn't work every time and that's it it actually finds a mapping for every word with this approach we'll call this new language glish which incidentally is how you say English English English English English so how does it do let's take a look at a sample paragraph the L ramp left is what we should use as a spom PR iaps we could just make up some draw words oh hey what B's this pre that I'm saying right now wait I've Dre been saying this in glit with zings this pre has Meed five EP Earth sibs than the sled um okay right it kind of does the thing the multiable words do in fact become new single syllables some words are not bad like sibs for syllable pre for paragraph some you could believe if you squint really hard like Dre for all r d and bed for about some are bad like Meed for 30 like WTF is that well I guess it technically matched my requirements so I'll give it a B minus and call it a a it does generally reduce the number of syllables in a paragraph by 30% I suppose it even has some kind of pro to it I was kind of hoping it would be mostly intelligible for English speakers without having to learn a bunch of new words but that's not exactly the case there's just too many words to make a mapping that's immediately obvious and unambiguous at least with my simple method most individual word mapping seem reasonable but it's not obvious what any given glish word actually means in English it does really help that most words are already monosyllabic I think of it like a sister language to English full of idiosyncrasies that you just have to memorize just like English I made a translator tool that you can type into and it spits out glish it converts the IPA symbols back into something that vaguely resembles English spelling so it's a bit easier to read the code is also available online now we just need dual lingo for glish