Goal: Make a version of English where every word is only one syllable

Inputs:

• words by frequency (optimize monosyllabification for more common words) inputs/word_frequency.txt
• words with pronunciations and split by syllables (CMU Dict syllablized) Note: multiple valid pronunciations for any given word, but all American english

Stages:

• syllablize.ts → convert CMU dict to JSON mapping of word → IPA split by syllables
• main.ts → load IPA syllables and generate new monosyllabic version of all words
• sonorityGraph.ts → data structure that helps generate new syllables following sonority sequencing.
• respellIPA.ts → convert IPA back into "readable" latin alphabet.

To run code to generate Glish language mapping,

• ts-node syllablize.ts to generate outputs/syllablizedIPA.json + syllableGraph + big list of randomly generated syllables
• ts-node main.ts to generate outputs/monosyllabic.json & other monosyllabic results

To run UI,

• cd ui
• npm run dev

Features: (enabled via: ts-node main.ts --features <feature1> <feature2> ... <featureN>)

• homonyms: Normally, words are assigned Glish syllables in an injective way. So, when a new word would be assigned to a syllable that has already been mapped to by a previously-assigned word, the new word is instead assigned to a worse syllable. This feature skips this check, allowing a non-injective mapping of words to syllables i.e. allowing Glish to have homonyms. This variant of Glish can be called "Ing" (as that's what "English" is translated to in this variant). The idea here is to trade distinctness of syllables for more ideal word-to-syllable correspondence. This is probably a worthwhile trade-off since English and many other language already trade semantics-to-(written or spoken)word injectivity for shorter/simpler words.