2 ms·
I have read the toki pona book. Its interesting but cheaty. You can't get rid of complexity by squeezing a balloon and pretending the enormous bulging out sha
by VLM 11y ago
I have read the toki pona book. Its interesting but cheaty.
You can't get rid of complexity by squeezing a balloon and pretending the enormous bulging out shared dictionary of nouns is "not really words".
English only has 26 letters, its really simple to learn English! Well, sure, sometimes you have to string together quite a few of those 26 letters in the correct order to get the point across, but its not really hard to learn English.
Likewise toki pona is an exercise in what amounts to stringing together long winded syllables.
It was disappointing, I was interested in the language with the idea of fooling around with natural language processing, which is complex, so a simpler language should allow easier to write simple toys. Perhaps I've found a silver bullet? Yet its actually a very complicated language because of the balloon analogy, you can't squeeze the complexity somewhere else and declare victory. It requires all kinds of memorization and life experience and humor and creativity and judgment calls about stringing together it's sylableWords into what in English would mostly be one word.
Even if its not a silver bullet, its still kind of cool. I ended up having more fun, while continuing to accomplish nothing, with a pidgin englishs. I ended up with a reasonable parser for butler english. I thought it would be funny to write a parser that turned butler english into coq and then replied back to me with what coq thought of my statements. Then I realized I was bored and moved on to something else. For at least awhile recreational natural language processing can be a lot of fun and is a wide open area where whatever you do, its probably a newly trodden path, at least if its weird enough. Here's a strange idea if I get back into the hobby. Can I make a butler english to toki pona automated translator? Hmm.
- sago 11y agoIt's interesting that you're not the only person here who's approached toki pona from a NLP perspective. toki pona is about the worst case for an NLP language, I would think. It puts all the communication burden on the semantics and pragmatics of the language. It doesn't make human expression, feelings, interests, creativity or community simpler, just the language. There are other conlangs designed to be unambiguous. toki pona was designed to be the opposite: maximally general. On the specific point of having to learn a compound lexicon - that is true a little, but not as much as you'd think. Sure there are conventional compounds for saying certain things, but you are very welcome (and aesthetically encouraged) to find other ways of expressing the same thing. You might find 'ilo nanpa' defined as a computer in a compound lexicon, and I use that sometimes, but my 'ilo nanpa' is also my 'ilo musi' and my 'ilo pali' (and sometimes my 'ilo pakala'). The key insight is that the language doesn't have to describe what a thing is, but what it means to you. That changes, it cannot be properly documented, and is almost impossible to parse. I take that as a feature, not a bug. Part of the pleasure of toki pona is the endless poetry of it: the joy of finding curious and evocative ways of expressing yourself.