more vowels than I feel comfortable pronouncing
Hah, yes, I realized most of you don't care about conlang, but I'm trying something cool and I wanted to document it.
Usually when I conlang, I have trouble generating affixes: I have trouble picking affixes that fit the sound I'm looking for, and I'm never sure if I've achieved the right balance of regularity and quirkiness. So for this project, I used IPP to generate a long list of polysyllabic words that I though sounded good. Then I cut all of them down to the last syllable, and in some cases the rhyme of the last syllable, and made a frequency table to look for patterns. Some of them are obvious (all the vowels are represented) but there's a few I didn't expect would occur so often.
Also, I figured out my main issue with IPP-generated words, though I'm not entirely sure how to fix it. Right now I've got it set up to generate words by calling a syllable table, and the syllable table calls seperate tables for onset, nucleus and coda. The coda table includes subtable calls for final clusters. This means that I occasionally get three-consonant sequences in the middle of a word, and I don't want that. Problem is, as the table is structured now, each syllable is generated independently; they can't "see" each other. Concievably I could fix this using the dice-rolling mechanism built into IPP, but I'd need to set up some kind of parameter-passing mechanism so that, say, if the coda table roll for syllable N is greater than X, where X is the lowest number needed to insert a cluster, then the nucleus for syllable N+1 is alway \z (IPP code for a zero value). I could probably do this in Ada if I tried hard enough, but I don't like the interface and don't want to end up generating the entire vocabulary of the language at once, so...
Usually when I conlang, I have trouble generating affixes: I have trouble picking affixes that fit the sound I'm looking for, and I'm never sure if I've achieved the right balance of regularity and quirkiness. So for this project, I used IPP to generate a long list of polysyllabic words that I though sounded good. Then I cut all of them down to the last syllable, and in some cases the rhyme of the last syllable, and made a frequency table to look for patterns. Some of them are obvious (all the vowels are represented) but there's a few I didn't expect would occur so often.
Also, I figured out my main issue with IPP-generated words, though I'm not entirely sure how to fix it. Right now I've got it set up to generate words by calling a syllable table, and the syllable table calls seperate tables for onset, nucleus and coda. The coda table includes subtable calls for final clusters. This means that I occasionally get three-consonant sequences in the middle of a word, and I don't want that. Problem is, as the table is structured now, each syllable is generated independently; they can't "see" each other. Concievably I could fix this using the dice-rolling mechanism built into IPP, but I'd need to set up some kind of parameter-passing mechanism so that, say, if the coda table roll for syllable N is greater than X, where X is the lowest number needed to insert a cluster, then the nucleus for syllable N+1 is alway \z (IPP code for a zero value). I could probably do this in Ada if I tried hard enough, but I don't like the interface and don't want to end up generating the entire vocabulary of the language at once, so...