On 11 Nov 2015 22:09, "Logan Kearsley" <[log in to unmask]> wrote: > > Version 1 of the new word generator is now working! > You can get it from https://github.com/gliese1337/Logopoeist > Instructions are included. I haven't had time to investigate those instructions, but I wonder if I might describe the generator I had been about to work on with Alex, before the rest of our lives supervened. (This was for Jundian, an artlang I'd begun work on.) Each combinatorial element is manually assigned a (phonological) complexity value, and the complexity values are summed (by the combinatorial component of the generator) to give an entire form's complexity value. When I want a word, the generator picks one randomly following a probability curve whose axes are probability and complexity and whose peak and higher inflection point are complexity values specified by me. (This anyway is my layman's understanding of what I want.) In other words, I specify a value that is the most likely complexity of the word and another value that the word's complexity very probably won't exceed. Could the generator incorporate this? It seems to me that it requires the ability to specify complexity values in the input, plus the simple complexity summing, plus the complicated bit that generates words in a fancily probabilistic way. --And.