(26-05-2026, 05:19 PM)Dunsel wrote: You are not allowed to view links. Register or Login to view.No, you're trying to make it look Voynich. Don't. Just try to make it look like a language. Create your own visually appealing language.
So, if I understand it correctly, you give me a set of rules (copy + mutate and the ledger) and you want me to add on top of this a set of my own preferences and then produce some text and you claim that this text would bear statistical resemblance to Voynichese no matter what my preferences are? This is clearly wrong. Suppose my preferences are: most common words should be 1-3 characters long and repeat often, there should be only a few of them, longer words should variate more but occasionally also repeat close to one another.
Seeds: daiin,sary,qokain,otedy,okol
Text: sar oko qokair oko daiir sar qotedy daiiny or qokainy sar oko san qotey o daiir otedysary oko saiin qotain or o sary daiir qotair sar oko otedy otedysar daiiny okol sary qotady sar oko okedy o daiir qokain daiin qokairy qotey o daiir otesy sar oko qokair.
And naturally, if I wanted to make it look like a real language, I'd make sure that "sar oko" and "or" and "o" keep repeating thought the whole text, while "ar" and "sa" do not appear often, because that's the way most (all?) languages work when it comes to short words. If you think about it, I have to fight against copy + mutate here, defaulting to plain copy in most cases and feeling unnecessary restricted by mutation rules to keep going, because copy + mutate is not a good way to create plausible looking language. Just creating a bunch of random common words and word patterns and sticking a new long random word now and then would be much easier.
Now, I think this looks much more similar to a real language than Voyncihese. Because most real languages have plenty of grammar words and particles that often repeat and have structures that also repeat. Languages like Latin have a lot of suffixes that repeat all over the place, etc. So naturally, if I wanted to imitate a language, I would make it look like a language and not like a weird codebook of number like sequences. Of course it goes without saying that statistically this won't look similar to Voynichese, and the distribution of tokens and types would be different.