Conlang Design- Challenges I’m currently facing

It is about 30-Day-Conlang time again, so I’m back to the drawing board. Maybe I should follow this up with the rules of the 30-Day-Conlang Challenge.

Quick review: Goal wise, I had in mind a conlang that would be extendable, simple and small, phonetically rich, but learnable with readily available resources that could be used as a home-lang, i.e. a conlang that you could use to talk to your significant other and children while at home. I had planned extensions to be implemented via “layers” where the content words take on a new set of thematically related meanings.

Structure wise, I had in mind a language that borrows all of modern Icelandic’s phonotactics but none of the words or semantic nets (i.e. no loan words, no clusters of words mean sheep=wealth or such). By small, I mean about 500 words, but unlike toki pona, it should have “layers” of homonyms– that is all the content words would have a basic meaning, a Buddhist jargon meaning, and a sex and romance meaning, so “foo” might mean, chair, nirvana and “to kiss by rubbing noses”. The community, if there were to be one, could add additional layers as they saw fit, maybe a Star Wars layer, a culturally Icelandic layer, so “foo” could also mean “fermented shark” and “wookie”

Culture wise, this language would have culture with a small “c”, meaning the basic set of words would have generic and abstract words, plus some specific words for household items, plants and animals common in East Coast North American cities. Beyond that, I don’t plan to plan a culture- this isn’t an art lang for use in a novel and it isn’t an art lang of the same recipe without a companion novel. The real culture of a fake language is the culture of language hobbyists who interact over the internet, so the culture with a big C is that culture and I can’t dictate that. At best, I can provide some of the grammar and vocabulary for “polite” language that may or may not get picked up, compare that to the toki pona community where “jan” as a substitute for “Mr”/”Mrs” was picked up, but proposals and attempts to use “sina” vs “sina suli” or even “ona suli” in a Frenchy sort of tu/vous never caught on.

Grammar Design
On one hand, I’d like to include an exotic feature, such as lexical basic word order, i.e. the basic SVO word order would vary depending on the verb. Some *transitive* verbs would be SVO and some would be OVS, some SOV etc. Because the set of content words is closed, and orders would be assigned at random, we would have a language with lexical syntax, something that is probably unattested in natural languages. The interesting question being, if a competent user of this language were to be given a novel verb, what word order would he use?

On the other hand, I’d like to include some wordbuilding mechanisms. The most bang for the buck (most words created for a set of rules) are going to come from polysynthetic morphology. But as soon as you go down that path, syntax becomes far less important. One way to put it: there are so many things describing the situation in a polysynthetic word that the situation being described is clear, even if subject-object-verb are arranged in random order. Some exceptions to that generalization are situations where the subject and object are too similar, i.e. both inanimate, both male, both human, etc.

So I can’t really put polysynthesism and lexical basic word order into the same grammar. I think lexical word order is going to have to wait for another project. Polysynthesism is too valuable for a oligosynthetic language to give up.

Meta languages
Describing grammar is sort of hard because you need a grammar of grammars. I barely understand meta languages (I sort of internally understand BNF grammars since I wrote C# & SQL code at the office all day), but I don’t think I would want to limit myself to using the constructs that BNF can model. On the other hand, BNF looks like it could model most of the cases. One nod to machine parsing I would like is that “function” words be members of only one word class, i.e. it should be easy for a parser to find a function word and infer that a certain construction is going on. A counter example would be the pop band “The The”, which uses a function word (the first “the” as a content word, the second “the”) Another challenge is that the machine parsable grammar IDEs are aimed at computer program writing, and have a very steep learning curve–I fell like I’d need a four year computer science degree for them. I have no idea if anything analogous exists for logic or other branches of math that could be used to model communication.

This entry was posted in 30DayConlang, conlang design. Bookmark the permalink.

One Response to Conlang Design- Challenges I’m currently facing

  1. Kevin B says:

    I really like the idea of homonyms with different meanings in disparate semantic areas, and the idea of randomized word order depending on the verb. I may try to experiment with those myself someday!

    Overall, I’m getting a lot of neat ideas from your blog. Keep it up!