cross-posted from: https://lemmy.world/post/44425631

Here’s a minimal multilingual conlang database idea.

All words in all languages use the same lexeme table and the same pool of IDs. There is a language column to indicate which language a word belongs to. The words and meanings are stored in separate tables and connected via a senses table. If two words from the same language share a meaning they’re synonyms. If two words from different languages share a meaning they can be translated back and forth.

The senses table has a rank column indicating how salient a particular meaning of a word is, with lower numbers being more salient. A sense with a rank of 1 is likely the meaning inferred by a listener without further context, and higher numbers refer to rarer senses of a word.

Importantly, lexemes and sememes can have zero or more entries in the senses table, meaning you can enter words or meanings on their own. Maybe you have a particular string of sounds but don’t know what they mean yet, or have a particular concept you want to have a word for but don’t know what that word is yet.

  • Venia Silente@lemmy.dbzer0.comEnglish
    1·
    5 days ago

    Oooh this would fit nicely with SQLite.

    This sounds interesting for working in the level of things such as designing elements of the language, but I feel like it lacks some important info (or a good solver) for once you get to the stage of actually “testing” a language. For example, unless I’m missing something obvious, you’d first have to know how to decompose a word into proper lexemes in the specific language you want before you are ever capable of asking what is a possible (or proper) word for a given meaning.

    We’ll be watching your progress with great interest if you implement a frontend for this!

    • early_riserOPEnglish
      2·
      5 days ago

      Not sure what you mean by “testing” a language, unless you mean testing it for completeness. I’m not aware of any conlanger who’s aiming for that level of expressiveness. At least I’m not. I make words as I need them. Sometimes it involves making multiple morphemes that I can reuse in other words later, sometimes I expand the meanings of an existing word, sometimes I make up a new word altogether. So I’m not “decomposing” the words, I’m building them up.

      For example, the word for “saboteur” in Commonthroat is sMpslPqmg. It contains three pre-existing morphemes. sMp to tread or step, slPq tail, and -m agentive suffix, literally tail treader.

      I can’t imagine making any kind of frontend for this. I’m bad at that sort of thing.