> Why is it so hard, and why hasn't anybody done it before? The whole problem seems trivial to me. You need a language where

> - first you need your words to have only one meaning

There's your answer. There is no such language, there are no such words. It's questionable if it's even possible - to me, it sounds like a category error. "Words" aren't keys in a hash table, and "meaning" isn't a discrete thing.

> now you have a language without any ambiguity. If there are some ambiguity in the objects/concepts that you are referring, then use your language more to specify the object.

Thing is, I don't think ambiguity comes from language. Concepts and categories are naturally fuzzy for us. Precision is a continuum.

If you need to express something more precisely than your language allows with single word or term, you just create a new word and attach a long-form description to it to position and connect it to other words in the latent^H^H^H^H^H^H^H concept space. That's how we always did it, and it was never hard.

I still don't get it. In the linguistic part the only problem is to make to word->object mapping unambiguous. And it is trivial problem. If the world is fuzzy (i don't think so) then is not a problem in linguistics.

Yet linguists seem to aim for a language that eliminates ambiguity (see TFA). Where am I wrong?

theoretically you can introduce new words to be more precise, for example eskimos have many words for snow, from (1) that is because they use polysynthesis: a base word is attached to many different suffixes which change its meaning. So, while in the English language we might have a sentence describing snow, fusional languages such as the Eskimo-Aleut family will have long, complex words.

So, you may solve the ambiguity problem by introducing a huge numbers of words, that could help a LLM, but humans need a relatively small vocabulary tailored to everyday use. So the tradeoff is reducing ambiguity and not increasing a lot the vocabulary.

(1) https://readable.com/blog/do-inuits-really-have-50-words-for...

A more common example is the meme about women and men colors. Women have magenta and salmon, while men call both pink. This has not much to do with linguistics or eliminating ambiguity.

edit: this was a strawmanning, colors are a non-issue. But there can be situations where there are 2 objects, you say a word to refer one of the objects, and your partner thinks about the other object, and neither of you think that it was ambiguous, so you don't clarify until it's too late. Like there is a dead bear and a bear cub, you say 'bring the bear' and you mean the cub, and they mean the dead bear.

Virtually every word has edge cases and gray areas, so there can be no 1:1 correlation between language and the world. Think of something as basic as "chair." Is a love seat a couch or a chair, or both? What about a bench? A stool? An ottoman? A small table? A tree stump? Reasonable people will give different answers, and some answers will be nonbinary ("it's very chairlike but not strictly a chair").

Once thought is encoded in language, it's in a lossy symbolic domain. It's in a messy shared vector space. And every human decoder unpacks the embeddings differently.

The bear... no, not that bear, the other bear--yes, I know it's a dead bear... what, you call that a corpse and not a bear? Okay, the new corpse of a former bear, the one that is right over there...

When does a bear stop (or start) being a bear? The 'bear goo' quickly boils down to "what you believe in" and that (surprise!) people have different views, even if we restrict "bear" to only living individuals of the family Ursidae--is it a bear from the time of conception, or even before that, or maybe only at or some time after birth, or always, or never?

Of course you could be more specific--pa vizyci'ojvemrojvesofyjvericparcribe, for example--but now it takes forever to say anything, and you'll lose some of the audience--"the magistrates however replied that they had forgotten the things which had been spoken at the beginning"--and you can probably still find someone for whom the "nearby young and dead and soviet and tree-climbing bear" is still a bit too fuzzy, or anyways there will be people who will disagree with (or ignore) whatever arbitrary definition of bear you've cooked up. Maybe there are two such bears, or they'll quibble that a corpse is clearly not a bear, or they'll happily modify your oh-so-precise definition to mean something else, if you know what I mean, nudge nudge wink wink say no more because The Algorithm plusgoods discussions about bears, and so they use 'bear' to talk about something else while still collecting all those sweet, sweet plusgoods.

Also, lojban does not eliminate ambiguity, zo'e co'e be do'e zu'i, but we're za'o into nunribypevykreborvi'u. A shaved bear would be less fuzzy?

The difficulty in languages is not in having good features, but getting people to use the language at all. Existing human languages mostly solved the adoption problem because you want to speak the same language as the people around you. I don't think Lojban has reached that critical mass yet.

Programming languages have other ways of increasing adoption, such as offering libraries that are tightly coupled with that language. If there are interesting literature that are only available exclusively in Lojban, that might have been a good motivation to learn the language, assuming that they can compete with the heaps of existing work written in languages we already knew.