31 y/o white passing mixed w/ Black monster woman from the Netherlands. Artist (occasionally). Writer (again, every so often). Prone to camera sniffing behavior, one of them θΔ folk.

AD: @blimpjackal


shel
@shel

While studying Mandarin I've been noticing that there are a lot of homonyms, including with the same tone. Most natural languages have quite a few homonyms. The logographic writing system used by the Chinese languages makes the homonyms pretty much a non-issue for students reading and writing, but I've yet to try having any sort of oral conversation with anyone.

Constructed languages almost never have homonyms. Most ConLangs want to be easier to learn and more regular and logical than natural languages, which makes sense, but this even applies to ConLangs that are meant to feel like natural languages in a fictional setting.

I think homonyms are one of those things that will naturally eventually occur in any language, whether its simple sound collisions like 期 and 妻 or more complicated homonyms like how in English "reservations" could refer to having a table saved at a restaurant but also to being hesitant. "Play" could mean playing with toys but also a theater performance. These words likely share an etymological origin but came to be used for such different things over time that they ended up being different words entirely that sound and look exactly the same. You become very aware of things like this when learning other languages like ASL or Mandarin where the "same word" in English gets exploded into three different signs or words in ASL or Mandarin. Semantics are an open pool of concepts and different languages draw different circles around those concepts when applying their sounds. Deciding that applying am ointment and applying to a job are kinda the same thing is something English did, but other languages say those are definitely not the same thing.

A natural sounding ConLang naturally would have ended up with strange homonyms as well, whether by running out possible sound combinations, previously different words colliding as people speak quickly and abridge longer words, or strange circles being drawn around multiple concepts the gradually become so different from each other it's hard to say it's really the same word anymore.

It's a natural by-product of the fact the ConLangs are constructed all at once at a fixed point in time. A snapshot of one part of its life. Natural languages have been constantly evolving for thousands of years. English today is one moment in a constantly changing river. English today will not exist tomorrow, when someone invents the word Gobslob, an oral sex technique one uses when you go goblin mode (the Oxford word of the year).

Laadan and Dothraki have no such river. They have no history and no future. Perhaps it's not necessary for the snippets you need for the story you are telling, but it still results in a strange consistency, a striking unnaturalness. Tolkien, a blot on his name, might be the only prominent conlanger who was mindful of this, having created multiple versions of Elvish across his constructed history.

Anyway just a thought for the morning 😊 good morning nerds


kda
@kda

This, here, is exactly why I'm subjecting every conlang I'm planning on making to at least one round of "here's enough changes to break mutual intelligibility with some original form of the language".

This is also why I'm likely not going to be even remotely approaching any meaningful sense of being "done" this process before 2025.


VeraLycaon
@VeraLycaon

As someone who's done a bunch of conlanging herself, yeah - both from looking at other people's work and what I've caught myself subconsciously doing, people really tend to avoid homophones a lot; along with similar situations (words distinguished by pronunciation but not spelling, even in very phonemic spelling systems; syncretism within cases and conjugations; blending in presentations between verbs, nouns, adjectives; etc) as well as some general quirks common enough in natural languages that we often think of as making them hard to learn (set expressions/poetic language as a result of now-obsolete grammatical features, non-phonemic spelling systems like English or Tibetan's that more closely reflect an older form of the language, etc).

Personally these days I kinda go out of my way to avoid doing this, which is actually surprisingly easy (accidental homonym? shit, just let it stick there as long as people'd be realistically able to distinguish between different meanings); one big part of that all though?

Actually using it. Seriously, just like... write sentences or small pieces of poetry or prose, translate lyrics, etc. etc; adjust things to your liking as you actually use it, let it evolve along the way. Even with only one user, actually bringing a language to life is the probably easiest way to make it more natural.


You must log in to comment.

in reply to @shel's post:

This is a perpetual frustration for me, as a long-time conlang enthusiast and degree-holding linguist, specifically with a focus on historical linguistics and phenomena of language change.

Accounting for language change is more common in specialist circles, where nerds like me used to (and probably still do) congregate in phpBB forums, but it's basically absent from the kinds of high profile conlangs that show up in popular culture.

I think the biggest challenge to having conlang homonyms is that most artistic conlangs are pretty vocabulary-poor, and most conlangs with practical aspirations are trying to avoid ambiguity. The exception of the month here is Toki Pona, which is interesting because it’s got so few words that many concepts are homonymic. This is so different from natural language homonyms that it feels like it needs a different word to describe it. homosemantics?

I’ve seen conlangs (or conlang ideas) that are meant to be semi-historical which run along the lines of
“What if Latin but P-Celtic? Brithenig
“What if Latin but Czech?” (no link)
and my favorite, “What if Japanese but Polish?” twitter thread

In principle, because they have a regular rootstock and a well-defined phonetic transformation rule, all of these would have “natural” polysemantic homonyms; the Japanese → Polish one in particular because it’s starting from a highly homophonic base.

One thing about ASL that I’m finding interesting is the way shape fractures signs for nouns. What’s the sign for “light” as in “turn on the light”? Welllll, that depends on where the light is and how it’s oriented. What’s the sign for “queue”? Well, first, a queue of what (vehicles? people?); second, what shape is it (how many files? does it curve? is it long?); third, how quickly is it moving? (Arguably this is productive enough as to count as a kind of inflection?)

And of course ASL has homophones, too, e.g.: SAME/ALSO/TOO, DIFFERENT/BUT, and the polysemantic WORSE. These are along the lines of your example of English “apply” (both senses you mention go back to a verb that meant “to put one thing in contact with another”). Japanese is notorious for this. What does KI 気 mean??

I'm learning both Irish and an Indigenous language from my area right now and there's some interesting contrast. Irish has quite a few homophones of different origin, while the other language has, as far as I can tell, less than a dozen homophonic words. This other language is polysynthetic, and on top of that agglutinative, almost purely so- you take a root, and then stick on affixes to build your meaning, and you basically don't modify anything. Irish, however, is fairly fusional, with basically everything being inflected for a number of different cases, with the inflected forms sometimes being hard to recognize from their base (the verb bí comes to mind).

in reply to @VeraLycaon's post:

Oh, yeah, that's one part of the conlanging process I'm really excited about? Actually getting to do things with these fictional languages.

I'm mostly coming up with them as context for a broader narrative, but giving them any ~life~ means applying them!

"words distinguished by pronunciation but not spelling" and "set expressions/poetic language as a result of now-obsolete grammatical features" are two of my favorite linguistic quirks, closely followed by "non-phonemic spelling systems that more closely reflect an older form of the language".

As a kid, still really sinking my teeth into English, I felt like there was a wealth of information left over in these quirks, I could feel an earlier version of the language, reconstructed from vestigial features. I still remember trying to read the Canterbury Tales for the first time, getting like maybe 20% of the meaning, then I read it phonetically and listened to myself speaking it, and realized it kinda just sounds like Scots and now I understood like 60% of it. It was a dramatic change but to be honest I was still pretty lost haha