That's very impressive! Also, your crazy for doing all that work without asking toady if he WANTS someone to write a language for him 😉
Well, I'd always had a bit of a hankering to write my own language, so it was really more of a creative hobby exercise. But I actually
did do a bit of messaging with the Adams brothers a few months ago, and here's some of what Toady had to say (because I'm pretty sure what he told me wasn't in confidence):
SixOfSpades: When Dwarf Fortress is finally complete, do you envision that there will still be one, true, dwarven language? Or will languages be procedurally generated as part of worldgen, perhaps a different language for each civilization?
Toady One: It's not entirely clear. I imagine some people would like continuing to have a fixed language, but I'm planning to do some conlang stuff and have some interest, education and materials along those lines.
. . . as people mentioned, it's useful to see what sorts of things need to be thought about and what players are interested in. Right now I'm just working from online stuff and Mark Rosenfelder's Language Construction Kit book, but when I finally get around to it, that might expand.
. . . And modern linguistic theory is often adaptable to the computer, though games just dip their toes in now and then. I'm not sure how into syntax trees I'll get, and if I'll fake a(n English-adjacent) universal grammar for the languages so we can translate them easily into each other and into English (and hopefully let modders do other languages with some difficulty and simplifications), but this is all on the table.
SixOfSpades: DF players are here for reasons other than parsing arcane grammatical riddles. ... As long as all the languages are visually distinct from one another, and the computer can translate if desired, that seems likely to be all the language functionality that players would appreciate.
Toady One: For DF generally, when we can, we've always tried to represent the breadth of things, and language has a lot of breadth; something to make the cultures fascinating and distinct. Grammatical riddles aren't the point; I doubt we'd even have players picking over that stuff manually.
That covers a lot of the same ground that you've all touched on.
Could we have languages whose rows might define different ways a feature might be in the game, such as the noun-adjective order, so that these rules might be generated ar start? We could even have evolution during play.
As Toady said, he's considering exploring different syntactical forms, so there may be some procgen variation there--then again, there might not. As for my own project, I'm saying
not--each of the 4 racial languages will (most likely) have their own unique grammatical forms, but there won't be any variation from those 4. People play DF in order to dig cool forts and/or play an ASCII adventurer, they don't play it for the language. As for defining the various languages' syntax order in the raws--yeah, it could be done, but it'd make coding the rest of the dialogue-parsing sections a lot more awkward--and only the most hardcore language geeks would even bother to look at it in the first place. As for
procedurally-generated language syntax; that almost certainly won't happen. To me, coding a language family whose very syntax varies randomly, or especially mutates into different dialects over time, sounds like one hell of a lot of work, for very close to
zero player impact.
I am composing a reply, but I'm trying to work out how to reliably construct a sentence. My first thoughts are that it's a pretty wordy language. having sounded out some simple sentences, it takes twice as long, even if you assume that dwarves will skip bits and fuse things like we do in other languages.
Well, the fact that you're making the attempt at
all is quite gratifying for me, and a credit to you. I must admit, however, that Dorwaf is currently very ill-suited to discuss
itself: It doesn't even have a word for "word", because I was in a bit of a hurry to publish: I gave it a full vocabulary for stones, elements, animals, and matches for most of the original dwarven vocabulary . . . but no words for a
language or its components. (Thankfully, that'll improve in future versions: One of my goals is to be able to teach how to speak Dorwaf entirely
in Dorwaf, just to prove that dwarves would be able to do it too.) Yes, admittedly it frequently
does take a lot of syllables to form the words that you might want to say--that's an unhappy side effect of having syllables that are so short. But as the number of root words increases, the average word length will go down.
Connections between actual languages and fantastical conlangs need not have any real connotation to anything. . . . as a semitic speaker, I don't see any offense in ideas from my language group being used to create a Dwarvish conlang.
Well, historical negative caricatures (visual and verbal) of Jews have depicted them as being short, ugly, unfriendly to outsiders, and extremely parsimonious--but skilled with gems & jewelry. So drawing any additional parallels between them and dwarves (especially Tolkien's dwarves) just feels too on-the-nose to me, like I'm walking in racist footsteps. But hey, you do you.
In general, I think patterning (while still mixing and changing things up) conlangs on real language groups is generally beneficial. One of the problems with many Conglangs for me is that many of them don't feel real. They feel like they were constructed by a single person, and not through aeons of development.
True, but in a world that you literally
watch being assembled, the belief that language developed naturally is never even an issue--you already
know that it sprang into being, fully grown at Year 1. So, to me, the language bearing clear signs of being artificially constructed (most likely by the gods) isn't a problem.
When you take a real language group, and try and take inspiration from that, you get something that feels more real.
Myself, I would try to emulate
traits from real languages, rather than try to emulate the language (families) themselves. For example, Human could conjugate its verbs, Elven words could be gendered, and Goblin could use cases. (Fun fact: I hear that Russian has NINE cases. They literally have a single word that means "to
begin to twinkle".) Sprinkle enough linguistic quirks around, and your languages will be quite distinct from all real-world tongues, as well as from each other.
But yeah, in our case, it's best to keep things rather simple. In my vision, the language would still be highly regular. It will have a few complexities, but you won't have multiple conjugations or declinations.
Agreed--we must always remember that, if we genuinely intend these languages to be used in DF, the responsibility of making sure the game can translate them falls squarely upon
our shoulders--and the more complex & disparate the languages are, the more difficult that task will be.
Since you can't expect a player to play on the same world for years, or even months, there is no reason for him to even pay attention to the languages, since they will be regenerated in his next world anyway. . . . If you keep the languages constant, on the other hand, then players have a reason to at least skim over things in the game's languages, since as time goes by, they will get more and more familiar with the languages, especially if they are kept rather regular/simple.
Yes, I mentioned that dilemma earlier in the thread; how having a single "true" dwarven language was incompatible with each civ having its own language. But I've found a compromise: Let the game procedurally generate one language per civilization during worldgen, just like "normal"--but don't assign
which civ gets
which language until the player actually starts a game, either in Fortress or Adventurer mode. Whichever culture the player chooses as their starting civ,
that's the civ that takes the "true" language (of whichever race was chosen), and their closest neighbor civ of each of the other 3 races uses the "true" language of their
own respective race. So, as you will still encounter individuals from other civs beyond just those 4, the procgen language variations are still present, but the "real" versions will also be predominant enough to make an impact, perhaps even enough for hardcore players to start learning phrases.