So, last night I was creating a language. I <3 the
Language Construction Kit.
These drawings by Llorengil (
Alessara the Centaur and
Portrait of Zsofiel) are supposed to be sketches from a travelogue/diary of an elvish explorer. I've been wanting to do more of these sketches, and have some back-story to go with them. Also, I want this to be a fantasy world, but not a direct rip-off of Tolkein, so I need to make my own Elvish language. So last night, I settled in to do some hefty world-building.
I wanted the world to have several interesting species of humanoid, so I figured those out, and their social niches in the world. Since the elves are the ones writing the sketchbook, I needed to hack out that language first to make it possible for Llorengil to write snarky notes, scribble down grocery lists in the margins, translate mysterious glyphs, label bizarre plants or machines, remind himself to buy new socks, etc.
So, I gathered up a bunch of reference books and started work. First, the phonemes. I want this Elvish to sound liquid and flowing, so I got rid of the low vowels and the palatal stops (basically getting rid of 'ow', 'uh', 'aa', k, g, q and x and while I was in there, I also removed the j). I added in a couple of phonemes we don't have in English, the 'zs' sound in French 'rouge' and the 'ny' sound in Spanish 'cañon'. I also set up a rule that a stop (p, b, t, d) can only be used once per word, and never at the end of a word. So, words like 'sail' or 'timorous' are okay, but 'sat' and 'powderkeg' are not.
I figure I can rape words from Hungarian, morph them against words in Greek and Sanskrit, and then shift the phonemes so it fits this scheme. With the Finno-Ugric base, it'll be pretty alien to Western ears familiar with Germanic and Romance languages, and it'll be just lovely to hear.
Groovy. On to orthographics and alphabet.
First, because Elvish is more sensible than English, every damn sound will have it's own glyph. None of this "our language has sixteen vowel sounds but we only have five vowels and no diacritics" bullshit. So, I have 29 distinct glyphs that I need. Also, I need to figure out how I'm going to transliterate this into English when necessary, and how to write it in IPA. And later, I need to develop ancient, modern, and cursive forms of the script. So I make some charts.
I am not even going to try to reproduce the charts here, since I'm not sure how well LiveJournal supports unicode, or how many of you even have IPA-supporting fonts anyways.
Now, time for an alphabet. I don't want hieroglyphs or pictograms, because that's just a pain in the arse. I don't want it to look like Chinese or any of the familiar Western languages. Another race is going to have a cuneiform-like script. I want something elegant and beautiful, like Devanagari, but a little different. And you know, the combining-vowel system of the south asian languages is pretty damn elegant...
And so I do some research at
ancientscripts.com. Tibetan is just sexy. So is Lepcha. I like hPhags-pa as an ancient-looking, blocky script. And something about the rounded tops of Oriya is very compelling. But I want some things that can swoop below the main line of text in the cursive, so Llorengil can have a nicely flourished copperplate style of handwriting. So, something like a flourished Lepcha, that acts a bit more like Tibetan, but with the rounded tops of Oriya, that can be reduced to square blocks like hPhags-pa for old relics.
So, what if all consonants have those Oriya rounded tops... and vowels are basically inverted 'cups'. That way, vowels can be combining, and can combine with the vowel placeholder in the same way, so they don't need extra glyphs. And consonants can combine by just not having a vowel-mark above the stand-alone consonant. Excellent. Researching Tibetan calligraphy, I like the way most of the meat of the glyph is in the top third, and some glyphs swoop below. That's nice. I'll steal that, too. So... after nattering with some characters, it looks like most of the glyphs can have a swoop or no-swoop version. So there are sets of pairs. Hey... In the IPA linguistic charts, a lot of consonants are in voiced/unvoiced pairs. So what if I have unvoiced be the short glyphs, and voiced get the swoopy bits? Ha! That's clever! So 'pa' and 'ba' will be nearly identical, but 'ba' gets the elegant swoop down from the main line. And look at all the pairs I have, if I stretch a few definitions... p-b, t-d, f-v, th-ð, s-z, š-zs, r-l, and n-m. Excellent! That only leaves three odd consonants, and I had three funky characters that didn't fit the pattern anyways. Sweet! Bwa ha ha!
So, let's make some vowel modifiers for the top of the glyphs... and those are in convenient pairs too, since there's essentially five pairs of long and short vowels. Gods damn, but that's clever. Doodley doodley doodley...
Needs a little cleanup, but this looks *good*. I like it a lot.
Now, to double-check and make sure it doesn't look too much like Tolkein's Tengwar.
OH FOR @#$% $%@#$@# @%$#.
Okay, I knew that Tolkein was a linguistics buff. I even knew that he based his Elvish languages in part on Finnish because the Finno-Ugric languages are unrelated to anything else in Europe (Hungarian and Estonian being the only other Finno-Ugric languages in Europe). And just looking at the script, I could tell that there was some south asian influence in the way he designed tengwar. But THE GODS AS MY WITNESS, I had never really tried to learn tengwar, so I'd never studied it closely. I did not realize that he had his glyphs paired up in voiced and unvoiced sets. ARGH. And I thought I was being so clever.
Well, screw it. I've put enough work into it, it is definitely mine. And, ::sigh:: it is recognizeably Elvish, and recognizeably a *differnt* Elvish than Tolkein's. So, I'm going to call it good, and start generating a lexicon and making up some grammar.
Oh yeah, and come up with a name for this language.