Making a language: Basics/phonology
It's also, um, a multi-part rant, probably with three or four different parts. So expect to see more chatter on the subject.
I've been making up conlangs since I was nine, when I tried to come up with a language for a group of intelligent, telepathic, spacefaring wolves. Then I read Tolkien when I was eleven, and got the idea of making up languages for human-like nonhumans as well as intelligent animals. Since then, it's always kind of hung around in the back of my mind, and I make up bits of them when I feel like it.
1) Decide your phonemes. Phonemes are the sounds of a language. They include not only sounds represented by a single letter of the Roman alphabet, such as b, but sounds represented by two, like sh or gw. And they can certainly include sounds that don't exist in English, and sounds not represented by the Roman alphabet.
And no, you can't just use whatever you feel like. Most languages will set limits on the sounds, called 'phonological constraints.' It may be broad or large, but it doesn't have to be enormous to get a lot of mileage. I believe English only has forty-something phonemes, and it still manages to coin a ton of words. (More on that in point 3).
Yet English has its limits. For example, according to English rules of sound, you cannot just glance at, and know roughly how to pronounce, pjishjl, because English has a silent rule that pj is not allowed at the beginnings of words, and that the consonant cluster shjl is not allowed at all; they have to have vowels in them for them to make sense. Try to pronounce them anyway, and you'll probably find yourself adding vowels whether or not you want to.
So, what sound combinations does your language not allow? This can be even more valuable than deciding what it will allow, since it'll force you to set limits and produce words that sound like they belong to the same family.
Here are some I've used for one of my own languages, Aril:
-Two consonants cannot begin a word unless they're combined with a liquid (that is, in Aril, l, r, rh, or ll). Crith, the word for a light snowfall or the beginning of winter, is fine. *Sca [the asterisk is a symbol used in dictionaries and the like to show that a word is uncertain and may not exist] is not.
-Three-consonant combinations in the middle of a word are only allowed when one letter is a liquid and one is a nasal (in Aril, m, n, ng, or ñ). Maldñal, 'to shout at angrily,' is an example.
-Three-consonant clusters can never begin a word; something like 'spring' is not a possible word in Aril.
-The schwa sound (e as in butter) does not exist. Every foreign word which has a schwa and that Aril adopts is promptly 'Arilized' into a word with a strong vowel sound.
2) Know what your own linguistic preferences are. This one, I will admit, is a bit weird, but I'm mentioning it here because I have preferences, and I suspect that other conlangers may have them, too. (Tolkien did.)
What sounds do you find beautiful? Which are all right? Which irritate you for inexplicable reasons? Which sounds don't bother you in English or other languages you speak or read, but grate on you like nails in a chalkboard when you try to put them into an invented one?
Finding all of this out will produce two good results. First, you'll know what you like, and can stop wondering why a lot of your languages sound similar. You can make a language beautiful or irritating on purpose. I've accepted that I just love l and r among the consonants, and the strong or long vowel sounds (especially the vowel system of Spanish). Why? I don't know. The latter probably comes from my years as a Spanish minor, but I have no clue about my love for l and r. And I don't think it really matters. If you love g and c and h and i and o, then at least you'll make a language that's pretty to you.
Second, when you want to design a language that's really different from the primary ones(s) in your story, you'll be able to. You can use sounds you wouldn't otherwise, restrict the frequency of others, and introduce vowel/consonant combinations that would be impossible in the primary language(s). I don't like w or q- 'cause it has w in it- and I don't particularly care for g, and the sound of the i in girl irks me. So I can make languages with those sounds featured prominently in them when I want them to sound really, really strange, even abhorrent or ugly, to the ears of a character in my story who speaks the l/r/strong vowel-dominated languages I favor.
3) Know how your language typically responds to new words/the need for new words. English has a lot of ways to respond to new words just come into the language, such as the names of places in non-English-speaking countries, new foods, drinks, and animals, and specialized phrases that are taken in because English has no existing synonym for them. Some are Anglicized, the sounds twisted and hammered into submission until they fit with English's sound system, because some of the original sounds are difficult or rare in English. (This may be the reason why the English name for Spain is Spain instead of a closer variant of España, though I'm not sure of that). Some are adopted whole or almost whole into the language, since English's sound system can handle it. (This seems to be the case with sake/saki). And others are left intact, even preserved with spelling that few other English words get, because they're specialized words or phrases and aren't used that often. (This is the case with many French phrases, such as folie à deux). In cases where a whole new word is required because a whole new concept has come up, a word can be coined from the ever-handy Greek or Latin roots, like radiotelegraphy, or it can come from a brand name, like Walkman, or from a person's name, like boycott, or from an acronym, like snafu.
Other languages have different means. Icelandic usually invents new words entirely to deal with new concepts, because it's been kept "pure," first by geographical isolation and then by dedicated effort, and as close as possible to the original language of the Scandinavian explorers who settled the island. Names of places may change when a colony becomes independent and decides to reject as much as possible of a conqueror's language for native terms. And, of course, the reverse of Anglicizing a word exists when a useful English word jumps into another language; French rosbif is cognate with, but not the exact same word in sound or spelling as, English roast beef.
So, what does your language do when new words walk up and stare it in the face? Factors to take into mind:
-How isolated the language is. A small group of speakers in a hidden valley may create a whole new invention, and make their own word for it, too, because they can't borrow anyone else's, and that word will sound like a lot of other words by sheer necessity.
-How dedicated people are to keeping the language "pure." Is there some equivalent of a Royal Academy that watches over that kind of thing?
-Political history. English is in part such a mongrel language (I read one estimate that it's composed of different parts of at least 50 languages) because of the Norman Conquest of England in 1066, which interbred Anglo-Saxon with Norman French, and because the British then became a conquering empire on their own, rather than remaining on one tiny island. That may also be the reason why English has so many different ways to respond when new words show up.
-How restrictive the language's sound system is. If a foreign word has a c/k sound in it, and the language's phonological constraints say NO C/K SOUNDS EVER, then the sound may get altered to g. If g doesn't exist, either, the word may not sound at all like itself by the time that language gets through with it.
4) This should be more than a relexification of English (or whatever other language you speak). A "relexification" is just switching English, or another language, into a conlang, as if by code. Thus the conlang will have a word for car in the worldbuilding notes, even if this is a fantasy world, because English has a word for car. And it will translate the sentence "The rabbit hopped over the frozen ground" with seven different words, even though there are languages that could translate it with five, or three, or four. (You could drop the, for a start. It's helpful, but it's not necessary in, say, the way that nouns usually are. Many real-world languages, including Latin, have gotten along quite happily without it). And it will have verbs with only five different forms at most- like sing, sings, sang, sung, singing- because that's what English has.
I want to shake people like this until their teeth rattle in their heads. It's not necessary, okay? Your native language is not the sum of all languages that were and are, okay? Make something that's a goddamned language and not a code, okay?
I'll get into more detail on this in the second part of the conlang rant, when I'll start talking about grammar. For now, though, know that just because your native language has a feature doesn't mean that that feature needs to wind up in your conlang.
I’m starting to shade into grammar, so I’ll stop here. But first, some useful links:
Ardalambion (Of the Tongues of Arda)- Massive, masterful site on Tolkien’s languages.
The Babel Text Archive- This is an old site, but its link to the new one doesn’t seem to work. It’s a collection of translations of the story of the Tower of Babel into various languages, natural and invented. Some links are broken, but in the ones that work, you can see how the languages function to translate the story.
Conlangs A-Z- An enormous list of constructed languages.
The Language Construction Kit- Enormously useful, detailed breakdown of the process of building an artificial language.
And that’s all for tonight.