Second part of the conlang rant; this deals mostly with grammar. And this will definitely have to be four parts, now, in order to fit everything in that I want to.
1) What kind of language is your language?
Some languages are isolating, like English. There’s very little inflection (changing the meaning or grammatical function of a word by means of additions like suffixes and prefixes), and many words have a one-on-one correspondence to specific concepts, such as English “boat” meaning a boat and not being a combination of, say, *bo and *at to mean “water-craft.” There aren’t that many verb formations in English, and there are only sixty or so irregular verbs. English has made a habit of shedding its inflections as it goes along. (However, not every language does this, and it’s a mistake to assume they do. There are languages in the world, like Spanish, that retain far more inflection than English).
These are the kinds of languages I most often see as fantasy conlangs; you can tell because they translate a seven-word English sentence with seven words, as I complained in the first part of the rant. Okay, you can have an isolating language if you want, but if so, please know what you’re in for, and pick it as a conscious choice rather than an unconscious one.
Agglutinating languages link units of meaning together to form long words, which sometimes can convey as much information as a whole sentence in an isolating language. If you had a language where at meant she, olo meant past tense, ili meant imperfect aspect (see “aspect” below), ni meant indicative mood (see “mood” while you’re at it), nas meant see, te meant him, and all words were assumed to be in active voice unless otherwise stated, atoloilininaste could mean “She was seeing him.” Literally, it would be “She-[past tense]-was -ing-[indicative mood]-see-him.”
These languages can be hard for someone used to speaking an isolating language to get a grip on as conlangs. However, one advantage is that you can declare each morpheme (meaning unit) has one and only one meaning. So at always means she, never her, and olo can only mean past tense, not past tense or present tense depending on how you pronounce it. This may ease the strain somewhat.
Inflecting languages are the ones I favor, the ones that alter the forms of their words by such things as prefixes and suffixes to change tense, alter grammatical function, show number agreement, and many other things. Inflections can and often do carry more than one meaning. So Spanish bailé, ‘I danced,’ has bail- from bailar, the verb ‘to dance.’ That means that that little é carries a formidable burden: first person, past tense, preterite aspect (the action is complete, or conceived of as a complete unit of time; you’d have to use a different form if you meant ‘I danced for a living’), indicative mood, active voice.
I like inflecting languages because I like altering the forms of words rather than constantly coming up with new ones, and, done right, they can really force a conlanger to stop thinking in terms of his or her native language (assuming that the native language does not inflect in a similar way, of course). And it’s fun to try to figure out exactly how you’d translate an English sentence like “I’m not going to do this” into, say, two or three words in an inflecting language.
2) Choose your basic sentence order.
There are six of them, classified as variants of V (Verb), S (Subject, the one who does the action), and O (Object, the one who receives the action). Not every sentence in every language will be the same, of course, and this says nothing about the placement of, oh, adjectives and prepositions. But it’s a good thing to figure out, so that you’ll know how to construct basic sentences. It’d be good to know if “The dog bit the man” in your language has the exact same meaning as the English sentence does, or if that sentence actually means the same thing as the English “The man bit the dog” because it’s an OVS language.
English itself is SVO, by the way: “Dolphins like fish.” In general, isolating languages will have to have fairly strict word order, so as to prevent meaning from wandering all over the place. A language that cannot communicate the difference between “Dolphins like fish” and “Fish like dolphins” because it has no inflection and no fixed word order is not going to do anyone any good. Go back and work on it some more.
Here is where inflecting and agglutinating languages have an advantage, because while you should still know the general order of a sentence, the changes to the words, or the morphemes, will tell you what they can and cannot mean. Our earlier example of atoloilininastecannot mean “He was seeing her” if at never means her and te never means he. Likewise, even if an inflected Spanish verb goes near the end of its sentence, most times it will contain enough information that you can still know who the subject of the sentence is.
It’s worth noting that Object-initial (OVS and OSV) languages are rare, and some linguists argue that they don’t really exist. Individual sentences in some languages, especially in poetry and questions, may put the object first, and there are some languages where it’s arguably more common than that, but the Subject-initial sentence structures are the most common, followed by the Verb-initial. Of course, you can take advantage of this fact to construct deliberately nonhuman languages, which is what some conlangers have done.
3) How do you emphasize a word?
The means in English (and some other languages, too) is stress. Certain syllables are spoken with more force than others. English stress rules are not necessarily tied to parts of speech; verbs don’t all receive stress on the first syllable because they’re verbs, for example. Spanish has more regular stress rules, and when it breaks them, it tells you so by sticking an accent over the affected vowel; avión, airplane, has to be pronounced with stress on the syllable the O’s in, but that’s not usual. (And yes, if you’re going to use ~ and é and è, please have them MEAN something? Not just be a fancy way of getting the language to look alien? Please and thank you).
In conlangs, varying emphases could convey various emotions, change in meaning (which is what a tone language does, by using various tones to distinguish the meanings of words), changes in parts of speech, social class, and other factors.
4) Where do the syllables fall?
This becomes especially important when you want to set up phonological constraints for your language. Perhaps no syllable can end with a vowel—in which case you have to make sure you’re not dividing the words so that there’s an “a” sitting all by its lonesome at the end of a syllable. And does that mean that all double or triple vowels are run together as diphthongs? That could be feasible with esain, and maybe even aleoith depending on how you pronounced it, but let’s see you try that with laaaoies.
What about consonants? Are there certain consonant sounds that must end a syllable/can never end it/may end it only in combination with something else? Do certain consonants get used more frequently for this purpose than others? Do the writing rules differ from the speaking ones? (I have known people who said the word “solution” as “soh-LOO-shun,” but you aren’t technically supposed to put the -tion all by itself on a line when you’re writing, even if you say it that way).
And yes, technically this pointer should have come before the stress one, but it doesn’t have to. Just because you decide every single word in your language takes its stress on the first syllable doesn’t mean that you have any idea what those syllables are yet.
5) Decide your parts of speech.
It may be possible to design a conlang without nouns and verbs. (I know there’s one called “AllNoun”). However, I cannot give practical advice on it, never having done it, so I’m going to recommend that you keep them, and give some ideas for handling them in the rest of this rant and the next.
The others can become more flexible. Pronouns? Well, sure, they’re damn useful, but that doesn’t mean they have to be just like pronouns in your native language. One thing that amuses me is how many conlangs have “he,” “she,” “it,” and then a pronoun meaning “he or she,” as if the writers are making some great grand attempt to show how very Non-Sexist they are. There are few with “he or she” and “it,” and even fewer that just collapse all the third-person singular distinctions into one pronoun, or decide that distinctions are going to be made on the basis of species or whether something’s animate instead of gender, or that that have pronouns referencing distance or duals. (Dual pronouns indicate twos: “we two,” “you two,” “they two.”) I wanted to see how far I could do without pronouns in Aril, and managed to reduce the subject ones to verb endings, and decided that the -y ending of the “it” pronoun could do just as well to represent “he or she.” If the subject is swinging a sword around, the listener is going to know it’s not a newt.
Articles? Piffle. You don’t need them. Or you could have articles that reflect varying degrees of distance, importance to the speaker, or emphasis—in effect saying the horse instead of just the horse.
Conjunctions? Sure, but they don’t have to be separate words. You can have cases, or bits of words attached to the words in front of them! Yay inflecting languages! (More on that next rant).
Prepositions? Well, some languages have postpositions, which mean the same thing but go after the words they modify instead of in front. Perhaps your language has a mixture of both, with some words caught in the transition from one to the other. I did read a claim, once, that languages which put adjectives before the nouns are more likely to have prepositions and languages which put adjectives after the nouns are more likely to have postpositions. I don’t know for certain if that holds for real-world languages (it was an older book), but it might make a good guide for a conlang.
Adjectives? Perhaps they’re in effect halved, since people usually follow the practice of making one adjective out of the other by the addition of prefixes and suffixes. Think Orwell’s doubleplusungood. This can be fun to play with, if only as a way of seeing what the cultural prejudices are. Is the opposite of green red, or is it blue? Or perhaps it’s soft, because the word for “green” is related to the word for “hard,” both having come originally from the word for “jade.” And what’s the usual conception of color/texture/smell/sound in your culture? When people say “green,” do they see forest-green, or the color of the second moon?
Adverbs? Don’t need to be separate words. You could form them by modifying the verb, or by modifying adjectives. This could easily lead to adverbial forms that, translated, aren’t at all common in English, like “greenly.”
And I’m sure that you can come up with your own classes of words that your language needs. For Aril, which is a language spoken by empaths who would feel blind if they didn’t have some indication of emotion in writing, there’s a system of dummy words that indicate such states as “I’m extremely angry about what follows.” There’s virtually no end to the invention fantasy conlangs permit in this direction.
These alone have so many considerations that I’m going to end the rant after I finish them. It will just get too long otherwise.
Also, I’m using lots of examples from Aril, because I think the verbal system is developed enough to give good examples, and also to show a few ways of dealing with these considerations (especially if you’re attempting an inflecting language).
Traditionally, this is divided into first (“I,” “we”), second (“you” plural and singular), and third (“he,” “she,” “it,” “they.”) The numbers used are singular, for one, and I, for more than one.
You can play with this. Spanish has separate ellos and ellas pronouns to indicate the gender of a group of people (ellas has to be a group of women), though that’s in the pronoun and not the verb. You can have a dual number in addition to a singular and plural. You can differentiate formal and informal, so that one “you” verbal form is polite, and the other is one that you would never use save around a close friend, and another is the form used by/to foreigners. Of course, this can build into astounding complexity quite quickly, so you might want not to play too much, lest you run the risk of designing the entire verb around person.
Aril has a system of suffixes to indicate person, number, and in a few places gender, the separate subject pronoun having been abolished long ago. Thus:
- an: I.
- it: you, singular
- o: he
- a: she
- y: it, also stands for he or she
- un: we
- ot: you, plural
- yr: they, used irrespective of gender.
Riallan, I sing, consists of rial, the verb “to sing” (all Aril verbs end with a liquid) + -l- (present tense, indicative mood, perfect aspect indicator…um, I’ll get to this in a moment) + -an= I.
Yes, yes, past, present, and future. But quite a few languages divide the past tense up. After all, are you talking about something you do only once, like dancing with your grandmother at the Shackleberry Ball, or something you did all the time, like going for a jog every day? And how far in the future are we talking? What does “present” mean, exactly? How do you say “I’m busy,” and mean it in the sense of “right at this moment?” Or does present mean something else entirely?
Aril has a pretty basic tense system (it gets complicated with other things): -l- for present tense, -d- for preterite (past actions done only once, or conceived of as a complete unit), -c- for imperfect (past actions repeated more than once, or over a long period of time), and -ch- for future. Here’s aamal, “to keep one’s promises,” conjugated in all four for “we”: aamallun, aamaldun, aamalcun, aamalchun.
(By now you may be a little green around the gills, and I do not blame you).
Aspect refers to the completion of an action. Is it in progress, or finished, or about to be begun? (Not all languages have all three as a compact part of a verb, and some make other distinctions). This may not seem like the most important distinction to make, but when your character is getting attacked by a dragon, there’s a world of difference between “I’m running” and “I just ran.”
Aril has only a progressive aspect, and it functions only in present and future tenses; all other tenses are assumed to be perfect, or finished. The aspect marker is an -I- infix attached just after the first syllable, unless the verb is irregular. Riiallan, I am singing.
As in, active and passive. I believe Latin or Greek, one of the two, also has a “semi-passive” that is sort of in the middle, but I didn’t study it in detail. And no, you don’t have to do things exactly like the English passive voice, which many people dislike for its ugliness.
Aril just assumes that all verbs are active voice, and you indicate passive voice by using an instrumental case on a noun, to show what the action was done with, and the third person singular verb: Osaldy meri timblaídh, or, literally, “Killed her with-a-sword,” where the actor is not known.
Or, the relation of the verb to reality and desire. Common are indicative (something just happening), imperative (a command: “Go!” “Run!” “Fly, you fools!”), and negative (“This is not happening.”) There’s also subjunctive, which is sometimes contrary to fact, as in English “I wish I were doing that,” and sometimes simply required by the language; Spanish makes you put “changes” in subjunctive mood when you say “I believe that it changes,” even though the belief can be sincere, because, well, the construction requires it. And this is fertile ground for playing about, just as gender and formality are with pronouns. There are languages with other moods (which I am not commenting on because I don’t know them as well) which can refer to things like, oh, eyewitnesses; did the person speaking seethe action happen, hear about it secondhand, know about it from rumor, or what?
Aril’s imperative form is distinct from the rest of the verb, dropping the liquid at the end: “Cara!” or “Give battle!” from caral, for example. Subjunctive usually means wishes or desires that won’t come true, and is achieved by changing the vowels in the personal endings, so that Cal crendulchain is “I wish that I could stand in your way,” while Cal crendulchan is simply “I will stand in your way.”
I left indicative as the basic verb form, because sometimes enough is enough, and negative is achieved simply by using an obverse sound of the usual tense markers. Cal crendulzhan is “I will not stand in your way.” And yes, Aril requires extremely keen ears to distinguish the similar sounds, which its nonhuman speakers have. When it becomes a human language, as it does in places around the continent, it tends to simplify.
Here’s a little factoid: Almost all languages have irregular forms of the verb “to be.” One theory is that the “to be” verb is often built up from other verbs, which gradually coalesce and collapse into that meaning, and in so doing leave traces of their own very different conjugations behind. So there’s a prime candidate for an irregular verb.
Verbs can move between regular and irregular; many, many more English verbs were once irregular than is presently the case. Some have left traces of their old past tenses behind in adjectives. “Molten” was once a past tense of “to melt,” and “wrought” the past tense of “to work.” A similar procedure in your own language could produce formal terms, untranslatable ones, poetic ones, new irregulars, or new regulars. The sky’s the playground here.
Jesus. Nouns and pronouns later. Damn thing.