Reposted (with alterations) from October 27, 2007
When I talk about constructing an artificial language to my friends and family members, the most common reaction I get is “are you crazy? You need to get out more! Why don’t you get a hobby or something?” Well, constructing languages (or ‘conlanging’) is my hobby, and it’s a terribly fun one if you enjoy things like linguistics or puzzles.
I started conlanging very early. Coming from both an English and a French background, I had a lot of experience to draw from (and often came out with sentences along the lines of “il vas rain beaucoup today”). I started by making up secret languages with best friends. This mostly involved babbling nonsense and pretending to understand each other, but some words were repeated and eventually acquired their own meanings.
In my early teens, I started experimenting with actual language construction that included grammatical systems and root words. I hadn’t yet learned about phonetic consistency, do I drew from words I learned from all over the world intermingled with nonsense syllables I came up with off-hand. My grammar was simplistic and most probably unusable (though at least I avoided the common phase new conlangers go through of creating an “English with different words” language). It wasn’t until the last year or so that I’ve started approaching this hobby in any kind of serious or methodical way.
I was lucky. Grammatical instruction in Swiss schools is very comprehensive and I had been learning other languages (to date including German, Spanish, Latin, and Russian) for most of my life. Knowing only one language will make creating a language much more difficult, mostly because the grammar will be largely invisible, intuitively felt rather than intellectually known, and because it will make it that much harder to creatively imagine how alternative systems might work. To quote Goethe: “He who knows one language knows none.”
For the purposes of this guide, I will stick to the process of language creation. I will explain a few grammatical concepts, but this will not be the focus. Rather, I will recommend studying foreign languages, reading grammar books (some are actually quite pleasant reads), and the Zompist Language Construction Kit (which provides an alternative method for language creation as well as many grammatical explanations).
Culture
My first step when creating a new language is trying to understand the culture I am creating it for. Since language and culture are closely related, this gives me a sense of direction. For example, cultures with a big emphasis on social hierarchy may have multiple pronouns to reflect the relative positions of speaker and the person she is speaking about. They may also include modifiers for people’s names to reflect social status. Cultures that make little distinction between male and female members may not have gendered pronouns.
Culture can have an impact on a phonetic level as well. For example, a rough Barbarian culture living in a desert wasteland might have harsher sounds, while a plump and happy culture living in a warm region that is protected from invaders and has ample food growing without the need of care may have softer or more melodic sounds.
Establishing a culture first will also help when entering the lexicon step of language creation, informing the sorts of words the culture is likely to have. For example, a hot desert culture may need to borrow a word to express “snow.” A communal and egalitarian culture may not know how to express ideas like “power” or “king.”
Phonemes
The next step is to create a phonetic system. There’s an easy way and a hard way to do this. The easy way is to pick an already existing language and use its phonetic rules. To give an example, I’ve created a language that sounds like French. The hard way is to actually sit down and come up with a totally new phonetic system (have fun coming up with aliens whose mouths are physically different from our own – which of our phonemes are impossible for them to pronounce and which new phonemes might they create?).
If you decide to make up your own, I recommend learning something about phonetics (including a phonetic notation system such as X-SAMPA or IPA). Some people have mentioned choosing a dominant part of the mouth the native speaker would use (some languages tend to use the front of the mouth, some are more nasal, etc…). If you plan to do this, I’ve found a website that may help.
Whichever method you use, here are a few things to consider:
- Are the sounds predominantly harsh, using a lot of Ks or aspirated Hs?
- Are the sounds soft, using mostly vowels and soft Hs, Ls, Ns, etc…?
- What are some of the favoured sounds of the language? Which letters are used more frequently and what letters exist but are hardly ever heard?
- What are acceptable combinations? For example, is it against the rules to use a EE sound after another vowel?
Grammar
Once you’ve decided on the phonetic system, it’s time to come up with a grammatical system. I’ve found this to be the most time-consuming, confusing, and difficult part of the entire process. The good news is that once the grammar is planned out, it’s fairly simple to take the phonemes from the last step and fit them together into the grammatical structure. It almost becomes a matter of just filling in the blanks. Of course, we have to get through this section first!
Just to cover some basics: Sentences are generally formed using a verb, subject, object, indirect object, and dependent clause. Words can be nouns, verbs, adjectives, pronouns (personal, demonstrative, and possessive), prepositions, conjunctions, interjections, or adverbs. There are other possibilities and not all need to be present (in fact, if you want to give yourself a challenge, try eliminating all adverbs from your constructed language). If you are unsure of what any of these terms mean, look them up before getting started.
The following is a checklist you may wish to use (you may of course add or subtract any elements you feel necessary):
- Word order (which I expand to include whether the language is inflected or not)
- Stress pattern (both for single words and for complete sentences) – where does the stress come? If you don’t know what a stress is, try adding “fucking” to a word in English. The first syllable after the addition is the one that would normally be stressed. For example: abso-fucking-lutly is absoLUTly.
- Grammatical gender – Some languages have none. Some use male, female, and neuter. Some use living object, human, inanimate object, and God. There are plenty of options, so feel free to play around a bit.
- Negating sentences – In English, we make a positive sentence and add the word not after the verb (as in “I do not want to go”). In French, two such words are added, one before the verb and one after (as in “Je ne veux pas“).
- Numbers – Just to give some examples of possibilities, the number 27 might be expressed in many different ways including twenty seven, two and seven, seven and twenty, seven and two ten, and so forth. Consider also that your language may use something other than base ten (Napoleon, for example, tried to switch French over to a base twenty system, creating numbers like quatre-vingt (four-twenty) to say 80!).
- Questions – Questions mostly come in two forms: yes/no questions and general questions. Yes/no questions simply require a yes or a no as an answer while general questions require a more elaborate response. Many languages formulate the questions differently based on the type of answer needed.
- Conjunctions – Conjunctions join two ideas together. The most commonly used in English are probably and, or, and but. In addition, there is when, because, after, while, if, and composite conjunctions like neither… nor… or both… and…. Consider also that these can be used differently than they are in English. For example, Latin uses one type of or to mean “one or both” and another type to mean “one or the other.”
- Definite and Indefinite Articles – We need one of these before every noun in English. Some languages will have one as the unstated default, only having a word for the other.
- Verb Conjugation – First decide how conjugation will take place. The most common way is for the verb to have a root that remains unchanged while an ending is changed to reflect tense, mood, person, gender, and number. Other options include having an ending that reflects person, gender, and number and a prefix for tense and mood. Use your imagination. Many languages will also include a special form to indicate that the subject performed the action on herself. To make things easier, I have identified the following tenses that I use for all my languages: Infinitive, Habitual, Past Perfect (a completed action), Past Progressive (an ongoing action), Past Subjunctive (wishful), Uncertain Past, Past Imperfect (an action that was not completed, “I was watching TV while Robert did my homework), Present Indicative, Present Subjunctive, Uncertain Present, Future Indicative, Uncertain Future, Future Subjunctive, and Imperative (a command).
- Pronouns – This is where all that work you did designing the culture will come in handy. Pronouns should reflect the gender and social distribution ideologies of your culture.
- Declensions – If creating an inflected language, determine how nouns and adjectives will decline.
Lexicon
The lexicon is the words (or, more accurately, the morphemes) used in your language. To make things easier, try thinking up basic words that can be used as stems for other words. For example, I created a language that used the stem “soul” to create all of the following words: divine, good (holy), self, etc.
Another interesting trick I’ve seen is to come up with word combinations. For example, assign a number value to each phoneme you came up with. Next, draw up a chart with every possible combination of one, two, three, four, or five phonemes (or more, but remember that longer words will be rarer). For example, if my language has B, Ah, D, T, Eh, and L, I might assign each a number in order. My chart might begin like this: 1, 2, 3, 4, 5, 6, 12, 13, 14, 15, 16, 23, 24, 25, etc. The next step is to go through all the results and cross out any combinations that are impossible. For example, Btld is a bit too difficult to pronounce, so I will reject it as a potential word. What you will end up with is a list of useable words that need only to be assigned meanings.
Real languages tend to be pronounceable and, where following the word would cause difficulty, words are made irregular. A great way to simulate this is to construct your language following all the rules you’ve drawn up and then say each word, conjugated verb, declined noun, etc. as quickly as you can 20 times in a row. You should naturally alter the sound to make it more pronounceable.
Alphabet
The final step is to come up with an alphabet. If your culture is a literate one, this might mean completely original letters. If not, you may simply use a romantic alphabet, transcribing your phonemes into something recognizable for an English audience. Remember that each phoneme does not require a separate representation. In English, for example, we use the letter A for all the following sounds: father, mat, make, march, main, taught, steal, manor, etc.
If you want to come up with an original alphabet but are having trouble, try closing your eyes and squiggling on a piece of paper with a pen. Then open your eyes and scan the squiggles for anything useable. You may need to do this several times before you have enough letters for a full alphabet.
Also, while coming up with letters, consider the culture of the language. A culture that has invented paper may have flowing letters while one that carves its words on wood is more likely to have vertical and diagonal lines (horizontal ones being too difficult to carve as they go against the grain of the wood).
Extra
Try to make the language as functional as possible. Remember that while apostrophes might look totally awesome, they are used far too commonly in fantasy languages/names and often have no discernible purpose.
As a final note, I would just like to point out that this is the process that I personally use when creating language and is not by any stretch of the imagination the only option. To find out some other methods, I’ve asked around GaiaOnline and here’s the result:
- NothingOf starts with an alphabet and then culture and grammar are determined together. She has identified her greatest difficulty in language creation as determining when and where to use her invented words in her creative writing and how to convey the sense that it is a real language and not just a collection of “cool-sounding noises.”
- Homurakitsune, creator of Niora, likes to experiment with different strategies. To date, she has tried coming up with a vocabulary first, with a grammatical system first, and is currently creating a language by coming up with both simultaneously. She says that she usually creates the language first and then bases the culture around it, altering when inconsistencies occur.
- Dopplegaanger created an alphabet first, then added sounds and consonant clusters. He then created the grammar and, finally, the vocabulary. For his latest project, he looked up cases on Wikipedia and turned them into a language, adding grammar as he went. He then created a culture. He hasn’t started working on the vocabulary yet.
- Xeigrich starts with the sound of the language (using an IPA chart to pick sounds that work well together and provide variety), then he makes some basic words before tackling the grammar. Finally, he fleshes out the lexicon. He finds making the grammar to be easy but has a lot of difficulty making up words.
- Eccentric Iconoclast likes to start with an idea (“what would happen if a language did this…”).
If you would like to be mentioned, please leave a comment!