How to Create an Alien Language

If you are writing a science-fiction or fantasy story and the setting is another world, whether on a different, physical planet or a world in imagination, it is likely that when your characters encounter intelligent life there, that intelligent life (we are the "aliens" there, of course) will have some system for communication: language. There are three ways to handle this situation.

1. You create that new and different language for those "others" to use (more on this below).

2. You have someone translate what they are speaking - but then you need a character established for that, and where does one get training in, ummm Martian or TXDFGYHUJ?

3. You render the "alien" language as paraphrase: Then the F'G'HiX told them where to find water.

While some authors have argued for the second or third ways as being "realistic" and/or easier on the reader (one that comes to mind is Isaac Asimov's introduction to the novel Nightfall, based on his 1941 story [link], where he states that to use "alien" terms would be ridiculous in context and an obstacle for readers; I mentioned this in the previous posting), I find the bafflement of a character who encounters a language unknown to him/her to be an essential part of the story. His/her confusion is what readers want and need to experience...to a degree.

Granted, I do not want to unduly tax my readers, yet I feel that having some "other-language" present in a relevant scene (not added willy-nilly) adds to the setting (some call it "flavor") and can also deepen the meaning of action in the scenes (e.g., our hero smells smoke, yells "Fire!" but none of the aliens understand him; one of the aliens, seeing his consternation, tells him not to worry because they are simply roasting a Dtguuuuggbi over a bonfire in his honor; yes, after the appetizer he will be the Huguyumm [= "main course"]!)

For me, the first way of solving the problem is the most interesting. As a professional linguist, I love studying languages and comparing them. Anyone who has ever taken a foreign language in school knows that what we learn as much as the language itself is how other people think. Language reflects how a culture thinks, how they see their world. And if a sci-fi/fantasy author is building a world, certainly the language of that world (or cultures there) is of primary importance. My formal language study began with French in high school and college. Then I went to Japan. Along the way I also dabbled in German and Russian. Indo-European languages had much in common compared to Japanese, so comparing all of them was an education. As my curiosity compelled me, I also found myself studying a host of other languages: Chinese, Icelandic, Italian, Irish, Korean, Turkish, Swahili, Sanskrit/Hindi, Thai, Arabic, Spanish, Finnish, Greek, Serbo-Croatian, Quechua, Inuit, and Klingon.

There are several things you need to do to create a language. 

First, choose the structure of communication, the sentence pattern. There are only three so this is the easy part, unless you wish to create a new one. The main two are:

Subject - Verb - Object     is what we use in English, for the most part. ("I love you.")
Subject - Object - Verb     is used in many European languages and sometimes in English ("I thee wed.")

A third pattern is:

Object - (particle) - Subject - Verb     is what is used in Japanese. The particle is not a word but an indicator of function of the preceeding word in the sentence ("wo," "wa," and "ga"). ("[implied object, implied subject] aisteru" = "you I love.")

You could also employ Verb - Subject - Object ("okka debu sofak") or even Verb - Object - Subject ("okka sofak debu") but how will you keep the meaning clear? You can use those "particles" - like this: "okka debu ro sofak" so that "ro" indicates the one doing the action (the subject "debu" doing action "okka" to object "sofak"). Then you can switch the words around any way you like and still have it make sense. Supposedly.

Or, you could create something altogether different, depending on the kind of society of the language. For example, if it were to be a polyamorous society then singular verb forms would not exist, or would be considered rude. They might say: "We love us." There would still be a subject and an object, just not singular pronouns.

The next thing you need is a lexicon, sometimes called the vocabulary. (Many linguists declare that the grammar comes before the words but that is rather like deciding whether the chicken or the egg came first; in the end you need one in order to do anything with the other.) I went through a dictionary and made a list of all the words I wanted to use. I broke some words into several words (e.g., "fly" became "fly like a bird," "fly like on an airplane," and "fly like a rocket"), and combined other words.

Of course, this began with the words in a few phrases already in the story. I invented what was spoken in the story, then back-engineered the grammar. I foresaw the need to create new sentences for these "alien" characters, so I began my project. I chose the basic Subject-Verb-Object pattern for Ghoupallean, but to make it "seem" like a real language I added some quirky grammar rules and a huge list of very detailed and specific pronouns, a mind-boggling compendium of pronouns, such that a non-native speaker would never be able to get them correct beyond the most basic level of fluency. For the northern warrior race, the Zetin, I decided on a consonant heavy language, distantly imitating the grunts of military Klingon. To compensate for the difficulty of pronunciation, I made the grammar easy.

In the Dream Land trilogy, the southern desert race, the Roue, is an interesting case. For their language, I wanted something melodious so I modeled it after Hawaiian, very vowel-heavy with glottal stops, catches in the throat between vowels. But it becomes difficult when you realize that Roue is based on numbers and they use a base-20 system. In other words, instead of counting 1 through 10 and starting over with 11 through 20, and so forth, they go straight from 1 through 20 before starting over. It gets worse. They don't just say "I love you" - no, each word is a number on a master list of words! If the word "I" is, say, number 5 and the words "love" and "you" were numbers 11 and 9 respectively, then when they say "I love you" they would be speaking numbers "five-eleven-nine"! (Ah! But how would they distinguish actual numbers from the words? Fair question; clever answer: There is a "number" that designates that the numbers which follow are used as numbers rather than words. Follow?)

Anyway, the words. It's all about the words. Just make them up. However, like real languages, words tend to be related. Related words have related spelling and pronunciation. (We shall forego discussion here of the various written scripts that an alien society may use to share their communication.)

Notice how words are constructed in English. Take the car, for example. Car comes from carriage; a carriage is a type of wagon, a conveyance used in the era before the combustible engine. You see how the society's needs force the creation of new words to describe new things? Aliens do that, too; make your words fit. Another name for car is automobile. The automobile is a moving conveyance which can move on its own power, not pulled by polar bears or giraffes or tyggfix: autocar could work just as well, yet notice how the longer word automobile, made of a root word mobile and a prefix auto, gradually comes to be shortened to just auto. However, I've not yet heard self-propelled howitzer called a self.

Today the word "app" (short for "application"; I suppose in our quickened society we cannot be bothered with uttering a couple more syllables or we'd miss the next tweet from people we don't know) stands in as a whole slew of other words. The next generation will not know that "app" once meant "application," which is what other people call "software" or, when computers first became consumer devices, a "program." Words have history; they change; their usage changes. Build that in when you create your alien language.

In English we have several categories of words, based on what function each has in a sentence: nouns, verbs, pronouns, prepositions, articles, conjunctions, adjectives, adverbs, interjections [chart]. In Ghoupallean, I simplified it down to four categories or "families": thing-family (nouns, pronouns), motion-family (verbs), flavor-family (adjectives, adverbs), and relational family (the helping words: articles, conjunctions, prepositions). You can decide what categories your alien language will have.

Next comes grammar. It's just a bunch of rules for putting the words together. Basically, it comes down to this question: Do the words change form or are they static? This mostly occurs with verbs, but also with other parts of speech when they must match the form of the verbs. In English, following the practices of languages that came before us, such as Latin, we tend to change the ending of a word to show a change in meaning. Take the verb "fly": depending on the situation, we can write or speak "fly" or "flew" or "flown," but we know it's the same word really. In your alien language, will you have similar changes (and the rules that keep it organized), or will you show the same changes perhaps by using other words (helper words)?

fly (present) - flew (past) - will fly (future)

In Ghoupallean we change the ending of the word to match the time (the ending en indicates a verb):

soren (present) - sorend (past) - ge-soren (future)

But if you did not want to have words change according to a pattern, then you need to have some way to indicate when the action occurs. Try some kind of particle:

Let's say that qu means the same as "fly" (what a bird does, not necessarily what an airplane does). Then you could have the following system:

qu (present) - iz qu (past) - me qu (future) ......or smash'em together: qu, iqu, mqu............

which is not too different from the Ghoupallean pattern. But suppose you changed the word completely just to show the change of time? You might then have something like:

qu (present) - fo (past) - vi (future)

which would be fine, except that as an author you'd never be able to keep it straight in your head. Keep it simple yet have a few strange rules (grammar) to make it feel real.

The last part is the pronunciation guide. How are these characters (granted, we must generally use the Roman alphabet in our stories) to be pronounced? Most phrase books will compare the sounds of the new language with English phonemes (a single vowel or consonant utterance). For example (using Ghoupallean):

S s          Sounds as [s] in “sassy” in all positions; when doubled and between vowel phonemes, it sounds as [ss] in “scissors”; when written with [h] in this dictionary transcription sounds as [sh] in “shoe” or “mesh”.

Š š (Sh)   Sounds as [sh] in “shoe” or “mesh” in all positions.

sz            Sounds as [s] in “has” or [z] in “buzz”; only found in final position.

szs          Sounds as /zh/ in “azure” or the French word “menage”; only found in final position.

For me, when this madness first began in childhood, I was limited only by what characters I could produce on a typewriter keyboard. I tried combining characters or using symbols as additional letters. When I upgraded to the IBM Selectric with the balls of typeface, I could switch them (to get different fonts, for one thing). I then had a vast library of letters to use. The computer and MS Word now allow me to bring in many additional letters and accent marks from languages used around the world.

But remember, an alien language is not very likely to use the Roman alphabet - unless you invent some causal explanation, such as long ago one of the aliens visited Earth and shared their language writing system with an old Phoenician scholar, etc. (Well, it could've happened!)

That's enough for now, perhaps forever. Grammar can be fun, especially if you get to make up the rules. Play around with your alien language, speak it aloud, invent a script (something like a secret code from childhood?), and have fun!

Rbfj[I nfadko[ jgvakoi koj nk nw guypphviphvi UHWE!