Brushing up on Na’vi, the Language of Avatar

December 1, 2009, 10:45 AM
Most languages evolved organically over the course of millennia. But James Cameron wanted a new one within a couple of years; a tall order, even from one of the most demanding directors in Hollywood. Still, the Na’vi—the race of 10-foot smurfs who inhabit the distant planet Pandora in Cameron’s upcoming, $400 million science fiction epic, Avatar—had to speak something. So the filmmaker called on linguist Paul Frommer, a professor of clinical management communication at U.S.C., to create the Na’vi language from scratch. Frommer hopes his extensive efforts won’t be wasted, and that his language will have a life beyond Avatar. Just like Cameron’s career, that depends largely on whether the fanboys embrace the film.

Engineered languages—even theoretically perfect ones—have a tough time catching on, especially when there is no homogenous culture or community to speak them. Just witness the phenomenal failure of Esperanto, whose universalist aspirations were reduced to a punchline. It’s a different story, however, with science-fiction languages. Frommer is holding out hope that Na’vi will follow the example of Klingon, which he calls the “gold standard for this alien language niche.”

Frommer spoke to me—mostly in English—about the challenges of creating an entirely alien language.

Julian Sancton: Before we get started, how would you greet someone who called you on the phone in Na’vi, if there were such things as phones on Pandora?

Paul Frommer: I would say, “Kaltxì. Ngaru lu fpom srak?” Which is kind of, “Hello, how are you?”


[In a subsequent email, Frommer elaborates: ‘Note the accented “i” in the first word—it represents the vowel in “sit” rather than the one in “seat.” (English doesn’t allow that vowel sound at the end of a word.) Also, the “tx” represents the ejective t-sound. The literal translation would be something like, “Greetings. Do you have a sense of well-being?”’]

How developed is this language?

It’s got a perfectly consistent sound system, and grammar, orthography, syntax, and at this point it probably has about a thousand words. That’s not a huge vocabulary, but it’s certainly something that could be developed further into something that hopefully you could use every day for conversation.

Something like Klingon, to compare it to another language that was developed for science fiction?

Yeah. Klingon is a gold standard for this alien-language niche, if you want to call it that. And that’s much more developed. At this point, it’s been around a long time. I have a translation of Hamlet, on my bookshelf, into Klingon.


Yes. This exists. You can actually buy it on Amazon. People have really jumped into that and developed it probably beyond the vision of the original creator, who was Marc Okrand. If Na’vi ever developed into something like that, that’d be quite a thrill.

Would fans help develop it? Is that how it works?

You know, I’m not quite sure how it happened with Klingon. I’m pretty sure it was developed by people beyond the originator. If this ever took off and it got to that point where other people began developing it, that would be great, as long as they did it correctly, within the guidelines of the grammar, of course.

What’s the structure of the Na’vi language? Is it based on the grammar of any human language?

No. It could be a human language in that all the components in the language are found somewhere in other languages. But it’s very eclectic, so it has a grammatical system which is found in some languages, and it has certain sound elements that are found in other languages, but it does not follow any one particular language.

And is there anything particularly alien about it? If I understand correctly, Noam Chomsky says that there is an almost instinctive capacity in people to understand the structure of grammar. Does Na’vi follow that same universal human grammar?

It does, and there are certain reasons for it. I could have come up with rules for this language which are not followed by any human language. The reason that wasn’t an option is that part of the [Avatar] story is that human beings have learned this language, and can speak it. And if a human is going to learn the language, then it has to be something within that capacity that Chomsky was talking about. So there are some elements in the Na’vi language that I think are pretty rare in human language, but nothing that couldn’t be found somewhere in some other language

How would you describe this language? What differentiates it? What are the characteristics of it?

Well, let’s see. We can talk about various aspects. In terms of phonology, you can talk about sounds that it does have, and sounds that it doesn’t have. There’s a whole class of sounds that it doesn’t have—it doesn’t have what we call voiced stops. So, it doesn’t have a “b,” a “d,” or hard “g” sound. What it does have is sounds that are kind of interesting, which I added in to add a little spice to the language: they’re called “ejectives.”

Which are?

They sound like, “tx-ooo,” “tx-aaa”

Like a clicking almost?

It’s not technically a click, but these sort of popping sounds that are produced with something called the glottalic mechanism. Ejectives are produced not with air from the lungs but with air trapped in the glottis, and so, if you say a “k” sound, for example, without breathing, and then add a vowel after that, you can get something like “kx-a.” So I transcribe those as “kx,” “px,” and “tx,” and they add a little interest to the language. So that’s in terms of the sound system. There’s also a major phonological rule called lenition, which changes certain sounds to other sounds in certain parts of the grammar.

Are verbs conjugated?

They are, but they’re not conjugated for a number and a person. So there are no endings that correspond to I, you, he, she, or anything. But they are conjugated for tense and aspect. The conjugation, I think, is interesting because rather than relying on prefixes or suffixes I relied exclusively on infixes. These are found in certain languages, for example, of the Philippines, where rather than tacking something on to the beginning or end of a verb, you kind of take the root and cleave it in the middle, and shove something in the middle of it. So, for example, the root for “hunt” is taron, but to say hunted you say tolaron, so the –ol goes in the middle, right after the t.

There would be no difference from the audience perspective—we wouldn’t know all this—so I’m assuming you did for yourself, to have fun with the language.

Yes, and thinking ahead, that if it ever took off, like Klingon, that that would be something that learners would have fun with as well. It has a very free word order, and that’s something which is radically different, from Klingon for example. It has a a case system, so that by looking at the form of a noun, for example, you can tell if it’s a subject, an object, a subject of a transitive verb, or a subject of an intransitive verb, which means that the word order is almost totally free. You can arrange a lot of elements in a sentence in many different ways and still have it mean the same thing.

And if there are any inefficiencies in the language, nobody would be able to point it out yet, right? You could just fix it without anyone knowing?

Well, I try to be consistent to the rules that I’ve constructed for us. At this point, nothing is—I don’t think anything is published yet on the language, which means that if I want to make a change, I can make a change. This inevitably happens when you’re using it with real people and actors that are trying to learn the lines. There are times when something didn’t come out exactly as I had intended, but if in fact what came out was consistent with the sound system of the language and the rules of grammar, then I said, “Guess what? That’s a word.”

Have you written all the rules out of this language? Do you have a compendium?

I’m working on it. I don’t have it yet. That is something that I’m going to turn in to Fox soon, because that’s sort of going to be the capstone of my work on it. I’ve been working on it since 2005, really.

How did James Cameron approach you?

His production company, Lightstorm Entertainment, contacted the linguistics department at U.S.C. asking for someone who could develop a language for a movie he was working on. At that point it wasn’t even called Avatar; it was called Project 880. Although I’m no longer part of the linguistics department, I have some friends there, so someone saw this and said, “This sounds like Paul.”

Why did your name come to mind? Are you a big science-fiction fan?

The person who forwarded me the email is my co-author on a book called Looking at Languages. It’s a workbook in elementary linguistics in which we had put together data from 30 different languages. So he knew that I loved to play around with data from languages and I had a little artificial language, which didn’t really have any development at all but was just something that I constructed for an exercise.

What is your background as a linguist?

I have a doctorate in linguistics from U.S.C. To backtrack beyond that, I was originally a math major as an undergraduate. I went into the Peace Corps after that. I spent two years in Malaysia, where I was teaching English and math. I taught in the Malay language, which was an interesting experience, and I realized that my real love was in the area of languages. Prior to that I had studied a number of foreign languages. I had studied Hebrew, French, Latin, a little German. So when I went to grad school, I decided to do it in linguistics. During that time I spent a year in Iran, so I learned Persian. And I wound up finishing a dissertation on a certain aspect of Persian grammar. So I had, at that point a somewhat extensive background in languages—not that I spoke all these languages fluently, by any means.

How many do you speak fluently?

One. But I have various competencies in other languages. Probably Persian is my best, although it’s a little rusty at this point. I was in Iran in the mid ’70s, so it was a long time ago.

Of the actors in Avatar, who would you say had a knack for picking up the language you invented?

They were all pretty good. I was quite surprised with the facility that people had. There were seven actors altogether who were speaking the language.

Sam Worthington, Zoe Saldana, Sigourney Weaver, and then …

Wes Studi, C.C. Pounder, Joel David Moore, and Laz Alonso. Of those seven, four were supposed to be native speakers and three were humans who had learned the language.

Is there any difference, in terms of accents, between the humans who had learned it, and the native speakers?

Yeah, well, some of the humans seemed to have more English-y accents than the others, which is good. Also, just in terms of the actual grammar that came out, there were some small grammatical errors that the humans made, which is entirely natural.

And was that intentional?

Not entirely, but some things came out, and I said, “You know, that’d be totally appropriate. That’s exactly what a second-language learner would probably do at this stage.”

What human language would you say it’s most similar to, if you had to choose one?

I really couldn’t. And that’s good, because I don’t want it to be identified with any particular language. You know, I didn’t start from absolute ground zero, because James Cameron had come up with, in the early script, maybe 30 words. Most of them were character names, but there were a couple of names of animals. So at that point I had a sense of some of the sounds that he had in his ear and it reminded me a little bit of some Polynesian languages.

I read in that New Yorker profile that he had gone scuba diving around there.

Yeah, he had spent some time in New Zealand and maybe had some Maori in his ear. But I expanded it considerably beyond that. And of course in terms of the orthography and the syntax and all that, I came up with that entirely on my own. So, I don’t think it can be compared in any aspect to any particular human language.