Paul Frommer on the Na'vi, Avatar and Creating a New Language

Art by Nick Lopergalo

Paul Frommer is the creator of the entirely new language for the fictional Na'vi in the critically acclaimed film Avatar. He is also a professor at the University of Southern California and a linguistics consultant.

SoR: As a mathematics major, what fascinated you so much about the Malay language as you were teaching in it in the Peace Corps that led you to become a linguist?

Frommer: To a linguist, even a future one, every language is fascinating—and there were lots of things about Malay that intrigued me. Before I knew the term “phonological rule,” I was tickled by the sound changes that often took place in Malay words, so that if you were looking up certain verb forms like, say, memandu (guiding), menari (dancing), mengata (saying), and menyatukan (uniting) in a dictionary, you’d find them under p, t, k, and s respectively. I was both fascinated and frustrated by the fact that Malay has a whole range of words for “you,” and that finding the most appropriate form of address in any given situation was far from straightforward. And I particularly remember how learning Malay helped me realize that every language is a unique window on the world. One example stands out in my mind: in Malay, the term for watermelon translates to “Chinese cucumber.” I had never thought of the relationship before, but in fact cucumbers and melons are in the same botanical family. Malay makes that a lot clearer than does English.

Malay has two totally different writing systems, the ubiquitous one based on the Roman alphabet and another, less commonly used, based on the Arabic. I became rather famous at my school for learning to read and write Jawi, the Arabic-based Malay orthography, and ironically, some of my students even came to me for help with their religion classes, where all the reading material was in Jawi.

Mostly, though, it wasn’t Malay in particular but rather the positive experience with a foreign language in general, regardless of which one—the challenge of learning it and the delight of being able to use it to communicate with people—that made me want to pursue linguistics in graduate school. I guess my Peace Corps experience also made me think I might have some native ability in that area.

SoR: Do you feel that certain languages are more effective at expressing certain themes or experiences as opposed to others given that some languages have a smaller vocabulary or more complicated pronunciations?

Frommer: I’m not sure that pronunciation has anything to do with expressivity, but vocabulary does. All natural languages are tied to culture, although some move out of their “home” cultures and adapt themselves to new ones. And languages develop so that it’s easy and efficient for speakers to talk about things that are important in their culture, environment, and experience. When I lived in Tehran, all the Americans there quickly learned the word joob, which refers to the ubiquitous paved ditch on either side of a roadway, right off the sidewalk, in which water flows by gravity down from the slopes of the mountains. It was a lot more efficient to say “His car got stuck in the joob” than to say it got stuck in the paved water-filled ditch by the side of the road.

In keeping with this idea, I tried to incorporate aspects of Na’vi culture and experience into the language. The vocabulary has words and phrases for things important to the daily life, spirituality, and worldview of the Na’vi—for example, fpom ‘sense of well-being and harmony with the natural world,’ tireafya’o ‘spirit path,’ waytelem ‘song cord,’ fpeio ‘ceremonial challenge,’ tsaheylu ‘neural connection or bond,’ Eywa ngahu ‘Eywa be with you.’ Also, since ceremonies are important on Pandora, I developed a ceremonial or honorific form of the language, with somewhat different personal pronouns and verb infixes. The Na’vi counting system was influenced by physiology rather than culture: since the Na’vi have four fingers on each hand, their numbering system is octal rather than decimal. (As one Na’vi enthusiast put it, “Now we’ve all got to carry around calculators!”)

SoR: What was it like creating an entirely new language that didn't resemble any known spoken language for James Cameron's film Avatar?

Frommer: Na’vi is indeed a new language, but that’s not to say it hasn’t been influenced by some existing human languages. For example, some of the original vocabulary that James Cameron had come up with—about 30 words—sounded a bit Polynesian to me, akin to words in Hawaiian or Maori. So that was where I started in creating the sound system. The “ejective” sounds I added, which sound to some people like “clicky” consonants, are found in certain indigenous languages of the western hemisphere (e.g. Quechua), Africa (e.g. Amharic), and Asia (e.g. the Caucasian family).

Some of Na’vi’s grammatical structures were inspired by various earth languages. For example, there’s a verbal construction that looks a bit like something in Persian; there’s a phonological rule that brings to mind similar processes in Hebrew and Irish; some sentence-final particles in Na’vi are reminiscent of certain things in Chinese; the case system is similar to one found in the Australian language Wanggumara. But there are also things in Na’vi grammar that, to my knowledge, aren’t found anywhere else. All in all, I think it’s safe to say that the particular combination of structures and processes found in Na’vi is unique.

SoR: How does someone create a new language in such a short period of time and how do you compare that to the actual evolution of human languages over thousands of years?

Frommer: Actually, there are some fairly well-defined steps in language creation. The first thing you typically try to nail down is the sound system of the language—what linguists refer to as phonetics and phonology. For Na’vi, I considered a number of possible elements and processes—which sounds to include, which to exclude, how sounds would combine with each other and perhaps change to other sounds in various situations. Some of these things wound up in the language and some didn’t. Initially I presented Cameron with three different “sound palettes” or possibilities for the overall sonic impression of the language—he chose one, and we were off. The next step was to decide on the morphology (meaningful elements and how they combine to form words) and syntax (rules for putting words together into phrases and sentences). For those, I was on my own. Since this was an alien language spoken on another world, I wanted to include structures and processes that were relatively rare in human languages but that could be acquired by humans. So, for example, Na’vi verbs change their form to reflect differences in tense and aspect (whether an action is viewed as completed or ongoing), but to show those differences you don’t use prefixes and suffixes, which are used in many familiar languages, but their relatively rare cousins, infixes. These are meaningful elements that are inserted inside a root. For example, the root of the verb ‘eat’ is yom; ‘ate’ (that is, eating has been completed) is yolom; ‘is presently eating’ is yerom; ‘will eat’ is yayom. So the infixes in these cases are ol, er, and ay respectively. Learning to use those elements naturally and spontaneously takes some practice.

And then it was time to begin constructing the lexicon—the vocabulary of the language. For that, I let the script drive my decisions about what words to come up with first. If a line of dialog had the word ‘run,’ for example, I obviously had to find a word for that. If there was no Na’vi line that included ‘sleep,’ I knew I could hold off on that one for a while.

I’d like to tell you that by the time Avatar came out, Na’vi was a completely functional language, but that’s not the case. It was certainly adequate for the needs of the film. But there were a number of grammatical structures—and of course a huge amount of vocabulary—still to be determined. Some of that has been completed, but there’s a lot more to be done. If anything has been driven home to me through this experience, it’s how extraordinarily rich and complex any language is, and how much information is unconsciously incorporated into the minds of every native speaker. For example, as English speakers we know that “big” and “large” mean pretty much the same thing. But there’s a difference between “my big brother” and “my large brother”—we can use the words interchangeably in some contexts but not in others. That gives you an idea of the kinds of decisions you have to make about thousands of words when you’re constructing a language from scratch. It’s a daunting task.

SoR: What do you think it was about the sounds of the ejective consonants in the preliminary design phases of the language that appealed to James Cameron so much that made him choose that direction over the others?

Frommer: That’s hard to say. I never really asked Jim why he liked the ejectives, the “clicky” sounds written px, tx, and kx. A lot of people do, though. I think they add a certain “exoticism” to the language, at least for people who have only been exposed to European-origin languages. Of course what’s exotic to one person is homey and familiar to another. It all depends on your background and experience.

SoR: What were some of the joys and some of the challenges that faced you while creating an entirely new language?

Frommer: Perhaps the biggest joy early on was discovering that the language I was creating, with all its quirkiness and complexity, actually worked. I found I was able to construct simple sentences pretty spontaneously that conveyed the meaning I wanted and felt “good in the mouth.” And the actors could reproduce them convincingly.

In retrospect, I think the biggest challenge in all of this was to avoid being unduly influenced by English. For example, suppose I needed to translate the sentence, “I thank you from the bottom of my heart.” The way NOT to do that is to come up with a word for “bottom” and a word for “heart”! The English expression is an idiom that’s certainly not common to most languages. Instead, one has to find an equivalent expression that would fit Na’vi culture and thinking. On the grammatical level too, it was important to avoid simply mimicking what English or other familiar languages do with various constructions.

SoR: How did it feel seeing and hearing people speak the language you developed?

Frommer: In a word, awesome. I’ll never forget the experience of seeing Avatar complete for the first time in a theater and hearing my words in the mouths of Jake and Neytiri and Eytukan and Mo’at. I felt very proud.

Sometime later came another memorable experience: my first time meeting fans of Na’vi who had learned enough of the language to be able to use it in conversation, in some cases with more facility than I could myself. That too was awesome.

SoR: I hear that you are still developing the Na'vi language. What direction is this heading and where do you want to see it go?

Frommer: Since Avatar was released, a worldwide community of Na’vi enthusiasts has emerged. who are learning the language. The fan-created web site learnnavi.org is the main meeting place online; its forums, in 19 languages, currently have a total of over 480,000 posts. I’ve met some of these people in person at Na’vi gatherings and have spoken to more of them online: they’re a wonderful group—bright, enthusiastic, and inventive. Some of them are extremely sophisticated linguistically, with extensive experience in foreign languages; others have developed an interest in language through Na’vi.

Some of the ablest Na’vi-ites in the community are working with me to help expand the lexicon, providing excellent suggestions for new vocabulary. I’m still the “gatekeeper,” in that only I can say what’s officially part of the language and what’s not. But I’ve found the input from the “Na’vi experts” extremely valuable.

As for where Na’vi is going . . . well, I don’t know for sure, but I guess my wish would be to see it develop into a fully functional auxiliary language that not only reflects the unique environment of Pandora and the culture of the Na’vi themselves but can be used efficiently in a wide variety of situations by speakers mì Rrta—here on earth. That will take some time, but we’re on our way. Right now there are already people who use Na’vi on a regular basis for genuine communication. I often receive e-mails written entirely in the language, and there’s at least one blog that’s exclusively in Na’vi.

SoR: Have you considered what the Na'vi language would look like if written? What would you envision it to appear as in literature?

Frommer: Na’vi can certainly be written down—it has its own version of the Roman alphabet and a well-defined spelling system. However, since it wasn’t a written language on Pandora (at least until the Sky People arrived), it didn’t develop its own indigenous writing system. But that hasn’t stopped some artistically minded fans from coming up with their own Na’vi orthographies, some of which are quite beautiful.

SoR: What is a real book you would love to see translated into the Na'vi language and why?

Frommer: Hmm. I think I might start with a version of Aesop’s Fables, but with the animals and plants found on Pandora. That would not only be fun to translate but might serve as a learning tool for students of Na’vi. We could have fables like “The Direhorse and the Stingbat” or “The Boy Who Cried Thanator.” I’d love to read that book!

SoR: Creating certain rules in the language must have appealed to you such as no voiced stops (like b d g) and no vowel length or tone. For someone who could have done anything he wanted, why did you choose to create the language as you did? Were there particular reasons, for instance, in creating the rules as mentioned above?

Frommer: Well, the no-voiced-stops rule came from James Cameron’s initial few words, which as I mentioned sounded Polynesian to me. Polynesian languages typically don’t have those sounds. Other familiar sounds in English not found in Na’vi are those represented by ch, sh, and th (both kinds). In determining your sound system, it’s just as important to decide which sounds you’re excluding as it is which ones you’re including.

You’re right that Na’vi doesn’t have distinctive vowel length or tone. That is, you can’t distinguish words by the length of time you prolong a vowel (as you can in, say, Arabic or classical Latin) or by the pitch of your voice and how it goes up and down (as in, say, Chinese, Vietnamese, and Thai). I presented those possibilities to Cameron, but he didn’t care for them for Na’vi. I think it was a good call. The actors had plenty to keep them on their toes with the ejectives and unusual combinations of sounds. Making them worry about vowel length and tone on top of that would have been too much to ask for.

SoR: What do you feel is the ultimate purpose of language and why do you feel there are so many differences and yet so many similarities in the multitudes of different languages across the globe?

Frommer: You could write books to answer those questions—and many people have! I’ll just say here that language serves several purposes. An obvious one is to function as a tool that can take a thought originating in one mind and put it into another. But language is also part of your identity: the way you speak and write ties you to communities large and small and says important things about who you are, where you come from, who you identify with, and how you fit into your society. Language also reflects, codifies, and transmits culture.

Differences arise because living languages are never static. Over time, words and expressions become obsolete or change meaning, and new words and expressions enter the language. Pronunciation changes too: vowels shift or get lost or are added; consonants drop or get inserted; one sound becomes another. And grammar alters as well. Some languages have undergone rapid change; others have changed more slowly. But none stand still. If you take a look at Shakespeare or the King James Bible, two great examples of Early Modern English, you’ll quickly find examples of usages that are no longer part of contemporary English. Going further back, trying to read Chaucer in the original—a prime example of Middle English—can be daunting, but deciphering it when listening to a recording made using authentic Middle English pronunciation, without the written text, is even more difficult. And the Old English of, say, Beowulf has to be studied like a foreign language.

It’s when groups of speakers of the same language become separated from each other without intercommunication that different languages arise. Since the languages of the two groups, which started out the same, each continue to change but not necessarily in the same direction, two different ways of speaking develop and diverge. When speakers from the two groups are no longer able to understand each other, we have two new but related languages, both “daughters” of the original “parent” language, which is now extinct.

Similarities among languages sometimes often reflect this kind of “familial relationship.” When the Roman Empire broke up, its widespread language, Latin, developed into different languages in different parts of Europe—French, Spanish, Italian, Portuguese, Romanian, Catalan, Occitan, and others—which we now call the romance languages. In the same way, Arabic, Hebrew, and Aramaic are related to each other as members of the Semitic language family. Beyond familial relationships, however, many linguists believe that all human languages share certain deep structural properties that reflect that fact that language is “hard-wired” into our brains from birth.

SoR: How do you feel the language you created for the Na'vi would impact the "culture" that James Cameron created for the species? Examples would be formal and informal use of language, slang...as you know, as a culture evolves, so does the language. Another example would be how in the Na'vi language, gender is optionally marked. So how would that affect the culture in usage?

Frommer: I’ve mentioned some of the ways that Na’vi reflects the culture of its speakers. As for formal and informal use of language, including slang—what linguists call “registers”—Na’vi has those as well. For example, when I translated the Na’vi dialog for the two Avatar-related video games, I was confronted with speakers and situations that hadn’t appeared in the film, including some rough, coarse talk on the part of certain warriors. For that I modified the language in various ways, lopping off parts of words and omitting certain words entirely, to come up with something akin to Na’vi soldier slang.

You’re right that Na’vi gives you more ways to leave gender unmarked than does English. The prime example is the third person pronoun. Our words “I” and “you”—the first and second person pronouns respectively—are not marked for gender: they can refer to either males or females. But in the third person we need to choose “he,” “she,” or “it.” There’s no common pronoun that means either he or she. In Na’vi, however, the most common third-person pronoun is po, which is gender-neutral. If you want to specify gender, however, you can do so: poan is he and poe is she. I’m not sure that reflects anything about Na’vi culture, however. If you look at languages here on earth, Malay and Indonesian have a universal pronoun for he/she, as does spoken (but not written) Mandarin Chinese. Do speakers of those languages look at gender differently than English speakers because of their language? I don’t know. In any event, I find it’s nice to have the freedom of specifying or ignoring such gender differences in Na’vi!