The Avatar alien language isn’t just cinematic flair—it’s a fully realized linguistic system that deepens the world of Pandora. In this episode, OH!CAST explores how linguist Paul Frommer crafted the Na’vi language for James Cameron’s Avatar, blending phonetics, culture, and creativity to build one of film’s most iconic conlangs. Discover how language shapes storytelling, and learn how other languages like Scottish Gaelic (OH!CAST anns a’ Ghaidhlig) are connecting communities through cultural expression.
The Unexpected Invitation
Paul Frommer’s journey into the world of Avatar alien language began unexpectedly. While teaching at the University of Southern California, he received an email from a colleague asking if he’d be interested in developing a language for a new sci-fi film. That film turned out to be Avatar, and Frommer’s background in linguistics and business communication made him the perfect fit.
Building a Language for Pandora
Creating the Na’vi language meant more than inventing words—it required designing a phonetic system that actors could speak naturally. Frommer worked closely with James Cameron to ensure the Avatar alien language felt authentic to the fictional Na’vi culture. He trained cast members like Sigourney Weaver and Zoe Saldana to pronounce the language convincingly, balancing linguistic complexity with cinematic clarity.
Linguistic Influences
Frommer drew inspiration from languages such as Hebrew, Amharic, and Malay. He incorporated features like ejective consonants and infixation to give the Avatar alien language a unique rhythm and structure. These choices added depth and realism, making Na’vi feel like a living language rather than a collection of made-up sounds.
A Growing Community
Since Avatar’s release, the Na’vi language has sparked a global community of learners, linguists, and fans. Online forums, dictionaries, and teaching materials have emerged, allowing people to study and speak the Avatar alien language. Frommer continues to engage with the community, supporting its evolution and expansion.
Language as Worldbuilding
The success of Na’vi highlights how constructed languages can enrich storytelling. The Avatar alien language adds cultural texture to Pandora, helping audiences connect with the Na’vi people on a deeper level. It’s a testament to how language can shape identity, emotion, and narrative in film.
Conclusion: Linguistic Imagination in Cinema
Paul Frommer’s work on the Avatar alien language showcases the power of linguistic creativity in worldbuilding. His dedication turned a fictional tongue into a global phenomenon, inspiring both filmmakers and language enthusiasts. As the Avatar franchise continues, Na’vi remains a symbol of how language can bring imaginary worlds to life.
🎧 Subscribe to OH!CAST for more behind-the-scenes stories from the intersection of language, culture, and geek cinema.
.
🎙️ Full Transcript Outline (Quick Jumps)
🎬 Full Interview Transcript
Cal Macdonald (00:01.383)
Hello everyone and welcome to OCAST, your island gateway to all things geek. I am your host, Cal MacDonald, and, well, as you’re looking at your screen, Fañch is here, which means this language is afoot, so I’ll hand over to Fañch for the rest of introductions. Fañch.
Fañch (00:17.958)
Should I say, thank you. So yes, we should indeed greet in Na’vi tonight. So, hello and welcome to OCAST indeed. We are delighted tonight; we have Paul Frommer, Dr. Paul Frommer, calling us from California. So, evening from us, morning for you.
Paul Frommer (00:21.154)
Very excellent.
Paul Frommer (00:28.492)
Whoa, I’m impressed. Very good.
Paul Frommer (00:46.668)
Indeed.
Fañch (00:47.398)
And we’re going to talk about the Na’vi language. For those who know and who don’t, Avatar by James Cameron, that’s the language that the Na’vi people speak in it. And Dr. Frommer has created that language. So we’re quite excited tonight with Kieran to discuss linguistics, going to that kind of nerdy things, which is totally appropriate for fantasy and sci-fi as we’re all Tolkien fans, and everything started when Tolkien created Elvish a hundred years ago. So, and I guess to begin with, I would like to ask you, Paul, then how did you get on board with that? What happened for a linguist that has worked on Farsi and Malay to end up creating a language for space?
Paul Frommer (01:36.822)
Yeah, well, it was something that was totally unexpected. I sometimes think it was a gift from the gods. So I can’t believe that my experience with Na’vi is now 20 years old. So back in 2005, I was a professor at the University of Southern California. But even though I’m a linguist with a doctorate in linguistics, I was not actually teaching in the linguistics department, but rather was teaching in the business school. And at the time, it turns out that James Cameron was looking for a linguist who could develop a language for a new science fiction film. We didn’t know it was called Avatar.
I think the code name was Project 880 or something like that. And so he and his production company, Lightstorm Entertainment, kind of the natural thing, they went to various universities in the area and solicited a linguist who might be able to do the job. So they contacted UCLA, University of California at Los Angeles. They contacted my university, USC. But they contacted the linguistics department.
And of course, at the time, I was no longer in the linguistics department; I was in another part of the university. So I would never have seen this solicitation, so to speak, if it weren’t for someone I’m very grateful to, a good friend, a mentor, a colleague, my first professor named Ed Finnegan, who saw this email and said to himself, this sounds like Paul.
And I’m very glad he said that because he forwarded me the email and I looked at it, and he said, yeah, I would love to do this. So essentially, I applied for the job. I sent Jim Cameron a copy of the linguistics workbook that Ed and I had put together. It’s called Looking at Languages. It’s a book of data, really, or actually a book of problems for students in an elementary linguistics class.
Paul Frommer (03:53.45)
As it turns out, one of the problems that I had constructed was in Klingon. I had learned just enough Klingon to be able to construct a problem in the syntax for students in, say, Linguistics 101 class. And so I kind of pointed that out to Jim Cameron, and I was very enthusiastic, and a week or so later, I got a message: ‘Why don’t you come and talk to Jim?’
So I had an absolutely magic 90 minutes one-on-one with James Cameron in his office in Santa Monica, California. And it was really quite incredible. He told me his vision for the language and for the movie, showed me some art, which was preliminary versions of the Na’vi themselves and of creatures.
I said, wow, you have an incredible art department. And he says, no, no, no, I drew these myself. I mean, the man is very, very talented. Anyway, we had a great talk and he seemed to think I could do the job and stood up, shook hands, and he said, welcome aboard. And that’s kind of how it started.
Fañch (05:10.68)
And my understanding is that he already had some words himself, had kind of dabbled with it at first. So you didn’t start from zero, did you?
Paul Frommer (05:18.174)
I didn’t start from absolute zero, because James Cameron had come up with, I think it was about 30 words of his own. These were mainly names of characters, names of animals, a few place names and so on. So for example, Eytukan, which was the name of the clan leader, that was from James Cameron. Neytiri and so on.
Palulukan, which is the name of this panther-like, six-legged fierce creature that you see at the beginning of the first movie. And so I looked at those words, and I kind of had a sense of the sound he had in his head. So I said, whatever I come up with should incorporate the sounds and the sound patterns in these words. Now, it doesn’t mean it had to be limited to that. In fact, it wasn’t. So what I did expanded greatly beyond those 30 words, but it allowed those 30 words to be part of the language. So it occurred to me that those words had kind of a Polynesian feel. And so I kind of went in that direction.
So, for example, Polynesian languages typically don’t have voiced stops. They don’t have b, d, g. They have the voiceless counterparts, p, t, k. So that’s something that I included in Na’vi, for example. So Na’vi has voiceless stops, p, t, k, but it doesn’t have b, d, g. At least the original dialect of Na’vi doesn’t have that. That’s a whole other story about how the reef dialect developed. But those are the kinds of things I had in mind.
Paul Frommer (07:10.136)
…trying to give Jim what he had in his ear. And I remember what I did was kind of come up with what I called a couple of possibilities for the sound, some sound palettes, I said. So one thing that I ran by him was the possibility of what linguists call consonant gemination, which is to say that you can make a difference in a word by either holding a consonant for a little bit or just not holding at all. So for example, in Italian, there’s a word nono, N-O-N-O, but there’s also a word nonno, spelled N-O-N-N-O, and then you actually, those are two different words with different meanings. And so I tried out things like atta versus at-ta. Okay, he didn’t like that very much.
Next thing I tried was distinctive vowel length. In many languages, if you hold a vowel longer, it has a different meaning than if you hold a vowel for a shorter period of time. One of my favorite examples is in classical Latin, m-a-l-a. If you hold that first vowel long, māla, it means apples.
If you shorten it, mala, it means evils. So I tried that with Jim. He wasn’t too happy with that. But then the third one was a charm. And that’s where I included these sounds that everyone seems to be interested in. They’re called ejectives.
Normal consonants found in many languages are found in parts of Africa—in Ethiopia, for example—they’re found in parts of Asia, they’re found in a lot of Native American languages. And they’re sort of popping consonants that sound like, [makes ejective sounds]. I wound up, orthographically, I wound up spelling them with an X. So the p-ejective is [makes sound], and it would be spelled p-x-a.
Paul Frommer (09:23.79)
By the way, are those sounds coming across? Because sometimes the software will filter the sound out and try to gloss over it.
Fañch (09:29.894)
Yes, they are, yeah, yeah, yeah.
MacFuz (09:30.687)
Yeah.
Paul Frommer (09:45.088)
So he seemed pretty happy with the sound pattern that had emerged. After that, I was totally on my own in terms of the morphology, which is to say the way you build up the verbs and the nouns and the adjectives and the adverbs, in terms of the syntax, the way you put words together into phrases and sentences. That was totally up to me. And so I kind of had fun doing that.
Fañch (10:11.53)
That sounds like—I mean, these sounds—I saw Kieran’s eyes light up when you mentioned Klingon. I guess Klingon has a lot of these, I guess, they sound very exotic—sounds for, especially for an English speaker. Certainly, when you talk about long consonants, we have them in Breton, my language, and long vowels are a staple of Gaelic that will also make a difference in meaning. So I guess it doesn’t sound as exotic as…
MacFuz (10:18.656)
Mm-hmm.
Paul Frommer (10:17.468)
Okay.
Paul Frommer (10:34.801)
Right.
Paul Frommer (10:36.963)
Uh-huh.
Fañch (10:40.366)
… as ejectives and clicks and all these sort of things. Kieran, did you want to come in on something before we continue?
Paul Frommer (10:41.963)
[laughs]
MacFuz (10:48.491)
So for one thing, obviously you’ve kind of answered this already, that Na’vi itself was inspired by Polynesian languages. Is there any other languages that kind of helped bring it out and bring it together?
Paul Frommer (11:05.55)
Yeah, so I mean.
Paul Frommer (11:11.37)
There has been this sort of almost internet meme that has gotten around that Na’vi is actually based on the Maori language or the Polynesian language of New Zealand. Okay, so that’s not true. I think the reason why this kind of got started is that Jim Cameron had recently come back from New Zealand.
MacFuz (11:20.757)
Mmm.
Paul Frommer (11:35.842)
And I thought that maybe he had some Maori in his ear a little bit. And so it’s true that some of the phonology reflects or is similar to the phonology of some Polynesian languages, but it’s way beyond that. The ejectives, the wealth of consonants, that simply does not exist in Polynesian. But there definitely are other inspirations.
…that went into the creation of the language. I think it’s inevitable that the languages that a language constructor is most familiar with will have some sort of influence on the language that he or she is going to create. So let me give you some examples. The first foreign language that I had exposure to was Hebrew
because if you’re a Jewish kid growing up in New York, you are sent to Hebrew school at an early age to learn how to pray and so on and so forth. So one thing about Hebrew is that it doesn’t have a verb to have. To say, “I have a book,” you say, “there is to me a book.” Yesh li sefer. So I thought that was kind of cool. And so,
I took that little part of Hebrew grammar and incorporated the idea into Na’vi. So in the same way, Na’vi doesn’t have a verb to have. To say “I have a book,” you say, “loeru puk,” which means “there is”—or “to me, there is a book.”
Paul Frommer (13:23.726)
One of the interesting things that I got out of studying Malay was this idea of consonant mutations, initial consonant mutations. Now, for Gaelic speakers, this is very, very familiar, I know. I was in Malaysia in the mid-60s, in the mid-1960s.
I graduated from university in 1965, and then not knowing what I would do with the rest of my life, I made a very good decision and I joined the United States Peace Corps, which had just been started a couple of years before. And I wound up in Malaysia and I was teaching maths and English. And the maths—turned out that I was sent to a Malay medium school where I had to teach in the Malay language. And so I
had to pick up Malay as fast as I could. I did pretty well. But one thing that fascinated me about the Malay language is that it has this series of consonant changes at the beginning of words. So for example, a form of the word to dance is menari. If you were to look up that word in the dictionary, you have to look it up under T, because the root of that word is tari.
T-A-R-I. And through some processes, the T has changed to an N, menari. The word for to unite is menyatukan. Under what letter do you look this up traditionally? Under S. Because the root of menyatukan is satu. Satu means one, and menyatukan means to make one. So S has changed to ny.
T has changed to N and so on. So this kind of fascinated me. And then many, many years later, jump cut to 1998, was my first and only trip to Ireland. I decided I was going to learn a little bit of Irish. Didn’t get very far, but in studying it, I said, hey, this language has some very interesting consonant mutations. And of course, the process is called lenition.
Paul Frommer (15:46.966)
I gather that Irish has another one called eclipsis, which is not found in Gaelic. But there too, consonants at the beginning of a word will change, sometimes rather radically. So I said, hey, that’d be cool to include in Na’vi. And so I did. So Na’vi also has a process of lenition. We actually call it lenition. So,
Fañch (15:50.074)
Yeah, in some dialects only.
Paul Frommer (16:17.72)
For example, the word for world is kiefkì. The word for in is mì. How do you say “in the world?” You think it’d be mì kiefkì, but it’s not. It’s mì hiefkì. So the initial consonant, the k, has changed to an h. And there’s a whole series of these changes that you have to learn. It has to become very natural to you. So there again, that’s an example of something in my own language experience I kind of liked, and I incorporated it into the Na’vi grammatical system. I studied Chinese for a little bit. So, Mandarin Chinese has a number of sentence final particles, which indicate things like encouragement and so on. In order to
tell someone in Mandarin, let’s get going, you’d say, wǒmen zǒu ba, and that ba at the end is a particle, and it means, kind of, let’s do it. Let’s go along with something. And so I said, that’s kind of interesting. And so I included something of that in Na’vi as well. So there’s this particle, ko, which is kind of equivalent to Mandarin ba. And so there are…
Basically, what I’m saying is there are a number of these things where I took a little bit from this language, a little bit from that language, and then there are things that are totally, as far as I know, original, which I’ve never seen in any other language. And yeah, I could give you examples of that if you’re interested, but, well, I mean, a kind of weird thing which occurred to me, and I don’t know of any other language that does this, to say,
Fañch (17:57.456)
Yeah, definitely.
MacFuz (17:59.337)
Yeah, yeah.
Paul Frommer (18:12.438)
you are in a certain place. It occurred to me that when you’re in a place, you’re kind of modifying that place. The place is no longer what it was before because your presence in it has somehow changed the place. I thought, wouldn’t it be interesting if there were a transitive verb, in other words, a verb that has an object, which is a way you would say where you are? So to say that
you’re in the forest, you use a transitive structure, which is, as far as I know, very unusual. So you say, well, ‘I’ and the verb, which is partì—okay, partì is a transitive verb that kind of means “to occupy a space with.” So ‘oe’ is a transitive agent, “I’m occupying a space in,” then ‘na’rengit,’ which means “forest” with an object marker. So I am occupying a space in the forest. So that’s a structure which I think is pretty unusual. I’ve never seen it before, but it seems to be working. So there are a lot of different threads that you can kind of bring together. But the important thing is that the result you get is absolutely unique. There is no other language anywhere
that has precisely the same structure as Na’vi does. And that’s the way it should be.
Fañch (19:44.752)
Did you, yeah, yeah. Did you ever have to hold back? Like, when you talk about all these things as a linguist, I get very excited. These are amazing features. And I’m just wondering, was there at any point in this process of creation where you actually had to hold back and you were thinking, I should add this, I should add that, and you kind of had to prevent yourself from overdoing it to an extent?
MacFuz (19:51.147)
You
Paul Frommer (19:51.383)
[laughs]
Paul Frommer (20:02.829)
That’s an interesting question. Basically, I had several things in mind in creating the language. I kind of anticipated that it would develop a following, which it certainly has. And, you know, it’s been one of the most gratifying things in the whole experience. But that meant that I had to kind of find a balance point between two extremes. One extreme
is making the language so incredibly complicated, which I could, that anyone looking at it would throw up his hands and say, there’s no way I could possibly learn this. Crazy. And the other extreme is to make it so easy that people would say, you know, it’s just not interesting. So to find that balance point and, as you see, it’s kind of a, in a way, it’s kind of holding back,
not to make it so complicated that it’s going to be crazy, but also make it complicated enough so that people wouldn’t say, wow, that’s interesting. And I think I could learn this. To find that balance point was important. And I’m pretty happy with the way that turned out.
Cal Macdonald (21:26.032)
So I’m just gonna bump in here. Were you on the set at all with the actors when you were teaching them their lines for the Na’vi language? Like, how involved were you at that point?
Fañch (21:26.406)
Alright.
Paul Frommer (21:38.514)
That was a big part of my job. And actually, it was a wonderful part of my job. I should tell you that I was tremendously privileged in that I was able to be on set, and I was invited to be on set, and to work with the actors. Of course, I worked with the actors ahead of time as well, providing them
Cal Macdonald (21:41.53)
Alright.
Paul Frommer (22:05.176)
quote unquote, phonetic transcriptions, providing them MP3 recordings so that they could put it on their phone, you know, listen to it while they were working out at the gym, that kind of thing. But I also had the experience of being on set, which was wonderful. And from one take to another, I could sometimes approach an actor and say, hey, that was really good. But just remember the vowel at the end of this word is ‘a’ and not ‘e.’
Okay. And so that really helped with the accuracy. I should tell you that sometimes people were very open to that. Sometimes a particular actor would see me coming towards her and she would kind of run in the other direction. But anyway, that worked out. So, yeah, so I actually worked with the actors. Now, the reason I say I feel privileged is that that’s typically not the case. Typically, what happens is that a language constructor like myself
Cal Macdonald (22:48.196)
Yeah
MacFuz (22:48.885)
You
Paul Frommer (23:04.5)
will develop the material and essentially send it in, and then someone else, typically a dialect coach, will work with the actor to nail the pronunciation. But I was able to do that myself, which is great. And I’ve got to tell you, to be able to sit down with Sigourney Weaver one-on-one and help her pronounce her Na’vi, that was just an experience I never thought I would have.
MacFuz (23:26.847)
Yeah.
Cal Macdonald (23:27.152)
Yeah, that’s…
Cal Macdonald (23:33.744)
Yeah. And I’m just wondering also, did the actors sometimes say some lines different and you thought, that’s a slight improvement to what I had. Did they help you refine the language? Did hearing it used conversationally like that for the first time help refine the language a little bit for you?
Paul Frommer (23:43.022)
Ha
Paul Frommer (23:50.958)
It did. And I’m thinking of one particular example. There was one actor who had to say something. And when I listened to the playback, I said, I don’t know what that is, but it’s not what I intended. OK, so I listened to it carefully. The good thing was that it actually fit into the Na’vi sound system. So I said, OK, well,
it sounds like this is a new word, and what could it possibly mean? So I took a look at what the translation was for the particular line and I said, aha, this word means ‘however.’ And so that’s how we got the word ‘however.’ And that actor does not know
that she actually coined that word, but there it is. So, yeah, there were times when I had to modify things. There were also times—very scary times—when Jim Cameron would come up to me and say, by the way, Paul, we just changed this little scene and so-and-so is going to be saying this, so how do you say that? And it’s something I hadn’t come up with, so I had to come up with something very quickly.
Cal Macdonald (24:47.28)
Yeah
MacFuz (24:48.849)
You
Paul Frommer (25:17.422)
The example that I have in mind is, this in fact did not make it to the movie, but at the time they were thinking of including a little conversation between Jake and some of the Na’vi warriors. And Jake was telling about an experience he had where one of these gigantic flying creatures attacked him. And he said,
‘It almost bit me on my big blue ass.’ Okay, so how do you say ass in… And I had no idea. So I said, give me a few minutes. And so I quickly began to coin some words and I had some actors who were sitting close by and I said, what do you think of this? What do you think of that? And we finally came up with the word ‘tìng’
spelled T-X-I-N-G. And so that is now the Na’vi word for rear end, so to speak. So those were some scary moments where you had to kind of modify the language on the spur of the moment.
Cal Macdonald (26:18.198)
You
MacFuz (26:18.507)
You
Fañch (26:19.056)
Cool.
Fañch (26:28.614)
Go for it, you’ve got that, yeah, Cal.
Cal Macdonald (26:32.284)
It’s just, and probably a burning question a lot of you—How is Jim Cameron to work for? Because we’ve all heard the rumors, the myths, the legends. I’m just very curious. Did you get on with him well? What was he like? I was quite curious about that when you agreed to it. It’s a question I wanted to know as well.
Paul Frommer (26:51.436)
Yeah. I think James Cameron and I had a very good relationship. It was a pleasure to work with him. I found him extremely intelligent and very, very interested in all different aspects of things. Needless to say, he’s a very demanding individual, and he knows what he wants—not just about the language, but in general, on the set. If he
doesn’t get what he wants, he lets you know it. But when he does get what he wants, he’s very happy. So yeah, so I think we had a pretty good relationship and have. It’s an ongoing thing. That’s the most amazing thing because it doesn’t stop. It’s the gift that keeps on giving. The third Avatar movie, which is the second sequel, Avatar: Fire and Ash, is coming out in December.
MacFuz (27:31.838)
You
Paul Frommer (27:45.67)
And it’s already been announced that there are two further sequels that are going to be made. So, hope I’ll be around for that. Fingers crossed. Yeah, right.
MacFuz (27:52.672)
Wow
Cal Macdonald (27:54.182)
So you’ve got work for a few years yet.
MacFuz (28:00.445)
I mean, sorry, just like a note on from that. So, I take it then, is there anyone who’s learning Na’vi under you, that you’re teaching to others that can potentially pass it on or?
Fañch (28:00.806)
One second. Go for it, Kieran.
Paul Frommer (28:14.208)
Yeah, the task, if you like, or the project of teaching Na’vi has really fallen on the Na’vi community, what we call the lì’fya ayllök, the language community, language clan, language tribe.
They’re a remarkable group of very, very dedicated fans and very, very dedicated people, some of whom have learned the language to the point where they—and this is not false modesty—they actually speak it better than I do. And they are more fluent, and they write it better than I do. And they come up with some incredibly wonderful stuff. And they have developed some remarkable teaching materials. We now have two different online sites.
One is called learnnavi.org and the other is called Kelutral, which means “home tree.” And each one has wonderful learning resources. And anyone who wants to learn the language can contact one of these sites and actually kind of hook up with a mentor and be channeled to some of these learning resources. And so…
Even though, a while back, I did a few preliminary lessons—you can find them on YouTube. Some of these were in person. I taught a few classes. But basically, the task has fallen to the community, and they’re doing a great job.
MacFuz (29:48.051)
That’s pretty good to know. Pretty neat.
Paul Frommer (29:50.733)
Yeah.
Fañch (29:52.758)
Thinking about tying together that aspect of things with working on set, I guess what’s interesting about the conlang is the cultural aspect, right? Like we’re talking about the business, countries, Gaelic here and so on. Its language doesn’t exist in the void, right? They exist in connection to a culture, to a history, to a way of life. How did that work for you when you were creating Na’vi? Were you creating it
on its own as its own object, or did you get to work with the designers that were creating the Na’vi as a people? How was that dynamic?
Paul Frommer (30:25.484)
Yeah, so I did not have any input at all into the creation of the Na’vi people. That was a given. That was a totally separate endeavor. But what I did have to do is see what kind of language these already created people would probably speak. One very important assumption
is that the vocal productive mechanism of the Na’vi is essentially equivalent to our own. And that meant that they would produce the same kind of sounds that humans would produce. And that was important because Jim didn’t want to modify any of the Na’vi through electronic means. What you hear is what the actor is actually producing.
Paul Frommer (31:24.014)
But I also had to take into account the environment in which they spoke. I had to take into account the other aspects of their physiology. I mean, one example I like to give is the Na’vi don’t have five fingers or five digits, they have four. And so as I was thinking about that, it occurred to me their counting system is probably not based on 10. It’s probably an octal—based on eight, an octal counting system. So I approached
Jim Cameron and said, what do you think? He said, absolutely. It’s based on eight. And so now the Na’vi counting system is based on base eight. And the way you count from one to eight is a’aw, mune, pxey, tsìng, mrr, pukap, kinä, and vol is eight. Nine is vola, which is eight and one. Ten is volmune, which is eight and two, and so on. So that’s an aspect of
Na’vi physiology, if you like, that I had to take into account. And then, you know, I had to think about what concepts are important to these people. What kind of vocabulary would they have for things that we might not have in any Earth language, like…
The concept of living in harmony with nature is kind of a one word thing. We don’t have the concept of the exhilaration and joy that you feel upon first bonding with your banshee, with your ikran. Okay, we don’t have a word for that, but surely they had a word for that. And so those were some of the kinds of things that I had to take into account.
So, I mean, the difference between the relationship, rather, between language and culture and language and environment—that was definitely an important thing. One of the most fun things was actually to come up with idiomatic expressions in the language that would reflect their experience, reflect their life on Pandora. And so we have…
Paul Frommer (33:37.81)
…we have idioms, we have proverbs. For example,
One proverb that I particularly like because it just sounds really cool to me is, ‘Fwewì ke fwefwi.’ ‘Fwewì ke fwefwi.’ What it means—a fwewì is a kind of mantis. It’s this kind of insect-like thing found on Pandora. Fwefwi is a verb, to whistle. And ke is negative. So, ‘fwewì ke fwefwi’ means “a mantis doesn’t whistle.”
Now, what does that mean? It means don’t expect someone to do something which is not within their natural capacity. So these kinds of things were really a lot of fun to come up with—to use a word play and similar sounds and so on. That was one of the most fun things about creating the language.
Cal Macdonald (34:23.354)
Okay.
Fañch (34:40.72)
Kieran, do you want to come in with something?
MacFuz (34:46.428)
Yeah, I’m just taking it all in. I’m just enjoying how the words just roll off the tongue right there and just how they play together. I’m just enjoying the sounds right now.
Fañch (34:46.746)
Taking it in.
Cal Macdonald (34:55.964)
One thing I’m curious about, Paul—like, you came up with a language in the first film, and then I think it seems to be, with each film we’re introduced to a new kind of tribe of the Na’vi. Have you introduced dialect into the language, like for the water people and I think it’s the ash people in the next film? Like have you…
Paul Frommer (35:15.638)
Yeah, well, I cannot say anything about the next film in that respect. But I can tell you something about the second film, Avatar: The Way of Water. So my assignment for that was specifically to come up with a different dialect of Na’vi. And that was a very interesting assignment. And so if you want to do something which is linguistically valid,
Cal Macdonald (35:21.179)
MacFuz (35:21.381)
Nice try.
Paul Frommer (35:45.985)
along those lines, you have to think in terms of history. Linguists use the terms synchronic and diachronic. So, synchronic is looking at a language at a certain point in time—for example, the present. This is what the language is like. Diachronic is to look at the language across time. And of course, since language, if it’s alive, is constantly going to change, you can think about what the language used to be at a time in the past, how it evolved
to the way it is right now. And so if you want to create another dialect or a parallel dialect, the assumption is that the original dialect that you’ve been working with and the new dialect—both of those are probably daughters of an earlier single parent. And so if you think in those terms, then you say, what would be the characteristics of that earlier form of the language?
How would it have evolved into the familiar form that we’re ready for and also into this other dialect? Because that’s kind of the assumption: that language is constantly changing. But if there were two groups which don’t have much interaction, then the changes in one group will not necessarily be parallel to the change in the other. And eventually you’ll get different dialects. If the changes get even more pronounced, so to speak…
eventually, you’re going to get two different languages. So I had to kind of look at the original dialect, which now we are calling forest Na’vi or Omaticayan Na’vi, and then the new dialect, which I was charged with creating, which is the reef dialect, and to see how both of those might have evolved from an original source. So a given example:
One thing I took a look at was the vowel system of original Na’vi. It turns out there are—let’s see, one, two, three, four, five—there are seven vowels. Is that right? Yeah. But it’s an asymmetrical vowel system in that if you look at the vowel charts, the vowels are not necessarily perfectly symmetrical front to back. So, you know,
Paul Frommer (38:10.674)
vowels are classified by how high or low your tongue is in the mouth, how front or back your tongue is and so on. Whether your lips are round or unround, there are lots of variations of these things. But in Na’vi, there’s a distinction in high front vowels. So we have that distinction in English very strongly. We have “i” versus “ɪ.” So
sit and seat are two different words. One has “ɪ,” which is called a lax vowel, and the other has “i,” which is called a tense vowel. So you have “i” and “ɪ.” But if you look in the back of the mouth, which would be the “u”-kind of vowels, there isn’t any comparable distinction in Na’vi, although there is in English. In English, we have a distinction between suit and soot—”u” versus “ʊ.”
But that distinction does not exist in Na’vi even though the “i” distinction does. So I said, OK, well, maybe our familiar form of Na’vi has lost the distinction in the back, whereas the new dialect would have retained that distinction. And so that’s exactly what I did. So in the new dialect, the reef Na’vi, there is an “ɪ” versus “i” by which you can distinguish words. But there’s also
“ʊ” versus “u” by which you can distinguish words. And so those are some of the things that I had to take into account. I mentioned earlier, I think, forest Na’vi, the one we heard in the first movie, does not have the sounds “b,” “d,” “g.” Well, it turns out that reef Na’vi actually does. And in certain situations in that dialect, the ejectives, the “px,” “tx,” “kx” kind of sounds,
each one will change to a parallel voiced sound under certain situations. Now, aside from the pronunciation, phonological differences, there are differences in vocabulary. I mean, there’s some vocabulary that is used pretty much by the reef people and not so much by the forest people, and so on. But it’s important to keep in mind that
Paul Frommer (40:40.206)
there is mutual comprehension with very, very little difficulty between the two of them, so they couldn’t be that different. But there are now people in the Na’vi community who really like the reef dialect and they do all this writing in the reef dialect, although I think most people are kind of sticking to the original one.
Fañch (41:04.478)
That’s exciting. It is really taking a life of its own, then. Talking about—you were just talking about writing, which leads me to a question. Something I was wondering about in conlangs and when it comes to sci-fi and fantasy: people like to create alphabets. That’s one of the big things, right? They want it not only to sound cool, but to look cool. Na’vi is spelled with a Latin alphabet. Was that part of the decision process, that it didn’t have its own writing system? How did that come about?
Paul Frommer (41:07.704)
Yeah.
Paul Frommer (41:33.815)
Yeah, well, that part was easy for me because the assumption from the original script was that the Na’vi themselves do not have a written language. And so if they had, then I would probably have to come up with—and probably team up with the art department to come up with—some kind of script which was indigenous to the language. But I didn’t do that. And so
if they now have a written language, it probably came from the sky people who arrived. And then there were interactions; of course, they actually had classes. And so the sky people, namely the humans, would probably have used something like the Latin alphabet to adapt it to the Na’vi language. So that really wasn’t a component
of what I had to do, other than come up with some sort of spelling system which would work—a modification of the Roman alphabet which would work for Na’vi, and it works pretty well. Now, some people in the community have come up with their own very beautiful-looking scripts and write many random things, and, you know, that’s fine, but that’s not canon, that’s not official at all.
MacFuz (42:52.288)
Mmm.
Fañch (42:57.766)
I have a kind of a follow-up, except I can’t remember what it is.
Paul Frommer (43:01.768)
Sure.
MacFuz (43:02.379)
I think the only one that comes to mind is, what would have been the trickiest or most difficult sort of hurdle to overcome when creating the Na’vi language and applying it to film?
Paul Frommer (43:18.102)
Most difficult thing, that probably varies from language creator to language creator, but I think a pitfall that’s very easy to fall into is—usually unconsciously—to come up with something in your created language which is simply a reflection of your own native language. In other words, I sometimes have to look back at what I’ve done and say, wait a minute,
is this too Englishy? Not in that the words sound like English, but the construction that I’ve come up with—I might not have realized originally, but you know what? That’s something that English does, and there’s no reason to think that Na’vi would do the same thing. Now, it doesn’t happen very often, but it’s something I found that I had to be very careful about. Because it’s just something, you know, your own language just seems like that’s the natural way of talking for us.
MacFuz (43:51.125)
Yeah.
MacFuz (44:03.957)
Yeah.
MacFuz (44:09.387)
Yeah, because…
Paul Frommer (44:16.258)
But it’s not. It’s only one out of a very large number of possibilities. So I had to kind of be on the lookout for that. I think basically it’s turned out okay. I don’t think there are many places, if any, where people can say, well, that’s just too much like English.
MacFuz (44:38.182)
Because that kind of ties in with the SVO structure that most languages have. So what kind of structure would that apply to in terms of Na’vi?
Paul Frommer (44:49.256)
You mean in terms of word order? SVO? SOV? Yeah, yeah, yeah. Well, that was a very important part of my thinking. And word order is something that I am quite interested in just in terms of general linguistics. And so I tried to come up with a language that really would not have any basic word order.
MacFuz (44:52.135)
Yes.
Paul Frommer (45:19.062)
And so, okay, so you have essentially these three major components of the language, of any language, right? Subject, object, and verb—S, O, V. And mathematically, there are three factorial, or six different possibilities of arranging these things: SOV, SVO, OSV, OVS, and so on.
It turns out that in Earth languages, these are not equally distributed by any means. And so the two biggies are SVO—subject, verb, object—English, Gaelic. And Gaelic is verb first?
Fañch (46:03.532)
Gaelic is verb first.
Yeah, VSO. Yeah.
Fañch (46:16.358)
So let’s say English and French.
Paul Frommer (46:16.362)
Actually, I remember that from the little bit of Irish I did. Thank you for that. But SOV is another big one: subject, object, where verb comes last.
Fañch (46:24.838)
It’s all right.
Paul Frommer (46:35.508)
Object-initial languages that begin with O are very rare, but they do exist. Klingon is an object-initial language, okay. But I tried to come up with a case system so that virtually any ordering of those three components would be grammatical. And in fact, that is the case. So,
We have this sort of robust case system where nouns and pronouns will have little endings on the end of them indicating the function within the sentence—whether they’re transitive subjects, intransitive (which is essentially agents), whether they’re intransitive subjects, whether they’re objects, whether they’re genitives, indirect, and so on. And so because of that, the word order is very flexible.
And that was actually good because if an actor was having a bit of a problem pronouncing the sentence in the order in which I originally gave it to them, I could often sort of change the order without changing the meaning, and it would be easier for the actor to pronounce. That was particularly useful when we did the songs because very often the—
In some cases, the music came first and then I had to kind of fit the language to the song, and being able to manipulate the word order was very useful. But this word order flexibility is something that permeates the Na’vi language. So, for example, we don’t have prepositions like “in,” “on,” “at,” “by,” “for.”
We also don’t have postpositions. We have adpositions. And so those words can come either before or after the noun. So, for example, the word for home is kelku, and the word for at is ro. That’s an adposition. So to say “at home,” I can put it at the end of the word: kelku ro,
Paul Frommer (48:46.606)
which means “at home.” Or I can put it at the beginning. But if I put it at the beginning, then it triggers lenition. So it’s not “ro kelku,” but “ro helku.” So those are things which I’ve kind of enjoyed working with in Na’vi.
Fañch (49:02.649)
It does sound pretty fun, actually.
MacFuz (49:04.028)
Uh-huh.
Paul Frommer (49:04.75)
It was tremendous fun. It really was.
Fañch (49:09.542)
We’ve been giving a lot of love to Na’vi because of just how much it exploded with the community and so on. But you did create another language in 2012, didn’t you? Barsoomian for the film John Carter. Can you tell us a little bit about that experience? And in particular, I’d like to know, did the experience of creating Na’vi inform you, maybe mistakes you didn’t make again with Barsoomian? How did that go?
Paul Frommer (49:38.031)
Yeah, that was really a totally different kind of experience and kind of assignment. And the reason is that the movie, by the way, the name of the movie was John Carter. I don’t know if everyone has seen it. It’s fair to say it didn’t do too well, but it has fans, and it’s still being watched.
Paul Frommer (50:06.126)
The script was based on a series of novels by Edgar Rice Burroughs, who is the creator of Tarzan. And it was early 20th century. And there are seven books. I can’t say I’ve read all seven, but the movie was mainly based on the first book, and I read that cover to cover. And what I discovered was that the author had already come up with something like 400 words.
And so that was a big start. But he was not a linguist by any means. And I had to kind of figure out if he wrote a word with CH, I had to figure out what he meant. And it wasn’t always clear. CH could be ch as in church, but it could also be k as in chorus, or it could be sh as in machine. And so I kind of had to make educated guesses as to what he meant. He had no grammar whatsoever. There was precisely one complete sentence, which consisted of one word, which was “sak,” which means jump. Other than that, just words. And so from that,
I had to sort of create a phonology based on the words and the spellings that he had given in the novel and then come up with the grammar. The other thing which had a tremendous influence on the way I developed that was that the contention in the book was that John Carter, who is the protagonist—he’s a 19th-century American who somehow or other found himself on Mars. It’s not clear how, but he found himself on Mars. And he listened to the local language and felt it was so easy and so simple that within two weeks he was fluent. So that told me it could be a very simple grammar. And that, you know, it should be a very simple grammar. And so I
Paul Frommer (52:16.93)
came up with a pretty simple grammar and fairly easy to pronounce phonological system. I didn’t really go beyond the needs of the movie. And so everything that was supposed to be in Barsoomian in the movie was in the language, and I think pretty well created, but it really didn’t develop a life of its own after the movie. And so I haven’t done anything with it beyond that.
Fañch (52:44.142)
Well, it’s very interesting. You’ve been mentioning Klingon and all these languages. There is a life now for conlangs. It seems that, you know, look, I mean, Game of Thrones, people creating Valyrian and all these sort of things. So there seems to be an interest beyond linguistic nerds into these languages. People really want to know about them. They want to…
Paul Frommer (52:57.292)
Mm-hmm.
Fañch (53:07.546)
…go farther, and for me it was—I’ve seen the Avatar films, but it was a discovery that, no, actually, Na’vi has an entire fan base just for the language as well, and that’s very heartening. That must have been, like, you said you were expecting it from the get-go that there would be some interest. The way it’s gone, has it surprised you?
Paul Frommer (53:14.828)
Yeah.
Paul Frommer (53:28.046)
It has surprised me and has delighted me and is something that was totally unexpected. I mean, Klingon has been around for, I don’t know, 40 years, something like that. And of course, it has a very large fan base. Hamlet has been translated into Klingon. You know, it’s amazing. So, and it has a large and continuing fan base. And I didn’t know if Na’vi would develop anything like that, but I was hoping that…
MacFuz (53:47.275)
You
Cal Macdonald (53:47.828)
Yeah, well.
Paul Frommer (53:58.061)
that maybe some people would be interested. But, Fañch, you’re absolutely right. The whole idea of a constructed language is something which got very little respect, say, when I first started out. And I’ll tell you something interesting.
As I mentioned early on, when this solicitation went out to various linguistic departments asking for a linguist who could take on this role, there was very little response. I think I may have been the only one person who responded. Why? Because linguists looked at it and said, this is nonsense. We’re not supposed to be making up languages. We are scientists. We are studying languages. We’re trying to bring
Cal Macdonald (54:34.544)
Really?
Paul Frommer (54:50.094)
extinct and dying languages, endangered languages back from the dead, we shouldn’t be doing this kind of stuff. Well, you know, one thing doesn’t preclude the other; you can kind of do both. Long story short, in the interim, languages have developed kind of a following. And one of the best indicators of this is that there are
linguistics courses in prestigious linguistics departments pretty much all over the world devoted to constructed languages, where the students in these classes as the semester project have to construct the beginning of their own constructed language. I’ve had the pleasure of speaking to some of these classes. And one of the most delightful things is that at the end of my talk, each student
would take a few minutes and tell me about his or her constructed language and say a few things in the language. There’s no better way, pedagogically speaking, to learn about something like linguistics than actually doing it yourself and seeing what the difficulties are, what pitfalls you fall into, why you have to do certain things one way and not another. It’s been quite amazing. So
So yeah, I think constructed languages, or conlangs as we call them, really are coming into their own. There was a time when someone with a science fiction script or a fantasy script where someone’s supposed to be speaking an alien language would say, you know, just make up some stuff, gibberish, nobody’s going to know. Who’s going to know if it’s right or wrong? Well, that
does not cut it at all anymore. I mean, Klingon was probably the major influence, but nowadays any producer or director worth their salt will actually contact a linguist and say, you know, we need a language—a well-constructed, linguistically viable language for this project. And so that’s been very encouraging.
Cal Macdonald (57:12.732)
I just want to ask, see the community you said you have online, how much do you guide them? Do they go to you as being the ultimate authority or do you just kind of let them progress their language on their own? Like, how much of a hand do you have in it?
Paul Frommer (57:22.626)
Yeah. I try to keep somewhat of a hands-off policy. I don’t want people to think that I’m always there looking over their shoulder. However, the language is never finished. Na’vi is not finished. We’re constantly adding new vocabulary. We’re constantly refining some of the grammatical points and so on.
At this point, I think we have somewhat over 3,000 words. That’s not a lot. But as needs come up, people will say, you know, could we have a word for this and so on. One thing that I think is unique in our community—I could be wrong, but I think we may be the only ones. Early on, people in the community would come to me and say, hey, we want to talk about such and such, so could you give me a word for this, for this, for this, for this?
And I said, OK, great. And I would try to come up with it. But I quickly realized that there are some people who had embraced the language to such an extent that they knew very well what could be a possible Na’vi word and what could not. They knew all the phonotactic constraints. They knew what consonants could be at the end of a word, what could not be at the end of a word, and so on and so forth. And I said to myself, you know what?
these people could probably not just request new vocabulary, but propose new vocabulary. Come up with a word yourself. And so this has continued. And we have something called the Lexical Expansion Project, where every month or every two months, I will get a document with maybe a dozen suggestions for new vocabulary. And these come from the community. It’s all submitted to me anonymously.
I can take a look at any of these terms and make one of three decisions. One is, I love it—it’s part of the language. Another way to go is, I think it has promise; I’d like to tweak it a little bit, go in a slightly different direction. And the third is, I’d rather do this in a different way. But what I’m saying is that I am still the gatekeeper.
Paul Frommer (59:40.843)
I’m still the only one who determines whether or not a word is an official part of the language or not. So, in no sense is the language now crowdsourced, but the suggestions that I’m getting from the community will very often determine the direction I go in creating new vocabulary. And I love giving people that sense of—it’s kind of a sense of ownership.
You know, who owns the language? Who owns English? Who owns Gaelic? Who owns the language that they speak? Well, we own it. It’s part of us. We’re the speakers of the language. And so to give someone a chance to sort of contribute to that is something that makes me very happy.
Cal Macdonald (01:00:29.358)
All right, yeah. Very good.
Fañch (01:00:30.758)
Do you ever speak Na’vi?
Paul Frommer (01:00:33.863)
I’m sorry?
Fañch (01:00:33.958)
Have you ever gotten to speak Na’vi with somebody?
Paul Frommer (01:00:38.222)
Yeah, I mean, people in the community will speak it all the time. There are get-togethers online where you must speak Na’vi; that’s the only thing you can speak. There is at least one situation that I came across where there were two individuals in the Na’vi community who did not share an actual Earth language.
And so the only way that they could communicate really was in Na’vi. My own spoken Na’vi—pronunciation is pretty good, but my fluency is not as great as the fluency of some of these people who use it a lot more than I do. And so that whole thing really has been very gratifying to me.
Cal Macdonald (01:01:04.92)
Yeah
Fañch (01:01:15.097)
You
Cal Macdonald (01:01:31.036)
That’s really cool. That has just come up on the hour mark there, so I will go to Kieran and Fañch for their final thoughts or questions or comments.
Fañch (01:01:31.201)
It sounds like it, yeah.
Paul Frommer (01:01:32.706)
Yeah.
Fañch (01:01:46.0)
Kieran, please go.
MacFuz (01:01:47.599)
Okay, well, I’ve just enjoyed just hearing you introduce your little aspects of how Na’vi came to be, as well as the pronunciation. I’m just here just to love listening to the pronunciations of how the words just all fall into place, as well as the history that helped you create it, as well as the impact it has on communities, which is pretty cool.
Paul Frommer (01:02:10.03)
Great.
Paul Frommer (01:02:16.172)
Yeah, really. Thank you.
Fañch (01:02:19.104)
It was fascinating and certainly I know that when I see the next film as well, I’ll be paying particular attention to that part of things. Yeah, no, it’s been a wonderful discovery for me to actually look into the community and to hear it from you was a privilege. So thank you, irayo, for everything.
MacFuz (01:02:26.091)
Paul Frommer (01:02:27.059)
Ha ha ha ha!
MacFuz (01:02:28.895)
Definitely.
MacFuz (01:02:37.419)
Hmm
Paul Frommer (01:02:38.542)
Oel ngati kameie, you’re quite welcome. It’s a pleasure. And I will leave you with the famous greeting, “I see you,” which I have to modify now because there’s a special way of referring to three people. In Na’vi, we have singular, we have a dual, we have a trial, and then we have a plural. The plural is four or more.
Cal Macdonald (01:02:40.848)
It’s
MacFuz (01:02:55.742)
Okay.
Paul Frommer (01:03:05.494)
So if I were speaking to one of you and said, “I see you,” it would be, oel ngati kameie. But to speak to three of you, I would say, oel ngat kameiem, which means “I see you all three.”
MacFuz (01:03:19.179)
All right.
Cal Macdonald (01:03:19.59)
Cool, very cool. For the people listening, where can they find you or reach out to you if they’ve got any more questions or want to know more about Na’vi?
Paul Frommer (01:03:20.206)
Okay.
Paul Frommer (01:03:28.494)
Probably the easiest way—I mean, I can just give out my email address, but you could find it online. It’s my last name, Frommer, F-R-O-M-M-E-R, at Marshall, M-A-R-S-H-A-L-L, dot U-S-C dot E-D-U. And if anybody would like to email me with a question, I will try to get back to you.
Cal Macdonald (01:03:58.278)
Thank you very, very much. It’s been really fascinating to us. And just a little…
Paul Frommer (01:04:02.082)
My pleasure. And I must say I have never spoken to people in the Outer Hebrides before, and it’s been a real pleasure.
MacFuz (01:04:08.907)
It’s a first for everything.
Cal Macdonald (01:04:09.102)
All right. Yeah, for sure. It’s just very interesting because this time next week we have the Klingon Pop Warrior on, which ties into this quite nicely. You know her? You’ve heard of her, have you?
Paul Frommer (01:04:18.51)
Okay.
Paul Frommer (01:04:24.789)
I don’t know if I’ve heard of her, but, you know, I know Marc Okrand, who is the creator of Klingon, and we’re kind of friends.
Cal Macdonald (01:04:28.058)
PISSED
Cal Macdonald (01:04:31.706)
Yeah, Klingon Pop Warrior just basically takes popular songs and translates them into Klingon and then she sings them. She’s on Spotify, so… You might want to tune in for that yourself. It’s about the same time next week, so might interest you.
Paul Frommer (01:04:38.478)
Cool, wow, I love it.
Paul Frommer (01:04:46.676)
Absolutely. Yeah, that sounds great. Wonderful.
Cal Macdonald (01:04:50.638)
All right, so, nothing else to say but just thank you very much for coming on and thanks to my co-hosts. It’s been fascinating, loved it, and we’ll see everyone next time on the show. Take care everyone. Bye.
Paul Frommer (01:05:00.662)
My pleasure, my pleasure. Kìyevame, take care.
MacFuz (01:05:01.515)
Bye for now.
