From Learn Na'vi Wiki
Jump to navigation Jump to search

The Na'vi language canon is the complete collection of information about the Na'vi language provided by authoritative sources, namely Paul Frommer and the creators of Avatar (James Cameron and Twentieth Century Fox).

The canon comprises two things:

  • words and phrases spoken or written in Na'vi
  • descriptions of the linguistic elements of Na'vi such as orthography, morphology, syntax, and grammar

The Na'vi words and phrases from canonical sources are presented or linked to on the Corpus page (where copyright allows). Documentation, explanation, and analysis of Na'vi linguistics are presented on other Learn Na'vi wiki pages (Phonology, Morphology, Grammar, etc.). This page serves to document the canonical sources themselves.

This page includes information dated Dec 2009 - Feb 2010. To access more recent information please see:

List of Canonical Sources

The canonical material we have available to date includes:

  • The film Avatar
  • The Activist Survival Guide
  • Paul Frommer's post to Language Log and his later reply to comments there
  • The media interviews Frommer has given (mostly around the release of the movie)
  • Private correspondence between Frommer and people interested in his language
  • The form letter Frommer emailed to his many correspondents in late 2009 and early 2010
  • The out-of-office auto-reply message Frommer set up during the semester break in December 2009
  • A partial page of Na'vi dialog from Avatar used by Sam Worthington ("Jake Sully"), as seen in close-up during a behind-the-scenes video released by Fox
  • "A Message From Paul" - his message to the Na'vi language community on January 21, 2010
  • The materials he prepared for the March 5 Good Morning America spot

Commercial Sources

The film Avatar (IMDB link) contains numerous examples of both human and Na'vi characters speaking in Na'vi. Some of the Na'vi dialog is subtitled in English (or in the local language where Avatar has been released in non-English-speaking countries).

Avatar: A Confidential Report on the Biological and Social History of Pandora aka "The Activist Survival Guide" or "ASG" (Amazon link, Browsable link from publisher) is an official companion book to the movie. It includes a Na'vi-English word list and two Na'vi songs with English translations. It also includes an account of animal and plant life containing some "Na'vi" words that violate what we know of Na'vi phonetics.

Na'vi Dialog for Jake

From a Fox-authorized video on YouTube, Avatar The Movie James Cameron Avatar The Movie Behind The Scenes Making The Movie 2. There's a brief video close-up of some of Jake's Na'vi dialog, "Na'vi Dialog for Jake - 3-13-2007." The stress accent is marked with the underline. See also this still-frame image from the video. See the Corpus page for a transcription. See also this forum thread for discussion. Since the dialog is Jake's, there is some concern the sentences may not be perfectly grammatical.

Interviews with Paul Frommer

Shortly before and after Avatar's release, Paul Frommer gave interviews to several media outlets in which he discussed the Na'vi language and its use in the film. He provided some examples of the language (written in print interviews, spoken in audio and video interviews). He also explained some of its unusual features and talked generally about how the language was developed and the work he did for the film. As the interviews and media stories were intended for general audiences, they provide little detailed information about Na'vi linguistics.

A complete list of Frommer's media appearances is on the Resources page (Print; Audio and Video).

Paul Frommer's Language Log Post

On December 19, 2009 (the day after Avatar's release) Paul Frommer made a guest post on the linguistics blog Language Log titled "Some highlights of Na’vi". In that post, he provided an extensive but by no means complete description of the Na'vi language covering phonetics and phonology, word classes and morphology, and syntax. The blog post attracted many comments, and on December 21 Frommer posted a reply addressing many of them, providing clarifications and some additional information about the language.

To date, Frommer's post remains the largest single direct source of information about Na'vi linguistics. A great deal of additional information has been learned by analyzing the Corpus, and Frommer has provided additional information in smaller amounts as described elsewhere on this page, but the Language Log post provided the foundation upon which our understanding of Na'vi has been built.

Public Email From Paul Frommer

In late December 2009 and early January 2010, Paul Frommer began replying to his many email correspondents with a form letter, reproduced here from this post on the Learn Na'vi forum:

Kaltxì, and thanks for your interest in Na’vi.

The enthusiastic response to the language has been very gratifying. I wish I could answer all your e-mails personally, but the volume has been staggering, so I hope you’ll forgive this generic message.

Irayo to everyone who thanked me for the effort and wished me well. It’s been a privilege to be involved in something as extraordinary as “Avatar,” and I couldn’t be happier that people feel my contribution added to the film.

For those who have expressed interest in learning the language, thank you! The way Na’vi will live and grow is for enthusiasts like yourselves to take it up and expand it beyond its present boundaries. I hope that not too long from now there will be learning tools available to make that possible. Film consultants like me, however, don’t own the rights to the products or services we provide, so I can’t put out any such materials on my own—they’ll have to be in cooperation with the people who brought you “Avatar.” We’ve already begun thinking about that, and I hope that some official form of “Learn Na’vi!” will soon be available to the public.

In the meantime, a number of people have put up unofficial web sites talking about various aspects of Na’vi. Some of those are quite good, others are only speculative, and still others are highly inaccurate. The Wikipedia article on Na’vi, although not a complete description of the language, is well done and reliable.

Kìyevame ulte Eywa ngahu. See you again, and may Eywa be with you.

Paul Frommer

Around the same time period, during semester break at the University of Southern California where he teaches, correspondents emailing Frommer received an out-of-office auto-reply message which included the following sign-off:

Ayftozä lefpom ayngaru nìwotx! Happy Holidays to You All!

A Message From Paul

On January 20, Paul Frommer sent a message to the Na'vi language community expressing his appreciation for the community's interest in the language. The message was sent as Na'vi text, English text, and audio of Frommer speaking the message in Na'vi. Frommer also provided a version of the text marked for stress.

To date, this message is the largest single authoritative example of written and spoken Na'vi (other than the film dialog, for which there are no canonical Na'vi transcriptions). Consisting of fourteen sentences, many of them lengthy and complex, the message contained previously unseen grammatical constructs and added many new words to the Na'vi lexicon. Frommer signed the message "Your friend and teacher, Paul."

The message can be read and heard in its entirety at the site The stress-marked version is on the Corpus page at A Message From Paul.

Private Correspondence With Paul Frommer

Many people interested in learning more about Na'vi have sent questions to Paul Frommer. Frommer is not always able to reply, but he has on several occasions given answers which have provided greater insight into the language. Some of those answers have been shared with the Na'vi language community; those answers are documented here.

Note: communication between Dr. Frommer and his correspondents is inherently private. How much of that communication to share, and whether to share anything at all, is a personal decision. No one should ever feel obligated to share anything, and no one should ever expect or demand that others do so.

Note also: Dr. Frommer is a party in these communications, and his privacy must be respected as well. If you have received something from Dr. Frommer and are considering sharing it with the community, do not do so until you are confident that he has no objections.

If you have new information from Dr. Frommer about the Na'vi language that has not been published elsewhere, and you are comfortable sharing it, and you are certain that he does not mind you doing so, then (and only then) please consider making it available to the Na'vi community so that it can become part of the canon and improve everyone's knowledge.

If you do publish your correspondence, please only publish the parts that are directly related to the Na'vi language, such as new words and phrases or new grammatical rules. Post excerpts only. Leave out anything that might be a personal matter for either you or Dr. Frommer. When possible, post direct quotations from Dr. Frommer so that we can all see what his exact words and intentions are. When you are not comfortable giving direct quotes, a summary or paraphrase is still valuable and welcome, but take care to not present your own interpretation as his words.

Extracts from various emails

Added by roger. Dates are date of answer by Frommer

Dec 15
Kaltxì ngaru.
Dec 18

on monosyllabic verbs

And yes, you can see where längu comes from. Lu, as you know, is ‘to be.’ So you have inflected forms like lolu, leru, layu, etc. Those are all with 1st position infixes. Of course you can also have forms like lereiu, lolängu, etc.
One thing I didn’t mention in the post is that whatever vowel receives stress in the root retains that stress in the inflected forms. So all of these forms of lu are stressed on the last syllable.
Dec 21
Glad you liked the x-digraphs. I'm not thrilled about digraphs in general (as I've indicated, I really wanted to use c instead of ts, as Pinyin does, and g instead of ng, as in Pago Pago), but using them here seemed to be the best choice. Since the apostrophe already meant a glottal stop, trying to use it for the ejectives as well would have resulted in ambiguities: would a sequence like hap'a indicate an ejective (i.e., ha-pxa) or a p closing one syllable followed by a glottal stop opening another (hap-'a)? They're not the same thing!
Dec 27

on why le-adjectives don't take a, and on ergative case with futa

The one time you don't need to use the adjectival a is with le- adjectives that follow the noun. (Having two consecutive markers in a row relating to adjectives seemed redundant. Also, I bet the a + le- combination would tend to decay over time on phonological grounds.) It's not wrong to put in the a in such cases, but it can be optionally omitted, and usually is.
If a le- adjective precedes its noun, however, you need the a.
ayftxozä lefpom
ayftxozä alefpom
*lefpom ayftxozä
lefpoma ayftxozä
As for futa and oel:
Futa is a contraction of fì'ut a. Fì'u means 'this thing, object, idea, fact, abstraction.' Here it has the objective (or "patientive") marker t.
The idea is that fpìl 'think' is transitive: you think SOMETHING, which is a thought. (It's transitive in the same way that "create" is transitive: you bring something, the object created, into being.) The a is just the usual "modifying a," as in the adjective examples above.
So the sense is, "I thought this thing, (namely) that . . . " And therefore oel is in the agentive or ergative case, while fì'ut is in the objective or patientive case.
Dec 28

I also take it that "N le-ADJ lu" is ambiguous between attributive adj. w existential vb and predicative adj. w copula. ... I assume <arm> is <a<er>m> PST.IMPV

Right. I don't think that's a problem, but if need be, a speaker could disambiguate by including the a (optional here) for the attributive case and postposing the adjective after the copula for the predicative case.

Are the esses in slu and sìltsan related?

As for slu/lu, I haven't developed anything like a consistent s-derivation. (I could think about that, though.) The sìltsan/nìltsan pair isn't related—nìltsan is simply a development of *nìsìltsan, which I think could be natural. I wanted to avoid a 3-syllable word for 'well'!
Good guesses about -arm- and -iv-. ;-)
Dec 31

In Tsakrr nga Na'viru yomtìyìng "then you will feed the People", shouldn't nga be in the ergative?

You'd think the verb is transitive, which would require the ergative case. But yomtìyìng 'feed' is from yomtìng, a compound: yom 'eat' + tìng 'give.' So there's no DO, and the verb is thought of intransitively, along the lines of 'engage in eat-giving.' Na'viru, as you know, is in the dative. So the whole thing is rather like, "Then you will engage in (or do) eat-giving to/for the people." There's a whole class of verbs that work on a similar principle.
To say "you feed the people meat," you need a different structure. It would probably just be with the verb tìng, where 'you' is ergative/agentive, 'meat' objective/patientive, and 'the people' dative.
A bit odd, perhaps, but I think it works.
Mipa Zìsìt Lefpom Ngaru!
Dec 31, on the SG
You're quite right that the words you've listed below are not at all in keeping with Na'vi phonology. I've never seen them before.
As for aylaberu, the b should be an h.
Jan 07
Ngaru irayo seiyi oe nìmun.
Jan 09
There's no -ev- infix. The line from the film "Let her walk among us . . ." is:
Tivìran po ayoekip . . .
(Walk: tìran. The infix -iv- is subjunctive, here an "optative." Po is used for he/she unless it's desirable to make a gender distinction--e.g. to avoid ambiguity in discourse.)
About the movie dialog: Keep in mind that there's variation in all languages. Some of it is due to dialectal differences, some to performance errors, some to fast-speech phenomena . . . In other words, it's not necessarily the case that everything you hear in the movie theater prefectly represents standard, canonical Na'vi. Also, remember that Jake, Grace, and Norm speak Na'vi as a second language! It would be especially unrealistic to expect them to be perfect.
As for ko, I don't know any Japanese but I did have in mind Mandarin ba, a sentence-final particle glossed by Li and Thompson as "solicit agreement." Typical translations of ba are: "Don't you think so?" "Wouldn't you agree?" "Let's . . ." "Why don't you . . ." "I'll do X, OK?" I've used ko for all such things, maybe a bit more widely than ba.
Jan 19
Sìpawmìri oe ngaru seiyi irayo.
Jan 20

Why is there no overt SBRD w modals?

That's how I've been handling want, can, and must: via a simple, shorter structure, where the subject in the main clause is considered intransitive, and there's no overt subject in the subordinate clause if they're identical:
Oe new kivä. 'I want to go.'
Oe tsun kivä. 'I can go.'
Oe zene kivä. 'I must go.'
(Different word orders are possible in all of these, of course: Zene oe kivä, etc.)
So in these cases, the -iv- form of the verb functions somewhat like an infinitive. (But never like a gerund!)
You can also use the longer form:
Oel new futa (oe) kivä.
But that's less common.
Jan 20
Oh, btw, for pxeeveng etc. . . . Yeah, you're right. It should contract to pxeveng. (Likewise for *meeveng > meveng.) I've gone back and forth on the contraction question when two identical vowels naturally come together as they do here. I should be consistent.
Jan 20

So, what if the verb after the modal is transitive?

If the 2nd verb has the same subject as the first, it's pretty straightforward:
Oe new yivom teylut. 'I want to eat teylu.'
(Back in the dark ages when I had my first syntax courses, that would have been termed an instance of equi-NP deletion.)

[section on when the subject differs removed for the time being]

Jan 21

Do we get that long form w other modals?

No, with tsun and zene, the plain subjunctive is the only possibility.
Jan 20,

on our proofreading of his message

Eylanur oeyä Kwami sì Prrton . . .
Ma Kwami:
Tìsteftxaw ngeyä flolä nìltsan! [ flä = 'succeed' ]
Just a couple of things to adjust:
yawn is not actually the verb 'love.' Rather, 'yawne means 'beloved.' (For the noun, the e has dropped. Happens a lot.) So . . . how would you say 'I love you?' :-)
'Work' needs some explanation, which I'll provide when I have a bit more time. Hint: kan is the verb meaning 'aim.'
'Best' is simply swey.
Rule is koren. If it were horen, then there would be no short form for the plural—you'd have to say ayhoren.
Ma Prrton:
'Ngal oeyä 'upxaret aysuteru fpole' a fì'uri, ngaru irayo seiyi oe nìtxan!
(Alternately: Furia ngal . . . fpole', ngaru . . . )
[I'm debating whether standard punctuation should include the comma.]
Jan 20
Eywa'evengä is OK. Eywa'eveng = Pandora, lit. 'child of Eywa.' Jim's coinage, not mine. (He has the right!) But it works as a compound.
Jan 21

on our guesses for 'I love you'

Re "love": The "indirect" way is correct.
I've been playing around with word orders to come up with the most euphonious version. For me, it's:
Nga yawne lu oer.
Five syllables (sentence stress on YAWN, of course), not too long . . . sounds good, I think. The laudatory infix (as someone has dubbed it) -ie- is possible but not necessary, and I've been omitting it here. "Love" itself is positive enough that it's probably redundant, and all other things being equal, shorter is better.
As for the sentence with tìyawn si, yes, that might mean something different. I'll have to think about it!
OK, here's a grammatical point I haven't yet talked about explicitly:
"Si constructions" have a special syntax. They're considered intransitive--a bit strange, perhaps, but reasonable, I think. That is, "X si" is thought of as "engage in the X-activity," an intransitive concept. What would normally be considered the object is then in the dative, along the lines of, "engage in the X-activity to/for Y."
So rather than *Oel ngati tìyawn si," we have Oe ngaru tìyawn si.
The literal, pidgen-English translation would be something like "I do loving to/for you."
As I said, I'll to think about the implications of that!
BTW, [...] I may have forgotten to mention this in the LL post, but the "pseudo-vowels" rr and ll only exist in open syllables. (If you see I've violated that anywhere, please let me know . . . it's not impossible.) [...] Lu ngeyä tìftxey!
Jan 21

But the stress shift in oe would result in the consonant clusters ?/mw/ and ?/pxw/ in the pronouns.

No new C-clusters! The phonotactics are [in]violable.
In the dual and trial, the stress on the oe element does not shift. So pxoengaru has 4 syllables.
In the 1st pers. dual inclusive, however, the problem doesn't arise, because there's no me- prefix. It's just "you (sg.) and me," i.e. naturally dual, so you don't need the marker. The forms are simply oeng (N.B.: no -a), oengal, oengat(i), etc.
Jan 21

How about Plltxe fko san X fyape nìNa'vi? or X sìk nìNa'vi plltxe fxo fyape? for "How do you say X in Na'vi?"

There's an easier and more concise way:
X nìNa'vi slu upe?
(As you know, slu = 'become')
It's an example of a neat structure I adapted from Persian:
X be-faarsi chi mishe? X in-Persian what becomes?
BTW, I've been thinking of san . . . sìk as correlatives, always occurring together. The only exception is if you're quoting someone and you end with material still in the quote. Then you don't have to say "unquote."

I'm not following.

Well, let's see . . .
Suppose the sentence is, "Eytukan said he would go, but I don't believe him."
Everything converts to direct speech, so it would be:
Poltxe Eytukan san oe kayä sìk, slä oel pot ke spaw.
(Note: *pollltxe > poltxe )
But if it's simply "Eytukan said he would go," and you end your speech turn there, then it's not necessary to have the concluding sìk—it's "understood."

Ah, okay. So "he said X" would use just san, "X he said" would use just sìk, right?

Yes, "He said X" would just use san if you ended there and didn't continue.
As for whether "X he said" would only use sìk and omit the initial san . . . I think that depends.
Suppose in a conversation in English you said the following:
(1) "I'm planning to sell everything I have and move to Nepal, where I'm going to take up mountain climbing and flower arranging," he said.
Since your listeners can't see the quotation marks, would it be clear to them from the start that you're talking about someone else's plans, not your own? It probably depends heavily on the context. Come to think of it, though, (1) would be very unnatural in conversation, wouldn't it? You see that structure in written material all the time, but people don't speak that way. So "X he said" questions might be rare in discourse.

And it looks like only direct speech is allowed?


In 'how do you say', was 'upe rather than upe a typo?

Not a typo, a goof! It is of course 'upe, with stress on the first syllable. Thanks for the catch.
The other form, as you know, is peu, without the glottal stop, stress on the second.
So you can also have:
X nìNa'vi slu peu?
Either one is fine.
Jan 27

on making a mistake:

Oel foru fìaylì'ut tolìng a krr, kxawm oe harmahängaw.
[hahaw = sleep, stress on 1st]
[-arm- is one of the fused infixes that combine tense and aspect.]
Irayo nìmun.
Jan 31

on asking how to say 'good night'

Let's make it txon lefpom.
lefpom = happy, peaceful, joyous (not for people)
A bit expected, perhaps, but I think it has a nice quality, with the two o's.
As you know, le- adjectives don't take "modifying a" when post-nominal.
The full version would be, "Txon lefpom livu ngar."
Feb 1

what's the diff tween 'vay' and 'mi'?

Vay (ADP-) means 'up to.' It can be used in several senses, not just temporal: "He counted up to 35" or "Follow the river up to the land of destruction" (a line from one of the video games).
Note vaykrr (CONJ): until.

how is dropping the 'a' from 'taluna' and 'taweyka' affected by swapping clause order?

Yes, good point--the 'a' can drop only if the construction is head first, with the reason clause following. There's precedent for that kind of thing, however. With le- adjectives, the 'a' drops when the adjective is post-nominal but not when it's pre-:
txon lefpom
  • lefpom txon
lefpoma txon
I hadn't thought of atalun vs. alunta! Interesting. Let me consider that, and whether it has implications for other conjunctions. For the time being, though, let's make the canonical form alunta.

Pronunciation of a

Quoted by Taronyu, January 25, 2010 (source).

In your Language Log, you show in the trapezium that [a] is the low back vowel, in contrast to [æ]. Did you mean to say [ɑ]? I am aware that you use an old American system.
I meant [a], a low central vowel; essentially the corresponding vowel in Spanish and Italian, or in the standard American pronunciation of 'hot.' So you're right: phonetically it's not really back.


Quoted by Taronyu, January 25, 2010 (source).

Is transitivity lexically marked: you've shown that 'promise' is ditransitive in the guide, but what about the other words? Will you be showing us how to do this soon?

Well, I've been pretty much leaving it up to the semantics to determine what's transitive and what's not, when it's clear. "Sleep" is clearly intransitive, for example; "kill" is clearly transitive. So I don't need to annotate those in the lexicon. For verbs like 'begin,' however, you have to be told which one you're talking about. So for those, I've indicated the type in the complete glossary. Until that's published, let me know what's not obvious and I'll get back to you.

Monosyllabic Stress

Quoted by Taronyu, January 25, 2010 (source).

I've noticed in a shot of a script that you stress monosyllabic words when they have an inflection (and therefore are polysyllabic), but not otherwise. Should all monosyllabic words be stressed normally?
Yes indeed. I underlined the stressed syllables in polysyllabic words for the actors, so they could pronounce their lines correctly. I've done that in the glossary too for all polysyllabic words.
The rule for inflected verbs is that whatever vowel is stressed in the root keeps its stress in the inflected form. So for example, the root for hunt, as you know, is taron, stress on the first syllable. So all the inflected forms retain stress on the original a of the root: tivaron, tolaron, tayarängon, etc. Cf. pängkxo "chat, converse," where the stress on the ultima. The stress stays there in the inflected forms: pivängkxo, polängkxo, payängkxängo.

sä- Prefix

Quoted by Taronyu, January 25, 2010 (source).

Is sänume indicative of a - nomilinalizer to show the instrument of the verb?
The - prefix is a bit of a loose end. So far I only have one example of it in the lexicon: nume "learn" vs. sänume "teaching, instruction." You're right: I was thinking of it as something like an instrumental affix: instruction is the thing BY MEANS OF WHICH you learn. On that basis, since mun'i is the verb "cut," sämun'i could be a word for a general cutting instrument. I'll need to think more about that, however, since I'm not sure these two cases are comparable. In the second case, you have a concrete instrument, which must be present in order to implement the verb. With nume, though, you can learn by other means than teaching: from experience, from trial and error, etc. So should sänume refer to any means whatsoever of learning something? I'm not sure yet. Thanks for the question!

Verb Phrases as Objects

Quoted by Taronyu, January 25, 2010 (source).

If a verb is understood to be transitive (depending on your answer above), when the direct object is replaced by a verb phrase, is the subject still marked as ergative? Essentially, which is more correct: Oe new pivlltxe or Oel new pivlltxe?

Good question.
First note that for "can" and "must," the subject is considered intransitive:
Oe tsun kivä. 'I can go.' (NOT *Oel tsun kivä.)
Oe zene kivä. 'I must go.' (NOT *Oel zene kivä.)
"Want" MAY follow the same pattern:
Oe new kivä. (That is, it patterns like a modal.)
But since want, unlike can and must, is a transitive verb, there's an alternate pattern:
Oel new futa (= fì'ut a) kivä.
And you also have sentences like:
Oel new futa Taronyu kivä. "I want Taronyu to go."
So in the sentences you asked about, Oe new pivlltxe is fine. In one with oel, though, insert futa.

Later in the same message:

BTW, let's take a look at your sentence again:
(1) Oel new pivlltxe nìNa'vi mì oeyä letrra tìrey, (2) slä oel tsun pivey (3) trrit a ngat taying (oe new tìying!) ayoe nì'ul aylì'u!
(1) and (2) are virtually perfect: just change oel to oe in both cases. I like letrr for daily! I'll add that to the glossary, with RL in the Source column. :-)
(3), though, needs a little help. Are you awaiting "the day that you will give us more words?" If so, the verb for "give" is tìng, so it should be:
trrit a nga tayìng (or, hopefully, tìyìng) ayoer(u) aylì'ut nì'ul.
But if I've misinterpreted you, please let me know.
Hope that helps!

In a follow-up message:

At the end of my previous e-mail:
Let's make that--
trrit a NGAL tayìng (or, hopefully, tìyìng) ayoer(u) aylì'ut nì'ul.
Kxeyeyri tsap'alute sengi oe. [kxeyey = mistake, error]

Auxilary verb SI

Quoted by omängum fra'uti, January 28, 2010

Can "si" be a verb on its own, or is it always paired with something else?
Your intuitions are right. Si is only an auxiliary verb--it needs to be paired with something else. If you ant the main verb 'make,' for example, you need a different lexical item, like txula (stress on 1st) 'build, construct' or ngop 'create.'
About the only time si stands alone is when the non-verbal element has already been mentioned and is understood from the discourse:
--Nga tsap'alute soli srak?
(Or: Srane, soli.)

Dative + copula possessive

Quoted by omängum fra'uti, January 28, 2010

We have seen examples of the dative and copula used for possession of abstract concepts and states, can it be used for physical possession too?
This can definitely indicate physical possession. In fact, that's the only way to express the concept of "having" in Na'vi--there's no verb 'have.'
But contrary to the general principle of flexible word order, "have" constructions usually begin with the verb. So "I have an ikran" is: Lu oeru ikran. Here, lu has the force of 'there is.'
BTW, the inspiration for this was Hebrew, which does something similar: Yesh li sefer (there-is to-me book), i.e. I have a book. The difference is that Hebrew has no copula in the present tense, so it needs a special word for there-is/are. Na'vi just uses the ordinary copula for this construction.

Krr: noun vs adverb

Quoted by omängum fra'uti, January 28, 2010

The uses of "krr" we have seen seem to use it as a temporal adjunct more than a noun, is this a correct analysis?
Right again. Krr is a noun, but it can also be used adverbially. So tsakrr isn't just 'that time'; it's also (and more commonly) 'AT that time' or 'then.' Likewise kawkrr 'never' and frakrr 'always' are commonly used as time adverbials.
If you add a modifying clause to adverbial krr, linked of course with a, you get a "when" clause:
Ngeyä Tireaioang zola'u a krr, law layu ngaru. 'When your Spirit Animal comes, you will know.'
(Lit.: At the time that your Spirit Animal has come, it will be clear to you.)


Quoted by omängum fra'uti, January 28, 2010

How do you express causality?
Longish story, which I'll tell another time. (There's a causative infix . . . )

Vocabulary and <iyev>

Quoted by suomichris, January 30, 2010 (source).

frrfen (stress on 1st) = visit (transitive verb)
Cf. frrtu = guest
teri (stress on 2nd) = about, concerning (ADP- . . . that is, an adposition that doesn't trigger lenition)
letsunslu = possible
The -iyev- infix, which I think you asked about before, is a "fused" form combining tense and subjunctive mood. This one is a future subjunctive.


Quoted by Nayumeie, January 30, 2010 (source)

Kaltxì ma oeyä 'eylan Nayumeie,
Ngeyä pxesìpawmìri ngaru seiyi oe irayo.
'Awa tìpawmìri 'iveyng oe set; aylari zusawkrr 'ayeyng.
[zusawkrr = future; in the future]
I have a nice complete chart, but it wouldn't be cool if I simply attached it, as I'm sure you understand. Let me convey the essence, however, in a different form.
The system, as you know, is octal:
'aw, mune, pxey, tsìng, mrr, pukap, kinä, vol
volaw, vomun, vopey, vosìng, vomrr, vofu, vohin, mevol
mevolaw, mevomun, . . ., pxevol
pxevolaw, pxevomun, . . ., tsìvol
zam ( = 64, or 100 octal)
vozam ( = 512, or 1000 octal)
zazam ( = 4096, or 10000 octal)
First line above: In disyllables, stress is on the 1st.
Second line above: In disyllables, stress is on the 2nd, except for mevol.
Third line: Stress on the final syllables, except for pxevol.
That should be enough for you to figure out the rest.
Kìyevame ulte Eywa ngahu,

Fused -iv- Infixes

Quoted by suomichris, January 30, 2010 (source)

Ngeyä txantsana tìpawmìri ngaru seiyi oe irayo.
[txantsan = excellent]
A number of people have asked about this, so let me explain:
There's a small class of "fused" infixes that combine the common -iv- subjunctive/dependent verb infix with those that indicate tense and aspect. It goes like this:
I think you can see what's happening here: the CONSONANT of the tense or aspect infix is shoved into the middle of the -iv- infix--an infix in an infix, if you will.
With the future subjunctive, there's a slight complication:
The problem is that although a syllable can end with r, l, or m, it can't end with y (unless it's part of a diphthong, which iy is not). That would violate the phonotactic constraints of Na'vi. So an epenthetic vowel comes to the rescue: -iyev-
An alternate form of this fused infix has arisen: -ìyev-. As an instance of vowel harmony (rare in Na'vi), the high front tense vowel i has become lax (ì) under the influence of the lax vowel e in the following syllable. Both -iyve- and -ìyev- are acceptable.
So that's a very long-winded explanation of why Kìyevame means: "May (we) See (each other again) in the future."
Sìlpey oe, oeyä tì'eyng law livu ngar!

Follow-up email correcting a typo, January 31, 2010 (source)

Ma oeyä eylan,
Tsap'alute, mì upxare a fpole' oel ayngaru trram lu kxeyey:
Both -iyve- and -ìyev- are acceptable
Both -iyev- and -ìyev- are acceptable
Irayo 'eylanur awngeyä Prrton a kxeyeyti rolun.

Subsequent email regarding "rolun", summarized by suomichris (source): "Just heard back from Frommer. run means 'find, discover.'"

Midsummer Night's Dream Vocabulary

Quoted by Kiliyä, January 31, 2010 (source)
Note: stress marking is by Kiliyä based on unquoted notes from Frommer. Kiliyä: "I have assumed that syllables in words start with a consonant, and that -wrr is the syllable's boundry in mungwrr (from munge + wrrpa, "brought outside"?) rather than mu.ngwrr."

Let me answer some of your vocabulary questions:
sweet = kalin
tongue = ftxì
air = ya
bud = prrnesyul
prrnen = infant, baby
syulang = flower
So "bud" is an obvious compound. (Components are often truncated in Na'vi compounds.)
sick = spxin
disease = säspxin
the state of being ill = tìspxin
yet, still = mi
except = mungwrr ADP- (that is, an adposition that doesn't trigger lenition)
game = uvan
to play = uvan si
metal (in general) = fngap
No words yet for horned, moon, favor, nor, judge/judgment. I'll be thinking about those. (I'm not sure if Pandora even has moons in its sky. Can you have a moon going around a moon?)
For "nor," you may be able to get around it by using "or" (fu) with a negative: to do neither A nor B is not to do A or B.
Hope that helps!
Feel free to share any of this...

Some Conjunctions and Adverbs

Quoted by Prrton, February 1, 2010 (source)

As for "because":
Talun [ta smukan Skyinou] is almost exactly one of the words I came up with! I had come up with taluna, from ta + lun (reason) + a, in keeping with the idea that the clause following "because" is modifying the head noun lun, so that "a" is required. But I can see that over time the final a could drop . . . so I'm happy to parenthesize it in the glossary: talun(a).
There's a related word oeyk (2 syllables, stress on the 2nd) meaning 'cause.' The corresponding conjunction is taweyka (with the obvious small sound changes). And in keeping with the above argument, I'll go with taweyk(a).
So it's talun(a) or taweyk(a), which are pretty much synonyms. Take your pick!
By the way, a related word is tafral, meaning 'therefore, because of that.' (The derivation is probably pretty clear.) It's an adverb, unlike talun(a) and taweyk(a), which are conjunctions.
Perfect analysis of sìlpey', comparing it to ralpeng. Yup, that's it! [(tìlaw ta Prrton)meaning that they inflect on the verbal stems pey and peng and not on sìl- or ral-.]
And speaking of Neytiri's and Jake's eveng . . . and holpxay ayzekwäyä feyä . . . that's one for JC to figure out. <g>
hol = few.

Reflexives and Naming

Quoted by Tiriuä, February 1, 2010 (source)

Kaltxì ma oeyä 'eylan,
Thanks for the questions, Jake. (Good name for someone who likes Avatar!)
I guess I haven't mentioned anything about the reflexive yet. Here's a quick explanation:
The reflexive is formed with an infix -äp- that's in "pre-first" position--that is, it comes before the familiar infixes in first position.
I see myself. = Oe tsäpe'a.
Note: There's no pronoun "myself"--all you need is the infix. And it's oe, not oel.
You've seen yourself. = Nga tsäpole'a.
Be careful to keep the stress in the original place. Since for tse'a the stress is on the final a, that's where it stays in all the inflected forms.
They'll see themselves. = (Ay)fo tsäpaye'a.
wash = yur
I wash myself. = Oe yäpur.
I'm washing myself. = Oe yäperur.
Tslolam srak? :-)
As for "My name is . . ."
You COULD say
Tstxo oeyä lu ___.
tstxo = name
But that's not the idiomatic way to say it. Instead, you say "They call me . . ." (which is of course common in many earth languages as well).
The grammar here is tricky:
Oeru syaw fko ___.
Things to note:
syaw 'call' is intransitive: you call TO someone. That's why it's fko, not fkol, and oeru, not oeti. (Fko, as I think you know, means 'one' or 'they' in the general sense.) The actual name as almost like an adverb in this construction: To say "My name is Jake," you actually say something like, "They call to me in a Jake way." :-)
The question, by the way, is:
Fyape fko syaw ngar?
Literally: How does one call you?
So a typical little conversation is:
--Fyape fko syaw ngar?
--Oeru syaw fko Neytiri. Ngaru tut?
--Oeru syaw Tseyk.
In the last sentence, "fko" is understood. You could actually leave it out in the 2nd sentences as well. Ngar and ngaru are used interchangeably--pick the one you think sounds better in its position. And "tut" is a "particle of continuation"--here you'd translate "ngaru tut" as "And how about you?"
Hope that helps! Feel free to share . . .

Art-related Vocabulary

Quoted by Skxawng (source)

Law lu oeru fwa nga mì reltseo nolume nìtxan!
[fwa = fì'u a
rel = image, picture
reltseo = visual art
law = clear]

and also:

Ma Pìraysì,
Rel oeyä na uniltìranyu lor lu nìngay! Reltseotu atxantsan lu nga!
[reltseotu = artist
txantsan = excellent]

A followup regarding tu,

About -tu vs. -yu: You're right--"yu" is an ending for a verb, converting it into the doer or agent, similar to English -er. "Tu," on the other hand, is short for "tute," person (as in Tawtute), and is generally used with parts of speech other than verbs.
So: for "reltseo" (as you know, rel 'image, picture' + tseo 'art'), we get "reltseotu," 'artist (lit. picture-art-person)'.
Likewise, from pamtseo 'music' we get pamtseotu, 'musician'.

Atan Is Illumination

Quoted by Prrton (source)

Atan is a noun meaning “light” in the sense of “source of illumination.” It’s Jim Cameron’s word, actually. I don’t yet have a word for “light = not heavy”—“light” and “heavy” would be good ones to come up with.


Quoted by roger (source)

Keruseya tute vs. ke rey a tute (don't attach the ke or a, since rey is a finite verb) is parallel to "dead (= non-living) person" vs. "person who doesn't live OR isn't alive" in English. Semantically they seem pretty much the same [...].
Note the Avatar line (which I don't believe made it into the final version of the film):
Hetuwongìl awngeyä swotut ska'a, fte kllkivulat keruseya tskxet.
'The aliens destroy our sacred place to dig up dead rock.'
Palulukan atusaron lu lehrrap. 'A hunting thanator is dangerous.'
Palulukan a teraron lu lehrrap. 'A thanator that's hunting is dangerous.'

February 16 2010 Combo

Quoted by Wm Annis (source)

Indirect Question "Where"

1. Is this correct: po kä a tseng(ne) ke tsìme'a oel I didn't see where s/he was going.

No, it's not. But it raises in interesting question.
The error is in the case of "tseng(e)." The structure has to be, "I didn't see the place that s/he was going (to)," so we're essentially dealing with a relative clause. "Place" is in the matrix sentence, so it should be marked as an object. Changing the word order for clarity (which doesn't really affect anything) . . .
Oel tsìme'a tsenget/tsengit [depending on whether you like tseng or tsenge for 'place'] a . . .
But what's the RC? The place THAT S/HE WAS GOING or the place THAT S/HE WAS GOING TO?
If it's the former, we simply have: a po karmä (I like indicating the tense and aspect here, although it's not obligatory.)
If it's the latter, then since we can't strand an adposition in Na'vi the way we can in English, the structure is similar to "the place that s/he was going to it."
For inanimate "it" you shouldn't use po but rather tsa: a tsane po karmä.
So the two possible versions of the sentence you want are:
(1) Po karmä a tsengit ke tsìme'a oel.
(2) Po tsane karmä a tsengit ke tsìme'a oel.
(2) is unimpeachable, I think. But I'd accept (1) as well. It's a bit loose but it seems like a natural development, and it has the virtue of brevity.

Multiple Vocatives

4. Are multiple vocatives clustered (ma smukan sì smuke) or not (ma smukan sì ma smuke)?

Clustered. "Ma smukan sì smuke."

Vocative with Modifiers

5. Does the vocative always come before the noun and all modifiers including adjectives and genitives, or is it just before the head noun? Is it always a particle or can it ever be used as a suffix/enclitic?

It's before the noun and all modifiers, not necessarily immediately before the head:
ma oeyä eylan: 'O my friends'
It's always before the NP, never a suffix/enclitic.

Toruk Makto

7. In toruk makto it seems like it should be maktoyu. Aside from "James Cameron Said So" is there a reason it is not?

So you guys noticed that, huh? :-)
JC and I had a bit of a discussion about "Toruk makto." I pointed out that according to the grammar, it should be maktoyu. And I was told in no uncertain terms that it was going to be "toruk makto," and I should figure out a way to make it work. The man certainly has the right . . .
So it's an exception--one of the iconic phrases in the language that developed unusually, for whatever reasons, and don't follow the normal rules. If we looked hard enough we could probably find parallels in English and in all other languages.

Enclitic Adpositions

1. Regarding adpositions. When they follow, they are written attached to the word. Can we assume this means they are enclitic, and have no stress accent of their own? Is the accent obliterated in two syllable adpostions, as in eyktanmungwrr?

Yes, they're enclitic, without their own stress. EYKtanmungwrr is stressed on the first syllable--a bit awkward, perhaps, but not too bad.

Attributive a

9. How are adjectives that begin or end with "a" such as apxa dealt with when the attributive is used? Do other vowels need any special treatment with the attributive?

In such cases the attributive "a" is swallowed up and disappears:
skxawng apxa, apxa skxawng
I can't think of any other cases where this happens, although aä and äa seem like rather unstable sequences. Until further notice, though, I think we'll allow them.

More extracts from various emails

Added by roger. Dates are date of answer by Frommer
Feb 13

Keruseya tute vs. ke rey a tute (don't attach the ke or a, since rey is a finite verb) is parallel to "dead (= non-living) person" vs. "person who doesn't live OR isn't alive" in English. Semantically they seem pretty much the same; I'll have to think if there are any systematic differences between these forms.

Note the Avatar line (which I don't believe made it into the final version of the film):

Hetuwongìl awngeyä swotut ska'a, fte kllkivulat keruseya tskxet.
'The aliens destroy our sacred place to dig up dead rock.'

To my knowledge, there's a parallel situation in French:

"Laughing Cow" (brand name of a kind of cheese) could be either:

1. la vache riante


2. la vache qui rit

In 1, riante is a (feminine) participle. 2 is literally "the cow that laughs," but that's apparently the more idiomatic way to say it, since it's the actual name of the cheese. (I think all of that is correct, but my French is shaky, so it should be verified by a French speaker.)

I'm not sure which is the more idiomatic expression in Na'vi, but if the meaning is clearly "dead," I'd go with kerusey.

The point of participles in Na'vi is that they're ONLY used attributively. If you want a gerund, use a tì- nominalization:

1. Tìtaron lu lehrrap. 'Hunting is dangerous.'


2. Palulukan atusaron lu lehrrap. 'A hunting thanator is dangerous.'


3. Palulukan a teraron lu lehrrap. 'A thanator that's hunting is dangerous.'
Feb 17
I asked William's question about caus. v.t.'s taking the dative, and F said,

Yes and yes. Good guesses.

Eytukanìl Neytirir yerikit teykolaron.
‘Eytukan made Neytiri hunt a hexapede.’

Cf. French:

(1) Pierre a fait construire un bateau à Paul.
Peter made Paul build a boat.
(2) Pierre a fait construire un bateau par Paul.
Peter had a boat built by Paul.

For the sentence corresponding to (2)—i.e. ‘Eytukan had a hexapede hunted by Neytiri’ (which sounds a bit odd)—Neytiri should not be in the dative but rather be the object of an adposition corresponding to Fr. “par.” A candidate would be fa [ADP-] = with, in the sense of “by means of.”


Eytukanìl fa Neytiri yerikit teykolaron.
‘Eytukand had a hexapede hunted by Neytiri.’
I then asked about the ambiguous valency of a causative transitive Vb, and whether there was a passive in Na'vi

Correct—I haven’t included a passive because I thought that any function it would serve was already satisfied by other mechanisms in the language.

For the equivalent of an agentless passive, you’re right—just use fko.

For passives with agents, the flexible word order of Na’vi should obviate the need for a passive, or so it would seem. There’s no semantic distinction between

(1) Shakespeare wrote Hamlet.
(2) Hamlet was written by Shakespeare.

but the two are not equivalent in discourse, where focus is a consideration. In Na’vi, however, the equivalent of (3) is fully grammatical and not at all strange:

(3) Hamlet Shakespeare wrote.

which should fulfill the functions of (2).

But maybe not entirely. Conjoining might cause a problem. “Hamlet was written by Shakespeare and is the most famous play in the world.” If that comes from “H was written by S” and “H is X,” could you drop the second occurrence of H in the Na’vi sentence? Probably not, since the two occurrences of H (assuming the first occurrence is in the form of (3) above) are in different cases. Maybe Na’vi just doesn’t allow that kind of conjoining. Doesn’t seem unreasonable. But I’ll think more about it.

I don’t think there’s a problem with valency not being indicated on the verb, but let me know if you find examples where not having such a marker would cause confusion.

I don't think Na'vi needs a passive, since it doesn't appear to have subjects, and passives are a way of keeping subjects aligned in discourse.
I asked about William's phrase (Ngeyä teri faytele a aysänumeri ngar irayo seiyi ayoe nìwotx), which F described as elegant, and attributive nouns, forgetting that it was an attributive PP:

Well, there’s no attributive noun here! It’s rather an attributive prepositional (or I guess I should say adpositional) phrase:

[ Ngeyä [ [teri faytele] a aysänumeri ] ]

Is that kosher? I think so. An attributive prepositional phrase results from the same process that yields attributive adjectives from predicate ones, parallel to an analysis that derives “good question” from “a question that is good.” Delete “that is” (when I studied lx, we used to call that WH-is, pronounced “whiz,” deletion) and then, for English, pre-pose the adjective.

In Na’vi:

*[tìpawm lu sìltsan] a tìpawm > lu sìltsan a tìpawm > sìltsana tìpawm

The middle form is perfectly grammatical, equivalent to “a question that is good.”

If you can do that, then it seems a small step to do this:

*[tìpawm teri faytele lu] a tìpawm > teri faytele lu a tìpawm > teri faytele a tìpawm
‘a question about these matters’

But your question about attributive nouns, or NPs in general, is relevant. What about:

*[tutean oeyä ‘eylan lu] a tutean > oeyä eylan lu a tutean >? oeyä eylan a tutean
‘the man who is my friend’

The final form seems questionable to me. There should probably be a hierarchy of acceptability here, with adjectives at one end of the scale, PPs in the middle, and NPs at the other end.

So, as we've seen, the genitive is what's used for attrib. nouns, not "a"; also, it's now confirmed that the "a" of RC's and the "a" of adj's are the same.
Feb 18

Sìpawm ngeyä ke lu ftue [...]! Slä lu sìltsan.

OK, let's see . . .

1. mesyalä ikran

Strikes me as a little odd, but I guess it's OK. "An ikran of two wings." Parallel to "a man of few words." Genitives are used for many purposes.

2. mesyalhu a ikran

I like that the best. Seems quite natural.

3. *mesyal a ikran

Right. That should be ruled out.

4. ?mesyal lu a ikran

This is the most interesting one. I don't think it should be considered acceptable, but it touches on wider issues.

Assuming the RC starts out this way:

*[ikranur mesyal lu] a ikran

[Digression: I've been preferring lu-initial order for "have" constructions: Lu ikranur mesyal. But that's not a hard-and-fast rule, so I'll keep the above order, which better reflects your example.}

The question, of course, is what do you do with the coref. NP in the RC--delete or pronominalize?

As I'm sure you know, languages have different hierarchies with respect to what happens to coref. NPs. Subjects are the most deletable, followed by DOs, followed by obliques, followed by comparatives . . . (It's been a long time since I looked carefully at this stuff, but I think that's correct.) So, for example, in English you can say "He's the man I spoke to" but in many other languages you have to say "He's the man that I spoke to him."

In the Na'vi case, coref. NP deletion only extends as far as DOs. So for datives, you need to pronominalize:

*[ikranur mesyal lu] a ikran > poru mesyal lu a ikran

Other word orders work too, of course: lu poru mesyal a ikran (I like that one the best), mesyal lu poru a ikran, etc.

Trials & transitivity

Wasn't this a Jane Austen novel?

Doesn't look like this was ever posted. Don't know why; it was held up for a while as part of the Ma Sempul project, but then F specifically said it was okay to release. —Roger

The regular pronouns are as follows, listed in the order Singular, Dual, Trial, Plural:
Regular pronouns sing. dual trial plural
1st person, exclusive oe moe pxoe ayoe
1st person, inclusive oeng pxoeng ayoeng/awnga
2nd person nga menga pxenga aynga
3rd person po mefo pxefo (ay)fo
Note the alternate forms of the 1st Pers. Incl. Pl. form. The advantage of the aw- forms (unrelated, by the way, to 'aw, one) is that most of the declined forms are 2 syllables rather than 3, or 3 rather than 4. Example: ayoengal (3 syll.) vs. awngal (2 syll.).
Formal forms: Only for 1st pers. excl. and 2nd pers.
Formal pronouns sing. dual trial plural
1st person, exclusive ohe mohe pxohe ayohe
2nd person ngenga mengenga pxengenga ayngenga
For inclusive forms, you use separate pronouns, e.g., ohe ngengasì (where is "and" cliticized to the second element).
Verb forms: The honorific/ceremonial -uy- is a second-position infix (e.g., occurs in the second syllable of disyllabic roots) in the same position as the other attitudinals like -ei-. (It can't co-occur with them; you can only choose one attitudinal, if any.)
So for taron, which I've been using as a model verb, you have forms like:
tolaruyon, tayaruyon, etc.
Keep in mind, though, that once you've used this form in discourse, you don't have to keep using it. It's "understood" from then on.

Then this came out end Jan.

As for "because":
Talun [ta smukan Skyinou] is almost exactly one of the words I came up with! I had come up with taluna, from ta + lun (reason) + a, in keeping with the idea that the clause following "because" is modifying the head noun lun, so that "a" is required. But I can see that over time the final a could drop . . . so I'm happy to parenthesize it in the glossary: talun(a).
There's a related word oeyk (2 syllables, stress on the 2nd) meaning 'cause.' The corresponding conjunction is taweyka (with the obvious small sound changes). And in keeping with the above argument, I'll go with taweyk(a).
So it's talun(a) or taweyk(a), which are pretty much synonyms. Take your pick!
By the way, a related word is tafral, meaning 'therefore, because of that.' (The derivation is probably pretty clear.) It's an adverb, unlike talun(a) and taweyk(a), which are conjunctions.
Perfect analysis of sìlpey, comparing it to ralpeng. Yup, that's it! [(tìlaw ta Prrton) meaning that they inflect on the verbal stems pey and peng and not on sìl- or ral-.]
And speaking of Neytiri's and Jake's eveng . . . and holpxay ayzekwäyä feyä . . . that's one for JC to figure out. <g>
hol = few.

Note we now have oeyk "a cause", eyk "to lead", and <eyk> CAUS.

Feb 15
A couple of things to note:
1. Mìso and neto are both adverbs meaning "away." The former is positional: Stay away! The latter is directional: Go away!
2. Imperatives, either positive or negative, have two forms: either the bare root or the -iv- form. At an earlier point in the history of the language there was probably a polite/familiar distinction (the -iv- form being the politer one), but that's no longer the case. They're used interchangeably. So to say "Go!" you can say either kivä or just kä.
This is an important grammatical point. (Feel free to share.) 'Change' is indeed "latem," but it's the intransitive version, as in "Everything is changing." To change SOMETHING, you need the transitive version, which is achieved through a causative: to change something is to cause something to change.
The causative infix is -eyk-, which is in "pre-first" position--that is, it comes before the familiar first-position infixes like -iv-, -ol-, etc. So for the imperative you can say either leykatem or leykivatem.
"Begin" works similarly: sngä'i is intransitive: Something is beginning. To begin or start something, you need the causative form: sngeykä'i. Oel sngeykolä'i tìkangkemit. 'I began the work.'


Quoted by Roger, February 19 2010 (source)

Ma oeyä eylan,

Fpìrmìl oel futa aynga natsew tsive’a fi’ut.

[ -ìrm- : past proximate imperfective – “was just thinking . . . “

-ats- : 2nd position infix indicating uncertainty or indirect knowledge: “you might want.” Could be used with “kxawm,” redundantly, for reinforcement.]

I was corresponding with someone from Wales, and the question of lenition in Celtic came up. I mentioned the connection to Na’vi and thought you might to see it too. No new content here, just history. :-)

“In 1998 I visited Ireland and of course wanted to try to learn a little Irish for the occasion. Wow! I consider myself a pretty sophisticated language learner, but Irish knocked me for a loop. (And they say English spelling is hard!) I found the mutations—if I remember correctly, eclipsis and lenition—very interesting. And lenition wound up influencing Na’vi. I had also studied Hebrew, where there’s a process of “spirantization” that has something of the same effect although in different environments, so that was an influence as well.

“The final influence was Malay/Indonesian, where certain prefixes mutate the initial consonant of the root. It always tickled me that in a Malay dictionary, you need to look up the word “menarik” under T, the word “memandang” under P, and the word “menyatukan” under S! And that situation, of course, has a parallel in Na’vi, with the “short plurals.” (You look up “hilvan” under K, “sute” under T, “fizayu” under P, etc.)



Reported by Prrton, Feb 19 2010 (source).

.. go with "tam."

As you know, the verb means "suffice, 'do'" as in, "That'll do."

In the film, I think Grace says "Tsun tivam" in response to Norm's (very good, if formal) Na'vi. Or at least that was once going to be in the film. (As I mentioned, I have a hard time keeping track of what wound up in the final version and what got cut.)

So saying "tam" for "OK" seems appropriate. It's easy to say...


Reported by roger, Feb 19 2010 (source).

Sìltsan nìtxan. Tìng nari nekll.

[That is, "See below", where he interlaced his answers w my questions, which I'll put in italics. Comments I'm adding now are in brackets.]

[embedded email]

Did I get this right?
Hém geyä zégke fkóru livú cá(y)hem a gáru pŕrte' ke lú.
[this was the Golden Rule]

Very fine. I’ve spelled it zenke, even though it’s pronounced with a velar nasal. And I realize I’ve been inconsistent in doing that, since I’ve spelled lumpe with an m rather than the underlying n. Was probably influenced by English (“impossible,” but “income”). But maybe a little inconsistency isn’t the worst thing in the world.

BTW, zenke and ke zene are semantically distinct:

Nga zenke kivä! ‘You must not go!’

Nga ke zene kivä. ‘It’s not obligatory (OR: it’s not a must) that you go.’

And how many people have adopted the “scientific” transcription? It’s nice to see it, although I don’t think it will ever become as popular as the “official” one. (Notice what happens when you write Tsu’tey in scientific. NOT a good association with that character. <g>)

and is zegke short for zene ke ?
This one seems clear:
lǐ'fyari leNá'vi Ŕrtamì vay sét 'almóg a frá'u zerá'u ta gŕrpogu
assuming Rrta for "Earth"; confirms grr "root" and past pfv. I assume future pfv would be 'alyog.

Right. As you now know, it’s ‘Rrta.

  • maw (ADP-) = after (time)
  • hìkrr (stress on 1st) = second, very short time
  • tätxaw (stress in 2nd) = return (intrans.)
  • Maw hìkrr ayoe tìyätxaw.

Sar, Fìtxan

Quoted by Prrton, Feb 18, 2010 (source).

*Fayl*ìri lEksel irayo seiyi oe ngar, ma oeyä 'eylan. Tsat sayar oel nìltsan.

Tewti! Ngeyä lì'fya leNa'vi txantsan lu nìngay. Fwa sute pxel nga tsun oeyä hì'ia tìngopit sivar fte pivlltxe nìlor fìtxan oeru teya si.

sar = 'to use'

fìtxan = 'so, to such an extent'

oeru teya si: 'fills me' (an idiom, English version courtesy James Cameron; understood: "with satisfaction, with joy")

Ngeyä kxetse lu oeru

From the newspaper contest, quoted by Prrton, Feb 22, 2010 (source).

Congratulations to Eri!

'Oleyng nga nìltsan, ma Eri! Ngari Nawma Sa'nok lrrtok soleiyi.

Your entry in the "You're Going Down" contest, "Oeru lu ngeyä kxetse--Your tail is mine" was the clear winner!

I've already submitted it to the Los Angeles Times editor. The only revision I made was to change the word order to place more emphasis on the "mine":

Ngeyä kxetse lu oeru.

(The end of the sentence is where the "punch" comes.)

And thanks to everyone who entered and voted.

Irayo ayngaru nìwotx!



Adjective Comparatives

Quoted by Prrton, Feb 23, 2010 (source).

..Is this correct? :
Oeru ke tsun livam ke'u lor to Eywa'evengä na'ring a lew säpoli fa prrwll,
kxawm mungwrr fìkifkey a lew säpìyevi fa fpom sì lì'fya leNa'vi.

Very fine.

Only one suggestion...

The verb säpìyevi: Why subjunctive? Why not simply säpìyi?

Argument in favor of the subjunctive: It's a world that MIGHT soon cover itself . . . etc.

Arguments in favor of the indicative:

(1) Your translation seems to indicate you're thinking of a world that WILL soon cover itself . . . etc. (2) I like the clear parallelism between säpoli and säpìyi. (3) The indicative is one syllable shorter.

So I'm somewhat more in favor of the indicative, säpìyi. But I think you could justify either version. Your call...


Oeri/oeru is often a toss-up. With "lam," though, I think the dative is marginally better, since it's natural for "seem" to take the dative:

"It seems to me . . . " But you could make a case for oeri as well.

The syntax of comparatives is perhaps a bit unusual.

To say "A is adj-er than B," you simply say "A to B lu adj."

Example: Po to oe lu sìltsan. (not "sìltsan nì'ul")

No "comparative degree" of the adjective is used.

Cf. Mandarin:

Ta bi wo hao.

Tag Question

Quoted by Prrton, March 1, 2010 (source).

Tewti, ma Prrton! Txantsana tìkangkem, txantsana aysäfpìl.

[As you can guess, säfpìl = idea, thought -- sä.FPÌL]

Lu awngar aytele apxay a teri sa'u pivlltxe...

« Teri » does not cause lenition. "Sa'u" is a short plural (short for aysa'u, of course): Literally: We have many matters that (we) may speak about THEM (or: THOSE THINGS)--i.e., we have a lot to talk about.

For the equivalent of "isn't that true?" "¿verdad?" "n'est-ce pas?" etc. let's go with "kefya [ke.FYA] srak?" or, as an equivalent shorter form, "kefyak?" (Derived, as we discussed, from "ke fìfya srak?")