Following Paul Frommer’s guest post on the ever-popular linguistics blog Language Log, is what started LearnNavi, and the Na’vi-learning community. Language Log’s follow-up post occurred 3 months thereafter: How Language Log helped jump-start a subculture.

Also of note: many enthused fans found Professor Frommer’s email (which I will not re-post here to preserve some level of privacy), and emailed him with questions about the language. After several days of what is assumed to be a constant flood of interest, Paulmanaged to respond to his newly found fans.

From: http://languagelog.ldc.upenn.edu/nll/?p=1977

---

Given the interest that’s already been shown in Na’vi, I’m grateful to Ben Zimmer for the opportunity to post a few highlights of the language to Language Log. As will be apparent, the information below is not intended to be anything like a complete description; the Phonetics and Phonology section is the most complete, but the Morphology and Syntax sections are mere sketches. Given my contractual obligations, a more thorough treatment awaits another venue. But I hope this sketch will answer a few questions and perhaps serve to counterbalance some of the erroneous information that has made its way to the Internet. Needless to say, comments are welcome.

Phonetics and Phonology

Na’vi has 20 consonants, 7 vowels, 4 diphthongs, and 2 syllabic “pseudovowels,” rr and ll.

1. Consonants

The consonants are (in the “official” Na’vi transcription):

Ejectives:	px	tx		kx
Voiceless Stops:	p	t		k	’
Affricate:		ts
Voiceless fricatives:	f	s			h
Voiced fricatives:	v	z
Nasals:	m	n		ng
Liquids:		r, l
Glides:	w		y

Note the following:

• The red consonants can occur as the first element of a syllable-initial consonant cluster.
• The blue consonants can occur in syllable-final position.

Note also:

• px, tx, kx, ts, and ng are digraphs representing the three ejectives, the affricate, and the velar nasal respectively.
• In the “scientific” transcription, ts is replaced by c and ng by g. For commercial purposes, however—and also for ease of reading by the actors—the “official” transcription is preferred.
• The letters b, d, j, and q never appear in Na’vi.

2. Vowels, Diphthongs, and “Pseudovowels”

Vowels

Na’vi has a 7-vowel system:

	i , ì				u
		e			o
			ä		a

Transcription and phonetics:

	i	[i]
	ì	[I]
	e	[ε]  Note: always lax
	ä	[æ]
	u	[u] or [U]
	o	[o]
	a	[a]

Diphthongs

Na’vi has 4 diphthongs: aw [aw], ew [εw], ay [aj], ey [εj].

3. Syllable structure and phonotactic constraints

Every syllable has a single vowel or diphthong at its center. Each vowel or diphthong in a word corresponds to a separate syllable. A single vowel or diphthong may be a syllable by itself.

Within syllables, Na’vi vowels and diphthongs can be preceded by either one or two consonants. They can also be followed by one consonant. That is, the syllable structure is (C)(C)V(C), where V represents a vowel or a diphthong. Restrictions on which consonants can occur in which positions are given below.

Initial consonants. Any consonant can occur at the beginning of a syllable.

Consonant clusters. Clusters of two consonants can occur, but only in syllable-initial position and only in the following combinations:

f,  s,  ts +  {p, t, k, px, tx, kx, m, n, ng, r, l, w, y}

There are thus 39 possible initial C-clusters, all of which are attested in the lexicon.

Final consonants. Only certain consonants occur in syllable-final position. These are:

	Ejectives:	px	tx	kx
	Stops:	p	t	k	’
	Nasals:	m	n	ng
	Liquids:		r, l

Pseudovowels. In CV syllables, the liquids l and r can replace the vowel. When they are syllabic they are lengthened (the r is very strongly trilled, the l always front and “light”) and written ll and rr respectively.

Note: Sequences of stop + liquid, though they cannot occur initially, may be found medially. In such cases, however, a syllable boundary intervenes. Example: ikran ‘banshee’ divides as ik-ran, not *i-kran.

4. Vowel clusters

Na’vi allows unlimited sequences of vowels in a word. If no glottal stop intervenes, the vowels in such clusters glide smoothly from one to another. Each such vowel represents a separate syllable.

Examples: tsaleioae (6 syllables), meoauniaea (8 syllables)

5. Phonetic detail and phonology

Voiceless stops are unaspirated. In final position they are unreleased.

Na’vi r is a flap, as in Spanish pero or Indonesian surat.

Word stress in Na’vi is unpredictable and distinctive. Stress must thus be specified for each word. (In learning materials only, the stressed syllable in a word is underlined.)

E.g. tute ‘person’, tute ‘female person’

Lenition. Following certain adpositions and prefixes, initial consonants mutate as follows:

px, tx, kx

⇒

p, t, k

p, t/ts, k

⇒

f, s, h

Glottal stop:

’

⇒

Ø

8 C’s participate in rule: px, tx, kx, ’, p, t, ts, k

12 C’s do not: f, s, h, v, z, m, n, ng, r, l, w, y

Word Classes and Morphology

1. Nouns

Nouns are inflected for case and number but only rarely for gender.

Number

Number (singular, dual, trial, plural) is indicated by prefixes, each of which triggers lenition:

Short plurals: When the plural marker ay- is prefixed to a word beginning with a lenitable consonant, it may be dropped after lenition has occurred.

Example : The plural of tokx ‘body’ is ay+tokx. Thus we have :

*aytokx → aysokx ‘bodies’

But now the plural is marked redundantly, first by the prefix itself and second by lenition of the initial consonant of the singular. So the ay- may be optionally dropped, yielding tokx ‘body’ vs. sokx ‘bodies’.

Case

Nouns and pronouns take six cases (counting Topical as a case): Subjective, Agentive, Patientive, Genitive, Dative, Topical. The case system is tripartite—i.e., it distinguishes between intransitive subjects (S), transitive subjects (A), and objects (P). Case morphemes are suffixes, generally with several allomorphs. Changes to the noun base sometimes occur with the Genitive.

The Topical form of a noun or pronoun establishes a loose semantic connection to the clause and has a wide range of uses. It may be translated along the lines of “with regard to,” “as for,” “turning to,” “concerning,” etc., but it can also appear where a genitive or dative might be expected.

Example:

Oeri	ta	peyä	fahew	akewong	ontu	teya	längu.
I-TOP	from	his	smell	alien	nose	full	is-NEG-ATTITUDE

‘My nose is full of his alien smell.’

2. Pronouns

Like nouns, pronouns exist in singular, dual, trial, and plural forms. In the first person dual, trial, and plural, a distinction is made between inclusive and exclusive forms.

3. Verbs

Verbs are inflected for tense, aspect, mood/dependency, and speaker attitude, but not for person or number. Verb inflections are effected exclusively through infixes, which are of two types—first position and second position.

With monosyllabic verb roots, first-position infixes simply come before second-position ones. With multisyllabic roots, however, first-position infixes occur in the penultimate syllable and second-position ones in the final syllable.

First-position infixes indicate tense, aspect, or mood; there are also participial and reflexive infixes in this position, the latter being in “pre-first” position so it can co-occur with other first-position infixes. Second-position infixes indicate speaker attitude—positive orientation, negative orientation, or uncertainty/indirect knowledge. Many of these infixes are optional on the sentence level. (In discourse, however, overt indication of tense or aspect may be required.)

Aspect is perfective or imperfective. Tense has five points on the time line: present, past proximate, past general, future proximate, future general. Verbs can be inflected for tense alone, aspect alone, or a combination of tense and aspect.

Selected examples:

Root: taron ‘hunt’

Note: English translations are only approximate and represent one of several possibilities.

	Tense only:
	taron	‘hunt’
	tìmaron	‘just now hunted’
	tayaron	‘will hunt’
	Aspect only:
	teraron	‘be hunting’
	tolaron	‘have hunted’
	Both tense and aspect:
	tìrmaron	‘was just now hunting’

Many more such forms exist.

Including second-position infixes:

tìrmareion ‘was just now hunting (and the speaker feels positive about it)’

tayarängon ‘will hunt (and the speaker feels negative about it)’

In the last two examples, the root is indicated in red. Such forms raise an interesting question: To what extent can a root be obscured by inflections and still be recognizable? When Na’vi listeners hear tìrmareion, for example, do they immediately recognize it as a form of the verb taron? By the same token, are speakers able to produce such forms spontaneously? I’d like to think the answer to both questions is yes, but the matter requires further study; we need more samples of discourse from Pandora!

4. Adjectives

Adjectives are invariant and undeclined. A derivational prefix forms adjectives out of other parts of speech.

5. Adpositions

These can either precede or follow their heads with no semantic distinction; in the latter case, they’re bound to the noun or pronoun. E.g., ‘with you’ = hu nga or ngahu.

Certain adpositions, when in pre-nominal position, trigger lenition. There’s no predicting which do and which don’t—they simply have to be learned. (Adpositions are marked in the lexicon as either ADP+ or ADP-.)

Because of the “short plural” phenomenon, ADP+ adpositions can yield ambiguous structures. Example: mì ‘in’ is ADP+; does mì sokx mean ‘in the body’ or ‘in the bodies’? The language has developed ways of dealing with these potential ambiguities.

Syntax

The most notable aspect of Na’vi syntax is the freedom of word order. The case system allows all 6 sequences of S, O, and V. Additionally, adjectives, genitives, and relative clauses can either precede or follow their heads.

Nouns and adjectives are tied together by the morpheme a, which comes between them and is attached as a bound morpheme to the adjective. For example, ‘long river’ is either ngima kilvan or kilvan angim.

There’s obviously a lot more to say about syntax—for example, how the language handles subordination and complementation. That will be for another time.