Details for: Aguaruna
Alternative names:
Aguajun, Ahuajun
Dialect of:
Genetic affiliation:
Western upper Marañon River area, Potro, Mayo, and Cahuapanas rivers
Number of speakers:
38,300 (2000) almost 0 monolinguals
Similar to Huambisa [hub], Achuar-Shiwiar [acu].
Syllable types:
(C)V(N) [N represents unspecified nas segment that surfaces as a nasal obstruent in a homorganic cluster or vowel nasality]. The nucleus may consist of any long vowel or diphthong, or a triphthong which must be /V1aV2/ [where V2 is not /a/] VERY RARE
Restrictions on single onsets:
A consonant forms the onset of a syllable if it is followed by a vowel. No branching is
permitted in the onset
Restrictions on onset clusters:
Restrictions on single codas:
All Cs except glides appear in codas
Restrictions on coda clusters:
Voiced-voiceless stop contrast:
Nasal contrast on consonants:
Nasal-voiced stop contrast:
Nasal-voiced stop contrast restrictions:
/m,n/ alternate with [b,d] (with some prenasalization) when followed by oral vowel and consonants within the nas domain.
Nasal-voiceless stop contrast:
Nasal-voiceless stop contrast restrictions:
Nasal-oral vowel contrast:
Nasal-oral vowel contrast restrictions:
Nasal assimilation:
Trigger segments class:
nasal (unspecified)
Trigger segments list:
Unspecified N
Conditions on trigger segments:
Note: when a vowel is elided from a position immediately following a homorganic NC cluster, the C is also elided and only the nasal remains, and it retains the place features of the elided stop. /ukumpɨ/ [ukum] ‘blackfly’; /ukumpɨ-ka-it/ blackfly-INTERROG-COP ‘is it a blackfly?’ The third vowel /ɨ/ gets elided because it is in a light syllable, and the /p/ is then unsyllabifiable and also gets deleted. So we get the surface form [ukumkait].
Target segments class:
Target segments list:
Conditions on target segments:
Nasal harmony:
Domain of spreading:
Trigger segments class:
EITHER there is /Ṽ/ OR underspecified nas /N/
Trigger segments list:
Conditions on trigger segments:
For more info on ambiguity, see REMARKS tab
Target segments class:
vowels, h
Target segments list:
ɨ, i ,u, a, h
Conditions on target segments:
Note that nasals that surface, with specified place features, never trigger nasality on adjacent Vs.
Transparent segments class:
Transparent segments list:
Blockers class:
Blockers list:
Prenasalization of blockers:
Postnasalization of blockers:
Nasal morphemes:
Commutation level:
Spreading source:
Floating nasal:
Stylistic conditions:
Adjacency conditions:
Contour segments:
Source of contour segments:
spreading of nasality
Examples of contour segments:
md, nd
Adelaar (LOTA:434) observes that radical final nasals sometimes correspond to pre-nasalized stops.
- Domain of spreading is phonological word
- Underlyingly nasal vowels alternate with VN sequences when followed by a stop or affricate.
- /m,n/ alternate with [b,d] (with some prenasalization) when followed by oral vowel and consonants within the nas domain
Consonant examples:
Oral vowel examples:
Nasal vowel examples:
[ɨtsã] 'sun' vs. [ɨtsantu] 'shine' (-tu is a vblz sfx) (note Ṽ and VN alternation)
Stress rule(s):
Stress source(s):
There is also a weird denasalising phenomenon, where a couple of homophonous suffixes of the shape /hu/ trigger obligatory denasalisation of a preceding nasal sequence. For example /nuwa/ 'woman' plus the 1sg possessor suffix /hu/ surfaces after final vowel deletion as [duwaŋ] (to further complicate matters, /h/ has the syllable final allophone [ŋ]).
[Overall, Personal communication 2010]

1. On the status of underlyingly nasalized vowels
2. On the status of glides and their status as allophones of vowels

== On the status of /Ṽ/s ==

Some forms show an alternation between Ṽ and VN:

(1) [ɨtsã] ‘sun’
(2) [ɨtsantu] ‘shine’ – where -tu is a verbalising suffix
So it seems to me one could analyse either as underlying nasal vowels with the following rule:
(3) Ṽ → VN / __ C
where C is a stop or affricate and N assimilates to the place features of C.
OR we could assume an unspecified nasal /N/ and say that:
(4) N → [αplace] / __ [C, αplace]
again C is a stop or affricate; and subsequently any N that remain underspecified are erased – but the [+nasal] feature is transferred to the preceding vowel:
(5) VN → Ṽ

--> PRO /Ṽ/: Many nasal vowels don’t alternate with VN, and in some cases it is impossible to tell which part of the domain is initiating the nasality. E.g.:
(6) [h̃ ̃ɰ̃ ã] < /hɨNɰa/ or /hɨɰaN/? ‘house’
Of course, this would also be a problem with /Ṽ/s but we would then simply say that nasality is a property of the nasal domain rather than a particular segment.

Also, there are some nasal vowels/domains that have arisen in Aguaruna contiguous
to the phoneme /h/, in particular where it comes historically from /r/. An synchronic
analysis that treats the /h/ as the source of nasality is untenable, and it seems strange
to posit that /h/ has triggered the insertion of /N/; better to say that /h/ has triggered
the change V → Ṽ.

--> PRO /N/: The Ṽ~VN alternation can be seen as assimilation rather than epenthesis of
a nasal. This is particularly appealing given in the light of examples where the C is
subsequently elided but the N retains its place features (see example 9 below).

= On the status of glides =
I analysed them as allophones in my dissertation, as there are clear examples of vowel-glide alternation in different morphological environments. But there’s always an alternative explanation isn’t there?

They could be independent phonemes with some positional neutralisation effects. Note that glides don’t behave consistently as consonants for elision effects. There are two processes of elision: apocope deletes the vowel of a word-final CV syllable and syncope deletes the vowel of some word-internal CV syllables. A sequence of glide+vowel does not count as CV for apocope, but it does for syncope.
Data provided by:
Simon Overall (input/research by Coler)
ml.coler at
Data based on acoustic evidence:
Data tested in laboratory:
Data collected:
Primary source(s):
Overall, Simon. 2007 A grammar of Aguaruna. PhD dissertation. La Trobe University. Available at
Secondary source(s):