Ndak Ta/Phonology

From AkanaWiki
Jump to: navigation, search

Phonemes and Orthography

Phoneme inventory


   labial      alveolar      velar    labiovelar
plain plosive p t k
voiced plosive b · bʷ d ɡ
affricate ʦ
fricative s
nasal m n ŋ ŋʷ
  approximant l w
trill r


  • p t k kw /p t k kʷ/
  • b bw d g /b bʷ d ɡ/
  • s /s/
  • ts /ʦ/
  • m n ng ngw /m n ŋ ŋʷ/
  • r l w /r l w/

When initial, final, or intervocalic, all of the above are always pronounced as shown. The only consonant allophony occurs in clusters (see cluster rules section below).


  front     central     back  
high i · ɪ̃ u
mid e · ɛ̃ o
low a · ɐ̃
  diphthong ai · ɐɪ̃ au · ɐʊ̃


  • Oral: a e i ai au o u /a e i ɑi au o u/
  • Nasal: â ê î âi âu /ɐ̃ ɛ̃ ɪ̃ ɐɪ̃ ɐʊ̃/

Syllable structure

There is a difference between what's phonetically syllabic and what the phonology treats as a syllable. The phonological rules of the Ndak Ta allow the following syllables:

  • initial: (N)(C)V(C)
  • medial: (C)V(C)
  • final: (C)V(C)(N)

That is to say, a word may have nasal+consonant cluster initially or a consonant+nasal cluster finally, or both. From a phonetic standpoint, however, such initial and final clustered nasals are syllabic themselves. Some examples to illustrate...

  • Legal words: [pap] [mam] [m=pap] [papm=] [m=papm=]
  • Illegal words: *[pmap] *[pamp]

All of these legal words would be considered monosyllabic for stress rule and any other syllabification purposes.

Cluster Rules

Only two consonants per cluster, no more. These are treated as coda+onset in all cases except initially and finally, where clusters are disallowed except for nasal+stop and stop+nasal respectively (see syllable structure section above for more on that).

  • NASAL+NASAL: All combinations allowed.
  • STOP+STOP: All combinations allowed. The first stop assimilates to the voiced/voiceless value of the second (/t/ + /b/ becomes [db]).
  • STOP+NASAL: Voiced stops assimilate into the corresponding nasal for their POA; if that matches following nasal, it results in a geminate. So /bama mba amba amab abm/ are all pronounced as shown, while /ab/ + /ma/ comes out as [amːa] (spelled amma) and /ad/ + /ma/ as [anma] (spelled anma). Voiceless stops before nasals all become [ʔ], such that /ampa/ = [ampa], while /ap/ + /ma/ = [aʔma].
  • NASAL+STOP: Stops assimilate to the POA of the previous nasal: /an/ + /ba/ becomes [anda].


  • /ʦ/ patterns as a stop.
  • Liquids may cluster with all stops and nasals, or with /s w/ or each other, in either direction.
  • /s/ assimilates to the voiced/voiceless value of any stop that follows it, but stops that preceed it become voiceless. /s/ remains voiceless when clustered with nasals or /r l w/.
  • /r/ is pronounced [ɾ] when clustered with a labial or a velar, and [r] when clustered with an alveolar (including /l/).
  • /w/ may follow any stop, nasal, liquid, or /s/.
  • /kʷ ŋʷ bʷ/ are phonetically equivalent to [k ŋ b] + [w], but I'm analyzing them as seperate phonemes because unlike other sounds, these "clusters" can follow a consonant. Thus [mbw] or [skw] or [lŋw] are all legal clusters while, e.g., [mpw] is not, but it's easier to call these [w] combinations phonemes than to complexify the coda+onset analysis of clusters. They also behave like separate phonemes when on occasion a morpheme boundary may result in /kʷ ŋʷ bʷ/ being followed by a consonant, in which case the [w] vanishes and they become [k ŋ b] - and because some word-final [k ŋ b] become [kw ŋw bw] when a vowel is suffixed, but not others.

The epenthetic schwa rule: When adding a prefix or suffix would create an illegal consonant cluster, an epenthetic [ə] is inserted if the affix isn't a nasal that can become syllabic. An epenthetic schwa is always spelled as a and identified by speakers as belonging to the phoneme /a/. Since this epenthetic vowel is not contrastive, there is some variation as to whether it is added between the consonants in question or at the edge of the word.

Vowel Hiatus Rules

Vowel hiatus is generally forbidden. The first exception is that when the vowel combination matches one of phonemic diphthongs (/au ɑi ɐʊ̃ ɐɪ̃/), it becomes that diphthong. The second exception is when the vowel combination is one that naturally causes a glide sound to occur while moving from one articulation to the other; these "implied" consonants [j] and [w] are sufficient to allow the vowels to exist in quasi-hiatus.

For example, /oa/ and /io/ would be allowed, because they'd be pronounced [owa] and [ijo] respectively, while /ao/ and /ɛ̃a/ do not involve an intervening glide and are thus disallowed.

There may be some variation between speakers as to precisely which vowel combinations do and do not involve a glide.

When morpheme combinations cause illegal hiatuses to form, the first of the two vowels deletes.


  • /ba/ + /us/ = [baus] (hiatus becomes diphthong)
  • /bi/ + /as/ = [bijas] (hiatus held apart by "implied" glide)
  • /ba/ +/es/ = [bes] (full hiatus disallowed; first vowel deletes)

Vowel Allophony

Remember the nasal vowels? It may be helpful to not think of them as "main" vowels but instead allophones of the oral vowels that, whoops, became contrastive/phonemic in certain limited cases, c.f. German /x/ splitting into /χ/ and /ç/ in a tiny handful of minimal pairs. I'll explain how it came about and its implications in a moment.

There are two sets of allophones for five of the seven main (oral) vowels:

  1. /a e i ɑi au/ have the allophones of [ɐ ɛ ɪ ɐɪ ɐʊ] before a syllable coda.
  2. /a e i ɑi au/ have the allophones of [ɐ̃ ɛ̃ ɪ̃ ɐɪ̃ ɐʊ̃] before a nasal (whether that nasal is a coda or not).

The other two main vowels, /o/ and /u/, nasalize to [õ] and [ũ] respectively before nasals, but do not exhibit any other allophony.

The nasalized versions of the first five vowels have become contrastive. There used to be a fifth nasal consonant, /ɴ/. The usual vowel allophony applied before it, as before any other nasal; but /ɴ/ was lost after /a e i ɑi au/ and merged into /ŋ/ after /o u/. A cîrcûmflêx is used to mark that vowels formerly followed by /ɴ/ are still nasalized; all other nasalized vowels can predicted by seeing if it's followed by a nasal.


  • /pa/ [pa] pa
  • /pan/ [pɐn] pan
  • /paɴ/ → /pɐ̃/ [pɐ̃]
  • /poɴ/ → /poŋ/ [põŋ] pong

Stress Rules

Stress is usually on the first syllable of the word stem, regardless of inflections. Keep in mind that this applies only to vowel syllables; initial and final syllabic nasals do not count as "syllables" for stress purposes. A few words have stress on some other syllable instead; when stress is elsewhere, it is marked with a gràvè àccènt.

Ndak Ta is a syllable-timed language. That is, each syllable, not including syllabic nasals, is pronounced in roughly the same amount of time as any other; syllabic nasals just kinda get "squished" in between.