Syllable Stress Allocation in Ayeri

This question has been bothering me for some time already and in the Ayeri Grammar in its current state (still only half-finished; I don’t know when I will go on with writing) it just says:

Stress in Ayeri is irregular for the most part, however, certain rules can be applied: Syllables ending in ŋ, as well as syllables containing a long vowel or a diphthong are usually stressed, and no two stressed syllables may follow each other.

This felt rather general and not sufficient as a definition to me. Some months ago, there was a posting on Glottoclast which dealt, among others, with how one of the author’s constructed languages, Cwindoià, applies stress: it uses syllable weight for stress allocation. This got me thinking again whether syllable weight based on the phonetic structure of a syllable might also apply to Ayeri. Thus, I did some research quickly the other day and found a paper on syllable weight strategies by Matthew Gordon of UCSB, called “A Phonetically-Driven Account of Syllable Weight,” which sounded just like what I had been looking for.

What I then did recently is to take all texts I’ve translated into Ayeri over the last 2 years to get a corpus made up of inflected words. I manually decomposed all approx. 530 unique words (so few!?) into syllables and classified their composition (e.g. P-A-​-​-N- for a syllable that starts with a voiceless plosive, has /a/ as its nucleus, and ends in a nasal; the dashes denote unoccupied segments) for the first 100 entries – after which I lost patience, as it was just too tedious to do the whole analysis manually and I had already realized the following (acutes are supposed to align over the macroned vowels if they don’t do so):

pa-pā-pang-pāng-pay-pāy-
-kapákapā́kapángkapā́ngkapáykapā́yka
-kāpakā́pā́kā́pangkā́pā́ngkā́paykā́pā́ykā́
-kangpakángpā́kangpangkángpā́ngkangpaykángpā́ykang
-kāngpakā́ngpā́kā́ngpangkā́ngpā́ngkā́ngpaykā́ngpā́ykā́ng
-kaypakáypā́kaypangkáypā́ngkaypaykáypā́ykay
-kāypakā́ypā́kā́ypangkā́ypā́ngkā́ypaykā́ypā́ykā́y

From this table, we can deduce the following:

  1. If both syllables are plain (i.e. contain a short vowel and optionally a final consonant other than /ŋ/), stress is on the penultimate syllable: páka.
    1. Long syllables attract stress.
    2. They override plain syllables: pakā́.
    3. They override syllables ending in /ŋ/: pā́kang.
    4. They override syllables ending in a diphthong: pā́kay.
    1. Syllables ending in /ŋ/ attract stress.
    2. They override plain syllables and diphthongs in final position: pakáng, paykáng.
    3. In non-final position, they override other syllable types only if they are long: pangkáy vs. pā́ngkay.
    1. Syllables ending in a diphthong attract stress.
    2. They override plain syllables and syllables ending in /ŋ/ in final position: pakáy, pangkáy.
    3. In non-final position, they override other syllable types only if they are long: paykáng vs. pā́ykang.
  2. Two long syllables will result in both having stress: pā́kā́, pā́ngkā́ng, pā́ykā́y.
  3. Two syllables both ending in /ŋ/ will result in only the latter having stress: pangkáng.
  4. Two syllables both ending in a diphthong will result in only the latter having stress: paykáy.

What can be observed furthermore – in general, not from this table – is that in words with more than two syllables, stress spreads backwards, alternating between unstressed and stressed. Syllables containing [ə] are unstressed by default and stress shifts to the preceding syllable as a result. Furthermore, the individual parts of compounds follow their individual stress pattern, so stress does not spread backwards unlimitedly. Although the table above lists combinations of two syllables where both syllables receive stress, this is, however, very rare in practice (the only word I can think of off-hand is bāmā ‘mum and dad’). In turn, multiple unstressed syllables in succession are not a problem.

In conclusion can be stated that after much work has gone into breaking down words and analyzing their syllable structure, things turned out to be probably easier than expected and it could be mostly confirmed and corrobated what was tentatively assumed earlier. Syllable weight may in fact not play much of a role in how Ayeri assigns stress to words. Instead, it seems to be iambic for words with an odd number of syllables and trochaic for words with an even number of syllables by default (so not totally irregular), with the exception being that certain phonetic features of syllables cause a shift.