Tahitian stress
Carleton L. Taylor
I notice Jacques Guy has mentioned Tahitian in several recent posts. Having
just acquired three books on this language I wonder if he, or anyone else,
can provide any definitive information about stress in this language. The
following works each state something very different about where the stress
falls within a word, i.e. on which syllable(s) or mora(e). Peltzer and
Tryon do agree, however, that stress is quite weak, and at times hardly
perceptible.
1. _Grammaire descriptive du tahitien_, Louise Peltzer, Editions Polycop,
1996
2. _Parler tahitien en 24 leçons_, D. T. Tryon, Editions Octavo, 1997
3. _Lexique du tahitien contemporain_, Yves Lemaître, IRD Editions, 1995
Generally speaking, is word stress a controversial, debated, or open topic
in Polynesian language studies?
Carl Taylor
Carleton L. Taylor
"JGuy" <jg...@dev.nul.nu> wrote in message news:3C9CF7...@dev.nul.nu....
[...]
> My Tahitian is awfully rusty,
> but the sound of the language was so different from what I had
> believed (Tahitian is supposed to be mellifluous and all that;
> well, it sounded awfully rough to me), that it stuck. I'd rather
> say that there is no stress, but that in some positions, short
> vowels become very short or even elided. I remember the
> pronunciation of my Tahitian name: Tihâti. The first i was
> elided, giving [tSa:ti] (/ih/ -> [iS], e.g. niho [niSo] "tooth",
> ihu --> [iSu] "nose"). Well, at least that was how our teacher,
> Mme Lenormand (100% Tahitian despite her name) pronounced them.
Yes, the /h/, especially when preceded by /i/ and followed by a rounded
vowel is rather strong. Watching the evening Tahitian language news in
Pape'ete it sounded to me like [C] as in German "ich" or even sometimes like
[H] as in Arabic "al Harb" (war). And there definitely is a goodly amount
of elision. When listening to hotel personnel reading a text as I looked
on, I'd get the impression quite a few syllables were disappearing.
Carl Taylor
Bart Mathias
> Generally speaking, is word stress a controversial, debated, or
> open topic in Polynesian language studies?
After reading two responses about Tahitian, I hope Ross Clark will
join the discussion.
As Hawaiian is taught nowadays, the rules are pretty simple, boiling
down to approximately: Stress all long vowels and the first vowel of
diphthongs (including those created at word boundaries in tight
phrases). Stress the first syllable of words of more than three
syllables, then arrange to have alternate syllables stressed down to
the penultimate.
I'm not quoting anything here, but I think it sums up what I was
taught--or rather, what I learned.
Bart Mathias
benlizross
Hey, that doesn't sound all that simple...
I've been away for a couple of days. Lovely weekend here in NZ.
Polynesian languages are known for having small phonemic inventories and
not very complex morphology. "No phonology" is the mental shorthand for
this situation. Nevertheless, there is more than one might think. The
well known -Cia suffixes have provided a bone for phonologists to pick
for more than 30 years now. And stress rules are surprisingly resistant
to neat formulation.
Disclaimers: My direct experience with Tahitian is nil. And phonology is
not my field. '
In Polynesian as a whole, the following will get you a long way: Long
vowels are basically sequences of two identical vowels. Stress alternate
vowels (morae) working left from the penultimate. This has as an
automatic consequence that all long vowels and diphthongs get a stress,
and that final short vowels never get stressed, which is the way the
rules are stated in many practical grammars. If there is a main word
stress it will be the last in the sequence (except in Maori). Al Schütz
has pointed out the difficulties that arise in trying to predict the
stress of long words (say more than 5 morae), and has argued for a
system where words are divided into foot-like stress units which are
stressed according to very simple rules. Unfortunately he says little
about where this division comes from. Clearly in part it comes from
morpheme boundaries. Later editions of the Pukui-Elbert Hawaiian
dictionary divide every word into units like this, which seems to me
largely unnecessary.
Goodness me, where am I? Must have fallen asleep muttering about
Polynesian stress rules.
The other Tahitian phenomenon mentioned by some posters is also
interesting. It seems to be a kind of bleeding of vowel quality across
the glottals (both /h/ and /?/), one conspicuous effect of which is to
produce a palatalized quality of /h/ when preceded by /i/. Something
else somebody could do a proper study on.
Ross Clark
Carleton L. Taylor
"benlizross" <benl...@ihug.co.nz> wrote in message
news:3C9DA4...@ihug.co.nz...
[...]
>
> In Polynesian as a whole, the following will get you a long way: Long
> vowels are basically sequences of two identical vowels. Stress alternate
> vowels (morae) working left from the penultimate. This has as an
> automatic consequence that all long vowels and diphthongs get a stress,
> and that final short vowels never get stressed, which is the way the
> rules are stated in many practical grammars. If there is a main word
> stress it will be the last in the sequence (except in Maori). Al Schütz
> has pointed out the difficulties that arise in trying to predict the
> stress of long words (say more than 5 morae), and has argued for a
> system where words are divided into foot-like stress units which are
> stressed according to very simple rules.
Ray Harlow writing in _A Maori Reference Grammar_ , Pearson Education NZ
Ltd., 2001 states:
"The stressed syllable is never more than four morae from the end. Words
longer than four morae in length are usually complex bases, that is, formed
from smaller units by means of prefixes, suffixes, or reduplication."
>Unfortunately he says little
>about where this division comes from. Clearly in part it comes from
>morpheme boundaries.
Which is what Harlow is saying.
> Later editions of the Pukui-Elbert Hawaiian
>dictionary divide every word into units like this, which seems to me
>largely unnecessary.
[...]
Carl Taylor
Carleton L. Taylor
"JGuy" <jg...@dev.nul.nu> wrote in message news:3C9CF7...@dev.nul.nu...
[...]
> I have only Tryon's and Lemaître's books. The two descriptions do not
> concur, but, if you read behind (yes, behind, not "between") the
> lines, they are not so very different. According to Lemaître, almost
> all the vowels are stressed (accentuées), the most frequent exception
> being short word-final vowels.
Peltzer simply says the stress occurs on the penultimate mora. A syllable
of form CV has one mora; those of form CVV two, where VV = a long vowel.
Her examples then sometimes contradict Tryon and Lemaître, as in tôtara ( a
rooster's crowing), which she rewrites as tootàra" to show the stress. Both
Tryon and Lemaître on the other hand seem to emphasize that long vowels and
diphthongs, (what Tryon calls a sequence of two different vowels), are the
primary recipients of stress, weak though it may be.
Carl Taylor
Bart Mathias
> In Polynesian as a whole, the following will get you a long way:
> Long vowels are basically sequences of two identical vowels. Stress
> alternate vowels (morae) working left from the penultimate. This
> has as an automatic consequence that all long vowels and diphthongs
> get a stress, and that final short vowels never get stressed, which
> is the way the rules are stated in many practical grammars.
I'm a little surprised that I didn't put it that way, because I can
remember formulating it so.
Problems as I hear Hawaiian pronounced include the word for "land,"
which, doubling for length (and indeed, Jacques, that doesn't make a
vowel long enough for Japanese to hear long vowels in Hawaiian--they
get katakanasized with short vowels) would be "`aaina." I went for
years saying "`AaIna" (caps representing two stresses) until I was
sure I was the only one. Apparently you can cram three vowels
("-aai") into one syllable, resulting in "`Aaina."
Re my contention that the first syllable of words of more than three
syllables is always stressed, one source (Pukui and Elbert, Hawaiian
Dictionary, 1971) has "..., except that words containing five
syllables [they leave 'syllable' undefined] without macrons are
stressed on the first and fourth syllables." But everybody but me
seems to say "kaMEhaMEha." (I would too, if there were an "`o," or
"me," or "i" or something in front to take a stress. Or should it
perhaps really be "KA(a)MEhaMEha"?).
Bart Mathias
Carleton L. Taylor
"Bart Mathias" <mat...@hawaii.edu> wrote in message
news:U1.81.iI2rE7...@hawaii.edu...
[...]
> Problems as I hear Hawaiian pronounced include the word for "land,"
> which, doubling for length (and indeed, Jacques, that doesn't make a
> vowel long enough for Japanese to hear long vowels in Hawaiian--they
> get katakanasized with short vowels) would be "`aaina." I went for
> years saying "`AaIna" (caps representing two stresses) until I was
> sure I was the only one. Apparently you can cram three vowels
> ("-aai") into one syllable, resulting in "`Aaina."
This "cramming" of three vowels into a single syllable apparently leads to a
phonemic distinction between short and long diphthongs. Such an analysis
yields the following phoneme inventory. (For the circumflex read 'macron').
Source: A. V. Lyovin's _An Introduction to the Languages of the World_ .
8 consonant phonemes: / p k ? h m n l w /
5 short vowel phonemes: / i u e o a /
5 long vowel phonemes: / î û ê ô â /
9 short diphthong phonemes: / iu ei eu ai ae ao au oi ou /
6 long diphthong phonemes: / âi âu âe âo êi ôu /
Total phoneme count = 33. Not bad for a Polynesian language, or any
language for that matter.
[...]
Carl Taylor
Bart Mathias
[In Hawaiian]
> [the] "cramming" of three vowels into a single syllable apparently
> leads to a phonemic distinction between short and long diphthongs.
> Such an analysis yields the following phoneme inventory. (For the
> circumflex read 'macron'). Source: A. V. Lyovin's _An Introduction
> to the Languages of the World_ .
> 8 consonant phonemes: / p k ? h m n l w /
> 5 short vowel phonemes: / i u e o a /
> 5 long vowel phonemes: / ī ū ź ō ā /
> 9 short diphthong phonemes: / iu ei eu ai ae ao au oi ou /
> 6 long diphthong phonemes: / āi āu āe āo źi ōu /
> Total phoneme count = 33. Not bad for a Polynesian language, or
> any language for that matter.
This analysis came up once before in this NG. I suggested there
might be a problem with diphthong phonemes (as in "wai" = "water,"
not necessarily the same example I used last time) being
phonologically identical to vowel phoneme sequences (as in "ka i`a" =
"the fish"). No one took the bait, so perhaps there's a strange gap
in my conception of "phoneme." (I would be tolerant of diphthong
phonemes in English.)
How many phonemes in "ka i`a i ka wai" = "the fish in the water"?
(Careful pronunciation would probably be KAi`a IkaWAi, but in rapid
speech might it not become KAi`AikaWai?) Similar things happen in
word compounding, such that adding a word with initial "a" to a word
with final "a" gets you an "aa" that sounds exactly like an "ā."
I confess I have neither asked Lyovin about this nor read his book.
Establishment of long diphthongs, phonemic or not, does solve my
"`āina" problem.
Bart Mathias
Carleton L. Taylor
"JGuy" <jg...@dev.nul.nu> wrote in message news:3CA09E...@dev.nul.nu...
[...]
>
> This is a novel application of the notion of phoneme...
>
I thought so too when I posted a query about it in this forum back on Sept.
8, 2001.
http://groups.google.com/groups?hl=en&threadm=9ndp5e%242q6%241%40bob.news.rc
n.net&rnum=1&prev=/groups%3Fq%3DDiphthongs%2Bas%2Bphonemes%2B%253F%253F%253F
%26hl%3Den%26selm%3D9ndp5e%25242q6%25241%2540bob.news.rcn.net%26rnum%3D1
Mark Rosenfelder replied in part:
You'd probably have to read Lyovin's source to find the reasons for treating
diphthongs as phonemes in Hawai'ian, but at least one is given in Lyovin's
sketch: the diphthongs "act as unit phonemes in regard to stress".
Carl Taylor
Bart Mathias
> You'd probably have to read Lyovin's source to find the reasons for
> treating diphthongs as phonemes in Hawai'ian, but at least one is
> given in Lyovin's sketch: the diphthongs "act as unit phonemes in
> regard to stress".
But they don't, really (as if it would be a good reason if they
really did).
Diphthongs are always stressed in Hawaiian. Ditto long vowels. But
all other phonemes are optionally stressed (if one accepts the notion
of stressed consonants), so multi-mora vocalisms are distinctly
different from "other unit phonemes."
Bart Mathias
Trinker
Bart Mathias wrote:
[...]
> Problems as I hear Hawaiian pronounced include the word for "land,"
> which, doubling for length (and indeed, Jacques, that doesn't make a
> vowel long enough for Japanese to hear long vowels in Hawaiian--they
> get katakanasized with short vowels) would be "`aaina."
(Apologies for the belated post, I'm reading sci.lang for the first
time in months.)
Thank you, Bart, for writing that. I'd been going crazy for years
trying to hear the "long vowel" in Hawaiian, and having problems
because compared to "long vowels" in Japanese, they're not long at
all.
(For those who don't know me, I'm a native speaker of Japanese whose
main daily language is English.)