tlahtoa / saltillo

[Joost Kremers]

hi all,

a few extra comments about the saltillo-discussion from a linguist
(albeit a syntactician, but one not entirely unknowledgable on matters
phonologic). basically, the account given by Galen Brokaw (see below) is
correct. in phonology, it is usually assumed that a word is stored in
the mental lexicon as a string of phonemes. when a word is pronounced,
these phonemes may undergo some alternation: the so-called allophones. a
good example of this is the spanish /b/, which, depending on context, is
pronounced either as [b] or [=DF] (=DF is the symbol for the so-called
bilabial fricative: sort of a v pronounced with both lips rather than
with the upper teeth and lower lips. note, by the way, that phonemes are
indicated with slashes, whereas square brackets indicate the actual
phonological realization.) in phonological terms: the phoneme /b/ has
two allophones in spanish: [b] and [=DF].

but this is not all of what is going on in the phonology of a language.
there is another set of processes, which do not apply to phonemes, but
to the actual phonological realization. for example, in RP (british
english), there is a rule called preglottalization: before a voiceless
stop (/p, t, k/) a glottal stop (/?/) is inserted. that is, a word such
as 'dent' is pronounced something like [den?t], with a glottal stop
before the /t/. (there is good reason not to see [?t] as an allophone of
/t/, but it would take too far to go into that.)

note that this category of rules may introduce phonological items that
do not exist as phonemes in the language: RP does not have the glottal
stop as a phoneme.

as for nahuatl, the case seems to be the following. (in this, i base
myself on the discussion on this list, as i have no first-hand
information.) there is an underlying phoneme, call it saltillo, that has
at least two allophones: as [h] in word-final position and between
vowels, and as a glottal stop (perhaps accompanied by aspiration) before
a consonant. if i'm not mistaken, (but i get all my information trough
the web, so for what it's worth:) it is even a velar fricative (close to
spanish j) in some dialects. the question what the underlying phoneme
is, cannot be answered easily. (from what i've read on the list so far,
i would argue it is probably /h/, not a glottal stop, but this is not a
critical question.)

apart from this, there is apparently a rule that inserts a glottal stop
after a word-final vowel. this glottal stop is not part of the phonemic
string that underlies a word, it is inserted afterwards. it is
comparable to the rule that inserts a glottal stop before word-initial
vowels in german (and other languages, e.g., dutch). i.e., it is not a
phonemic, but a phonological phenomenon.

so we have:

tlahtoah: underlyingly /Lahtoah/ (i'll just use L to indicate the
lateral plosive, for want of a better way...). allophone assignment
makes the first /h/ a /?/, and the second a /h/. no further rules apply,
so we get [La?toah].

tlahtoa: underlyingly /Lahtoa/. allophone assignment makes the /h/ a
/?/. after that, another rule, say 'post-glottalization', adds a /?/.
so, the word comes out as [La?toa?].

note that native speakers of a language are usually aware of the
presence of phonemes, (although they may not recognize the allophones),
but they are not aware of the inserted elements. for example, a speaker
of spanish recognizes the presence of a phoneme /b/ in 'mover' (he will
say the word consists of the sounds m, o, b (or v), e and r). but a
speaker of (british) english will not be aware that there is a glottal
stop in his pronunciation of 'dent'. that means that if the descriptions
given in this list of actual speech from native speakers is correct,
they should consider plural quicuah as containing one more sound than
singular quicua.


here ends my lecture :-) as i said, the conclusion reached in the
discussion is quite correct. i just added my bit to explain how
phonologists would probably look at the matter.

Joost Kremers

-------------------------------------------
Joost Kremers

University of Nijmegen - The Netherlands
Department of Languages and Cultures of the Middle East
POBox 9103
6501 HD Nijmegen - The Netherlands
phone: +31 24 3608226



Galen Brokaw wrote:
> =

> John,
> You are right. I think my last email crossed with yours in cyberspace. =
In the
> dialects we are discussing, the glottal stop is not a phoneme. But the =
closing
> off of the throat is a nonphonemic glottal stop which occurs at the end=
 of vowel
> final words. I think that was the original issue that Fran's post raise=
d in my
> mind. So, if I may sum up here just to see if we are all on the same pa=
ge. And
> please Fran, Mary, John, et al please correct me if I'm wrong:
> 1. The saltillo was a glottal stop in Classical (I know that from what =
John has
> said he may disagree with this)
> 2. In many (most?) modern dialects the glottal stop evolved into an asp=
iration
> /h/
> 3. In these dialects the glottal stop is no longer a phoneme but exists=
 as a
> phonetic characteristic of world final vowels.
> 4. In these modern dialects, the indicative singular verbs end with thi=
s
> nonphonemic glottal stop
> 5. In Classical, the indicative plural verbs ended with a glottal stop,=
 which
> was referred to as a saltillo (see #1)
> 6. In these modern dialects, the Classical saltillo of the indicative p=
lural
> verbs has evolved into an aspiration /h/ (see #2 and #5).
> 7. Most linguists continue to refer to the modern /h/ as a saltillo/glo=
ttal stop
> for historical reasons even though phonetically it is not.
> =

> I do have one question for those who may know. Are there differences of=
 opinion
> among linguists about what the saltillo was phonetically in Classical N=
ahuatl?
> Also, I should say that although this discussion may seem pedantic to m=
any, it
> has been very informative for me and I appreciate the dialogue with Joh=
n,
> Michael, and Mary.
> Galen
> =

> John Sullivan Hendricks wrote:
> =

> > Galen,
> > I'm sorry for the lack of precision with my terminology, but I'm not =
a
> > linguist and the last course I took in linguistics was about 20 years=
 ago.
> > However, my lay description of the phoneme/allophone distinction was
> > correct, although we disagree on its applicability in this case.  Mov=
ing
> > along, when I wrote, "The pronunciation of the vowel which ends the s=
ingular
> > form, terminates with a closing off of the throat.  But this is true =
of all
> > word final vowels," what I should have said was, "this is true of all=
 word
> > final vowels in Nahuatl."  I never meant to refer to "all languages. =
 With
> > this said, I agree with Mary's suggestion that the closing off of the=
 throat
> > after word final vowels in Nahuatl MIGHT only mark the absence of the=

> > saltillo. Don't you think it's a bit suspect that the only criterion =
for
> > distinction between the two phonemes you postulate (glottal stop and
> > aspiration) is in the case of singular vs. plural?  O are you aware o=
f any
> > word-internal distinctive pairs which contrast the glottal stop and t=
he
> > aspiration?
> >         John Sullivan
> >         Universidad Aut=F3noma de Zacatecas
> >
> > -----Mensaje original-----
> > De: owner-nahuat-l at server2.umt.edu [mailto:owner-nahuat-l at server2.umt=
=2Eedu]En
> > nombre de Mary Clayton
> > Enviado el: Jueves, 10 de Febrero de 2000 01:57 p.m.
> > Para: Multiple recipients of list
> > Asunto: RE: tlahtoa / saltillo
> >
> > Galen,
> >      I think you've given an admirable account of the concept of
> > phonemes and allophones. I think that the 'disagreement' here may act=
ually
> > involve an unstated assumption, which doesn't directly involve 'fact'=
, but
> > (as is usual in linguistics at all levels of analysis -- even phoneti=
cs)
> > *interpretation* of fact.
> >      I think that the question that one would need to ask is: Is ther=
e any
> > *third* way for words to end [that don't end in other consonants, of
> > course]? That is, do the h and the glottal stop form a three-way cont=
rast
> > with a simple final vowel which is followed by NEITHER of these? Beca=
use a
> > (not unreasonable) linguistic assumption might be that one of these i=
s
> > distinctive (as are other consonants) while the other just marks its
> > absence, that is, just draws special attention to the fact that the w=
ord
> > in question ends in a vowel rather than glottal stop (or h, depending=
 on
> > which way you want to argue). Seen this way, the question would be: C=
an
> > words end in
> >       -h,
> >       -? ( =3D glottal stop)
> >       -C ( =3D any consonant other than glottal stop)
> >       or -V (=3D any vowel)?
> >
> > Or are the choices just
> >       -h and -V as "allophones"
> >       -?
> >       -C
> >
> > or    -? and -V as "allophones"
> >       -h
> >       -C
> >
> >      Most people take a historical position on this question, calling=
 the
> > [historical] glottal stop a 'glottal stop' or 'saltillo' regardless o=
f
> > whether it is pronounced as [h] or [?] in the dialect in question.
> > Whichever the pronunciation, this segment shows up in the morphologic=
ally
> > expected places, which is one reason to keep the name the same even
> > thought the pronunciation of presnest-day dialects may differ.
> >      A way to 'explain' the modern distribution of sounds for those
> > dialects which have  *real* glottal stops in the singular is that
> > the *historical* glottal stop has > [h], and the *new* glottal stop j=
ust
> > signals its absence. Taken too literally, this sounds like double-tal=
k, of
> > course: [?] =3D nothing and [h] =3D glottal stop, but it's a natural =
way of
> > thinking for people who work with both classical and modern.
> >
> >      Let me mention two personal observations which are relevant to t=
he
> > issue, and certainly don't make it any simpler:
> >
> >      -- One of Joe's young friends (from Oapan) who is heavily
> > Nahuatl-dominant, when speaking Spanish very frequently ends vowel-fi=
nal
> > words with a glottal stop. Spanish "si/" is [si?] and "no" is [no?]. =
I
> > don't think he's exceptional. He's just the one I've heard the most (=
on
> > tape -- his tapes frequently accompany us on car trips.)
> >
> >      -- There is a characteristic of general Mexican Spanish which ma=
y
> > play a part in the confusion. I'm talking about Mexican speakers who =
know
> > no Nahuatl, many of whom have probably had little or no exposure to i=
t. My
> > Caribbean-hearing ear (I'm from Tampa and grew up around Cuban-type
> > Spanish) hears enough trailing-off of voice at the end of vowel-final=

> > words in Mexican Spanish to qualify them for -s final words in Cuban.=

> > That is to say, many pronunciations of singular nouns would be heard =
as
> > plurals by people who turn s > h, and third-singular verbs sound like=

> > tu-forms.  But Mexicans don't *mean* or *intend* this 'aspiration', i=
t's
> > just the way they frequently pronounce vowels in final positions.
> >      This may well influence the way 'outsiders' who are very familia=
r
> > with Mexican Spanish, whether native spekers or not, perceive final v=
owels
> > or final -h's in Nahuatl. I'll stop short of speculating what effect =
it
> > may have on Nahuatl speakers.
> >
> > Mary
> >
> > On Thu, 10 Feb 2000 brokawg at mail.lafayette.edu wrote:
> >
> > >
> > > John,
> > > I have to respectfully disagree. The linguistic terms you refer to =
are
> > > phoneme and allophone. The phoneme is the phonetic element which ma=
y have
> > > various allophones depending on the environment. In Spanish, for ex=
ample,
> > > the phoneme /n/ is pronounced one way when it is intervocalic and a=
nother
> > > when it is followed by a [g] or [c] and another when followed by [t=
], etc.
> > > The way you determine whether or not two sounds are separate phonem=
es or
> > > allophones of the same phoneme is to try and find two different wor=
ds that
> > > are exactly the same except for the two sounds in question. If two =
such
> > > words exist, then the sounds are separate phonemes. I don't have en=
ough
> > > experience to be able to make judgments about all modern dialects o=
f
> > > Nahuatl, but my impression is that the aspiration and the glottal s=
top are
> > > two separate phonemes. I would support this by providing a contrast=
ive
> > > pair of words consisting of the singular and plural of the indicati=
ve.
> > > Now, John, I know that you disagree that the singular ending of the=

> > > indicative is a glottal stop. It is tough to argue these kind of th=
ings
> > > over email without the benefit of speech, and you have more access =
to
> > > native informants than I do, but I still think that what you descri=
be as a
> > > closing off of the throat is a glottal stop. And I disagree that al=
l words
> > > in any language must end by closing off the throat. In English when=
 we
> > > pronounce a word that ends in a vowel, we don't close off the throa=
t at
> > > the end. The word "go" for example doesn't end by abruptly interrup=
ting
> > > the vibration of the vocal chords by closing off the throat. The vo=
cal
> > > chords just quit vibrating and we cease to expel air. Glottal stop =
is
> > > defined precisely as a closing off of the throat using the glottis =
and
> > > consequently an abrupt stopping of vocal chord vibration. The nativ=
e
> > > speaker with whom I have had experience had a very clear glottal st=
op at
> > > the end of verbs in the indicative singular. The difficulty I had w=
as in
> > > determining if there was an aspiration at the end of the plural or =
not. I
> > > like to think there was a faint aspiration just as you have noted i=
n the
> > > speech of Huastecan Nahuatl. So, if this is the case, then we have =
a
> > > contrastive pair in the third person singular and third person plur=
al
> > > indicative verbs such as quicua [kwa?] and quicuah [kwah] where the=

> > > [?]=3D3Dglottal stop. (The phonetic symbol is actually an upside do=
wn
> > questio=3D
> > > n
> > > mark with no dot, but I can't make that go through on the email.) I=

> > > conclude therefore that the glottal stop and the aspiration are two=

> > > separate phonemes.
> > > I submit this argument humbly and ask any of the professional lingu=
ists to
> > > correct my reasoning if it is flawed.
> > > Galen=3D20