Re: Syllabic consonants (was: Re: Beek)

>[snip]
> > In it, the word <karm>, 'shield" is pronounced in two syllables.  (The word
> > should properly be written with an accent over the <a> to indicate stress
> > on the first syllable.)  As a matter of fact, the /m/ is syllabic here
> > because liquid plus nasal clusters in the syllable coda are illegal, to the
> > illegal cluster is broken up by making the m syllabic.
>
>Mmmm, I like syllabic /m/. :-)
>
>Does this mean that the /m/ is usually consonantal, but turns syllabic
>here because of phonological constraints?
Yes.  I'm not sure that I realized this when I first made up the word, but
it is the case that the /m/ is syllabic due to phonological constraints.  I
made up the phonological rule sometime after I made up the word.  In the
post that I wrote dealing with this before I saw your post you can see some
impromptu examples of this /m/ responding to syllable structure constraints
and being realized as either syllabic or non-syllabic.

> > Other words from the same language with syllabic consonants: <tovl>, 'to
> > instruct' and <tovlm>, 'instruction' (both of which should have accents
> > over the <l> to indicate that it is stressed
>
>So is /l/ syllabic in both these instances, or is it consonantal in
>/tovlm/?
The /l/ in /tovlm/ is syllabic.  The word should be divided into syllables
thus: to-vlm.

>[snip]
> > the /l/ is actually the nucleus of the syllable.  In forming a present
> > active participle of this verb, the /l/ ends up in the syllable onset, and
> > the word is stressed on the second syllable, which contains the
> > /l/.  <tovleis>, 'instructing';  <toerevl>, (stressed on the second to last
> > syllable) means 'war"; and then there is the minimal pair <mta> and <mta>,
> > one of them stressed on the syllabic /m/ the other stressed on the
> > <ai>.  (I don't know what either word in the minimal pair actually means,
> > but I think that they are probably verbs.)
>
>Interesting. Usually, I'd think of syllabic /m/ and consonantal /m/ as
>being two totally different phonemes altogether. Your language seems to
>alternate between them freely, which is rather interesting.
To me it just seemed to be a logical thing to do as soon as I had made up
the phonotactic constraint that liquid+nasal clusters were illegal (I
already knew that the language had syllabic sonorants.)  BTW, in the native
orthography, AFAIK, it is not notated whether a sonorant consonant is
syllabic or not.  It should be transparent to speakers/readers when to
pronounce them syllabic and when to pronounce them consonental.

I wonder if any natural language alternates them like this?

To give a couple more examples using the two case endings that I made up
for my last post: <toerevl> 'war', 4 syllables, divided as to-e-re-vl, the
/l/ is syllabic; <toerevlab> 'very many wars', 4 syllables again, divided
as either to-e-re-vlab or to-e-rev-lab (depending on whether /vl/ is a
legal syllable onset or not.  I just don't know yet.  Like I said, this is
a *very* new language: all three of mine are.), the /l/ is not syllabic
here but has moved into the onset of the syllable; and <toerevldim> 'a few
wars', 5 syllables, divided as to-e-re-vl-dim, the /l/ is syllabic here
because the sequence /vld/ is not a legal syllable onset by any stretch of
the imagination (/vld/ is legal as a *syllable*, but *not* as an onset.)

Now, if the behavior that I have outlined above is correct, that has some
ramifications (which I had not thought through before now) for my minimal
pair verbs <mta> which begin with syllabic a syllabic sonorant.  If the
syllabicity(?) of a sonorant consonant is conditioned by its phonological
environment then it is possible for those /m/'s to turn consonental when
preceeded by a vowel.  As a matter of fact, if the /l/ in <tovleis> turns
consonental because it can in that context, then I cannot think of any
reasonable way to for the /m/ in <mta> not to do exactly the same
thing.  (I think it would be completely unnatural if it didn't behave in
the same way.  Can you think of any reasoning to the contrary?)  Here is an
example: the way to negate a verb is to add the negative prefix emi- to
it.  So you get <emimta> 'to not...whatever mta means.'  This would have to
be broken up into syllables as e-mim-ta, and stressed either e-MIM-ta or
e-mim-TA, depending on which of the two verbs it was.

I suppose that an alternative would be to trash the idea of the sonorant
consonats alternating between syllabic and consonantal realizations and
have <tovleis> pronounced as to-VL-eis and <emimta> pronounced as either
e-mi-M-ta or e-mi-m-TA.  (This would probably also require the addition of
a syllabicity diacritic to the transcription of the language to avoid
confusion.)  Which is the better one to go with do you think: phonemically
distinct syllabic and consonantal sonorant consonants, or syllabic and
consonantal sonorant consonants as allophonic variations conditioned by
context?

I'm glad that I happened to write about this.  I wasn't sure that anyone
would the post or care in the first place, but I've already come out of it
with two new case endings, a negative marker (I almost knew the form of the
negative marker before today, but it wasn't finalized), and a lot of food
for thought about syllabicity.

>[snip]
> > the unstressed -i is added.  (I wrote these words out and showed them to my
> > husband, and he thought that it looked atrocious, or at least unappetizing
> > or unappealing.  It turned him off, in any case.  I think there were too
> > many consonants together and that he didn't like the look of accents over
> > consonants.)
>[snip]
>
>Heh, he might get a fright out of Ebisedian then. The LaTeX orthography of
>Ebisedian has circumflex consonants and vowels with multiple diacritics. A
>single vowel can have a total of 4 diacritics: the "teardrop" accent (a
>superscript hook), an acute, a macron, and a subscript tilde. When written
>out in ASCII, it's even more atrocious; e.g., /`yy~'/ for
>y-macron-teardrop-acute-tilde (the ` and ' are part of the vowel).
I am certain that my husband would be fairly put off; he likes for
pronunciations to be transparent.  He says that if he's going to have to go
to that much trouble to learn the transcription system, he'd rather just
learn the original glyphs.  The ASCII transcription really looks unwieldy
to me.  I'm sure that it must look fine in LaTeX, though.  (What is LaTeX?)

Isidora

Replies