Quoting Christophe Grandsire <[log in to unmask]>:

> En réponse à "Thomas R. Wier" <[log in to unmask]>:
> >
> > Note that I was talking about *phonology*, not phonetics.  That's
> > important: we're talking about the distribution of sounds in the
> > language, because that affects what kinds of distinctions the
> > speaker *needs* to make. When a language has one series of stops
> > at labial, alveolar and velar positions, we can call them whatever
> > we want, because there is ipso facto no distinction, and
> > phonetically these languages in fact tend to have very short
> > voice-onset times, either positive or negative (i.e., they will
> > sound very much like plain [p t k]) because they do not need to
> > make those differences.  When you introduce a second series, the
> > speakers tend to make these just distinct enough to hear, and tend
> > to make them equidistant in terms of positive or negative VOT.
> > All of this means that voicing (i.e. distinctly negative VOT) is
> > typologically more marked, and that a system which has distinctly
> > voiced segments but lacks a unvoiced counterpart is itself
> > typologically marked.
> Then how do you explain the Classical Arabic system, which always render
> the bilabial stop as [b], never [p]?

Unfortunately, the Classical Arabic system is beside the point.
The job of a typologist is to infer from many different points of
data from various sources (geographical, familial, etc.) how
languages across the world tend to behave.  Classical Arabic
is one data point, which is fine, but IIRC statisticians believe
that you need to have at least *30* data points to form a normal
curve, and typically linguistic typologists look at far many more
data points than that (usually over 150, depending on the amount
of data that has been published).

> IIRC, saying that phonemically this is a /p/, which just
> happens to be phonetically a [b], only to agree with an
> unclear idea that voiced consonnants are more typologically
> marked than unvoiced ones

But I wasn't talking about Classical Arabic's obstruent
system, but rather obstruent systems in general.  There is
a robust amount of evidence to support my claim.  I suggest you
look through the UCLA Phonological Segment Inventory database
if you believe that I am in error. (I do not have ready access
to this right now; otherwise, I would quote it.)

> (something which is not proven, since until now, nobody has been
> able to come with an explanation of typological markedness other
> than frequency of appearance, which makes the reasoning circular
> since frequency of appearance should be the consequence of typological
> markedness, not the explanation.

But that's because linguists are trying to bring empirical evidence
to bear on the question.  They are not trying to use deductive logic
at all, but inductive.  Linguistics is a young science, and there
are many aspects of typology that have not been explained.  So, what
linguists do as a second best effort is to amass a vast body of data
about obstruent systems the world over, and find the patterns in them.
This is no different from physics in the 17th century: Newton himself
was criticized for invoking an "occult force" called gravity for which
he gave no explanation other than that it explained the data well.
Centuries later people came to understand that gravity was not just
an occult force, but a result of the geometry of the universe. The
same holds true for linguistics:  our "occult force", markedness,
drives languages to behave in certain ways. No doubt at some point in
the future we will begin to discover what that actually means.

> > > On the other hand, languages with back unrounded vowels lacking
> > > the corresponding back rounded ones are frequent (take simply
> > > Japanese, which has /M/ - high back unrounded - without its rounded
> > > variant /u/).
> >
> > I wouldn't say that at all:  [u] as a value for /u/ is very,
> > very frequent, with roughly 1/4 of the languages of the world
> > having an [i e a o u] system alone.  I would not say that back mid
> > or high unrounded vowels are rare, exactly, but they are certainly
> > not "frequent" except in an absolute sense of having hundreds of
> > languages with them.  For the same reason, front mid or high rounded
> > vowels are not rare but not relatively common, either:
> > unroundedness seems to be a marker of frontness, while roundedness
> > seems to be a marker of backness.
> It was frequency relative to that of rounded front vowels lacking
> their unrounded counterparts. You can't deny that while you probably
> can't find examples of languages having rounded front vowels lacking
> their unrounded front counterparts, you can *more* easily find
> languages with unrounded back vowels lacking their rounded
> counterparts

No, I would deny that, because one data point is not sufficient
to prove typological patterns -- in fact, is completely incompatible
with the notion of "typology", since typology ipso nomine implies
more than one point of data. Again, I would refer you to the UCLA
Phonological Segment inventory database, which says that front
rounded vowels and back unrounded vowels are statistically speaking
about equally common in languages of the world.  (If you do not have
access to this, I believe Ladefoged's introductory textbook on
phonetics says about the same thing.)

> (Japanese is an example with its /M/, but
> across Asia this is a common phenomenon).

But if we're aiming about the typology of a human language --
which was our topic of discussion, not an Asian language -- this
information matters only if you put it into its larger context.

Thomas Wier <[log in to unmask]> <>

             "...koruphàs hetéras hetére:isi prosápto:n /
Dept. of Linguistics  mú:tho:n mè: teléein atrapòn mían..."
University of Chicago "To join together diverse peaks of thought /
1010 E. 59th Street   and not complete one road that has no turn"
Chicago, IL 60637     Empedocles, _On Nature_, on speculative thinkers