Lots of Questions About Tones (more questions)

On Thu, 10 Jul 2008 03:16:29 -0400, John Vertical
<johnvertical@...> wrote:
>[snip]
>I'm under the impression that "glide" tones are included under contour tones,
>as long as the "glideness" is phonemic. Er, tonemic.
>[snip]
Well, "contour" is used in three different senses.

Sense 1
In one sense, "contour" is used opposite to "register".

In a "register tone language", the absolute pitch at which a syllable is spoken
is its tone, and has phonemic and/or lexical and/or morphological significance.

But, in a "contour(sense 1) tone language" what matters is whether, and by
how much, the pitch changes.
So a "contour(sense 1) tone language" can have at most one level tone; that
is, all level tones mean the same thing.  If it has four relevant pitches (say 1
and 2 and 3 and 4), there is a choice of three falls and of three rises;
21, 32, and 43 are all one possible kind of fall;
31 and 42 are another possible kind of fall; and
41 is a third possible kind of fall.
These languages can tell 41 apart from 31 and from 42, but can't tell 31 apart
from 42; they can tell 41 apart from any of 21, 32, or 43, but can't tell 21 or
32 or 43 apart from each other.

Sense 2
In another sense, "contour" is used opposite to "level".

In this sense, all rises, all falls, all peaks (rise-falls), all dips (fall-rises), and
anything more complicated (peak-dips, dip-peaks, and, if such things exist,
peak-dip-peaks and dip-peak-dips) are all "contour tones".

Sense 3
In a third sense, "contour" is used opposite to "level or glide".  Rises and falls
are grouped as "glide tones", minimally more complicated than level tones;
while peaks, dips, peak-dips, and dip-peaks are "contour tones".  If you want
to rank tones by complexity, one terminology is:
level tones;
glide tones (rises and falls);
simple contour tones (peaks and dips); and
complex contour tones (peak-dips, dip-peaks, and anything more complicated).

---------------------------------------------------------------------------

According to Encarta (or Britannica?), most tonal languages don't have more
than two tones of "the same shape", and very few have more than three.

So, what's a shape?

If the intervals count -- that is, if 12 and 13 and 14 are all different -- but the
absolute pitches don't -- that is, 12 and 23 and 34 are all the same -- then,
with four relevant pitches, there could be up to three "shapes" of rises,
three "shapes" of falls, nine "shapes" of peaks, and nine "shapes" of dips.
Rises:  12 = 23 = 34; 13 = 24; 14.
Falls:  21 = 32 = 43; 31 = 42; 41.
Peaks:  121 = 232 = 343; 131 = 242; 132 = 243; 231 = 342; 141; 142; 143;
241; 341.
Dips:  212 = 323 = 434; 313 = 424; 312 = 423; 213 = 324; 414; 413; 412;
314; 214.

--------

But, if only the direction of motion of the pitch counts, and the distance it
moves isn't important, all rises are (phonemically or morphemically) equal, all
falls are equal, all peaks are equal, all dips are equal, etc.

---------------------------------------------------------------------------

Are any natlangs sensitive to the timing of the rise or fall?

For instance, if a language has two relevant pitches (Low = 2, and High = 4),
can there be a (phonemic or morphemic) difference between a 224 rise and a
244 rise?  Or between a 442 fall and a 422 fall?

Or, if it has three relevant pitches (adding Middle = 3), can there be a
difference between 234 and either 224 or 244? Or between 432 and either 422
or 442?

What if it has more?  Can there be a difference between 135 and either 125 or
145? Can there be difference between 125 and 145?