Re: [tied] Laryngeal theory as an unnatural

From: Miguel Carrasquer
Message: 18888
Date: 2003-02-19

On Sat, 08 Feb 2003 16:55:10 +0100, Miguel Carrasquer <mcv@...>
wrote:

>... /m/ and /n/ are not the same phoneme, and from the point of view
>of historical linguistics, I would claim that /m/ is "more nasal" than
>/n/ (I'm not sure how that translates into phonetical terms).

I was toying around with the "Speech Tools" from SIL (www.sil.org),
which come with a nice little Speech Analyzer, and sound files for all
IPA symbols ("IPA Help").

Thinking about what I said above, I ran the files for intervocalic /m/
(Cons13b.wav /ama/) and /n/ (Cons15b.wav /ana/) through the analyzer
tool. The difference between the two is the following: the first
/a/'s are much alike (F1 at ca. 700, F2 at slightly above 1200 and F3
at 2600 Hz). In the case of /ama/, there is an abrupt transition of
F1 from the /a/ value to the nasal frequency (ca. 250 Hz), while F2
drops smoothly to the /m/'s (weak) F2 at ca. 1000 Hz. The second /a/
starts almost immediately with F1 back at ca. 650, but the F2
transition is again gradual, even more so, from 1000 back to 1200 Hz.
For /ana/, the situation is reversed: the F1 transition is smooth
(falling from 700 to 250 and rising back to 700), while F2 jumps
suddenly from 1200 to ca. 2000 Hz (the /n/'s F2) and back again to
1200.


2000 - - -

1200 F2 --- -- ---- ----
- --
1000 - - -

700 F1 ---- ---- ---- ----
- -
250 ------ ----
/ a m a / / a n a /


One of the acoustic properties of nasalized vowels is a lowered F2 as
compared to the oral vowel, which means that the transition from/to
the vowel's second formant to/from /m/'s second formant at 1000Hz can
more easily develop into full nasalization than in the case of /n/
which lacks a low F2 formant, having it at ca. 2000 Hz. This also
explains the tendency of /mi/ and /im/ to be heard as (or become) /n/
(/i/ has an F2 at slightly above 2000 Hz), while /un/ and /nu/ tend
towards /m/ (/u/ has an F2 slightly under a 1000 Hz).

The raising effect caused by nasals in general (e > i, o > u, a > &)
is caused by the low F1 which /m/ and /n/ share (together with the
other nasals).


If you download the SIL Speech Tools, don't forget to listen to the
phrase "you will order the thick new clothes" in Usila Chinantec
(which is transcribed as /a ?m.? ?m.? ?m. m.? m./).


=======================
Miguel Carrasquer Vidal
mcv@...