re: "length"

Most correlate the feature "length" with duration, yes....
however, there are cases where phonemic versus phonetic realizations may
not be identical with regards to duration. The Finnish and Japanese cases
exhibit duration differences between short and long vowels, maintaining
vowel quality (differences are negligible). Other examples might include:
Thai, Mongolian, Kyrgyz/Tyvan, etc.
Hungarian and Cantonese, on the other hand, maintain phonemic
"length" distinctions, e.g. Hungarian - a vs á, e/ë vs é, u vs ú, ö vs o
(o"), ü vs u (u"). However, while there is slight durational difference,
the shift has gone from quantity (duration) to quality (vowel identity),
such that Cantonese /a/ vs /aa/ is really [?] (turned script a; low-mid
central unrounded vowel) vs [A], Hungarian /a/ vs /á/ is [?] (turned a; low
round back vowel) vs [A], etc.

As Peter pointed out, American (Anglophone?) dictionary
assignments of "long" versus "short" have rationale... primarily that of
pronunciation distinctions of identical glyphs, which have their basis in
historical values, cf. "The Great Vowel Shift".

cheers,
-Patrick

>Nicholas Bodley wrote:
> > Shockingly-ignorant question, but this is a very civil
> > place: In this context, does "long" imply relatively-
> > long time duration? I assume so.
> > It seems to me that sometimes the vowel in "pot" is
> > sometimes called "short", and that in "oaf", "long", but
> > I suspect such usage is by the phonetically-untrained
> > (which I am, for the most part). I also have Japanese
> > vowels in mind.

At 10:50 AM 6/18/2004, Peter Daniels wrote:

>There really was a reason, long ago, why "dictionary pronunciations"
>assigned the labels "long" and "short" to the vowels the way they did, but
>those uses for quality-of-vowel aren't used in linguistics; "long" and
>"short" to us really do refer to the durations of segments.
>
>In Finnish (ObQalam) they make it easy for you by actually writing two
>vowels or two consonants when they want you to pronounce them long.

[Non-text portions of this message have been removed]