> On Thu, Oct 21, 2010 at 5:12 AM, Lars Finsen<[log in to unmask]>  wrote:
>> Den 21. okt. 2010 kl. 09.45 skreiv Philip Newton:
>>> I'm reminded of a complaint by a resident of New York that her niece
>>> was asked to "underline all the short vowels" in a piece of text at
>>> school... which was supposed to include the "o" in "dog". Only in the
>>> niece's native 'lect, that word has a *long* vowel (I suppose that's
>>> where the "dawg" respelling sometimes used in slang comes from).
>>> Fun when instructional materials try to generalise across multiple
>>> different accents of the same language....
>> Interesting. Where in the English-speaking world is that vowel short? It's
>> short in _doggie_, but much longer in _dog_ the way I think I've heard it -
>> comparable to _door_. Even in _dogs_ the vowel is much longer than in _digs_
>> the way I think I've heard it.

Many accents, including England and S Hemisphere have LOT in _dog_, and don't merge LOT with any other phoneme.

Note that the "underline short vowels" task would be phonological not phonetic: it means "underline vowels that can occur only in closed syllables".

Larry Sulky, On 21/10/2010 19:40:
> In my 'lect, which is pretty close to GenAm, the "o" of "dog" and the "o" of
> "doggie" are identical, and are short. They are also the same vowel as the
> "a" in "father".

If the "a" in "father" is the same phoneme as the "a" in "bra" and "spa", then it'd be by definition long (in the phonological sense of the long/short distinction pertinent to the vowel-underling exercise), since it can occur in stressed word-final positions.