Mon, 21 Apr 2008 10:22:56 -0700: Evan Kirshenbaum
: in sci.lang:
>It's not just that they're stored differently.  They're pronounced
>differently.  In American English, most (in cases like this, all) of
>the distinction between voiced and voiceless stops is carried on the
>preceding vowel, which is held significantly longer before
>(phonemically) voiced stops.  So "latter" (/l&tR/) is pronounced
>[l&*R], while "ladder" (/l&dR/) is pronounced [l&:*R].  The same vowel
>length distinction is used to distinguish, e.g., "cap" and "cab", when
>they come at the end of a sentence and the final stop is unreleased.
>
>If an American speaker listens to a tape of someone saying "latter",
>and segments from the middle of the /&/ are duplicated and inserted,
>at a certain point they'll start hearing "ladder".  Similarly if you
>take "ladder" and chop out segments, it will change to "latter".

Agreed, seems quite plausible.

>British speakers don't pay attention to this vowel length distinction
>and therefore get confused when hearing Americans.  Similarly, British
>speakers don't make the vowel length distinction 

I think they do. I hear them do it, anyway. I don't if they're paying
attention to it, though.

>and American speakers
>tend not to pay attention to the actual voicing (for intervocalic /t/
>and /d/) and so get confused when hearing British speakers.

Probably, yes.

-- 
Ruud Harmsen  
http://rudhar.com