On Tue, 19 Jan 2010 10:06:52 +0000, Peter Bleackley
<[log in to unmask]> wrote:

>What binary feature distinguishes [&] from [a]?

It depends on what [&] and [a] are being used for, is how I'd answer that. 
[a]'s often used for a mid low vowel, you know, given that IPA doesn't have
a proper symbol for that; and in this event [&] is just the front
counterpart, and that's the distinction.  This is e.g. Darin Flynn's
handling in
(though he only distinguishes two degrees of frontness).  

On Tue, 19 Jan 2010 13:49:47 +0000, Peter Collier
<[log in to unmask]> wrote:

>& [-cons +son +syll -fr -bk -hi +lo -
>[&] is more lax than [a]. I use:
>& [-cons +son +syll -fr -bk -hi +lo -str]
>a [-cons +son +syll -fr -bk -hi +lo +str]
>where [+/-str] = +/-strident

That does seem like a strange way to do it.  I'm used to the tense-lax
distinction being attributed to [+-ATR] or the like, certainly not

For that matter, when I was reading up on featural phonology recently, I
didn't find the case for the existence of [+-strident] particularly
convincing, especially not its application to different points of
articulation, where one might instead set up a different feature for each. 
Flynn didn't provide any good examples of its spread between different
places or anything.  He did say that it was defined acoustically, but I'm
skeptical of that too (sound change by acoustic similarity and for
articulatory reasons seem separate processes and a priori there's no reason
one should want a single theory to encompass both).  Can anyone defend it?