> From: Peter Bleackley <[log in to unmask]>
> What binary feature distinguishes [&] from [a]?

--- On Tue, 1/19/10, 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

Non-concur. Strident is usually used only as a feature of fricative consonants.

Phonetically, I think it's a question of [&] having "retracted tongue root" vs. "advanced" for [a].

IMO differentiation between [&]/[a] (or better /&/ /a/) in distinctive feature terms depends on the language-- since this is a matter of phonology (and the contrasts involved in a given language), not phonetics .

In the analysis of most languages, it makes most sense to consider /a/ either a central vowel [-Cons +Voc -hi -fr -ba +lo] (esp. if there is another higher central V like /@/) or else back unrounded [...-hi -fr +ba +lo -rnd] (if there is no other central V). I've seen Engl. /a/ described both ways. An advantage to calling /a/ central [-fr -ba] is that you don't then have to specify [rnd] 

/&/ OTOH can almost always be described as [...-hi +fr -ba +lo]. That works for English.

In the unlikely event that /&/ and /a/ contrast _as front vowels_ in some language (and I'd disagree...), then I think you'd have to introduce some other feature, perhaps /&/ [-tense or +lax/open] vs. /a/ [+tense or -lax]. Note tthat this doesn't work for Engl. since /&/ and /a/ are both -tns, in that they can't occur in an open CV syllable. (OK, /a/ marginally in "ma, pa, la-di-da" but that's affective vocab.)

Take French (please :-) )-- /a/ can be called a front V, , thus the front vowels are:
i [+fr +hi -rnd]
y [+fr +hi +rnd]
e [+fr -hi -lo +tns -rnd]
2/ [+fr -hi -lo +tns +rnd] 
E [+fr -hi -lo -tns -rnd? +/-nas] 
9/ [+fr -hi -lo -tns +rnd?] (I'm not sure the E/9 contrast is really phonemic--conditioned by open/closed syllables?)
a [+fr -hi +lo +tns +/-nas]

Note that for many speakers, nasalized /E~/ is realized as [&~], but that's purely a performance issue. At least prescriptively, presse and prince are supposed to contrast only in nasalization [pREs] vs. [pRE~s]. (It may be, too, that /i/ and /e/ need to include [+/- nas] but I'm not sure, since they both lower to [E~] and /a~/ resp.)

(Strikes me that this makes for a very lop-sided system, since there are far fewer back vowels /u o O? O~/ (and /A/ in older analyses). (I'll expect Christophe to weigh in on this...:-)))))

Whereas in German, it makes more sense to call /a/ [+ba] because of the umlaut rule, which fronts back vowels.