>>>> Is there also a /@:/ phoneme, or do "bird" and "bud" sound the same
>>>> to Americans?

>>>  In my idiolect of American English, "bird" and "bud" are two very
>>>  distinct words.  "R" is alive and well for many (not quite all)
>>>  Yankees.

>> Ah, so these would be /b@Rd/ and /b@d/? Or would "bird" be /b@:Rd/? 

> Well, I'd vote for /bRd/ and /b@d/,
> but as you can see there is no consensus.

>> I imagine that distinctions between long and short vowels in
>> monosyllabic words are similar in American and British English.

> Yes and no.  bird feels longer than bud to me, but in general length
> is less important in AE than in BE, and to my ear, there's a lot of
> variation in length from variety to variety here.  One of my
> coworkers, who was brought up all over the country, often pronounces
> stressed "short" (=lax) vowels much longer than I ever would.

The question of central vowel phonemes in English has a long history in linguistic literature. The Trager-Smith phoneme system from the early 50's recognized and provided useful symbols for a host of possible central vowels.

The problem is that the central vowel space in the mouth is not chopped up with anything close to the precision of the English front and back vowels: in general, front vowels, for instance, distinguish

• tense vowels /i, e/ from lax vowels /I, E/; and
• high vowels /i, I/ from mid vowels /e, E/ from low vowel /ae/;

But there are usually just two central vowels: /a/ and /@/. And /a/ is often classed as a back vowel, to parallel /ae/ and make front and back symmetric. So where's all this variation coming from?

Everywhere. English has this habit of reducing almost every unstressed vowel to somewhere in the shwa region. Sometimes this sounds a little like the vowel that's supposed to be there (most likely with /i/ and /u/), and sometimes -- especially when we're speaking rapidly, (i.e, naturally) -- it doesn't, as is almost always the case with /a/ and with the lax vowels, which mostly come out as normal shwa /@/.

All this is unstressed, OK? Now, what about stressed? That gets more complicated. When stressed, all English vowels are lengthened, and vowel length is also used to signal other things in the language, like the voicing value of a following consonant (which is how we really distinguish ad from at), or whether there's supposed to be a consonant there (the way fur is pronounced in RP, for instance).

And in rhotic American dialects there's a retroflexed mid-central phoneme that's a semivowel, like /w/ is the semivowel of /u/, and /y/ is the semivowel of /i/. This retroflexed (tongue turned backwards) mid-central semivowel is often written as a shwa with an r-like hook to its top right; let me represent this symbol here as <@r>. Then there are two central vowel phonemes in Rhotic American English (at least -- I'm describing my own dialect here, and your mileage may vary): /@/ and /<@r>/

And they are distributed very widely in this vast, airy central vowel space because they don't contrast with one another in space -- they contrast in whether the tongue is curled. In the front or back of the mouth, you have to be very precise in pronouncing stressed vowels, but in the middle, you can roam through enormous phonological space for which the IPA has at least a dozen symbols and you'll still be understood, because there's nothing to contrast with.

Systematically, that is. Individually, personal contrasts abound, for individual reasons. It's all part of the massive and inclusive art.

Historically, this is one way that sound changes happen. If any good contrasts catch on, the space begins to be fragmented, and after some threshhold is crossed, we begin to revise the system to take it into account. But this takes a long time. Unconscious generations of it.