The vowels in ''father'' and ''bother''?

>>If "spa" is /spA:r/, then is "calm" /kA:rm/? You might even go for /spar/ and /karm/, with [A:] just an allophone of /a/ before /r/, though you'd have to work out why "sari" and "carry" (or "curry", depending on what you think /a/ is) don't rhyme. <<

For the record, I was only referring to long vowels resulting from non-rhotic-ness, and not *all* vowels that just happen to be realized as long.

Bump

<<For the record, I was only referring to long vowels resulting from non-rhotic-ness, and not *all* vowels that just happen to be realized as long.>>

So are you suggesting that "spa" and "spar" should be transcribed differently (phonemically) even if they're pronounced the same in all contexts? That seems artificial to me.

If you're not suggesting that, then could you explain what you are suggesting, and how you would transcribe "spa" and "spar" in a dialect with both intrusive and linking R?

JHJ, note that I am referring specifically to dialects that don't insert [r\] between *all* words ending and starting in vowels, which I should have made clearer earlier. In the case of intrusive [r\], well, it depends on whether it shows up only at word boundaries or also at morpheme boundaries as well, and on whether it delengthens any long vowels and monophthongizes any diphthongs linked to *just* non-rhotic-ness (and not all long vowels and diphthongs that have [@] as an offglide) in places where it shows up.

<<JHJ, note that I am referring specifically to dialects that don't insert [r\] between *all* words ending and starting in vowels, which I should have made clearer earlier.>>

I don't think any dialects have intrusive R after *all* vowels. The usual pattern is that they add [r\] after certain vowels, usually long lax vowels, centring diphthongs and schwa. So "law" (with /O:/), "idea" (with /I@/), "India" (with /@/) and "spa" (with /A:/) get the intrusive R if the next word begins with a vowel, just like "for", "near" "father" and "spar", with the same vowels arising from loss of /r/.

<<In the case of intrusive [r\], well, it depends on whether it shows up only at word boundaries or also at morpheme boundaries as well,>>

It often does show up at morpheme boundaries - e.g. in the middle of "drawing".

<<and on whether it delengthens any long vowels and monophthongizes any diphthongs linked to *just* non-rhotic-ness (and not all long vowels and diphthongs that have [@] as an offglide) in places where it shows up. >>

I don't understand this. I don't think intrusive R ever delengthens any long vowels.

Could you give some examples of your preferred transcriptions, for example of the words I've mentioned above?

JHJ, the matter is that the preferred transcriptions would matter on the exact specifics of the phonology of the dialect in question, specifically with respect to how non-rhotic-ness is handled and how vowel-vowel morpheme and word boundaries are handled in different cases. Therefore, I really cannot generalize to all non-rhotic dialects, as their behavior differs in these respects.

JHJ:
<<I don't think any dialects have intrusive R after *all* vowels. The usual pattern is that they add [r\] after certain vowels, usually long lax vowels, centring diphthongs and schwa. So "law" (with /O:/), "idea" (with /I@/), "India" (with /@/) and "spa" (with /A:/) get the intrusive R if the next word begins with a vowel, just like "for", "near" "father" and "spar", with the same vowels arising from loss of /r/.>>

The underlying rule is that intrusive R appears after non-high vowels (which includes the centering diphthongs, since they end in a non-high vowel). Thus it is used after [A:], [O:], [@], [E@], [I@], and [U@], but not after [i:]/[i], [u:]/[u], [aI], [aU], or [oI].

Travis:
<<and on whether it delengthens any long vowels and monophthongizes any diphthongs linked to *just* non-rhotic-ness (and not all long vowels and diphthongs that have [@] as an offglide) in places where it shows up.>>

Again, I'm not sure what you're saying here. There *are* no long vowels or diphthongs that are linked to *just* non-rhoticness. As I mentioned earlier, I have the centering diphthongs in my speech, and I am fully rhotic. And as JHJ said, the intrusive R doesn't change the quality of the preceding vowel. It would never monophthongize a diphthong or delengthen a long vowel.

Okay, so with this sort of intrusive R in this given phonemic system, one could probably get away with claiming that there really was no /r/ there to start with, by having such be just a phonological rule triggered by said types of vowels. Of course, if there are any non-rhotic dialects with linking R but without intrustive R, this would not apply to them at all, as they would still need the phoneme /r/ to mark where linking R "should" be placed (like how French phonemically still keeps many word-final letters which are elided when words are spoken in isolation, as they show up if the word is followed in turn by another word starting with a vowel).

Are you trying to say that the last syllable of "father" sounds exactly the same in RP as in does in Brooklynese because they are both non-rhotic?
That is insane. Although both accents are non-rhotic, Brooklynese uses a pronunciation which sounds more like "fAtha" while RP uses a pronunciation in which the 'r' is not really pronounced BUT which maintains the integrity of the "er" sound at the end of the word without making sound like "ah". Likewise the second syllable of "pepper" in RP would sound like a non rhotic version of the word "purr" (like what a cat does) while in Brooklynese (also non-rhotic) it would sound more like "peppa" or "peppah". In Brooklynese, the two words mentioned above are pronounced as if they never had an 'r' at the end while the RP pronunciation, while still maintaining non-rhoticity, hints that there is an 'r' at the end. Thus, a semi-literate Anglophone or a foreigner who knows but very little English and who is trying to transcribe the word based on the Brooklynese pronunciation is likely spell "fatha" or "fathah" while the RP pronunciation is more likely to yield the correct spelling.

I guess that what I'm trying to say is that THERE IS MORE THAN JUST ONE PHONETIC REALISATION OF R-DROPPING, especially when the word in question ends in 'r' and comes at the end of a sentence or phrase. It would be so much easier if I could actually pronounce both words in front of you in both accents but I can't being that I'm sitting in front of a computer screen typing this.

Chris, and of course, such implies that in, say, RP, the phoneme /r/ is still there, but is usually just preserved as a degree of [r\]-coloring of a preceding vowel except where another vowel ends up cominga fter it.

Who ever mentioned the insane Brooklynese? :-)

The vowel is realised differently from group to group, even within the RP collective. But typically, in RP, the phoneme /r/ in such positions is NEGLECTED in fleeting speech, so that for example, "battered" and "batted" are homophones, as are "kata" and "carter". However, in careful speech these words are distinguished by vowel, to hint at the written "r".

<<But typically, in RP, the phoneme /r/ in such positions is NEGLECTED in fleeting speech, so that for example, "battered" and "batted" are homophones>>

Well from what I've read, most British people still preserve the [I] in verbal "-ed" endings, so "battered" and "batted" would be at least slightly different, as [b{t@d] and [b{tId].