Do you contrast BARD and BARRED?

[anteallach]

Speaking of bald and bold, this is actually a genuine example in my speech. The two words are either merged or very nearly so, but bawled is definitely different and retains a normal THOUGHT vowel.

What phoneme do you merge BALD-BOLD into? Do you merge COT-CAUGHT?

My phonemic intuition thinks bald and bold are GOAT, though I’m not entirely sure there isn’t a subtle difference. No I don’t merge cot/caught; that’s almost unheard of in England.

Edit: I've now looked at some spectrograms, and I'm pretty confident that bald and bold really are merged. The realisation is something like [öu]: a diphthong whose starting point is close to (perhaps slightly back of) my usual more-or-less monophthongal GOAT vowel and then glides upwards and backwards into the dark /l/.

[anteallach]

I am familiar with the gulf-golf merger, where certain cases of historical /ʌl/ before a fortis obstruent (e.g. multi-, ultra-, ultimate, mulch, insult, bulk, etc.) merge with /ɔːl/ (RP /ɒl/) as [ɒo̯] myself, which affects the dialect here outside of spelling pronunciation. Conversely, bald and bold never merge here, while bald and bawled are homophones with /ɔːl/ [ɒːo̯], as bold has /oʊl/ [o̞ːʊ̯].

I merge GULF-GOLF too, but into /V/. Do you have any words with /Vl/? What vowel do you have in DULL?

I preserve /ʌl/ when not followed by a fortis obstruent; e.g. I have /ʌl/, realized as [ʌ(ː)ɤ̯], in dull.

What about dulled? (Or culled, mulled, etc.?)

[Travis B.]

I merge GULF-GOLF too, but into /V/. Do you have any words with /Vl/? What vowel do you have in DULL?

I preserve /ʌl/ when not followed by a fortis obstruent; e.g. I have /ʌl/, realized as [ʌ(ː)ɤ̯], in dull.

What about dulled? (Or culled, mulled, etc.?)

These all have /ʌl/ as [ʌːɤ̯] for me.

[linguistcat]

Well first off, my dialect (idiolect?) is rhotic. I have the START vowel in both for quality but bard is short and barred is long. Otherwise, at least as far as I can tell, they are pronounced the same. I''d try to record it but I'm pretty sure I'd be too conscious of it now and over emphasize the difference, which wouldn't make for good data.

What vowel do you have in BAR?

As far as I can tell, it's the same as START but short, so like bard in this situation.

[jcb]

My phonemic intuition thinks bald and bold are GOAT, though I’m not entirely sure there isn’t a subtle difference. No I don’t merge cot/caught; that’s almost unheard of in England.

Wow, that means that depending on the accent, BALD could have any of /A V O o/. No wonder why some English learners just give up on learning which vowel goes with which word or spelling and just randomly pick one.

/o/ = British
/O/ = General American
/A/ = GA + CAUGHT-COT merger
/V/ = GA + CAUGHT-COT merger + BARD-BARRED split
Edit: And back to /o/ again:
/o/ = GA + CAUGHT-COT merger + BARD-BARRED split + HULL-HOLE merger

[TomHChappell]

In my dialect, ….
(4) Does this or something similar happen in your dialect?

No; not in my locolect.
There is no contrast between BARD and BARRED around here (Greater Detroit-Windsor Metropolitan Area).

Or, at least, I don’t remember ever hearing one.
I could be wrong.

[Travis B.]

I should also note that for me bulb has THOUGHT.

[jcb]

I preserve /ʌl/ when not followed by a fortis obstruent; e.g. I have /ʌl/, realized as [ʌ(ː)ɤ̯], in dull.

So, you have /V/ in VULGAR, and /O/ in VULTURE?

I have /ʌl/ in vulgar while both /ʌl/ and /ɔːl/ are acceptable to me in vulture.

What about in VULNERABLE?

I used to have /o/ in it, but I've now (hyper-)corrected myself to have /V/ instead.

Aside from other cases of historical C/aɪ/D/ər/ where D is a lenis plosive, the dialect here also has raising in tire, Ida, idle, idol, Idaho, and Midol, for instance, and in all of these words the raised vowel is long.

For me:
TIRE: [V]. FIRE: [V]. SIRE: [V]. DIRE: [A]. (DIRE sounds strange with [V] to me. Maybe because it's a rarer word and the shift isn't complete yet?) BUYER: [A]. (The morpheme boundary protects it.)
IDA: Not sure. This word is rare.
IDLE: [V]
IDOL: [V] (IDOL is identical to IDLE.)
IDAHO: [V] or [A]
MIDOL: Not sure. Never heard of this word.

Maybe it's another following syllable in general that causes [Aj] to raise to [Vj]?

I should also note that for me bulb has THOUGHT.

So, do people with the GULF-GOLF merger and the CAUGHT-COT merger, but not the BARD-BARRED split have /A/ in words like BULB?

[Travis B.]

So, you have /V/ in VULGAR, and /O/ in VULTURE?

I have /ʌl/ in vulgar while both /ʌl/ and /ɔːl/ are acceptable to me in vulture.

What about in VULNERABLE?

I used to have /o/ in it, but I've now (hyper-)corrected myself to have /V/ instead.

I have /ʌl/ in vulnerable myself.

Aside from other cases of historical C/aɪ/D/ər/ where D is a lenis plosive, the dialect here also has raising in tire, Ida, idle, idol, Idaho, and Midol, for instance, and in all of these words the raised vowel is long.

For me:
TIRE: [V]. FIRE: [V]. SIRE: [V]. DIRE: [A]. (DIRE sounds strange with [V] to me. Maybe because it's a rarer word and the shift isn't complete yet?) BUYER: [A]. (The morpheme boundary protects it.)
IDA: Not sure. This word is rare.
IDLE: [V]
IDOL: [V] (IDOL is identical to IDLE.)
IDAHO: [V] or [A]
MIDOL: Not sure. Never heard of this word.

Maybe it's another following syllable in general that causes [Aj] to raise to [Vj]?

If it is another following syllable that does it, then it has to be in monomorphemic cases, as of course rider does not have raising.

One thing to note is that I don't have raising in titanium or dichotomy, for instance; however it seems to be the metrical structure of these words that is blocking raising.

I should also note that for me bulb has THOUGHT.

So, do people with the GULF-GOLF merger and the CAUGHT-COT merger, but not the BARD-BARRED split have /A/ in words like BULB?

I honestly don't know how widespread /ɔːl/ in bulb is ─ from a cursory look on online dictionaries they only give /ʌl/ for this word.

[jcb]

If it is another following syllable that does it, then it has to be in monomorphemic cases, as of course rider does not have raising

My RIDER lacks raising, but WRITER has it. What do you have for WRITER?

I just realized something: Maybe the raising of /A/ in words like BARD is related to the raising of /{/ in some words like ALCOHOL and CALCULATE that I have? Both are low vowels before /l/, after all.

Examples:
[{] (unraised) = ALBERT, PAL, BALLAST, MALLOW, TALLOW, TALLY, DALLY, CALIFORNIA, GAL, RALLY
[E] (raised) = ALCOHOL, ALFALFA (both), ALPS, CALCULATE, CALCIUM, MARSHMALLOW

It's hard to tell, because there's few instances of /{l/ in English to begin with.

[Travis B.]

If it is another following syllable that does it, then it has to be in monomorphemic cases, as of course rider does not have raising

My RIDER lacks raising, but WRITER has it. What do you have for WRITER?

I just realized something: Maybe the raising of /A/ in words like BARD is related to the raising of /{/ in some words like ALCOHOL and CALCULATE that I have? Both are low vowels before /l/, after all.

Examples:
[{] (unraised) = ALBERT, PAL, BALLAST, MALLOW, TALLOW, TALLY, DALLY, CALIFORNIA, GAL, RALLY
[E] (raised) = ALCOHOL, ALFALFA (both), ALPS, CALCULATE, CALCIUM, MARSHMALLOW

It's hard to tell, because there's few instances of /{l/ in English to begin with.

For me /æl/ is [ɛ(ː)ɤ̯], which I have in all these words except for marshmallow, which has shifted to /ɛl/, for which I have [ɜ(ː)ɤ̯] with a notably centralized vowel (which is more central than my normal [ɜ(ː)] for /ɛ/).

[jcb]

So you have similar vowel length allophony to that I am familiar with here

BTW, do you know how common this vowel length/tone rule is in America? Where is it regionally located?

[Lērisama]

So you have similar vowel length allophony to that I am familiar with here

BTW, do you know how common this vowel length/tone rule is in America? Where is it regionally located?

It's present in SSBE – although it's more that vowels before fortis consonants are shortened. In combination with the historical vowel length¹, there is now a phonetic 4-way length distinction²

¹ Supplimented by lengthening of vowels before historic coda /ɹ/
² No mininal pairs I can thnnk of though. The closest I can find are 3-way ones like [b̥ɛ̆ʔ b̥ɛt̚ b̥ɛːt̚] ⟨bet bed beared⟩

[Travis B.]

So you have similar vowel length allophony to that I am familiar with here

BTW, do you know how common this vowel length/tone rule is in America? Where is it regionally located?

Vowel length allophony is found throughout English, even though its details vary from variety to variety.

For instance, in the dialect I speaks it operates under the rule that all vowels are long unless there is an obstruent phoneme before the next vowel, and the next obstruent phoneme is fortis, or a vowel is in an utterance-final syllable and it is not followed by a lenis obstruent in its coda; this rule is applied across word boundaries and before vowel lengthening phenomena due to elision and assimilation. Note that unlike in some NAE dialects neutralization of fortis-lenis pairs due to flapping, assimilation, or elision does not actually neutralize vowel length in the dialect here, so latter takes a short vowel while ladder takes a long vowel here.

SSBE also has vowel length allophony, but it operates differently from that here. Rather, all vowels have an inherent vowel length but are shortened ('clipped') before fortis obstruents in syllable codas.

[jcb]

So you have similar vowel length allophony to that I am familiar with here

BTW, do you know how common this vowel length/tone rule is in America? Where is it regionally located?

It's present in SSBE – although it's more that vowels before fortis consonants are shortened. In combination with the historical vowel length¹, there is now a phonetic 4-way length distinction²

¹ Supplimented by lengthening of vowels before historic coda /ɹ/
² No mininal pairs I can thnnk of though. The closest I can find are 3-way ones like [b̥ɛ̆ʔ b̥ɛt̚ b̥ɛːt̚] ⟨bet bed beared⟩

What's the fourth vowel in this 4-way distinction?

[Lērisama]

BTW, do you know how common this vowel length/tone rule is in America? Where is it regionally located?

It's present in SSBE – although it's more that vowels before fortis consonants are shortened. In combination with the historical vowel length¹, there is now a phonetic 4-way length distinction²

¹ Supplimented by lengthening of vowels before historic coda /ɹ/
² No mininal pairs I can thnnk of though. The closest I can find are 3-way ones like [b̥ɛ̆ʔ b̥ɛt̚ b̥ɛːt̚] ⟨bet bed beared⟩

What's the fourth vowel in this 4-way distinction?

• Clipped¹ short vowel. I notate it as [V̆]
• Un-clipped² short vowel. I notate it as [V]
• Clipped long vowel. I notate it as [Vˑ]
• Un-clipped long vowel. I notate it as [Vː]

Edit: given jcb's surprise in the other thread, I should probably briefly explain the phonemic length system of SSBE. Short vowels are TRAP, STRUT, DRESS, LOT/CLOTH, KIT, FOOT & the schwa. Their lengthened versions³⁴ are PALM/BATH/START, SQUARE, THOUGHT/NORTH/FORCE, NEAR, CURE⁵ & NURSE. Everything else is a diphthong of some kind

¹ I.e. before a fortis coda
² I.e. not before a fortis coda
³ Excepting TRAP, which doesn't have one (TRAP+/ɹ/ backed to merge with PALM/BATH)
⁴ LOT/CLOTH-THOUGHT/NORTH/FORCE, STRUT-PALM/START & COMMA/LETTER-NURSE have significant quality differences; the rest are very similar
⁵ Leaving aside the fact that CURE has mostly merged into THOUGHT/NORTH/FORCE now

[jcb]

It's present in SSBE – although it's more that vowels before fortis consonants are shortened. In combination with the historical vowel length¹, there is now a phonetic 4-way length distinction²

¹ Supplimented by lengthening of vowels before historic coda /ɹ/
² No mininal pairs I can thnnk of though. The closest I can find are 3-way ones like [b̥ɛ̆ʔ b̥ɛt̚ b̥ɛːt̚] ⟨bet bed beared⟩

What's the fourth vowel in this 4-way distinction?

• Clipped¹ short vowel. I notate it as [V̆]
• Un-clipped² short vowel. I notate it as [V]
• Clipped long vowel. I notate it as [Vˑ]
• Un-clipped long vowel. I notate it as [Vː]

Edit: given jcb's surprise in the other thread, I should probably briefly explain the phonemic length system of SSBE. Short vowels are TRAP, STRUT, DRESS, LOT/CLOTH, KIT, FOOT & the schwa. Their lengthened versions³⁴ are PALM/BATH/START, SQUARE, THOUGHT/NORTH/FORCE, NEAR, CURE⁵ & NURSE. Everything else is a diphthong of some kind

¹ I.e. before a fortis coda
² I.e. not before a fortis coda
³ Excepting TRAP, which doesn't have one (TRAP+/ɹ/ backed to merge with PALM/BATH)
⁴ LOT/CLOTH-THOUGHT/NORTH/FORCE, STRUT-PALM/START & COMMA/LETTER-NURSE have significant quality differences; the rest are very similar
⁵ Leaving aside the fact that CURE has mostly merged into THOUGHT/NORTH/FORCE now

(1) Interesting, I never knew that you could pair the vowels like that. (Of course, doing so makes no sense in my dialect, so i would never come up with this system myself.)
(2) So CURE rhymes with PORE???
(3) Are L2 English learners taught this system of organizing English vowels?

[Lērisama]

What's the fourth vowel in this 4-way distinction?

• Clipped¹ short vowel. I notate it as [V̆]
• Un-clipped² short vowel. I notate it as [V]
• Clipped long vowel. I notate it as [Vˑ]
• Un-clipped long vowel. I notate it as [Vː]

Edit: given jcb's surprise in the other thread, I should probably briefly explain the phonemic length system of SSBE. Short vowels are TRAP, STRUT, DRESS, LOT/CLOTH, KIT, FOOT & the schwa. Their lengthened versions³⁴ are PALM/BATH/START, SQUARE, THOUGHT/NORTH/FORCE, NEAR, CURE⁵ & NURSE. Everything else is a diphthong of some kind

¹ I.e. before a fortis coda
² I.e. not before a fortis coda
³ Excepting TRAP, which doesn't have one (TRAP+/ɹ/ backed to merge with PALM/BATH)
⁴ LOT/CLOTH-THOUGHT/NORTH/FORCE, STRUT-PALM/START & COMMA/LETTER-NURSE have significant quality differences; the rest are very similar
⁵ Leaving aside the fact that CURE has mostly merged into THOUGHT/NORTH/FORCE now

(1) Interesting, I never knew that you could pair the vowels like that. (Of course, doing so makes no sense in my dialect, so i would never come up with this system myself.)
(2) So CURE rhymes with PORE???
(3) Are L2 English learners taught this system of organizing English vowels?

1. Why yes it is. As a bonus fact, this length may¹ be best analysed as coda /ɹ/
2. Not for me personally – I preserve all CURE following a yod, but for less conservative speakers, yes²
3. No, that would be much too simple. They have to make do with the analysis for 1960s RP, which practically noöne speaks and is particularly unsuitable for it

¹ I'm undecided
² This does sound very odd to me as well

[Travis B.]

• Clipped¹ short vowel. I notate it as [V̆]
• Un-clipped² short vowel. I notate it as [V]
• Clipped long vowel. I notate it as [Vˑ]
• Un-clipped long vowel. I notate it as [Vː]

Edit: given jcb's surprise in the other thread, I should probably briefly explain the phonemic length system of SSBE. Short vowels are TRAP, STRUT, DRESS, LOT/CLOTH, KIT, FOOT & the schwa. Their lengthened versions³⁴ are PALM/BATH/START, SQUARE, THOUGHT/NORTH/FORCE, NEAR, CURE⁵ & NURSE. Everything else is a diphthong of some kind

¹ I.e. before a fortis coda
² I.e. not before a fortis coda
³ Excepting TRAP, which doesn't have one (TRAP+/ɹ/ backed to merge with PALM/BATH)
⁴ LOT/CLOTH-THOUGHT/NORTH/FORCE, STRUT-PALM/START & COMMA/LETTER-NURSE have significant quality differences; the rest are very similar
⁵ Leaving aside the fact that CURE has mostly merged into THOUGHT/NORTH/FORCE now

(1) Interesting, I never knew that you could pair the vowels like that. (Of course, doing so makes no sense in my dialect, so i would never come up with this system myself.)
(2) So CURE rhymes with PORE???
(3) Are L2 English learners taught this system of organizing English vowels?

1. Why yes it is. As a bonus fact, this length may¹ be best analysed as coda /ɹ/
2. Not for me personally – I preserve all CURE following a yod, but for less conservative speakers, yes²
3. No, that would be much too simple. They have to make do with the analysis for 1960s RP, which practically noöne speaks and is particularly unsuitable for it

¹ I'm undecided
² This does sound very odd to me as well

The problem with analyzing length as coda /r/ is COMMA/LETTER, because COMMA/LETTER is short even though it typically receives intrusive-r in varieties which have intrusive-r (e.g. most EngE varieties). Consequently, you end up with NURSE having coda /r/ and COMMA/LETTER not having coda /r/ despite the fact that both receive intrusive-r.

[anteallach]

(1) Interesting, I never knew that you could pair the vowels like that. (Of course, doing so makes no sense in my dialect, so i would never come up with this system myself.)
(2) So CURE rhymes with PORE???
(3) Are L2 English learners taught this system of organizing English vowels?

1. Why yes it is. As a bonus fact, this length may¹ be best analysed as coda /ɹ/
2. Not for me personally – I preserve all CURE following a yod, but for less conservative speakers, yes²
3. No, that would be much too simple. They have to make do with the analysis for 1960s RP, which practically noöne speaks and is particularly unsuitable for it

¹ I'm undecided
² This does sound very odd to me as well

The problem with analyzing length as coda /r/ is COMMA/LETTER, because COMMA/LETTER is short even though it typically receives intrusive-r in varieties which have intrusive-r (e.g. most EngE varieties). Consequently, you end up with NURSE having coda /r/ and COMMA/LETTER not having coda /r/ despite the fact that both receive intrusive-r.

There's also a question about how you analyse the contrast between the short and long vowels before intervocalic /r/, e.g. merry vs. Mary, if the long vowel found in Mary is underlyingly /ɛr/.