Do you contrast BARD and BARRED?

[jcb]

In my dialect, BARD and BARRED have different vowels qualities, despite how I would describe them as having the same phonemes: /bArd/. Both have /A/, but BARD has [V] and BARRED has [A]. This is because /A/ has an allophone of [V] when before RC (where R is /r/ or /l/, and C is any consonant except /r/ or /l/), but apparently not when there's a morpheme break.

Summary:
[A] = BAR, BARS, BARRED, CARL
[V] = BART, BARD, BARF, ARSON, BARN

(I just realized, maybe CARL is still [A] because it's 2 syllables /kAr.l=/, despite the spelling?)

[A] = BALL, BALLS, BALLED, BALLER
[V] = BALTIC, BALD, BALSAM, BALM, ALTERNATE, ULTIMATE, ULCER, HULL, HULK
[o] = BOWL, BOLD, HOLE (I don't merge HULL-HOLE. I wonder if some of those people with it also have this [A]/[V] allophoning, and how they interact.)

(1) Is there a symbol to use in morpheme transcriptions to indicate a morpheme break?
(2) Am I wrong to say that BART and BARD have /A/ instead of /V/? Am I letting spelling influence me? Is it wrong for me to be motivated to be able to still say that only tense vowels /i e A o u/ can appear before /r/? (Note that I merge MARY-MARRY-MERRY into /e/.) After all, /l/ has no such limitation on lax vowels appearing before it.
(3) Note that I merge COT-CAUGHT into /A/. I dunno if this is relevant.
(4) Does this or something similar happen in your dialect?

[Travis B.]

I don't have this particular split, and hadn't heard of it before. Rather, I am familiar with a raising of START [ɑ(ː)ʁˤ] to [ʌ(ː)ʁˤ] before fortis plosives, inconsistently before fortis fricatives (e.g. in scarf, hearth and marsh but not in Martha or farce, and only sporadically in parse), and in the specific lexeme target, except if there is an /r/ in the next syllable (e.g. not in Carter or partridge).

[linguistcat]

I have a difference in length, but not quality.

[anteallach]

I hadn't heard of this split before, but it doesn't particularly surprise me; there are several examples of allophonic splits where the normally word-final allophone continues to be used before a suffix.

One reasonably well-known example is in the Scottish Vowel Length Rule, where vowels are normally short before /d/ (unlike other types of vowel length allophony in English) and long word-finally, and the long form is used before the /d/ suffix, so for example rowed would have a longer vowel than road.

(Perhaps /d/ being treated as triggering shorter vowels is somehow what's happening here. If Travis's raised allophone of START before fortis consonants also applied before /d/, but not suffix /d/, then that would give the pattern you report.)

IIRC some south-east English varieties have diphthongisation of the NORTH/THOUGHT/FORCE vowel in word- or morpheme-final position, so you get e.g. sword with [oː] but soared with [oə]. And I also remember reading about a days/daze distinction caused by a word-final allophone of FACE.

[Travis B.]

I hadn't heard of this split before, but it doesn't particularly surprise me; there are several examples of allophonic splits where the normally word-final allophone continues to be used before a suffix.

One reasonably well-known example is in the Scottish Vowel Length Rule, where vowels are normally short before /d/ (unlike other types of vowel length allophony in English) and long word-finally, and the long form is used before the /d/ suffix, so for example rowed would have a longer vowel than road.

(Perhaps /d/ being treated as triggering shorter vowels is somehow what's happening here. If Travis's raised allophone of START before fortis consonants also applied before /d/, but not suffix /d/, then that would give the pattern you report.)

IIRC some south-east English varieties have diphthongisation of the NORTH/THOUGHT/FORCE vowel in word- or morpheme-final position, so you get e.g. sword with [oː] but soared with [oə]. And I also remember reading about a days/daze distinction caused by a word-final allophone of FACE.

My guess is that this is what is happening here. (Note that what I have is not allophony but rather a phonemic split of START, as in many cases the inconsistency of application before fortis fricatives is lexicalized, and the word target in particular stands out as an example of an unconditioned shift.)

[jcb]

I don't have this particular split, and hadn't heard of it before. Rather, I am familiar with a raising of START [ɑ(ː)ʁˤ] to [ʌ(ː)ʁˤ] before fortis plosives, inconsistently before fortis fricatives (e.g. in scarf, hearth and marsh but not in Martha or farce, and only sporadically in parse), and in the specific lexeme target, except if there is an /r/ in the next syllable (e.g. not in Carter or partridge).

I have [V] in all of those words.

Perhaps I just have a more progressed version of the change you're describing.

I have a difference in length, but not quality.

Please describe this more.

I have length differences too, but I skipped them, because they're orthogonal to quality (for me, at least). I guess I'll explain them now:

I have short vowels before unvoiced consonants, and long vowels before voiced consonants (or null). The vowel is still affected even if there's a R in between.

(There's also tone differences that go along with the vowel length. Short vowels have a sharply rising tone, and long vowels have a gently falling tone.)

short = BOT, BUT, FART, FARTS, SCARF
long = BOD(Y), PAW, PAWS, PAUSE, BAR, BARS, BARRED, BUD, BARD, BARDS, SCARVES

Summary:

Code: Select all

┌──────────────────┬─────────────────────────┬─────────────────────────────────────────────┐
│ Quality / Length │          Short          │                    Long                     │
├──────────────────┼─────────────────────────┼─────────────────────────────────────────────┤
│ [A]              │ BOT                     │ BOD(Y), PAW, PAWS, PAUSE, BAR, BARS, BARRED │
│ [V]              │ BUT, FART, FARTS, SCARF │ BUD, BARD, BARDS, SCARVES                   │
└──────────────────┴─────────────────────────┴─────────────────────────────────────────────┘

I hadn't heard of this split before, but it doesn't particularly surprise me; there are several examples of allophonic splits where the normally word-final allophone continues to be used before a suffix.

Yes, the Scottish vowel length rule sounds like it operates in a similar way.

What other examples are you thinking of?

NORTH/THOUGHT/FORCE

This makes me think I should give up on trying to say that BARD and BARRED have the same vowel /A/, because describing NORTH/THOUGHT/FORCE as a set sounds silly to somebody like me, where NORTH and FORCE have /o/, and THOUGHT has /A/.

I wonder if some of those people with it also have this [A]/[V] allophoning, and how they interact.

To describe it explicitly, are there any people that merge BALD into BOLD?
The route: /A/ →(BARD/BARRED split)→ [V] →(HULL/HOLE merger)→ /o/

[Travis B.]

So you have similar vowel length allophony to that I am familiar with here (rather than following the Scottish vowel length rule), and your raising of START does not pattern with vowel length (whereas my raising of START only occurs when it has a short vowel, even though there are cases for me of unraised short START, except in the isolated case of target).

What you have with raised START is definitely interesting, as it seems like you likely raise all cases of START before another consonant within the same morpheme (e.g. in your scarves). This is very distinct from the raising of START I am used to, which is more an extension of Canadian/American raising.

[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.

[Travis B.]

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].

[linguistcat]

I have a difference in length, but not quality.

Please describe this more.

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.

[jcb]

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?

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 have a difference in length, but not quality.

Please describe this more.

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?

[Travis B.]

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.

[jcb]

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.

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

[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.

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.

[Richard W]

(1) Is there a symbol to use in morpheme transcriptions to indicate a morpheme break?

I've seen the hash symbol ('#') used. It's frequently needed in precise phonemic level transcriptions, simply because morpheme boundaries can affect the realisation of allophones, just as syllable boundaries ('$') can.

[jcb]

This is very distinct from the raising of START I am used to, which is more an extension of Canadian/American raising.

At first I thought that this split might be related to the RICE-RISE allophony/split (which I have), but then I realized that the rules were different, being unrelated to length.

(1) Is there a symbol to use in morpheme transcriptions to indicate a morpheme break?

I've seen the hash symbol ('#') used. It's frequently needed in precise phonemic level transcriptions, simply because morpheme boundaries can affect the realisation of allophones, just as syllable boundaries ('$') can.

I thought a syllable boundary was marked with a period: '.', like I did above with CARL: /kar.l=/

[Travis B.]

This is very distinct from the raising of START I am used to, which is more an extension of Canadian/American raising.

At first I thought that this split might be related to the RICE-RISE allophony/split (which I have), but then I realized that the rules were different, being unrelated to length.

American raising is not necessarily linked to vowel quantity; e.g. in many American dialects there is raising in words such as spider and tiger, and in the dialect here these undergo raising even though the vowels are long.

[Travis B.]

(1) Is there a symbol to use in morpheme transcriptions to indicate a morpheme break?

I've seen the hash symbol ('#') used. It's frequently needed in precise phonemic level transcriptions, simply because morpheme boundaries can affect the realisation of allophones, just as syllable boundaries ('$') can.

I thought a syllable boundary was marked with a period: '.', like I did above with CARL: /kar.l=/

What I have seen is '.' to mark syllable boundaries, '$' to mark morpheme boundaries, and '#' to mark word boundaries.

[jcb]

This is very distinct from the raising of START I am used to, which is more an extension of Canadian/American raising.

At first I thought that this split might be related to the RICE-RISE allophony/split (which I have), but then I realized that the rules were different, being unrelated to length.

American raising is not necessarily linked to vowel quantity; e.g. in many American dialects there is raising in words such as spider and tiger, and in the dialect here these undergo raising even though the vowels are long.

I raise the vowel in SPIDER and TIGER too, but there's a rule that explains it: They're both followed by /r/.

I think the vowel in both is short too for me. I didn't know that some people don't also shorten the vowel after first raising it.

[Travis B.]

At first I thought that this split might be related to the RICE-RISE allophony/split (which I have), but then I realized that the rules were different, being unrelated to length.

American raising is not necessarily linked to vowel quantity; e.g. in many American dialects there is raising in words such as spider and tiger, and in the dialect here these undergo raising even though the vowels are long.

I raise the vowel in SPIDER and TIGER too, but there's a rule that explains it: They're both followed by /r/.

I think the vowel in both is short too for me. I didn't know that some people don't also shorten the vowel after first raising it.

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.