Conlang Random Thread

[Richard W]

Part of this is that I can understand SSBE and GenAus perfectly fine, and the idea that they really are radically different phonologically from the English I am familiar with here seems... wrong.

Don't these varieties generally prohibit stressed short vowels from appearing without a coda? (A related, syllabification-agnostic question: don't they generally prohibit word-final stressed short vowels?) Doesn't contrastive vowel length in these varieties come entirely from the elision of a coda consonant?

What's contrastive vowel length? I get the impression that final rhotic codas have vocalised and then assimilated into the previous final vowels, so I'm not sure that that is 'elision'.

While candidate vowel length differences are generally reinforced by qualitative differences in SSBE, the quality difference between SQUARE and DRESS can be slight. Also, the non-standard monophthongal BYRE, if it still exists, differs little in quality from TRAP, but here we see the simplification of a triphthong, as with the more resilient merger of non-standard POWER and START.

Stressed short vowels in SSBE require a coda or ambisyllabicity of the following consonant. However, widespread non-standard British English can drop a normally mandatory glottal stop (/t/).

[Lērisama]

Part of this is that I can understand SSBE and GenAus perfectly fine, and the idea that they really are radically different phonologically from the English I am familiar with here seems... wrong.

Don't these varieties generally prohibit stressed short vowels from appearing without a coda? (A related, syllabification-agnostic question: don't they generally prohibit word-final stressed short vowels?) Doesn't contrastive vowel length in these varieties come entirely from the elision of a coda consonant?

My big quibble is with treating things like PRICE, GOOSE, FACE, GOAT, GOAL, PRICE, MOUTH, and CHOICE as vowel-consonant sequences rather than diphthongs -- the analyses of SSBE and GenAus get much more sane when they are considered as diphthongs (PALM/START, SQUARE, THOUGHT/NORTH/FORCE, NEAR, and CURE can be treated as rhotic diphthongs) because then you avoid distribution problems with what vowel and what semivowel can go where (and because they then look far less radically different from other English varieties they are fully crossintelliglble with).

Part of this is that I can understand SSBE and GenAus perfectly fine, and the idea that they really are radically different phonologically from the English I am familiar with here seems... wrong.

Don't these varieties generally prohibit stressed short vowels from appearing without a coda? (A related, syllabification-agnostic question: don't they generally prohibit word-final stressed short vowels?) Doesn't contrastive vowel length in these varieties come entirely from the elision of a coda consonant?

Yes, they do, with the exception of some interjections. The question is whether you think it is more elegant to say that all stressed syllables end in a consonant, and attract a following consonant into their syllable, and that long vowels and diphthongs are VC, or that it is only short stressed vowels that to which those apply, so an analysis as diphthongs and long vowels. I lean towards a VC analysis because it make hiatus¹ much easier to explain and it seems odd to only apply obligate codae to short vowels in a language that otherwise cares very little about syllable weight.

I also maintain that since phonemicisation is internal to each speaker², there is no problem with different mututally intelligable dialects having different phonemic systems, so long as the speakers can understand all dialects using their own system. There is nothing that requires that rules that produce similar output be similar. It may be the simplist option, but id there is good reason to think otherwise, that doesn't matter. If you want a cross-dialectial pseudo-phonemicisation, that is what lexical sets are for.

¹ Although not [VrV], which means that is is actually the long vowels from historical Vr for which there are better arguments for being phonemic
² Or should that be listener

[Travis B.]

Don't these varieties generally prohibit stressed short vowels from appearing without a coda? (A related, syllabification-agnostic question: don't they generally prohibit word-final stressed short vowels?) Doesn't contrastive vowel length in these varieties come entirely from the elision of a coda consonant?

My big quibble is with treating things like PRICE, GOOSE, FACE, GOAT, GOAL, PRICE, MOUTH, and CHOICE as vowel-consonant sequences rather than diphthongs -- the analyses of SSBE and GenAus get much more sane when they are considered as diphthongs (PALM/START, SQUARE, THOUGHT/NORTH/FORCE, NEAR, and CURE can be treated as rhotic diphthongs) because then you avoid distribution problems with what vowel and what semivowel can go where (and because they then look far less radically different from other English varieties they are fully crossintelliglble with).

Part of this is that I can understand SSBE and GenAus perfectly fine, and the idea that they really are radically different phonologically from the English I am familiar with here seems... wrong.

Don't these varieties generally prohibit stressed short vowels from appearing without a coda? (A related, syllabification-agnostic question: don't they generally prohibit word-final stressed short vowels?) Doesn't contrastive vowel length in these varieties come entirely from the elision of a coda consonant?

Yes, they do, with the exception of some interjections. The question is whether you think it is more elegant to say that all stressed syllables end in a consonant, and attract a following consonant into their syllable, and that long vowels and diphthongs are VC, or that it is only short stressed vowels that to which those apply, so an analysis as diphthongs and long vowels. I lean towards a VC analysis because it make hiatus¹ much easier to explain and it seems odd to only apply obligate codae to short vowels in a language that otherwise cares very little about syllable weight.

I also maintain that since phonemicisation is internal to each speaker², there is no problem with different mututally intelligable dialects having different phonemic systems, so long as the speakers can understand all dialects using their own system. There is nothing that requires that rules that produce similar output be similar. It may be the simplist option, but id there is good reason to think otherwise, that doesn't matter. If you want a cross-dialectial pseudo-phonemicisation, that is what lexical sets are for.

¹ Although not [VrV], which means that is is actually the long vowels from historical Vr for which there are better arguments for being phonemic
² Or should that be listener

Wouldn't a simpler way to handle this would be to introduce a rule whereby all stressed syllables are required to be heavy or superheavy (excepting interjections)?

[bradrn]

Doesn't contrastive vowel length in these varieties come entirely from the elision of a coda consonant?

Counterexample: the BAD/LAD split in Australian English.

[Travis B.]

Doesn't contrastive vowel length in these varieties come entirely from the elision of a coda consonant?

Counterexample: the BAD/LAD split in Australian English.

Also, what about SQUARE versus DRESS versus TRAP before /r/?

[bradrn]

Doesn't contrastive vowel length in these varieties come entirely from the elision of a coda consonant?

Counterexample: the BAD/LAD split in Australian English.

Also, what about SQUARE versus DRESS versus TRAP before /r/?

TRAP is a different quality from the other two, and I think SQUARE comes under the heading of ‘elision of a coda consonant’.

[Travis B.]

Counterexample: the BAD/LAD split in Australian English.

Also, what about SQUARE versus DRESS versus TRAP before /r/?

TRAP is a different quality from the other two, and I think SQUARE comes under the heading of ‘elision of a coda consonant’.

Ignore my mention of TRAP here, but then would you analyze Mary and merry as something like /ˈmɛrrɪj/ versus /ˈmɛrɪj/? Needing a phonemic geminate /r/ to explain this particular case does not smell right to me.

[bradrn]

Also, what about SQUARE versus DRESS versus TRAP before /r/?

TRAP is a different quality from the other two, and I think SQUARE comes under the heading of ‘elision of a coda consonant’.

Ignore my mention of TRAP here, but then would you analyze Mary and merry as something like /ˈmɛrrɪj/ versus /ˈmɛrɪj/? Needing a phonemic geminate /r/ to explain this particular case does not smell right to me.

Oh, I see what you meant… an underlying coda consonant, not a historical one. This is indeed why I don’t use this analysis.

[Nortaneous]

Part of this is that I can understand SSBE and GenAus perfectly fine, and the idea that they really are radically different phonologically from the English I am familiar with here seems... wrong.

Don't these varieties generally prohibit stressed short vowels from appearing without a coda? (A related, syllabification-agnostic question: don't they generally prohibit word-final stressed short vowels?) Doesn't contrastive vowel length in these varieties come entirely from the elision of a coda consonant?

My big quibble is with treating things like PRICE, GOOSE, FACE, GOAT, GOAL, PRICE, MOUTH, and CHOICE as vowel-consonant sequences rather than diphthongs -- the analyses of SSBE and GenAus get much more sane when they are considered as diphthongs (PALM/START, SQUARE, THOUGHT/NORTH/FORCE, NEAR, and CURE can be treated as rhotic diphthongs) because then you avoid distribution problems with what vowel and what semivowel can go where (and because they then look far less radically different from other English varieties they are fully crossintelliglble with).

In AmEng a diphthongal analysis makes schwa insertion look silly, whereas a VC analysis means you can just say coda /jl/ is banned like /lm/ is in other varieties, which also explains the fill-feel and fell-fail mergers. Distribution problems aren't problems - this is common in Sino-Tibetan, and it'd be silly to say /it/ is a diphthong just because there's no /et/. Similarly for the distribution problems in Lahu, where sporadic vowel copying causes onset contrasts that would otherwise be allophonic variation, or in the varieties of English with mergers of short vowels before /l/.

[Lērisama]

Also, what about SQUARE versus DRESS versus TRAP before /r/?

TRAP is a different quality from the other two, and I think SQUARE comes under the heading of ‘elision of a coda consonant’.

Ignore my mention of TRAP here, but then would you analyze Mary and merry as something like /ˈmɛrrɪj/ versus /ˈmɛrɪj/? Needing a phonemic geminate /r/ to explain this particular case does not smell right to me.

We have already had this conversation. I remain unconvinced. Also, couldn't you have the following:

• Merry: /ˈmɛɪj/ [ˈmɛɹ̠ɪj], with intrusive [ɹ̠]
• Mary: /ˈmɛrɪj/ [ˈmɛːɹ̠ɪj], as the stress attracts the /r/ to the coda, where it lengthens the vowel, + intrusive [ɹ̠]

Although that also has disadvantages r.e. analysing a short stressed vowel in an open syllable.

+ What Nort just ninja'd me with.

[Travis B.]

Don't these varieties generally prohibit stressed short vowels from appearing without a coda? (A related, syllabification-agnostic question: don't they generally prohibit word-final stressed short vowels?) Doesn't contrastive vowel length in these varieties come entirely from the elision of a coda consonant?

My big quibble is with treating things like PRICE, GOOSE, FACE, GOAT, GOAL, PRICE, MOUTH, and CHOICE as vowel-consonant sequences rather than diphthongs -- the analyses of SSBE and GenAus get much more sane when they are considered as diphthongs (PALM/START, SQUARE, THOUGHT/NORTH/FORCE, NEAR, and CURE can be treated as rhotic diphthongs) because then you avoid distribution problems with what vowel and what semivowel can go where (and because they then look far less radically different from other English varieties they are fully crossintelliglble with).

In AmEng a diphthongal analysis makes schwa insertion look silly, whereas a VC analysis means you can just say coda /jl/ is banned like /lm/ is in other varieties, which also explains the fill-feel and fell-fail mergers.

This may work in some NAE varieties, but in very many NAE varieties FLEECE and GOOSE are unambiguously monopthongal, so to propose a coda /j/ and /w/ for them to just explain schwa insertion is introducing a much bigger phonological problem (i.e. justifying underlying glides that do not surface) to explain some limited cases in some dialects.

Distribution problems aren't problems - this is common in Sino-Tibetan, and it'd be silly to say /it/ is a diphthong just because there's no /et/. Similarly for the distribution problems in Lahu, where sporadic vowel copying causes onset contrasts that would otherwise be allophonic variation, or in the varieties of English with mergers of short vowels before /l/.

It is one thing to have incidental gaps in distributions -- this is very common and need not require positing diphthongs -- it is another thing to have systemic correlations between nuclei and glides that cannot be explained in terms of incidental gaps.

[Travis B.]

TRAP is a different quality from the other two, and I think SQUARE comes under the heading of ‘elision of a coda consonant’.

Ignore my mention of TRAP here, but then would you analyze Mary and merry as something like /ˈmɛrrɪj/ versus /ˈmɛrɪj/? Needing a phonemic geminate /r/ to explain this particular case does not smell right to me.

We have already had this conversation. I remain unconvinced. Also, couldn't you have the following:

• Merry: /ˈmɛɪj/ [ˈmɛɹ̠ɪj], with intrusive [ɹ̠]
• Mary: /ˈmɛrɪj/ [ˈmɛːɹ̠ɪj], as the stress attracts the /r/ to the coda, where it lengthens the vowel, + intrusive [ɹ̠]

Although that also has disadvantages r.e. analysing a short stressed vowel in an open syllable.

Okay, you're proposing phonemic coda /r/ for all long vowels... yet also proposing non-phonemic intrusive-r applied to short stressed vowels in open syllables (the very existence of which you'd need to justify) to explain cases where short vowels precede [ɹ̠V]. This does not pass the smell test for me -- it is making things way more complicated in a fashion that does not solve the problem that phonemic coda /r/ for all long vowels was originally intended to solve in the first place.

[Lērisama]

Okay, you're proposing phonemic coda /r/ for all long vowels... yet also proposing non-phonemic intrusive-r applied to short stressed vowels in open syllables (the very existence of which you'd need to justify) to explain cases where short vowels precede [ɹ̠V]. This does not pass the smell test for me -- it is making things way more complicated in a fashion that does not solve the problem that phonemic coda /r/ for all long vowels was originally intended to solve in the first place.

Yes… actually thinking about that for now than 10 seconds suggests it is not my best idea…

[Travis B.]

Even the English here does not permit short stressed vowels in open syllables even though I don't analyze it as having VC stressed syllables. (Rather, I analyze it as having single consonants, and even some consonant clusters such as /nt/ and /st/, before unstressed vowels and following vowels, whether stressed or unstressed, or when forming the coda of a final syllable and followed by a word beginning with a vowel function as ambisyllabic and function differently from both onsets and codas, as evidenced by the flapping of /t d n nt/ which only occurs when ambisyllabic. Note however there are some weird cases, such as /tl/ in Atlantic, which behave as ambisyllabic in this regard.)

[bradrn]

Distribution problems aren't problems - this is common in Sino-Tibetan, and it'd be silly to say /it/ is a diphthong just because there's no /et/. Similarly for the distribution problems in Lahu, where sporadic vowel copying causes onset contrasts that would otherwise be allophonic variation

Can you say more about these?

[jal]

I want to politely remind everyone here that this is the Conlang Random Thread, and there are other threads that deal with English and other languages.


JAL

[Travis B.]

I want to politely remind everyone here that this is the Conlang Random Thread, and there are other threads that deal with English and other languages.

But being an English-language language-related board, discussions about English linguistics can infect any thread on it, no matter off-topic they are!

[jal]

But being an English-language language-related board, discussions about English linguistics can infect any thread on it, no matter off-topic they are!

I fully understand, hence the reminder :D.


JAL

[Glenn]

Apologies for jumping so far back in the thread...

In the olden days (a few boards ago) we had some pretty complete conlangs like Itlani and Siwa, and what's that language created by that dude that even had a flag made and waving in his yard - damn, was it something with an "A"? I'm getting old...

You're thinking of Alurhsa (i.e., the conlang with the creator who made his own flag). As you can see, his site is still up and running.

[jal]

You're thinking of Alurhsa (i.e., the conlang with the creator who made his own flag). As you can see, his site is still up and running.

I am! Thanks for the link, I was going crazy not remembering it :).


JAL