Anglic sound changes

[Rounin Ryuuji]

My understanding is that all surviving dialects had a boil-bile merger, but there was then an orthographically conditioned split. (English is bedevilled by literacy.)

Is that really possible?

Yes. Look at all the common words that have spelling pronunciations. As a result of an 'o' in the spelling, one, won, none, ]trouble and wonder all have variously extensive and deprecated pronunciations with [ɒ]. /h/ has largely been restored in words of Latin origin (it's a fight to keep it out of heir and hour), and there are many Americans who resurrect /l/ in calm and walk.

I don't find reading pronunciations themselves that far-fetched (English is plagued with a terrible orthography), but I do find it difficult to swallow that a massive sound change could be totally reversed by the orthography alone, without extremely common words showing some resistance (note that one itself has avoided a spelling pronunciation, despite only, alone, with the same morpheme, surviving with the expected vowel). I also wouldn't expect the exact same diphthong to be reintroduced. If I have no [ɔɪ]. long o is [oʊ], and terminal -y in a monosyllable is [iɪ], I'm probably pronouncing joy, boy as homophones of Joey, Bowie. For [ɔɪ] to be still there, I think it must've survived in some words but not in others in the dialects that had the boil-bile merger, with the orthography likely helping the [ɔɪ] pronunciation of the boil words to become prestigious again, alongside the massive number of homophones it created making interdialectal borrowing for the sake of disambiguation more likely.

Of course, the dialects that resupplied the [ɔɪ] might all be dead by now, but I doubt very much that the sound itself was ever totally dead in English.

[Richard W]

(note that one itself has avoided a spelling pronunciation, despite only, alone, with the same morpheme, surviving with the expected vowel).

In some speech, yes. For me and many others, one and wan are homophones, unless I'm deliberately setting an example. To restate it, the spelling pronunciation is for one to rime with gone rather than with done.

I also wouldn't expect the exact same diphthong to be reintroduced. If I have no [ɔɪ]. long o is [oʊ], and terminal -y in a monosyllable is [iɪ], I'm probably pronouncing joy, boy as homophones of Joey, Bowie.

Before the merger, there were two diphthongs spelt <oi>/<oy>. What happened after the initial merger was that words pronounced with the vowel of bile but spelt with <oi>/<oy> came to be pronounced with the distinct vowel spelt with <oi>/<oy>. This sound change changed the distribution of vowels, but did not add another vowel to the repertoire.

[Rounin Ryuuji]

I also wouldn't expect the exact same diphthong to be reintroduced. If I have no [ɔɪ]. long o is [oʊ], and terminal -y in a monosyllable is [iɪ], I'm probably pronouncing joy, boy as homophones of Joey, Bowie.

Before the merger, there were two diphthongs spelt <oi>/<oy>. What happened after the initial merger was that words pronounced with the vowel of bile but spelt with <oi>/<oy> came to be pronounced with the distinct vowel spelt with <oi>/<oy>. This sound change changed the distribution of vowels, but did not add another vowel to the repertoire.

I mentioned there being two diphthongs initially; I must've mistaken you somewhere, though — I thought you were saying that all the diphthongs in question had ended up /ai/ and then spelling pronunciations had changed them. As far as every [ɔɪ] goes, all of them being restored after merging with /ai/ on the sole basis of the orthography still strikes me as implausible.

[Travis B.]

I also wouldn't expect the exact same diphthong to be reintroduced. If I have no [ɔɪ]. long o is [oʊ], and terminal -y in a monosyllable is [iɪ], I'm probably pronouncing joy, boy as homophones of Joey, Bowie.

Before the merger, there were two diphthongs spelt <oi>/<oy>. What happened after the initial merger was that words pronounced with the vowel of bile but spelt with <oi>/<oy> came to be pronounced with the distinct vowel spelt with <oi>/<oy>. This sound change changed the distribution of vowels, but did not add another vowel to the repertoire.

I mentioned there being two diphthongs initially; I must've mistaken you somewhere, though — I thought you were saying that all the diphthongs in question had ended up /ai/ and then spelling pronunciations had changed them. As far as every [ɔɪ] goes, all of them being restored after merging with /ai/ on the sole basis of the orthography still strikes me as implausible.

This sounds more like a case of two different sound changes occurring in two different sets of dialects, the merger of /ai/ and /ɔi/ in some dialects and the merger of /ɔi/ and /ui/ in other dialects, with the latter coming to replace the former through the latter dialects influencing the former dialects, without spelling pronunciation necessarily being at work (the latter dialects being more prestigious than the former dialects is all that is needed here). Of course, orthography could definitely help this along, but there are phonemic distinctions which are reasonably stable in spite of orthography (e.g. whether medial <ei> marks /ei/ or /i/~/iː/, even though the contrast between the two where both have been represented by <ey> finally has not historically been stable).

[Travis B.]

I sometimes feel like spelling pronunciation is overstated; e.g. in the case of -day, where former /-di/ became /-dei/. In those cases it could have been explained just as well as analogy with day /dei/ by itself. If anything, invoking "spelling pronunciation" often seems like a way to look down on the proles/Americans/etc. who don't "know better".

[Rounin Ryuuji]

This sounds more like a case of two different sound changes occurring in two different sets of dialects, the merger of /ai/ and /ɔi/ in some dialects and the merger of /ɔi/ and /ui/ in other dialects, with the latter coming to replace the former through the latter dialects influencing the former dialects, without spelling pronunciation necessarily being at work (the latter dialects being more prestigious than the former dialects is all that is needed here).

This is my understanding of what actually happened.

[bradrn]

There is also the possibility of ‘almost-mergers’: sound changes which don’t quite merge things completely, and hence are reversible. Can’t quite recall my source (Trask’s Historical Linguistics, it might have been?), but I do distinctly recall several examples of this from English. I do think it implausible that the demerger could have been triggered by spelling only.

[Rounin Ryuuji]

As I understand it, Middle English /ai/ had merged into /ei/ probably sometime in the late 1300s or mid 1400s, however historic /iː/ became a diphthong, probably [əi] or [əɪ]. In some dialects [əɪ ɔɪ] merged before whatever the result was shifted to [aɪ], in others [əɪ] became [aɪ], and [ɔɪ] merged with [ʊɪ].

I have, during the course of the whole boil-bile discussion, been somewhat perplexed by something — I would've expected boil to NOT merge with bile because it ought to have had [ʊɪ] rather than [ɔɪ] (note French bouiller, bouillon, etc., not *boiller, *boillon (if I'm not much mistaken, at least). There may have been dialects where [ʊɪ] > [ɔɪ] happened before [ɔɪ] > [aɪ]. I'm not totally sure about when all this happened, but I do distinctly remember some old dialects represented in Disney cartoons have /bail/ for boil (I do also wonder if this is some sort of exaggerated speech, however).

"Boil" (as in the thing you can get on your skin) would, I think, be expected to merge with bile in dialects with the [ɔɪ aɪ] merger, but "boil" (the thing that happens when water gets hot enough) wouldn't.

[Travis B.]

As I understand it, Middle English /ai/ had merged into /ei/ probably sometime in the late 1300s or mid 1400s, however historic /iː/ became a diphthong, probably [əi] or [əɪ]. In some dialects [əɪ ɔɪ] merged before whatever the result was shifted to [aɪ], in others [əɪ] became [aɪ], and [ɔɪ] merged with [ʊɪ].

I have, during the course of the whole boil-bile discussion, been somewhat perplexed by something — I would've expected boil to NOT merge with bile because it ought to have had [ʊɪ] rather than [ɔɪ] (note French bouiller, bouillon, etc., not *boiller, *boillon (if I'm not much mistaken, at least). There may have been dialects where [ʊɪ] > [ɔɪ] happened before [ɔɪ] > [aɪ]. I'm not totally sure about when all this happened, but I do distinctly remember some old dialects represented in Disney cartoons have /bail/ for boil (I do also wonder if this is some sort of exaggerated speech, however).

"Boil" (as in the thing you can get on your skin) would, I think, be expected to merge with bile in dialects with the [ɔɪ aɪ] merger, but "boil" (the thing that happens when water gets hot enough) wouldn't.

Boil the verb is from OF bolir, from Latin bullire. Of course, what OF <o> represents here, i.e. was it before or after /o/ > /u/, is a good question.

[Rounin Ryuuji]

As I understand it, Middle English /ai/ had merged into /ei/ probably sometime in the late 1300s or mid 1400s, however historic /iː/ became a diphthong, probably [əi] or [əɪ]. In some dialects [əɪ ɔɪ] merged before whatever the result was shifted to [aɪ], in others [əɪ] became [aɪ], and [ɔɪ] merged with [ʊɪ].

I have, during the course of the whole boil-bile discussion, been somewhat perplexed by something — I would've expected boil to NOT merge with bile because it ought to have had [ʊɪ] rather than [ɔɪ] (note French bouiller, bouillon, etc., not *boiller, *boillon (if I'm not much mistaken, at least). There may have been dialects where [ʊɪ] > [ɔɪ] happened before [ɔɪ] > [aɪ]. I'm not totally sure about when all this happened, but I do distinctly remember some old dialects represented in Disney cartoons have /bail/ for boil (I do also wonder if this is some sort of exaggerated speech, however).

"Boil" (as in the thing you can get on your skin) would, I think, be expected to merge with bile in dialects with the [ɔɪ aɪ] merger, but "boil" (the thing that happens when water gets hot enough) wouldn't.

Boil the verb is from OF bolir, from Latin bullire. Of course, what OF <o> represents here, i.e. was it before or after /o/ > /u/, is a good question.

Apparently, there was some variation in the pronunciation of [ɔɪ] and [ʊɪ] within the various dialects of Middle and Early Modern English, which is not at all surprising.

[Estav]

The usual hypothesis is actually that /ui/ merged with original long /i:/, while /Oi/ remained separate. Previous posts seem to have got this reversed. Compare the general unrounding of short u that occurred in nut etc.

[Nortaneous]

Also note that historical */auk/ derived from an earlier /uːk/ does not seem to occur in native words (though I imagine some people pronounce /alk/ in this way now).

GVS didn't apply to ū before noncoronals.

[Man in Space]

Is this something if possible use? Or is it too informal?

[bradrn]

Is this something if possible use? Or is it too informal?

Looks even less reliable than Wikipedia to me. (It doesn’t even cite any sources!)

[bradrn]

Actually, that reminds me — I’ve been meaning to make a mockup of the web interface for the Middle English sound changes, those being the only ones which are anything like complete at the moment. It’s escaped my attention due to my being utterly swamped with other work at the moment, but hopefully I’ll get around to it sometime in the next month or two.

…and, as promised, here’s the mockup: http://bradrn.com/files/english-mockup.html.

Cross-posting this here in case anyone has any comments on the sound changes.

[Rounin Ryuuji]

Is there any reason Middle English /iː uː/ are listed as having become [ei ou]? I thought the conventional understanding was that the shift was [iː uː] > [əɪ əu] > [aɪ aʊ] (for me and some others > [ɑɪ æʊ]). Also, [æ] > [ɑ] was frequently triggered by /l/ (calm, palm), and almost universally by /r/. In some dialects this was also triggered by voiceless fricatives, note some UK pronunciations of path, half.

[bradrn]

Is there any reason Middle English /iː uː/ are listed as having become [ei ou]? I thought the conventional understanding was that the shift was [iː uː] > [əɪ əu] > [aɪ aʊ] (for me and some others > [ɑɪ æʊ]).

Because Wikipedia’s page on the subject is incredibly confusing and this was what I decided made most sense. If you have a better source please do share!

Also, [æ] > [ɑ] was frequently triggered by /l/ (calm, palm), and almost universally by /r/. In some dialects this was also triggered by voiceless fricatives, note some UK pronunciations of path, half.

Thanks, if you have a source to cite I’ll add that! (And I have the same pronunciation for BATH, so it’s not like I’m unaware of that fact… but it’s still an unpredictable change!)

[Rounin Ryuuji]

Because Wikipedia’s page on the subject is incredibly confusing and this was what I decided made most sense. If you have a better source please do share!

...

Thanks, if you have a source to cite I’ll add that! (And I have the same pronunciation for BATH, so it’s not like I’m unaware of that fact… but it’s still an unpredictable change!)

I remember reading a book about it while I was in grad school, but I can't recall the title. There was some discussion in classes and things about the uncertainty of when the diphthongisation of /iː uː/ actually occurred. It might've been in my copy of the Canterbury Tales, but that's in storage somewhere. Some of the Wikipedia articles do seem to list /iː uː/ > [əi əu], incidentally.

[anteallach]

Is there any reason Middle English /iː uː/ are listed as having become [ei ou]? I thought the conventional understanding was that the shift was [iː uː] > [əɪ əu] > [aɪ aʊ] (for me and some others > [ɑɪ æʊ]).

The Wikipedia article on the Great Vowel Shift mentions both the [əi]/[əu] and [ei]/[ou] developments as possibilities and cites sources supporting each. I think the outcomes in regional speech, at least in southern England, suggest the former, and that's also more consistent with how they interacted with the ME diphthongs; in particular /iː/ merged with /ui/ (as mentioned earlier in this thread) but not with /ei/ (or however you want to transcribe the ancestor of the EModE vowel in main, wait etc.).

Also, [æ] > [ɑ] was frequently triggered by /l/ (calm, palm), and almost universally by /r/. In some dialects this was also triggered by voiceless fricatives, note some UK pronunciations of path, half.

The shift before /lm/ is already listed (under "/al/, /ɔl/ mutations"). In fact my suspicion is that the same development occurred, at least in some dialects, before /lf/, which would explain why calf and half have long vowels in standardised northern English (unlike typical BATH words). You also sometimes find spellings like cauf and hauf (e.g. Love's Labour's Lost Act V Scene 1, where they look like descriptions of the actual contemporary pronunciation) which suggest that the /l/ didn't simply disappear. (If I'm right, then the modern American pronunciation would have to derive either from a later shortening or a dialect which just lost the /l/ without diphthongisation or lengthening.)

The shift before /r/ is also listed (under "Short vowel changes"). On this, I think the environment should be when the /r/ is followed by a consonant or a word or morpheme boundary (star, start, starry but not carry), and the schwa should be long, giving the length of the modern NURSE vowel.

I'm not sure that the origin of the long vowel in father is actually understood: when it developed or why.

[anteallach]

On the /h/-loss section, this is a very messy sound change with some unpredictable outcomes, but:

- I think I would separate the simplification of /hl/, /hn/ and /hr/, mainly because the rest of what went on is so messy and it's a rather different environment. Also, there's no variation that I'm aware of in modern dialects.
- The list of outcomes includes /o/ and /ɔ/ separately; I don't think English had a vowel conventionally transcribed /o/ at this stage.
- The list of inputs includes the ME diphthongs. I'm not sure these occurred before /h/ other than by diphthongisation caused by the /h/ itself not long before it disappeared: are there any examples, say, with /iuh/? All the examples I can think of of /f/ from older /h/ would, I think, have initially had /a/ (laugh, /ɔ/ (cough) or /u/ (tough, feeding the FOOT/STRUT split).
- The diphthongisation/lengthening of short front vowels before the [ç] allophone of /h/ seems to be missing. It's not very clear, but the Wikipedia article on ME phonology does mention the lengthening of the vowel of night (feeding into the GVS).