Regarding PRICE and MOUTH

[PVER•PVERVM•AMAT]

Once (and remaining this way in many accents) PRICE and MOUTH had homogenous starting points; a central [ä], yes? But in SSBE (and GenAm for MOUTH) they have different starting points: a front [a] for MOUTH, a back [ɑ] for PRICE. My question: why did the diphthong with the front vowel ending point back its starting point, and the diphthong with a back vowel ending point front its starting point? It seems to me that it ought to have been the other way around.

[linguoboy]

Maybe because dissimilation is a thing?

I would also look at how the /oy/ and /ow/ diphthongs shifted over time in the varieties in question, since maintaining a distinction between them and their low counterparts would be pretty important. (Though not decisive, since IIRC some Irish English accents merge /ay/ and /oy/.)

[PVER•PVERVM•AMAT]

You do mean /oj/, right? No English diphthongs have rounded front vowel ending points (and I think that the lack of distinguishing them in some Irish accents is a resistance to a split, not a merger). CHOICE has only changed to match the less open THOUGHT vowel, while GOAT has centralized its starting point except before dark L, in which environments it has back vowel starting point comparable to LOT (usually creating a split).

[linguoboy]

You do mean /oj/, right?

No, I mean /oy/. "I'll take American Structuralist notation for a thousand, Alex!"

[PVER•PVERVM•AMAT]

Another thing: dissimilation is adequate for explaining this, but why has PRICE resisted dissimilation in GenAm but not MOUTH?

[----]

It only takes one change for dissimilation to occur. If you have A and A and then you have A and B they're already dissimilar, right?

[Travis B.]

In Inland North dialects the two dissimilated in the opposite fashion, such that /aɪ̯ aʊ̯/ became, e.g. in my dialect, [ae̯ ɑɔ̯] ([əe̯ ɑ̝o̯] with Canadian Raising). However I have read things that implied that this might actually be the historical pattern.

[Nortaneous]

Another thing: dissimilation is adequate for explaining this, but why has PRICE resisted dissimilation in GenAm but not MOUTH?

Has it? I've heard people who have [ɑɪ̯].

For dialects where it's not, it could be explained by the general tendency of back vowels to front. Do dialects with a fronted first element in /au/ also front their back vowels?

[PVER•PVERVM•AMAT]

In Estuary, BATH (and the lexically equal PALM and START) has fronted a bit from the old fully back quality, but GenAm shows no signs of fronting its fully back LOT or PALM, although it seems to be fronted in START sometimes.

[Travis B.]

In Estuary, BATH (and the lexically equal PALM and START) has fronted a bit from the old fully back quality, but GenAm shows no signs of fronting its fully back LOT or PALM, although it seems to be fronted in START sometimes.

My dialect (an Inland North dialect) is the opposite, where it has definitely fronted /ɑ/ to [a] for LOT and PALM (except when adjacent to /r h w kw gw/), but clearly has [ɑ] in START.

[Imralu]

Australian English is a bit more extreme, with MOUTH starting around the area of [æ] or even higher (pretty much the same starting point as /eI/) and the beginning of PRICE being right at the back and sometimes even rounded, driving /OI/ up higher, with an onset around the [o] area.

[PVER•PVERVM•AMAT]

In Estuary, BATH (and the lexically equal PALM and START) has fronted a bit from the old fully back quality, but GenAm shows no signs of fronting its fully back LOT or PALM, although it seems to be fronted in START sometimes.

My dialect (an Inland North dialect) is the opposite, where it has definitely fronted /ɑ/ to [a] for LOT and PALM (except when adjacent to /r h w kw gw/), but clearly has [ɑ] in START.

Travis, I'm fairly certain that the NCVS isn't considered GenAm. If you were contrasting your own accent with GenAm, however, ignore this.

[Tropylium]

It only takes one change for dissimilation to occur. If you have A and A and then you have A and B they're already dissimilar, right?

In Inland North dialects the two dissimilated in the opposite fashion, such that /aɪ̯ aʊ̯/ became, e.g. in my dialect, [ae̯ ɑɔ̯] ([əe̯ ɑ̝o̯] with Canadian Raising). However I have read things that implied that this might actually be the historical pattern.

Hang on, this sounds like different sort of dissimilation. What has been my understanding of this explanation is that this would have been dissimilation of each diphthong's components wrt/ each other: [aɪ] is mid open to front close, which dissimilates to [ɑɪ]: back open to front close. Equivalently, [aʊ] is mid open to back close, and dissimilates to [æʊ]: front open to back close. This would thus be a similar change as the Old French dissimilation of *ɔu to ‹eu›.

I don't see much reason why there would be "dissimilation" of the two diphthongs from each other? They're already distinguished by their offglides. The pronunciation described by Travis instead sounds like an assimilation of the diphthongs' components.

---

Another explanation entirely might be an attraction of the diphthong's 1st component to the existing short vowels. AmEng has no /a/, so it makes sense to e.g. shift MOUTH to a direction that can loosely speaking equal a short vowel + /w/ (in this case, TRAP).

(Of course, this approach fails to explain why FACE has not become ˣ[ɛɪ] and GOAT has not become ˣ[ɔʊ] or ˣ[ɑʊ].)