H/æ/lloween or H/ɑː/lloween, again

[Lērisama]

Interesting. I would have thought the main difference between the BATH vowel and the BUT vowel was length.

Certainly in modern Southern Bristish English, the length and quality are about equal: I'd hear [kʰɑ̆ʔ]* as either ⟨cut⟩ or ⟨cart⟩ depending on context. I will certainly test this at the ealiest opportunity. We've even innovated better pairs through non-rhotacism, such as the three-way [b̥ɛ̆t̚ b̥ɛt̚ b̥ɛːt̚] (that is bet, bed, beared). (I haven't found a four way vowel length minimal pair, I think because SQUARE is quite rare. This is part of why /ɑː/ really doesn't make sense to me in Halloween

*That is, the length of STRUT (ultra short because a fortis plosive follows) and the quality of PALM/BATH/START

[Raphael]

Interesting. I would have thought the main difference between the BATH vowel and the BUT vowel was length.

Certainly in modern Southern Bristish English, the length and quality are about equal: I'd hear [kʰɑ̆ʔ]* as either ⟨cut⟩ or ⟨cart⟩ depending on context. I will certainly test this at the ealiest opportunity. We've even innovated better pairs through non-rhotacism, such as the three-way [b̥ɛ̆t̚ b̥ɛt̚ b̥ɛːt̚] (that is bet, bed, beared). (I haven't found a four way vowel length minimal pair, I think because SQUARE is quite rare. This is part of why /ɑː/ really doesn't make sense to me in Halloween

*That is, the length of STRUT (ultra short because a fortis plosive follows) and the quality of PALM/BATH/START

Thank you!

[Travis B.]

One thing to remember is that (near-)minimal pairs are overrated. For instance, if one relied on (near-)minimal pairs my dialect has a three-way contrast in vowel quantity (and a two-way contrast in consonant quantity) and a contrast in vowel nasality. However, this breaks down when one actually analyzes things, as many of these distinctions are clearly largely allophonic and are due to a widespread combination of elision and assimilation. For instance, there is no vowel quantity contrast in intact morpheme-final syllables that do not have an obstruent in their coda -- rather it is clearly allophonic there. Likewise, forms without elision and/or assimilation readily resurface if one takes careful speech and higher registers into account, and are practically never "wrong" -- implying that the elision and/or assimilation does not exist in the underlying forms. (Note that all vowels are nasalized before nasal consonants -- except if the nasal consonant was originally a /b/ or /d/ that has itself assimilated to another nasal, where then they never are, and the parsimonious analysis here is that these are still underlying /b/ or /d/.)

[Emily]

I pronounce "Halloween" like in "hallow". The other pronunciation may have originated by analogy with "hollow", although the word "Halloween" has no etymological relation to "hollow".

In cot-caught merged varieties, it may also result from spelling influence. "Hall o ween"

The dialect here isn't cot-caught merged. My suspicion is that the [w] in Halloween exhibits a retracting influence at a distance.

this doesn't seem likely to me, or at least not as likely as the general avoidance of /æl/. historically this combination (well, /al/, with /a/ of course being the early ancestor of today's /æ/) became /ɑul/ in most environments, which later became /ɔl/, and to this day most english words with /æl/ are borrowings. halloween is a native term, but turning a less common combination into a more common combination seems likelier to me than influence from a /w/ that's two syllables away (and for many speakers already sort of merged with the preceding /o/). note also that reinforcement from the word "hallow" is probably decreasing as that word becomes less commonly used as well

[Sol717]

The dialect here isn't cot-caught merged. My suspicion is that the [w] in Halloween exhibits a retracting influence at a distance.

this doesn't seem likely to me, or at least not as likely as the general avoidance of /æl/. historically this combination (well, /al/, with /a/ of course being the early ancestor of today's /æ/) became /ɑul/ in most environments, which later became /ɔl/, and to this day most english words with /æl/ are borrowings. halloween is a native term, but turning a less common combination into a more common combination seems likelier to me than influence from a /w/ that's two syllables away (and for many speakers already sort of merged with the preceding /o/). note also that reinforcement from the word "hallow" is probably decreasing as that word becomes less commonly used as well

Words with historic /al/ → /æl/ (TRAP+/l/) only have alternate pronunciations with /ɒl~ɑl/ (LOT+/l/) when a back vowel historically followed (e.g. follows, making it likely that this was the conditioning factor; note the pronunciations of scallop (/ˈskæləp~ ˈskɒləp/) and fallow, tallow (RP /ˈfaləʊ̯ ˈtaləʊ̯/, GA /ˈfæloʊ̯ ˈtoləʊ̯/, but British English dialectal /ˈfɒlə ˈtɒlə/) in contrast with e.g. ally, dally (/ˈæli ˈdæli/). However, your theory is still possible since in my opinion, the least learned words with /al~æl/ are those where it precedes a back vowel, leaving those with a front vowel more vulnerable to modification under literary influence.

Incidentally, I've heard a pronunciation of Halloween with STRUT (~[ˈhɐɫɜwɪĩ̯ːn] or somesuch) rather than the TRAP usual in New Zealand; this could be a imitation of the American pronunciation with LOT given the recent spread of American-style trick-or-treating, but it could be a genuine phonetic development (I have STRUT in yellow, umbrella ~[ˈjɐɫɜɵ̯ m̩ˈbɻʷɐɰɐ]; note that historic /æl ɛl/ tend to coalesce in NZE).

[Travis B.]

The dialect here isn't cot-caught merged. My suspicion is that the [w] in Halloween exhibits a retracting influence at a distance.

this doesn't seem likely to me, or at least not as likely as the general avoidance of /æl/. historically this combination (well, /al/, with /a/ of course being the early ancestor of today's /æ/) became /ɑul/ in most environments, which later became /ɔl/, and to this day most english words with /æl/ are borrowings. halloween is a native term, but turning a less common combination into a more common combination seems likelier to me than influence from a /w/ that's two syllables away (and for many speakers already sort of merged with the preceding /o/). note also that reinforcement from the word "hallow" is probably decreasing as that word becomes less commonly used as well

Words with historic /al/ → /æl/ (TRAP+/l/) only have alternate pronunciations with /ɒl~ɑl/ (LOT+/l/) when a back vowel historically followed (e.g. follows, making it likely that this was the conditioning factor; note the pronunciations of scallop (/ˈskæləp~ ˈskɒləp/) and fallow, tallow (RP /ˈfaləʊ̯ ˈtaləʊ̯/, GA /ˈfæloʊ̯ ˈtoləʊ̯/, but English dialectal /ˈfɒlə ˈtɒlə/) in contrast with e.g. ally, dally (/ˈæli ˈdæli/). However, your theory is still possible since in my opinion, the least learned words with /al~æl/ are those where it precedes a back vowel, leaving those with a front vowel more vulnerable to modification under literary influence.

I have never heard LOT in scallop or tallow, and I'm an American. (I have TRAP [ɛː] in both of those words, and I haven't heard an American with anything other than TRAP in them.)

Incidentally, I've heard a pronunciation of Halloween with STRUT (~[ˈhɐɫɜwɪĩ̯ːn] or somesuch) rather than the TRAP usual in New Zealand; this could be a imitation of the American pronunciation with LOT given the recent spread of American-style trick-or-treating, but it could be a genuine phonetic development (I have STRUT in yellow, umbrella ~[ˈjɐɫɜɵ̯ m̩ˈbɻʷɐɰɐ]; note that historic /æl ɛl/ tend to coalesce in NZE).

One thing I should note is that the dialect here has DRESS [ɜː] in marshmallow, which is very close to your STRUT (but not my STRUT, which would be [ʌː] in the same position).

[Sol717]

I have never heard LOT in scallop or tallow, and I'm an American. (I have TRAP [ɛː] in both of those words, and I haven't heard an American with anything other than TRAP in them.)

To be clear, LOT in fallow, tallow is a geographically-restricted traditional dialect feature that's presumably extinct or nearly so, but in scallop LOT is standard (as attested by the obsolete spelling scollop) and still reasonably common; personally I have both [ˈskʰɔl̴ɘp] (LOT) and [ˈskʰɛ̞l̴ɘp] (TRAP), with the latter being more usual.

[Lērisama]

One thing to remember is that (near-)minimal pairs are overrated. [Lots of interesting examples]

Oh yes, it's just frustrating that the best pairs /ɛ ɛː/ /ɪ ɪː/ lack minimal pairs for historical reasons*, while the worse pairs /ʌ ɑː/ /ɔ oː/ have them in spades. (there's also /ɵ ɵː/, but /ɵː/ is marginal anyway

* The fir/fur/fer merger means that the long versions have to come from long vowels (really diphthongs now) followed by /ɹ/, and minimal pairs need a following consonant, and there was lots for Middle English shortening of vowels following rC

[Travis B.]

One thing to remember is that (near-)minimal pairs are overrated. [Lots of interesting examples]

Oh yes, it's just frustrating that the best pairs /ɛ ɛː/ /ɪ ɪː/ lack minimal pairs for historical reasons*, while the worse pairs /ʌ ɑː/ /ɔ oː/ have them in spades. (there's also /ɵ ɵː/, but /ɵː/ is marginal anyway

* The fir/fur/fer merger means that the long versions have to come from long vowels (really diphthongs now) followed by /ɹ/, and minimal pairs need a following consonant, and there was lots for Middle English shortening of vowels following rC

For me it is often difficult to find true minimal pairs in vowel quantity except before fricatives and nearly impossible to find three-way vowel quantity minimal triplets. For instance, while it is not too hard to find pairs like hence [hɜ̃nts] versus hens [hɜ̃ːnts], there are only a limited number of minimal pairs like matter [mɛːʁˤ] versus madder [mɛːːʁˤ], latter [ʟ̞ɛːʁˤ] versus ladder [ʟ̞ɛːːʁˤ], and mill [mɘːɯ̯] versus middle [mɘːːɯ̯].

[Darren]

hens [hɜ̃ːnts]

I'm surprised you have a [t] in there

[Travis B.]

hens [hɜ̃ːnts]

I'm surprised you have a [t] in there

/ns/ and /nz/ in my dialect are normally both pronounced with an epenthetic stop inserted in between, and /z/ is devoiced finally unless the next word in an utterance begins with a vowel (which also devoices the epenthetic stop). The two are distinguished by the length of the preceding vowel. Note that I have a distinction between mints [mɘ̃ʔts] and mince [mɘ̃nts], for instance, because the nasal is not elided before the epenthetic stop and the epethetic stop is not preglottalized, unlike with normal /t/ in the same position.

[Darren]

/ns/ and /nz/ in my dialect are normally both pronounced with an epenthetic stop inserted in between, and /z/ is devoiced finally unless the next word in an utterance begins with a vowel (which also devoices the epenthetic stop). The two are distinguished by the length of the preceding vowel. Note that I have a distinction between mints [mɘ̃ʔts] and mince [mɘ̃nts], for instance, because the nasal is not elided before the epenthetic stop and the epethetic stop is not preglottalized, unlike with normal /t/ in the same position.

That's interesting; I have the mints-mince merger but "hens" would just be [henˑz]

[Travis B.]

Note that properly I should have said that /s/ and /z/ become affricates after /n/ rather than that an epenthetic stop is inserted, because the stopping is much shorter and weaker than a normal stop (e.g. the [t] in mince is shorter and weaker than that in mints).