Pronunciations you had to unlearn

[jal]

I would think that "an" before an h would be a slip up for aitchless speakers? I've never seen "an" before [y] myself. Might be an overgeneralisation of a spelling rule (which isn't really a spelling rule to begin with).


JAL

[Salmoneus]

I meant the first u. "Oovular", I only saw it because it was written "an uvular stop". Probably he got the second u still correct.

Are you sure the person wasn't British? British style uses "an" before /h/ and words written with a vowel but pronounced with /j/.

As far as I know, there is no common style of English that uses "an" before words written with a vowel but pronounced with /j/. I have heard scattered reports of this usage, but I don't know of any style or usage guides that countenance it.

Likewise, as a British person, this is not something I'm aware of. "An yellow book"? I can't say it doesn't happen anywhere, in any dialect - to me, it sounds like something that might happen in an old rural dialect up north or out west - but it would sound very weird to me.

And I think the use of "an" before /h/ is not restricted to British style: there are also Americans who do this, although I think it is unpopular (in both the US and the UK). From what I understand, an + /h/ is usually only accepted as a "valid" (if eccentric) variant usage before words starting in an unstressed syllable, so "an historic" is more commonly accepted than "an history".

It's a little complicated.

The problem arises with words, originally all loanwords from French, that historically had no /h/ except in their spelling - hospital, hotel, history, herb, horrible, etc. These naturally took "an". However, in the 19th century, there was largescale h-dropping in Cockney, followed (or accompanied) by largescale h-restoration, including hypercorrection; h-dropping became a shibboleth, so h-hyperrestoration became more widespread than the original h-dropping had been - basically, everyone was desparate not to sound like a cockney. Eventually, cockney h-hyperrestoration itself became a widely-mocked shibboleth*, so people tried to to put too many /h/s in - but they used spelling as a guide, and didn't know which words with spelled <h> weren't actually meant to have pronounced /h/.

This created a class-based schism: upper class people said "an 'orrible 'ospital" and "an 'istoric 'erb", but "a hat"; lower-class people in London said the same things for the first two, but "an 'at" for the third; middle-class people had the "right" way of saying the third, but incorrectly analogised to the first two cases, so said "a horrible hospital" and "a historic herb".

BUT! The upper classes tried to make this a shibboleth as well, so everybody was taught that it's not "a historic moment", it's "an historic moment". But the poor middle-classes were so confused by this point that they didn't know what they were meant to be learning. By now it didn't occur to them that 'historic' might not have an /h/ in it, so they thought the rule was just "historic" takes "an". This was then generalised to other words with initial /h/.

However: I don't think it was every fully generalised; as time went on, it became more and more restricted in application. For my general, the rule pretty much ONLY applied to the word 'historic'.


Now, as I understand it, the situation is:
- most <h> words have /h/ and take 'a' for everybody (a hat, not an hat)
- the words 'historic' and 'historical' have /h/ and 'a' for most people, but no /h/ and 'an' for some better-educated and older people, while a lot of middle-aged people THINK there's a rule to have /h/ and 'an' but they only sporadically actually follow that rule in speech.
- a couple of other words follow the same division, but more rarely. The only semi-common one I can think of is "an hotel", which you do still sometimes see in writing, though you'd probably only hear from upper class old women.

[Tropylium]

Some people here have restored /ju/, or shall I say, [i̯ʉ] or even [i̯y] in words like new and stupid through fronting and breaking of /u/ after a coronal.

This would then presumably include also words like too, do, noose?

[Linguoboy]

Now, as I understand it, the situation is:
- most <h> words have /h/ and take 'a' for everybody (a hat, not an hat)
- the words 'historic' and 'historical' have /h/ and 'a' for most people, but no /h/ and 'an' for some better-educated and older people, while a lot of middle-aged people THINK there's a rule to have /h/ and 'an' but they only sporadically actually follow that rule in speech.
- a couple of other words follow the same division, but more rarely. The only semi-common one I can think of is "an hotel", which you do still sometimes see in writing, though you'd probably only hear from upper class old women.

And what's the deal with herb? Since USAmericans basically behave like UK Middle Classes in most things, it's very odd to me that we've preserved this case of historic h-lessness while all of non-Cockney Britain seems to have restored it.

[Travis B.]

How does this relate to the situation in NAE, which unlike English English does not appear to have undergone processes you mention? Yet at the same time almost all the <h> words here have /h/ and take a not an, aside from herb and sometimes homage (and I don't mean just the Frenchy loan-pronunciation /oʊˈmɑʒ/ that has taken root in some circles), so if many of these words originally did not take /h/ they must have undergone widespread spelling pronunciation to arrive at the current pronunciations. Yet this certainly could not have been under the influence of Cockney, since 19th-century Cockney pronunciations are unlikely to have influenced NAE pronunciation.

[Travis B.]

Some people here have restored /ju/, or shall I say, [i̯ʉ] or even [i̯y] in words like new and stupid through fronting and breaking of /u/ after a coronal.

This would then presumably include also words like too, do, noose?

Fronting of /u/ occurs after all coronals, and breaking, with or without partial unrounding, either occurs in general, not at all, or, for some individuals like myself, when /u/ is preceded but not followed by a coronal. I personally have [tʲʰʉ̯u(ː) tʲʉ̯u(ː) nʲʉs], as I normally lack the unrounding. But for the people who do have the unrounding it does not pattern with the historical distribution of /ju/, e.g. there are people who pronounce two as [tʲʰi̯ʉ(ː)] or [tʲʰi̯u(ː)].

[Estav]

As far as I know, there is no common style of English that uses "an" before words written with a vowel but pronounced with /j/. I have heard scattered reports of this usage, but I don't know of any style or usage guides that countenance it.

Likewise, as a British person, this is not something I'm aware of. "An yellow book"? I can't say it doesn't happen anywhere, in any dialect - to me, it sounds like something that might happen in an old rural dialect up north or out west - but it would sound very weird to me.

The original post by Zaarin specified that "an" would only be used before words written with a vowel, which I assume excludes words like "yellow", and only includes things like "an European" or "an unicorn". This could be phonologically conditioned in a dialect where the original "iw" diphthong had not merged with "juw" or "jiw" word-initially, but I don't know of any area in North America or England where the lack of this merger is a part of standard English. (Maybe it might occur in Welsh English? I don't know; Wikipedia only mentions a possible retention of an "iw"/"uw" contrast after consonants.) In a dialect with the merger, it would be a spelling-conditioned distribution, which of course doesn't seem very likely to be an organic usage.

However: I don't think it was every fully generalised; as time went on, it became more and more restricted in application. For my general, the rule pretty much ONLY applied to the word 'historic'.


Now, as I understand it, the situation is:
- most <h> words have /h/ and take 'a' for everybody (a hat, not an hat)
- the words 'historic' and 'historical' have /h/ and 'a' for most people, but no /h/ and 'an' for some better-educated and older people, while a lot of middle-aged people THINK there's a rule to have /h/ and 'an' but they only sporadically actually follow that rule in speech.
- a couple of other words follow the same division, but more rarely. The only semi-common one I can think of is "an hotel", which you do still sometimes see in writing, though you'd probably only hear from upper class old women.

It definitely seems like in practice, the use of "an" + /h/ has been associated with specific words (of which "historic(al)" is the most common). But the stress-conditioned rule that I mentioned is given in various prescriptive sources such as Fowler (1926):

A is used before all consonants except silent h (a history, an hour); an was formerly usual before an unaccented syllable beginning with h (an historical work), but now that the h in such words is pronounced the distinction has become pedantic, & a historical should be written; similarly an humble is now meaningless and undesirable. A is now usual also before vowels preceded in fact though not in appearance by the sound of y or w (a unit, a eulogy, a one).

Fowler does give the example of "an humble" where the original absence of /h/ seems to be attributable to the French origins, as you state. While he might be wrong about the role of stress in conditioning the use of "an" before other words, it doesn't seem too implausible to me that a general loss of /h/ word-initially before unstressed vowels might have once been common, even in words that were taken directly from Latin or Greek and not via the medium of French; /h/ is still often lost in such positions word-internally, as in vehicle, annihilate, prohibition (vs. vehicular and prohibit).

[mèþru]

I gave up on this long ago and just always pronounce a written <h>. It's English orthography's fault, not the non-native speakers'.

[Vlürch]

I gave up on this long ago and just always pronounce a written <h>. It's English orthography's fault, not the non-native speakers'.

Same, except with homage. That's probably the only word where I wouldn't pronounce the /h/ and would use "an" with, at least more likely than the opposite. Well, there may be a few other words, but I can't think of any right now. I don't think I have a problem with recognising most French loanwords that "an" could/should be used with, but using "an" with them would seem really posh and borderline Brit-LARPing so I don't see the appeal in doing it unironically.

Speaking of H-dropping, though, I started wondering if there's any dialect/accent with all of H-dropping, intrusive /r/, trap-strut merger, fronting of /a/ and reduction of unstressed vowels to schwa. I know it's probably really unlikely for all of those features to coexist in a single dialect/accent, but just imagine if you overheard someone exclaim "BLADDER 'ELL!"

[Zaarin]

Speaking of H-dropping, though, I started wondering if there's any dialect/accent with all of H-dropping, intrusive /r/, trap-strut merger, fronting of /a/ and reduction of unstressed vowels to schwa. I know it's probably really unlikely for all of those features to coexist in a single dialect/accent, but just imagine if you overheard someone exclaim "BLADDER 'ELL!"

That reminds me of the time I heard someone from Rhode Island say "Your odds are harder": [jə ɑʔdz ɑː hɑːɾəː].

[Pabappa]

Late reply but ... yeah, this person wasnt British and I agree that it would be wrong to say "an uvular" even in British English and even though the first syllable carries the stress. But I suspect it's a common mistake. It may be more common among people whose L1 is English but have studied foreign languages and tend to hyper-Latinize Latin words. I've made analogous mistakes with puerperal, spurious, and perhaps others I cant remember right now. My mother goes the other way and until recently always said /'pju.tṇ/ for Putin.

Herb is a name, but I suspect the loss of /h/ in AmE that word is just due to chance, not due to confusion with the name. herb with an /h/ was college slang in the 90s, and UD claims it may have originated at Burger King:
https://www.youtube.com/watch?v=U3d_uUNuLWc
But the evidence is pretty weak.

[Zaarin]

Herb is a name, but I suspect the loss of /h/ in AmE that word is just due to chance, not due to confusion with the name. herb with an /h/ was college slang in the 90s, and UD claims it may have originated at Burger King:
https://www.youtube.com/watch?v=U3d_uUNuLWc
But the evidence is pretty weak.

As far as I'm aware, in American English herb (a leafy plant) has no /h/ while Herb (a short form of the name Herbert) does have /h/...

[Pabappa]

sorry, i knew i shouldbve written that better. there was a college slang term "herb", not marijuana, but a mildly derogatory insult for a person, similar to "nerd" but not very well known. there are a few traces of it on urbandictionary .

[Zaarin]

Ah, hadn't heard that term before.

[Ryusenshi]

The original post by Zaarin specified that "an" would only be used before words written with a vowel, which I assume excludes words like "yellow", and only includes things like "an European" or "an unicorn". This could be phonologically conditioned in a dialect where the original "iw" diphthong had not merged with "juw" or "jiw" word-initially, but I don't know of any area in North America or England where the lack of this merger is a part of standard English. (Maybe it might occur in Welsh English? I don't know; Wikipedia only mentions a possible retention of an "iw"/"uw" contrast after consonants.) In a dialect with the merger, it would be a spelling-conditioned distribution, which of course doesn't seem very likely to be an organic usage.

I remember an old post by J.C. Wells talking about that (Tuesday 3 April). Apparently, some Welsh accents would indeed have something like [ɪʊ] in unicorn, and use an before such words. But probably most instances of an unicorn, an European are due to spelling (especially for non-native speakers).

Speaking of H-dropping, though, I started wondering if there's any dialect/accent with all of H-dropping, intrusive /r/, trap-strut merger, fronting of /a/ and reduction of unstressed vowels to schwa. I know it's probably really unlikely for all of those features to coexist in a single dialect/accent, but just imagine if you overheard someone exclaim "BLADDER 'ELL!"

Is there any native variety that merges TRAP and STRUT?

Related question: is there a variety of English that merges THOUGHT with GOAT? Many non-native speakers do, but I've never heard of such merger in a native accent. With all the weird mergers that happen in English, this is almost surprising.

[Linguoboy]

Related question: is there a variety of English that merges THOUGHT with GOAT? Many non-native speakers do, but I've never heard of such merger in a native accent. With all the weird mergers that happen in English, this is almost surprising.

Offhand I can't think of any unconditioned mergers of historically long vowels with the corresponding short vowel phonemes.

The closest I think you come for THOUGHT and GOAT are those varieties of Scotland and Ulster where the first is [ɔː] and the second [oː]. These probably don't sound very close at all to a native Francophone, but to other native English-speakers who are used to having a diphthong in the second case and low vowel in the first, they can become easily confused.

[Travis B.]

Related question: is there a variety of English that merges THOUGHT with GOAT? Many non-native speakers do, but I've never heard of such merger in a native accent. With all the weird mergers that happen in English, this is almost surprising.

Offhand I can't think of any unconditioned mergers of historically long vowels with the corresponding short vowel phonemes.

The closest I think you come for THOUGHT and GOAT are those varieties of Scotland and Ulster where the first is [ɔː] and the second [oː]. These probably don't sound very close at all to a native Francophone, but to other native English-speakers who are used to having a diphthong in the second case and low vowel in the first, they can become easily confused.

I often hear THOUGHT in EngE and AusE varieties as GOAT, because they are close to my native monophthongal (when not prevocalic) GOAT (which is usually [o]) and not very much like my own THOUGHT (which is [ɒ], much closer to their own LOT), and I have a hard time telling [ɔ] and [o] apart too, whereas at the same time I hear their GOAT as also being GOAT, because I am used to people from elsewhere having markedly diphthongal and often at least partially centralized GOAT.

[Vlürch]

Is there any native variety that merges TRAP and STRUT?

Well, googling it I found this pdf about Aberystwyth English, which also mentions other possible dialects/accents/whatever with "overlap" between the vowels but concludes that the merger in at least the focused-on dialect/accent/whatever is apparently not complete, and also mentions that they're native speakers of Welsh rather than English, so... I guess it doesn't count. Apparently RP is heading in that direction, too, but I'm kind of sceptical... I mean, how could that happen in practice unless RP stopped being a prescriptive pronunciation? And if RP stopped being a prescriptive pronunciation, would it even be RP anymore?

Also this about Gibraltarian English, which is unfortunately missing a lot of pages; it does mention that the merger was only "present in one case", and same for a BATH-TRAP-STRUT merger, so it's probably not common even there. Kind of relevant rant: it's annoying that not all research is publicly available for free like it IMHO should be, but that could well cause a collapse of societal order since few people would be motivated to do any kind of research if they couldn't sell their findings, leading to stalling of progress and an even sharper decrease in the quality of education, so...

[Travis B.]

Kind of relevant rant: it's annoying that not all research is publicly available for free like it IMHO should be, but that could well cause a collapse of societal order since few people would be motivated to do any kind of research if they couldn't sell their findings, leading to stalling of progress and an even sharper decrease in the quality of education, so...

You do realize that the people who write in journals make practically no money off of it, and practically all the money goes to the publishers - even the peer reviewers for the journals are normally not paid for the work they do.

[anteallach]

Is there any native variety that merges TRAP and STRUT?

Well, googling it I found this pdf about Aberystwyth English, which also mentions other possible dialects/accents/whatever with "overlap" between the vowels but concludes that the merger in at least the focused-on dialect/accent/whatever is apparently not complete, and also mentions that they're native speakers of Welsh rather than English, so... I guess it doesn't count. Apparently RP is heading in that direction, too, but I'm kind of sceptical... I mean, how could that happen in practice unless RP stopped being a prescriptive pronunciation? And if RP stopped being a prescriptive pronunciation, would it even be RP anymore?

Also this about Gibraltarian English, which is unfortunately missing a lot of pages; it does mention that the merger was only "present in one case", and same for a BATH-TRAP-STRUT merger, so it's probably not common even there.

I've seen references to a sporadic merger in RP or similar accents before; as that article says it's mentioned as a possibility in Wells's Accents of English. However, I don't think it was ever very common, and if you read Geoff Lindsey's article on STRUT he suggests that STRUT has reacted to TRAP's move towards [a] by moving away from [a] towards what he describes as a triangle defined by [ə], [ɑ] and [ʌ]. (Note the recording of the future Queen in 1940 with something very [a]-like for STRUT, saying what I hear as "jast", "lav" and "cams".)

BTW I think RP is better thought of as a sociolect than a "prescriptive standard". And Daniel Jones's RP isn't John Wells's RP, and John Wells's RP isn't my generation's RP. (My generation's RP is pretty much what Lindsey is describing, I think.) In one of his other articles, Lindsey says he met a student who was from London, but had been educated at a boarding school in the North of England and had managed to lose both the TRAP-BATH and FOOT-STRUT splits. Things like that might make you question whether RP as it traditionally existed actually still does...