Talk:Pronunciation of English ⟨a⟩/Archive 1

Talk from the old æ-tensing page

Name

Does anyone have a reference for this name for this phenomenon? It doesn't quite parse to me; and the only Google references for "pan tensing" are Wikipedia clones. -- Smerdis of Tlön 16:11, 23 Feb 2005 (UTC)

This name for it is a neologism, but there is no well-known name for the phenomenon. Some linguists also call it "short a tensing" or "[æ] tensing", but neither of those gets too many Google hits either. John C. Wells calls it "bath raising", two words which occur adjacent to each other frequently enough in Google, but only rarely with the intended meaning. The most common name for it appears to be æ-tensing; shall I move the page to that title? At any rate, it's definitely a real phenomenon, and it's much more complicated than laid out here so far. I'll work on it when I get some time. --Angr 12:57, 4 Mar 2005 (UTC)

"Accents of the Midwest and West, as well as Canadian English, do not have æ-tensing."

Is this saying all Midwestern accents lack [æ]-tensing? Because that is just plain wrong, considering most people I've seen in southeastern Illinois (including myself and my family) and southeastern Indiana have accents with [æ]-tensing (some people even "sound Southern"). --/ɛvɪs/ 18:24, Apr 15, 2005 (UTC)

Well, there's æ-tensing and æ-tensing. What Midwestern accents lack, AFAIK, is the phonemic contrast between tense and lax æ that Philadelphia and NYC have. But lots of Midwestern accents are affected by the Northern Cities Shift, which makes /æ/ have a rather tense realization in many environments, but always predictable (allophonic) ones. --Angr/_comhrá 19:13, 15 Apr 2005 (UTC)

Ohhhhhhhhh. I thought it was saying that Midwestern accents didn't have the nazalization of [æ] before [n] or [m]. I might also add that around here it sounds more like [æ̃] than [eә], so I assume that must not be a tense [æ]. --/ɛvɪs/ /tɑːk/ /kɑntɹɪbjuʃ(ə)nz/ 19:11, August 11, 2005 (UTC)

I'm from New York myself, and the chart given is hardly representative of my pronunciation. I pronounce the æ tense only before nasals. In the chart, 'man' and 'ham' are the only words I pronounce with the tense æ. (Actually, I sometimes pronounce 'manage' that way as well.) Furthermore, words like 'ran' I do pronounce with the tensed vowel. I think the influence of nasal sounds on the vowel is underestimated here.

Are you from New York City or upstate New York? The chart is supposed to be representative of NYC, not upstate. Anyway, even if you are from the city, the chart is based on what published linguistic research has shown to be typical of the local accents of Philadelphia and NYC; that doesn't mean that everyone who lives there exceptionlessly uses that pronunciation. --Angr/_{tɔk tə mi} 06:03, 26 October 2005 (UTC)

Edits, 16 April

My edits of 16 April were reverted by Angr with the comment "not an improvement". They were an improvement for the following reasons:

• Are there any accents in which tensing occurs only before nasals in the same syllable? The default pattern tenses /æ/ before all nasals, tautosyllabic or not; other patterns tense before tautosyllabic nasals and voiceless fricatives. I don't think there are any that tense before tautosyllabic nasals only.
• Saying "In accents that have undergone the Northern Cities Vowel Shift, there is no contrast, but..." the "there is no contrast" is redundant (since the sentence goes on to say there is only one realization of short a). Moreover it carries a false implication that it's notable that there's no contrast in the NCVS; on the contrary, the normal situation is for there to be no contrast; contrast only exists in a couple of accents, and it's notable there.
• Most accents of the Midwest and West do have some degree of æ-tensing. The NCVS region has tensing in all environments; the rest of the Midwest and the West have the default before-nasals-only pattern.

AJD 14:34, 16 Apr 2005 (UTC)

There are accents (NYC-Philadelphia) where there is no tensing before an intervocalic nasal or fricative unless there's a word boundary or class II morpheme boundary after the nasal. In these accents the a of family, planet, Spanish, classic, passive, etc. is lax; while the a in plan, man, class, pass, mass, path, calf is tense. (But the tense vowel remains when a clitic or class II suffix is added, so that plan it, mannish, classy, passing have tense vowels.)

It's important to distinguish between the purely allophonic æ-tensing of the NCVS and the potentially phonemic tense /eə/–lax /æ/ contrast of Philadelphia-NYC, which is also known as æ-tensing. I wouldn't consider either type of æ-tensing a default pattern, since they're both regionally restricted. Both kinds are notable and need to be carefully distinguished, since the majority of English speakers (even the majority of Americans) don't have either kind of æ-tensing. The slight raising of /æ/ before nasals found in the West and non-NCVS Midwest isn't really usefully considered æ-tensing in the sense discussed on this page, since the raised /æ/ is still more open (has a higher F1) than /ɛ/. The significantly tensed /æ/'s of the NCVS and New York-Philadelphia are more close (have a lower F1) than the /ɛ/ of those accents; IMO this page should be about those significantly tensed [eə]-sounds (whether a separate phoneme from /æ/ or an allophone of /æ/). --Angr/_comhrá 15:25, 16 Apr 2005 (UTC)

So, are there any accents in which tensing occurs only before tautosyllabic nasals? Note that the example you give of accents where tensing is dependent on syllable structure has tensing before nasals or voiceless fricatives. I don't think there are any accents in which, as the article says, "it occurs only before nasal consonants in the same syllable".

I agree that it's important to distinguish between purely allophonic æ-tensing and phonemic æ-tensing, and the article doesn't distinguish between them clearly enough. But allophonic æ-tensing is ordinary; phonemic æ-tensing is the exception, not the rule; and the article's current treatment of the NCVS tensing misleadingly suggests the opposite.

Labov describes nasal-only æ-tensing as the default system for a couple of reasons: one, it occurs in a variety of regions that have no other dialectal affiliation with each other, such as eastern New England, northern New Jersey, and the "Midland" region which extends from Ohio to Kansas. Two, it's the pattern most often developed by non-native or non-American speakers who are trying to acquire (or unintentionally end up acquiring) an American-sounding accent without acquiring whatever the stigmatized features of the region they live in.

I think you underestimate the degree of æ-tensing across the U.S. /æ/ before nasals is, I think, significantly higher than /ɛ/ in the West, and certainly in the Midland, the non-NCVS North, and New England. AJD 16:10, 16 Apr 2005 (UTC)

I'll try to improve the part about tensing before tautosyllabic nasals, taking morphemics into consideration, and I'll try to distinguish better between phonemic and allophonic æ-tensing. But I stand by my assertion that prenasal /æ/ is more open than /ɛ/ (even for speakers without the pin-pen merger) outside of the northeastern quadrant of the country.

I think you're misunderstanding my point about tautosyllabic nasals. Morphemics isn't relevant (to my point, I mean). My point is that there is no dialect in which (a) syllable (or morpheme) boundaries have any effect whatsoever and (b) nasal consonants are the only environments that condition tensing. The sentence "in some accents, it occurs only before nasal consonants in the same syllable" is false, and it would remain false if you added an exemption for morpheme boundaries. As for the degree of tensing in the Midland and the West, I'll try and track down some of Labov's research on the subject. AJD 18:55, 16 Apr 2005 (UTC)

Yeah, you're right. Condition (b) would have to include fricatives as well. Anyway, I've done a pretty major re-write which I hope clears up what's going on. Let me know what you think. --Angr/_comhrá 19:56, 16 Apr 2005 (UTC)

Merged article

I can see that it makes sense to have these features on the same page, but the title is a bit cumbersome. Any ideas for an improvement? "Sound changes affecting English short A" or something like that?--JHJ 12:11, 22 December 2005 (UTC)

Also, I think trap/bath should be first - it's the oldest split, and I'm pretty sure it's the best known.--JHJ 12:12, 22 December 2005 (UTC)

I agree, a snappier title is called for. Phonological history of /æ/, perhaps? And you're right about trap/bath coming first. Not only is it oldest, but it's important in Britain, America, and Australia, while the other two are more local. --Angr (t·c) 12:25, 22 December 2005 (UTC)

My suggestion is "Short A splits in English" or "Splits of short A in English". "Phonological history of /æ/ is, I think, not technically possible, and probably wants us to discuss other events surrounding /æ/. —Felix the Cassowary (ɑe hɪː jɐ) 13:55, 22 December 2005 (UTC)

Is the problem with my suggestion the slashes? The problem with your suggestion is that æ-tensing isn't always a phonemic split. --Angr (t·c) 14:28, 22 December 2005 (UTC)

I'm not keen on having the phoneme symbol in the title, for a number of reasons (a technical symbol in the title might put some people off; the splits mean it isn't a single phoneme any more in certain accents; it's not called /æ/ in all sources - descriptions of Scottish and, sometimes, other British accents, e.g. new edition OED transcriptions, call it /a/). Another suggestion might be "Phonological history of English short A". --JHJ 14:45, 22 December 2005 (UTC)

I'm unmerging the articles, until we can figure out if they should be merged and what the title should be. I don't think they should be merged, because they're very historically independent and occur in very different regions. Foogol 14:58, 22 December 2005 (UTC)

"Phonological history of English short A" or "Phonological history of the trap vowel" would be possible titles. Foogol 20:52, 22 December 2005 (UTC)

I could live with Phonological history of English short A. Article titles shouldn't have italics in them, because links to them tend to break: Phonological history of the ''trap'' vowel. --Angr (t·c) 21:54, 22 December 2005 (UTC)

Glad to see that most of us agree that it makes sense to have these merged. Yes, Foogol's has a good point that they are historically and geographically independent however they do involve the same phoneme. Just trying to tidy the place up a little but my solution is not necessarily the best nor do I expect it to be final. I also agree that the title wasn't the best but I thought I'd just call it that for the time being until we came up with a better one. Jimp 27Dec05

"Phonological history of the trap vowel" is inappropriate because broad A is not the trap vowel, by definition. It's the bath vowel. I'm happy with "... of English short A", though. —Felix the Cassowary 12:48, 27 December 2005 (UTC)

"Phonological history of the trap vowel" is appropriate, because broad A is not what the article is primarily concerned with. Broad A is just one of the possible outcomes of what was originally the trap vowel.

Also, I personally suspect that æ-tensing and the bad-lad split are historically related. --Ptcamn 15:43, 27 December 2005 (UTC)

Well, another problem with saying "the trap vowel" is that the term is very Wells-specific, and correspondingly only well known in Britain. The traditional name "short A" is understood by more people. --Angr (t·c) 16:07, 27 December 2005 (UTC)

I agree: "saying 'the trap vowel' is that the term is very Wells-specific". Personally I'm not keen on the traditional names "long A", "short A", etc. historically they made sense but then there was the Great Vowel Shift and long I is no longer a longer version of short I but a completely different animal. Anyhow, this doesn't really matter because Wikipedia is not the place to introduce new terminology. Of course, another problem is that the BAD vowel is not short (for those of us with the bad-lad split) this is the whole point but again this is a problem with the traditional terminology which we're not here to change. Moreover, this doesn't matter because, like the BATH vowel, it split from the TRAP. I'm thinking of merging them back again I s'pose "short A" is best unless I can get any nods for Phonological history of the low front vowels which I suggest in order to fit in with the Phonological history of the ... vowels articles. Jimp 18:20, 8 January 2006 (UTC)

Yet More Americentrism

The opening paragraph describes the 'broad A' as a US-specific pronunciation, mentioning nothing of the broad A in English used in Britain. In fact the whole article mentions nothing of this. Either the broad A page is a bad redirect, or someone has neglected to make this article global. Boothman 19:36, 29 December 2005 (UTC)

It's a U.S.-specific phenomenon, so it's kind of difficult to talk about what happens in other parts of the world. And the opening paragraph describes æ-tensing and makes no mention of "broad A" at all. As you suspect, the redirect from broad A is what's wrong, and I'll fix it. --Angr (t·c) 19:55, 29 December 2005 (UTC)

Moved from Talk:Trap-bath split

Why moved?

"Broad A" is a much better title for this article than "trap-bath split". "Broad A" is a term actually used by people (dialectologists, elocutionists, lexicographers, and the general public); "trap-bath split" is found almost exclusively within Wikipedia, and only has twelve Google hits. I'm going to move the article back unless someone can give me a good reason not to. AJD 20:19, 30 May 2005 (UTC)

Fine with me. There were originally two pages, and I merged them together under Broad A. An anonymous user then cut-and-pasted all the text from Broad A into Trap-bath split (which is Wells's name for it, so it's not just Wikipedia). But I agree Broad A is a better place for it. --Angr/_{tɔk tə mi} 04:46, 31 May 2005 (UTC)

Trap-bath split is a better name for the article than broad a, because broad a does not only refer to the a sound used in the bath words, but it can also refer to the a sound in father, which all dialects use, so the title broad a wouldn't be completely accurate.

Actually father often has the CAUGHT vowel in Hiberno-English. This is not the father-bother merger; the only other affected word I can think of is R (the name of the letter: hence the spelling ar for this is confusing in Ireland). What other words are not in trap-bath, not in father-bother, with the broad-A vowel not followed by R? Joestynes 14:53, 8 December 2005 (UTC)

Wells' keyword is PALM; calm and psalm could be here too. But then some people pronounce these words with /ɔlm/. PALM words often have the same vowel as TRAP in Hiberno-English (psalm and Sam then being homophones). It's actually very hard to find a word that has the "broad A" in all accents of English (or even all "major" accents of English) not followed by R. American English has it in loanwords like pasta and drama, but many English and Canadian speakers have /æ/ there (sort of an inverse trap-bath split). Sonata maybe? --Angr (t·c) 15:40, 8 December 2005 (UTC)

None of the minimal pairs on this RP list are homonyms in my idiolect, though the have-halve / ant-aunt difference seems purely one of vowel length. It's because there are so few such pairs that the phonemic status of "Broad A" is so marginal for me; without them the BATH words' vowel would be an allophone of /æ/ (and the PART words' a different allophone). Joestynes 16:40, 8 December 2005 (UTC)

Presumably accents with pairs like Sam and psalm the same (like many Scottish and northern Irish accents) just don't have the "broad A" as a phoneme distinct from /æ/ or /a/ or whatever you want to call it? I think this is the usual analysis of Scottish English. --JHJ 17:22, 8 December 2005 (UTC)

I've heard father pronounced with short [a] (which is the trap vowel round here), though I don't do so myself. The words I have with the "broad A" (which I actually call the "long A" except when communicating with Americans) are father and rather; half, calf and their relatives (though half past is usually [haf past]); can't and shan't; several words with alm in the spelling as mentioned above; and a number of loanwords, including drama and sonata but not pasta. I believe this is fairly standard for the northern half of England.--JHJ 17:22, 8 December 2005 (UTC)

Origin of the split

Two things. First, the article (in the fourth paragraph) implies that the vowel change was [æ] to [ɑː]. Is there enough evidence on the historic sounds to justify those particular transcriptions, given the variability of both sounds in present day accents?

Second, the Gupta (2005) paper that I added as a reference for the geographic distribution of the split in England (version available online here) suggests that the variation in words like aunt, dance and chance is considerably older than that in words like bath and grass, in fact suggesting that it dates to the 14th century for the former words, and quotes a reference Dobson (1968).

--JHJ 17:51, 8 December 2005 (UTC)

Southern drawl

In the Southern United States, the pattern most characteristic of Southern American English does not employ æ-tensing at all, but rather what has been called the "Southern drawl": /æ/ becomes in essence a triphthong [æjə].

Is this for real? It sounds more likely to be how the pronunciation is percieved by speakers of other dialects rather than how they actually say it. --Ptcamn 21:41, 8 February 2006 (UTC)

wow

It took me forever (I've been wondering for about 8 years) to finally figure out why dictionaries didn't have the correct phonetics for things such as ham, can, jam, etc. I always thought that everyone says these words this way. It's definately how I hear them. I just don't see how, say "plan" and "plack" can be considered to have the same vowel sound. Wierd. This adds nothing to the discussion I suppose. I just had to express my amazement. Also, I guess I should note that I was reared in Southwest Missouri and have never been to New England, so it may be in this region as well.

This was regarding bad-lad split, sorry.--68.89.209.197 04:59, 23 February 2006 (UTC)

Questions from my accent

I think I speak with a fairly standard southern British accent. Reading this and reflecting on my accent, I have the following questions:

• In what words do we use /ɑː/ before /nd/?

Demand and command are two which come to mind.--JHJ 20:40, 26 February 2006 (UTC)

• /ɑː/ is also used in words such as palm and calm. Is this part of the same trap-bath split or is it historically different?

I don't think so, no. I'm from the North, and don't have the trap/bath split ([a] in both), but I do have a long vowel in the alm words, and I don't think I've ever heard them pronounced otherwise. I think the vowel in these words is due to the influence of the disappeared /l/. (Yes, I know it's still there in some accents.)--JHJ 20:40, 26 February 2006 (UTC)

• I'm interested by the idea that /æː/ may be used before /l/. As in Pall Mall? When else is this combination used?

I'll let an Aussie answer this.--JHJ 20:40, 26 February 2006 (UTC)

I use a long /æː/ in pal, and I think in (old) gal, though I never actually say that word; but short in Al, Hal, shall, which is predictable from their being either contractions or grammatical functions.

Gailtb 06:09, 25 February 2006 (UTC)

Coloradan æ tensing

I'm from CO, and I add a /j/ offglide to my æ's before voiced velars (/g/ and /N/). Does anyone know of anywhere else where this occurs?

"Father-gather" split?

I don't think this is an authentic split in short a. Father had long A in Middle English, not short A; the interesting question is why it dodged the Great Vowel Shift, but that has nothing to do with the topic of the article. Calm and words like it didn't undergo a split in short A either (the modern vowel in them didn't develop from short A alone but from short A plus L), but at least that's related to the phonological history of short A. AJD 22:40, 19 August 2006 (UTC)

Father having /a:/ in Middle English would indicate that it had the same vowel as in name and cake, which mostly became /e:/ during the Great Vowel Shift, but for some reason didn't in father where it remain /a:/. The modern phoneme /A/ found in RP and North American father comes from three sources, the word father failing to undergo the Great Vowel Shift along with the other words, the loss of /l/ in words like calm, palm etc. along with the lengthening of the /a/ in those word, and the lengthening of historic /a/ before /r/, in words like car, park, star etc. The vast majority of the sources for this present phoneme therefore are from a lengthening of /a/ (Middle English short a) to /a:/.

I'd agree that "father-gather split" is not the best title for the development of this phoneme, as "father" had a long vowel in Middle English. Perhaps "Development of the /A/ phoneme" would be better? "Development of the modern /A/ phoneme"? However, it's a real phenominum as the Modern English /A/ phoneme didn't exist in Middle English. Voortle 04:14, 20 August 2006 (UTC)

I've renamed it to "Development of the /A/ phoneme", and edited it a bit. Voortle 04:23, 20 August 2006 (UTC)

Thanks. I did some light prose editing. AJD 04:41, 20 August 2006 (UTC)

What's the evidence father had a long A in Middle English? It certainly had a short æ in Old English. User:Angr 06:14, 20 August 2006 (UTC)

Apart from the fact that regular sound change (open-syllable lengthening between OE and ME, I mean) would produce long A in that environment? The reference in my notes is to Jespersen (1954); but I don't have the book itself, just the attribution. AJD 14:17, 20 August 2006 (UTC)

But if regular sound change is the reason, why is father the only word that did it? Gather and rather had short A in Middle English. And it just seems too unlikely that open-syllable-lengthened [a:] would have escaped the GVS in this word (cf. bathe). I rather suspect father had a short A in Middle English (at least the earliest stages of it) and it didn't get lengthened until after the GVS had already started moving old [a:] to [ε:]. Perhaps it was the first word to get the pre-fricative lengthening of the trap-bath split, early enough that it spread to all accents, even the ones that are otherwise "flat A" accents. Do we know what words it rhymes with in Chaucer, Shakespeare, and everyone in between? User:Angr 14:40, 20 August 2006 (UTC)

Gather is gadrian in OE—no open-syllable lengthening because it was a closed syllable. As for rather, well, that seems to have had it both ways: note that in some modern dialects it rhymes with father and in some with gather. (The broad A in rather can't be the "trap-bath" split; that doesn't occur before voiced fricatives. And anyway, broad-A in rather is much more widespread than in grass, dance, etc.) And you say it's unlikely that father would escape the GVS there, but note that the same thing happened to rather and water. (Water: long a in ME; missed GVS, so broad A in early post-ME; then the same rounding which affected watch and wad and so on. That's why water has /ɔ:/.) So it looks like a regular exception to the GVS, occurring before a coronal obstruent plus syllabic R. Which seems phonologically weird, but it seems to have affected all three of those words.

As far as I can tell, neither father nor rather is ever rhymed in Chaucer. Water is rhymed with (Latin) mater, which is long A. Shakespeare rhymes father with gather, though; I don't know what to make of that. Jespersen cites dialects in which father did get GVSed. AJD 16:03, 20 August 2006 (UTC)

Yeah, I just looked at the OED, and it mentions dialectal "faither" too. But there's a lot of variation on how widespread broad A in "bath" words is in general, not just "rather". "Grass" and "dance" have flat A in northern England, but I think "half" and "can't" both have broad A there; so "rather" isn't alone. I dunno. I think the Neo-Grammarians definitely didn't have English in mind when they claimed that all sound changes are regular. User:Angr 16:18, 20 August 2006 (UTC)

Yes, half and can't are frequently pronounced with what we call "long A" (i.e. /ɑː/) in northern England, although there is some variation, especially with half. But are we sure that the long vowels in half and calf aren't in fact due to the vocalisation of the original /l/ (like in calm, but with later shortening in some/most varieties) and not the trap-bath split? There are also local dialect forms of half with [eː] or the like, which may be relevant.--JHJ 09:45, 21 August 2006 (UTC)

Well, L-vocalization after a in English resulted in [ɔː] at least before /k/ (e.g. talk, walk, chalk, the original pronunciation of falcon before spelling pronunciation screwed it up). But the lengthening in calm must be separate from the lengthening in half, because calm lengthened early enough that American English has the "father" vowel in it, while half has the "trap" vowel. I'm pretty sure in Shakespeare's time half was [hæf] and didn't get lengthened until the same time as laugh and staff did. User:Angr 09:56, 21 August 2006 (UTC)

But by the same sort of argument, the lengthening in half seems to be separate from that in staff, based on the northern English forms, and at least the difference between calm and half in American English can potentially be blamed on the following consonant (a possible theory is that /a:/ from earlier /al/ before labials shortened before /f/ in an ancestor of modern American English, or something like that). As for Shakespeare's pronunciation, there's a bit in Love's Labour's Lost about silent letters which actually mentions half, and uses the alternative spelling hauf, which seems to suggest something other than [hæf] (would Shakespeare have had [æ] or [a] there?).--JHJ 11:24, 21 August 2006 (UTC)