Template talk:/æ/ raising in North American English

Citations

Nardog: If I add relevant citations to this, would you be willing to restore the template to articles you deleted it from? Most of the information comes from the ANAE, with a map covering all of North America and giving a pretty comprehensive overview on p. 182. I can get into specifics, but it'll take some time to find all the precise page numbers. Is there anything else that you feel makes this "poorly sourced and all over the place"? I'm happy to also reduce the number of example words. Wolfdog (talk) 00:33, 12 April 2020 (UTC)[reply]

Of course. Yes, we need more page numbers and the words should be cut down to a handful per row (if that). p. 182 doesn't provide verification for what exactly each symbol indicates, the quality of the vowels, or the claims that were hidden in the words column (which we should get rid of, or at least incorporate into the structure of the table itself, if we want to keep the table a summary). Nardog (talk) 10:54, 12 April 2020 (UTC)[reply]

Sounds good. Unless you have some specific reason otherwise, I'm going to go back to the "free" and "checked" format, which mirrors what the ANAE actually uses. Wolfdog (talk) 17:26, 12 April 2020 (UTC)[reply]

Oh, and regarding the fact that p. 182 doesn't provide verification for what exactly each symbol indicates, neither really does the rest of the book. There are often descriptions of vowel qualities, but these can be fairly broad too, such as, for a New Yorker The tense /æh/ tokens are concentrated in upper-mid position. Is "upper-mid" what we on Wikipedia call "close-mid" (i.e. [e]) or [ɪ] or something else? And the fact that there is an an in-glide to me means a diphthong should be represented, ending in a schwa. Anyway, these kinds of reasons are why I tried to keep to a simple system here of [eə] for extremely tense, [ɛə] for intermediately tense, and [æ] for lax. Any thoughts on transcription here? Wolfdog (talk) 18:17, 12 April 2020 (UTC)[reply]

I'm going to go back to the "free" and "checked" format Why? I haven't heard of this usage of "checked" and "free" and it's not even the sense described by our article, Wells (1982), or ANAE itself on p. 12. /æ/ is always "checked" in that sense, and even when it's applied to /æC/ it still doesn't mirror the usage of the earlier version of this template. Nardog (talk) 10:49, 13 April 2020 (UTC)[reply]

[ɛ(j)ə] on the Southern/AAVE column indicates your system was not strictly followed (also, does it mean a possibility of a triphthong or two syllables?). What about using plain descriptions like "tense" or "raised", "intermediate", etc. instead? But either way, it's not clear whether the raising is phonemic in any of the cells, which IMO should definitely be indicated. Nardog (talk) 10:56, 13 April 2020 (UTC)[reply]

Well, ANAE literally uses the terminology "closed" and "open", and Checked and free vowels seemed to be our closest equivalent; I hear you though that /æ/ is always "checked" in that sense. ~~We could link here. Having a simple "either/or" system just made the chart more concise and cleaner-looking, in my opinion; see the working example at my Sandbox.~~

True, regarding the Southern/AAVE column (and also the Canadian etc. column). We could use the "tense" or "raised", "intermediate", etc. method, but then we don't get the phonetic qualities of the vowel, which can differ pretty substantially; for example, some allophones are merely raised, while others are raised plus diphthongized/lengthened. I'd be happy to go either way. There are of courses sources I could look into beyond ANAE that could provide more specific IPA transcriptions. (That's for example, where the odd-looking Southern transcription came from: a source by Erik Thomas.) I'd be happy to find a way to represent phonemic vs. not too. That Sandbox is where I'll continue to try out stuff. Thanks for the input; keep it coming. Wolfdog (talk) 16:12, 13 April 2020 (UTC)[reply]

Checked and free vowels seemed to be our closest equivalent Not remotely. If that dichotomy was applied to /æC/, /æC#/ would be considered a free environment, not checked. "Open" vs. "closed" isn't without problems either, as not only does it depend on the approach to syllabification (Merriam-Webster, CEPD, and LDP would all give you different answers), but also e.g. random is syllabified as /ˈræn.dəm/, i.e. with an open /æC/, according to MOP-based theories, but it seems to belong to the "closed" group. "Prevocalic" and "non-prevocalic" are what they truly mean by "open" and "closed" as far as I can tell.

Either way, if you're going to mention specific phonetic qualities they must be cited, to a T. Nardog (talk) 01:53, 14 April 2020 (UTC)[reply]

Well, I did my best. And in the future, I'll do what I can. Clearly, phonetics is no realm of perfection. Wolfdog (talk) 17:55, 14 April 2020 (UTC)[reply]

My major edits and additions of citations are basically done. Feel free to discuss any other desired changes, suggestions, or lingering questions here. Wolfdog (talk) 15:44, 15 April 2020 (UTC)[reply]

@Wolfdog: Where did you get the notation ⟨eːʲ⟩ from? I don't see it on Boberg (145). Nardog (talk) 16:33, 15 April 2020 (UTC)[reply]

I generally hear Canadians use [εː], which I see is supported on p. 145, but only as one variant; still, I think that makes the best sense. Boberg 145 says this allophone before voiced velars, /g/ and /N/, ha[s] up-glides, toward [j]. This glide reversal causes a change across subsystems, so that short /æg/ effectively becomes long, up-gliding /eyg/, merging with the /eyg/ of plague, vague. Specifically, though, the description seems to come from Labov 181 (sorry about that) for the haggle-Hegel merger: another variant of course. Wolfdog (talk) 20:46, 15 April 2020 (UTC)[reply]

Thanks. Then wouldn't a more accurate representation be ⟨ɛː~ɛj⟩ or something? Any reason you've used ⟨e⟩ over ⟨ɛ⟩? Nardog (talk) 21:51, 15 April 2020 (UTC)[reply]

Labov and I think Boberg too tend to use the term "mid" position, which means something between [e] and [ɛ], so I just picked one. But it wasn't totally arbitrary; Boberg describes Canadian /æ/ before nasals and /æ/ before velars as beginning the same but gliding in different ("opposite") directions; so I wanted to represent that similarity by being consistent and using [e]. But I don't really care either way. Go for it; sounds truer to how I hear most Canadians say it, anyway. Wolfdog (talk) 00:38, 16 April 2020 (UTC)[reply]

the description seems to come from Labov 181 (sorry about that) Why the apology? Actually, can we dispense with Boberg (2008) and cite the entire thing to ANAE? That would be simpler, as the articles where this template would be used most likely already cite ANAE, but not Boberg. Nardog (talk) 02:49, 18 April 2020 (UTC)[reply]

Having two columns for "some Southern US" is definitely not helpful. Can we make them more specific? Nardog (talk) 17:10, 15 April 2020 (UTC)[reply]

This is one of those situations where the fluidity of accents rears its ugly head. Erik Thomas gives the more "General American" phones verbatim (let's call this Type A), though certainly acknowledges Labov's variants. Meanwhile, Labov et al. have found the "Southern-only" column pronunciations (I'll call this Type B), but his map reveals a significant degree of Thomas's system in the South as well. I knew you wouldn't like this, ha, but that's the phonetic reality in the South: some speakers use Type A and some Type B. The best patterns I can give you are this: Type A appears more in the western half of the South and is perhaps more coastal (kinda?), especially robust in Atlanta and Charleston; whereas Type B appears more in the eastern half of the South, in the more inland South (sorta?), and Dallas (Labov 182). Open to any suggestions. Wolfdog (talk) 20:46, 15 April 2020 (UTC)[reply]

My immediate suggestion would be to use descriptive text over phonetic transcription where there's too much variation, as in the GA/etc and South columns. Barring that, I'd suggest lumping all Southern realizations into one column like "Southern US | [eə~ɛjə] | [æ~æɛə]". Nardog (talk) 21:51, 15 April 2020 (UTC)[reply]

Either way is fine. Are you talking about descriptive roll-over text? I think with your latter suggestion, I felt some intuitive desire to avoid the tilde symbol this time around because it can become a slippery slope for transcribing infinite possible phonetic variants. Wolfdog (talk) 00:38, 16 April 2020 (UTC)[reply]

Oh sorry, by "over" I meant "in preference to, instead of". Like "Lax" or "Not raised" (or "Intermediate" or "Variable") instead of "[æ]" and so on. It seems we're still using [æ] and [eə] as shorthands for lax and tense in some cells and more literal transcriptions in others, and I don't think we should. Nardog (talk) 01:22, 16 April 2020 (UTC)[reply]

OK, gotcha. I think it would look weird for one column to use descriptions, breaking the whole IPA look. But it might be for the best. However, if those are the options, personally, I prefer the slippery slope look with the tildes after all. When you say we're still using [æ] and [eə] as shorthands, which columns concern you? Wolfdog (talk) 13:09, 16 April 2020 (UTC)[reply]

The first three (NYC/NO, Mid-Atlantic, GA&c) and the last (Inland North), mainly. I find it would be quite a coincidence if the qualities of the lax and tense vowels were that consistent across accents and environments.

What about "Roughly [æ]" or something then (where we currently have the footnote)? Nardog (talk) 02:49, 18 April 2020 (UTC)[reply]

I've split "Canada, Northern Mountain US" and "Upper Midwest US". If some or all of "pp. 82, 123, 177, 179" provide citation for [æ] in the former column, please split or copy the citation. Nardog (talk) 17:10, 15 April 2020 (UTC)[reply]

Just to be clear, the split you made ("pp. 82, 123, 177, 179") really refers to the eastern North Central region, rather than the whole North Central region. Wolfdog (talk) 20:46, 15 April 2020 (UTC)[reply]

I was asking if any of those pages also provided citation for the lack of raising in Canada, Northern Mountain, and other parts of Upper Midwest.

Then can we use "Western/Eastern Upper Midwest" or "Western/Eastern North Central"? Nardog (talk) 21:51, 15 April 2020 (UTC)[reply]

1) I don't recall off-hand (... but probably not in particular). Is it worth me going back to check? 2) How would you use "Western/Eastern Upper Midwest"? Are you worried about too many footnotes? Wolfdog (talk) 00:38, 16 April 2020 (UTC)[reply]

Nah, if a source verifies that it's raised in those specific contexts in those specific places, then presumably it also verifies the lack of raising in other places as well, so don't worry about it and I'll repeat the citation.

Like this. I'm not worried as much as we should represent anything that can be represented in the table in the table. Otherwise it defeats the purpose of the chart. Speaking of which, I find the reference to the possible /æɡ/–/eɪɡ/ neutralization superflous as it pertains to more than just /æ/. Nardog (talk) 01:22, 16 April 2020 (UTC)[reply]

I'm cool with everything you said above. Wolfdog (talk) 13:09, 16 April 2020 (UTC)[reply]

What are some pairs of words that rhyme except in New York and/or Mid-Atlantic accents (other than the ones already mentioned in the notes)? Nardog (talk) 17:26, 15 April 2020 (UTC)[reply]

Lass/mass, tanner/banner, mannish/Spanish, etc. These kinds of pairs are fairly forced for the most part. Wolfdog (talk) 20:46, 15 April 2020 (UTC)[reply]

Is it contrastive only before /m, n/? Nardog (talk) 22:02, 15 April 2020 (UTC)[reply]

No, it's contrastive before all the possible phonemes that lead to tensing. However, a lot of the contrasts invoke nuanced rules I haven't even bothered to include here ("exceptions to exceptions" I remember reading somewhere the last few days); for example caf (short for cafeteria) is lax because it's an abbreviation, but laugh follows the normal tensing rules; also, halve and have are a minimal pair, but the possible pool of words ending in nonprevocalic /v/ is minuscule and I decided probably not worth adding to this already detailed table; words like national can be mixed; words learned later in life tend to use a lax /æ/ regardless of the following consonant; and so on. Wolfdog (talk) 00:38, 16 April 2020 (UTC)[reply]

Then isn't it misleading to suggest tensing in New York and Mid-Atlantic accents depends on phonetic environments, which the table currently does? But given the rarity of the exceptions, I'm beginning to think this might actually be the place in which footnotes help more than splitting cells. Nardog (talk) 01:30, 16 April 2020 (UTC)[reply]

That works for me. The tensing does largely depend on phonetic environments, but footnotes can eventually fill in the many (if rare) exceptions, sure. Wolfdog (talk) 13:09, 16 April 2020 (UTC)[reply]

It suggests it invariably, not largely, depends on enviroments. Nardog (talk) 02:49, 18 April 2020 (UTC)[reply]

Isn't /æ/ before /ŋ/ raised along with /æɡ/ in western accents, especially Pacific Northwest? Nardog (talk) 20:53, 15 April 2020 (UTC)[reply]

Sounds right, but I didn't bother to tackle that consonant right now. I can in the future. Wolfdog (talk) 00:38, 16 April 2020 (UTC)[reply]

So according to a figure on Labov 183, /æ/ before /ŋ/ is: noticeably high in GenAm (including the West, Midland, etc.) and Canada; it's moderately high, though still basically on par with the other environments, in the South (though there is no accounting for the two Southern systems) and Great Lakes; and it's noticeably low in NYC and Philly/Balty. Wolfdog (talk) 00:30, 17 April 2020 (UTC)[reply]

Where does ANAE say New Orleans follows NYC and Baltimore follows Philadelphia? Nardog (talk) 06:26, 18 April 2020 (UTC)[reply]

Labov discusses the City of New Orleans dialect on p. 260. Baltimore's on 238 and 239. Now some questions/concerns for you. What happened to /dʒ/ (ok, I see it now as a footnote; I'll think on that... maybe that is for the better)? Is there really any evidence that Pittsburgh deserves its own column? I know you're not aesthetically pleased by the overstuffed GenAm column; maybe we can just split that column according to the footnote... [æ], [ɛə], and [eə] (or [ɪə]), perhaps. Honestly, though, the best option seems to be to leave Pittsburgh and GenAm as one unified column. Why does the /g/ example words column only have one example now? /æ/ before /ŋ/ doesn't tend to have the inglide of other raised allophones (if anything, it may have an upglide); I'll try to find sources confirming that. Wolfdog (talk) 14:40, 18 April 2020 (UTC)[reply]

Magic, imagine are not non-prevocalic (or "closed" in their terminology). I know you're not aesthetically pleased by the overstuffed GenAm column I don't know where you got such an idea... I just followed ANAE p. 183: "The Inland North, the Midland, the South, the Mid-Atlantic States, and western Pennsylvania show that the vowels before the velar nasal are significantly lower than before the apical nasal" (which means I should have split Midland too). I don't think we need more than one example of words that don't differ after /æ/. Nardog (talk) 07:07, 19 April 2020 (UTC)[reply]

Did you get my replies above (at 02:49, 18 April)? I have a hunch you didn't given you added sources when I'd suggested reducing them! Limiting the references to ANAE (or at least works of its co-authors from around the same time period, i.e. bar Baker et al.) also allows us to offer a snapshot based on a specific source of information rather than a comprehensive chart, which will never be complete and can lead to a creep of disparate sources.

But more important, I'm looking for a way to get rid of the lengthy note.

Also, how about indicating only the nuclear qualities? That would free us from accounting for Southern breaking, diphthongization vs. lengthening, etc. Nardog (talk) 20:03, 19 April 2020 (UTC)[reply]

Ah... no, I did not see those responses! It's true that exact phonetic transcriptions are rarely given in ANAE or other sources, and yet realizations certainly differ by dialect, context, speaker, etc. (with some definite patterns having emerged). I'd be less than happy whittling everything down to just the nuclear qualities, given all the research I've been putting into this, but that's just a personal feeling and maybe the nuclear-only idea is indeed for the best if you feel unconvinced about exact phonetic notation. (For instance, I believe I saw one source openly admitting to the range of possibilities in the Great Lakes area, verbatim including the three representations [ɪə], [eə], and [ɛə] in their description.) I basically find your ANAE-only idea compelling (though I do think if some more nuanced source, say a more intensely regionally-focused one, finds evidence that overwhelmingly contradicts ANAE, that's nothing to sneeze at; I'm not sure if that's the case though anyway, except a general trend of speakers nationwide to be moving toward a [pre-]nasal /æ/ system). What other loose ends remain? What's your main suggestion for being rid of the lengthy note (I assume you mean the unwiedly section of notes about Philly, NYC, etc.)? Wolfdog (talk) 15:19, 20 April 2020 (UTC)[reply]

Oh, and by the way, I wonder if /d/ deserves its own row or two, since, within the areas where the quality of /æ/ is... more mid before /b, d, ɡ, dʒ/, it's really "before /d/" that appears categorically as mid (that is: equal to or higher than all other non-prenasal phonetic environments) across all of ANAE's charts for regional speakers in this camp, including those from Columbus (Midland), Boston, and Pittsburgh, plus (in non-ANAE sources), New Jersey and Cooperstown, NY. Other than briefly noting this phenomenon with /d/, however, the ANAE doesn't dedicate a lot of writing to it. Wolfdog (talk) 15:38, 20 April 2020 (UTC)[reply]

I was referring to the "Roughly [æ]" idea. Nardog (talk) 17:08, 20 April 2020 (UTC)[reply]

It stands out as different from the rest, but I'm fine with it to simplify the table. Wolfdog (talk) 20:10, 20 April 2020 (UTC)[reply]

@Wolfdog: you feel unconvinced about exact phonetic notation It's not a matter of being convinced, we just can't say anything the source doesn't. Again, is [ə] in [eə, ɛə] in each cell verified? Can it be? If not, we have no standing to keep it. Nardog (talk) 15:01, 23 April 2020 (UTC)[reply]

Again, exact symbols are rarely given, but fairly precise descriptions are given such as "upper mid", "inglide", etc. In some cases, there are certainly sources outside the ANAE that provide symbols, but you wanted to try to stick to the one source. Wolfdog (talk) 12:29, 24 April 2020 (UTC)[reply]

@Nardog: I'm sorry if I missed something, but are we removing [eɪ] from the chart because it's not from ANAE? Wolfdog (talk) 17:38, 24 April 2020 (UTC)[reply]

The question isn't whether the source provides verification in general, but whether [ə] in each cell is verified. Not just that but more because of the inconsistency in singling /ŋ/ (and /ɡ/) out in providing the quality of the offglides. If [ə] is verified then I won't object to restoring [eɪ], but I highly doubt what we currently transcribe as [æ] is consistently monophthongal if the same level of rigor was applied. Which is why I suggested whittling it down to the nuclear qualities. Nardog (talk) 20:18, 25 April 2020 (UTC)[reply]

Well, here's where I feel we're establishing a catch-22. There are non-ANAE sources that establish [ə]; ANAE says "inglide," referring to a movement toward the center. Many other sources confirm that notationwise. And as you could see, I backed up my "singling out" of /ŋ/ with a source (though admittedly not ANAE). I thought it was helpful to distinguish the phonetic qualities; the chart certainly accommodated it in terms of size. If you want to stick to nuclear qualities only, I understand how that's what ANAE shows best. Or, we can incorporate in other sources after all. I think the range of phonetic transcriptions (expanding the use of the tilde) is certainly supportable with a few decent sources outside ANAE. Wolfdog (talk) 00:45, 26 April 2020 (UTC)[reply]

Please add them then. Nardog (talk) 01:03, 26 April 2020 (UTC)[reply]

Australian English

Can't we add it to the table, per Ae tensing#Australian English? Kbb2 (ex. Mr KEBAB) (talk) 06:16, 18 April 2020 (UTC)[reply]

I'd say no. We already have more than enough complications. Once we start including varieties outside North America who knows how big the table is going to be. Nardog (talk) 06:40, 18 April 2020 (UTC)[reply]

@Nardog: Which other non-American and non-Canadian varieties could be included here? I thought only Australia had a (simple, non-phonemic) /æ/ raising system. Kbb2 (ex. Mr KEBAB) (talk) 06:42, 18 April 2020 (UTC)[reply]

Cockney has [ɛɪ] for /æ/ before voiced consonants, especially /d/. This approaches FACE ([æɪ ~ ɐɪ ~ aɪ] in cockney) in production, much like in NAE. Sol505000 (talk) 04:02, 12 July 2021 (UTC)[reply]

More Citations

Lembit Staan: Hi, can you name your exact problems please? You can see that Nardog and I have already had a long discussion about relevant sources above. Thanks. Wolfdog (talk) 00:54, 7 July 2021 (UTC)[reply]

I explained in the edit summary: "... not footnoted, unlike several other similar comments in the table." Please see notes B and D. These statements have footnotes to the sources. I.e., my problem is purely formal. Lembit Staan (talk) 00:59, 7 July 2021 (UTC)[reply]

/æ/ raising before /ŋ/ in New York City

@Nardog, would it be okay if the footnote said "Some New Yorkers may have [ɛə] instead of [æ] before /ŋ/ (...)", instead of what it said earlier? This is closer to the source; my writing "raised [ɛə]" was a mistake. Tyrui (talk) 07:34, 20 August 2023 (UTC)[reply]

Sure, but I suggest we cite something newer and peer-reviewed like this instead of Becker & Wong (2010), which is a working paper. Nardog (talk) 16:15, 20 August 2023 (UTC)[reply]

Presumably, too, this chart is referring to a genericized classic New York accent, which we may want to keep in mind. Traditionally, NYC keeps a tense/lax distinction, but in reality the younger generations are rapidly changing the whole /æ/ situation and, in fact, the whole dialect. Wolfdog (talk) 17:55, 21 August 2023 (UTC)[reply]

canadian /æg/ raising

Moved from Talk:General American English

– Nardog (talk) 05:06, 13 May 2024 (UTC)[reply]

the Canadian column of the /æ/ raising chart seems incorrect to me, specifically the bag, dragon sets (pre-vocalic and non-prevocalic /g/). I read the source (Boberg 2008) and it seems to be to be showing a use of ɛ mainly in the prairies, not in "ontario, quebec, or the maritimes" which imo is specific enough to not be the best transcription for a broad general Canadian such as in this chart. I don't feel comfortable changing it because I know there are a lot of regional Canadian accents and I might stand to correction but anecdotally I've never heard this pronunciation in my life. (note this is distinction from leg, egg as /ejg/ which IS certainly common). Does anyone have any other sources on this? Phantasmofaman (talk) 19:45, 9 May 2024 (UTC)[reply]

What's the transcription for what you impressionistically hear? Wolfdog (talk) 12:09, 13 May 2024 (UTC)[reply]