American Ash

First Page of BeowulfIt’s time for us to talk about ash.

‘Ash’ refers not to the product of burnt charcoal, but rather the ‘short-a’ vowel symbolized by æ in the International Phonetic Alphabet.  In both the British Received Pronunciation and General American dialects of English, this vowel is found in words like ‘trap,’ ‘cat,’ ‘bad,’ and ‘knack.’  The phoneme dates back to the earliest days of the language we call English, conveniently represented by æ in the orthography of Old English.

As indicated by its symbol, ash is pronounced between the /ɛ/ in ‘dress’ and the /a/ in Spanish ‘para.’  In English, however, ash exhibits a lot of variation, to the extent that I find it one of the most relative of vowels.  In older types of Received Pronunciation (i.e. the ‘Queen’s English’), for example, it was arguably closer to the ‘e’ vowel that modern Britons use in ‘dress.’  In the trailer for the film Atonement, Keira Knightley emphasizes this feature in the line, ‘Come beck to me.’  In many modern British accents, we see the opposite: ash has moved toward /a/ or sometimes toward the vowel in ‘strut.’

America doesn’t quite exhibit the diversity of accents and dialects that the UK does.  I’ve come to accept that.  In terms of ash, however, I think we have may give England a run for its money. Just take a look at the linguistic landscape. In Northern states like Wisconsin and Michigan, extreme accents may render ‘fad’ so that it is homophonous with British RP ‘feared (i.e. IPA fɪəd).  In some Southern accents, ash may become a triphthong: ‘bad’ becomes ‘bay-uhd’ (IPA bæjəd).  In New York and Philadelphia, this vowel has split into a breathtakingly complex pair of phonemes.  Over in California, meanwhile, young people with the most advanced features of the California Vowel Shift may pronounce this with a more ‘cardinal’ /a/ vowel, so that ‘tap’ overlaps with the way many other Americans pronounce ‘top’ (IPA tap).

It’s worth noting that in American English, ash might almost be thought of as a ‘long’ vowel.  For the novices out there, that means most Americans pronounce ‘bad’ with a vowel greater in duration that the vowel in ‘bid’ (or other short vowels).  One study* found that ash is, on average, pronounced longer by American English speakers than any other vowel, at least in their Midwestern sample.  This differs from many British accents, which have traditionally been thought to have a short vowel for ash.

Likewise, ash is often described as ‘tenser’ in American accents.  ‘Tenseness’ is a difficult concept to describe (and subject to a lot of debate), so I’ll offer myself as an example here.  In my own accent, my tongue, jaw, and lips are fairly taut and rigid when pronouncing this vowel; when I imitate someone who speaks British RP, on the other hand, I instinctually relax these muscles.  Why we Americans tend to be more laborious with ash is unclear.

Then there are the many quirks of American ash, which are too numerous to fit into one post.  There’s the way Americans pronounce ash before nasals with a raised vowel, as in ‘Pam’ and ‘pan;’ the way many Westerners rhyme ‘bag’ with ‘vague;’ the way older New Englanders maintain a slight hint of the British ‘bath’ vowel; the way some Southerners rhyme words like ‘can’t with ‘ain’t.’  Ash is a vowel that’s all over the map.

What is your ash like?

*Hillenbrand, J., Getty, L. A., Clark, M. J., & Wheeler, K. (1995). Acoustic characteristics of American English vowels. Journal of the Acoustic Society of America, 97, 3099-3111.


About Ben

Ben Trawick-Smith launched his dialect fascination while working in theatre. He has worked as an actor, playwright, director, critic and dialect coach. Other passions include linguistics, urban development, philosophy and film.
This entry was posted in American English and tagged , , . Bookmark the permalink.

36 Responses to American Ash

  1. Charlie says:

    Also, raised monophthongal ash is a feature of some varieties of African American Vernacular English. Making /æ/ somewhere up near /ɛ/.

  2. Thomas says:

    I didn’t think you cared 🙂 Well mine does tend to be longer in duration than what I typically hear in the British Isles and slightly raised [æ̝ ː] in most environments. When it comes before nasal consonants it is slightly more raised and possibly diphthongal [ɛ̞ˑə]. Raising it much higher than that before nasals sounds feminine to me, so maybe this is a gender difference, where I’m from at least. Raising it higher in general, especially along with diphthongization, sounds “Inland northern”.

    Sometimes I seem to get it confused with DRESS pre-nasally. For example, I pronounce vendetta as if it were spelled vandetta and no one around me has ever said anything about it. But its weird, because if I were saying it with DRESS in the first syllable then it would sound like vindetta. Maybe I learned this word from people who didn’t have the PIN-PEN merger and I was trying to imitate their pronunciation of the word. I have no idea. There are lots of other weird little things like that in my idiolect though. In catch I have DRESS, like some Irish people according to what I’ve read.

    Before the velar nasal, as in bang, my ash reminds me a lot of my FACE vowel. I don’t know if it actually is the same or if I just hear it that way. I’m from the Midland region (page 30/146 here) if that means anything. Yeah, not terribly interesting I know. Okay, nap time!

    • Ellen K. says:

      I find it interesting that you (Thomas) say that “Before the velar nasal, as in bang, my ash reminds me a lot of my FACE vowel.”. Describing myself, I would say that is is the face vowel, rather than an ash. I can’t vouch for nuances of how I actually pronounce it, but I definitely think of it as the FACE vowel, and am surprised to now discover that according to dictionaries, it’s an ash.

      I too am from the Midland region (well, sort of), and also have the DRESS vowel in ketch.

      Seems like before ŋ (ng) in English there’s only 4 vowels, ing, ang, ong, ung. With e (Beng) appearing in Chinese names but not in actual English words (blends with -ang too much.).

      • Dw says:

        Length, lengthen, lengthy, strength, strengthen, ginseng.

        • Ellen K. says:

          Curious that those two words (not counting the one from Chinese) both have th after the eng.

          They do not, for me, have either the DRESS vowel, or the eng. The have the same vowel as BANG for me, and a regular n.

          So, make that 4 or 5, with one (where 5) being rare.

    • trawicks says:

      I merge ‘bang’ with ‘face’ as well. I’m like you in regards to ‘catch’ as well–it’s ‘ketch’ for me. Raising of ‘trap’ in the environment of a velar is still a salient feature in Belfast, BTW, which means that accent has perhaps the most allophonic variation of TRAP of any English accent: ranging from [ɛ] to [ɒ]!

      • Diarmuid says:

        Everything I’ve read says it ranges even further: from [ɛ] to [ɔ]. I can’t believe speakers think of those as the same phoneme! What’s amazing to me is if you look at a word like man and how it can be pronounced in English as a whole, the Inland North of the US and Boston would be at one extreme with [me̝ən] and Belfast vernacular at the other with [mɔˑən]. Everywhere else is in between. Even non-native speakers would probably say it in between those 2 extremes.

  3. IVV says:

    My ash is pretty GenAm standard, as best I can tell. It’s an open unrounded fronted vowel. I come from California, and the opening of /æ/ to /a/ is more coastal than inland, I find. Or more affected, for the “valley girl” sound.

    Although is the “short o” sound of “top” really /a/ and not /ɑ/? Or something more centered?

    • trawicks says:

      “Short o” in American English varies a lot, even within the same city! I’d say the most common realization is a slightly advanced and laxed /ɑ/. I find that this occasionally ranges to /a/ in many speakers, but it’s much more consistently [a] in the Upper Midwest, Upstate New York, and some parts of Western New England.

  4. Amy Stoller says:

    I think you’ll find that in some NYC accents phonemic ash has a three-way phonetic split: something close to the “pure” IPA vowel before voiceless consonants, and two different diphthongs – one before voiced consonants except /n/ and /m/, and one before /n/ and /m/ – /ŋ/ being a special case.

    It’s my impression that all or nearly all Americans have some variety of different ash sound before /ŋ/, though this special sound differs from region to region. I would have thought this a natural result of velar “pull”, except that I have heard [æ] or even [a, ɑ] before /ŋ/ in words like “thanks” in some accents of the British Isles – so perhaps not.

    I’ve been told my ash is low, but that was by someone who has a diphthongal ash in all environments. I think mine is actually pretty ashy – that is, fairly close to the IPA “pure” vowel, though I sometimes have a slightly diphthongal quality before /n/ (not /m/), and always a slightly raised sound (diphthongal? probably, but I can’t identify it) before /ŋ/.

    This brings to mind what I think of as a related quirk: the tendency of some midwesterners to have a very low /ɛ/ (and perhaps a slightly raised /æ/?) before /r/ so that Garrison Keillor’s joke about Harold Starr, editor of the Herald Star treats Harold/herald as genuine homophones.

    • IVV says:

      I think the Harold Starr joke is that Midwest (and West, like mine) accents don’t have /ær/ at all, but only /er/ (“air”) or /ɑr/ (“are”).

      • Peter S. says:

        How about in “pirate” and “tyrant”? Don’t some Americans pronounce that close to /ær/?

        • IVV says:

          I think most of Midwestern and Western accents will say /aɪr/. I know I do.

          Part of the issue is that I can’t think of any English syllables that end in ash, with the single exception of the word “Yeah.”

    • trawicks says:

      I’ve actually noticed something of a low ash among New Yorkers in the ‘lax’ environments. Although I find that this perception results from New Yorkers retaining a lax vowel in pre-nasal environments, where most Americans would use a tense vowel. New York ‘can’ (as in “I can do it”) SOUNDS very low to me, but that’s because I’m so used to hearing the raised vowel typical of American English.

      There also seems to be a gender discrepancy between men’s and women’s ashes: I typically find women’s lower than men’s (at least in GenAm).

      • Geoff says:

        “I’ve actually noticed something of a low ash among New Yorkers in the ‘lax’ environments.”

        Yes, I’ve noticed that too. In the “lax” environments, it sounds way too open to me and in the “tense” environments, it sounds way too close. My ash must be somewhere in between the two.

        Wells, whose name is mentioned here every day it seems, does mention a quality of [aː] or [aə] for New York City on this page of his Accents of English along with other qualities. And you can also look at the map on p. 82 of the Atlas of North American English (blurry version). In case you can’t read it, it’s showing the F1 of /æ/ and for the NYC/Mid Atlantic region it’s showing the F1 of the lax phoneme. There are many blue and green circles (highest F1 values) in that region, reflecting the very open quality of the lax phoneme there. They extend to Providence too, but there is a different short-a system there.

        “Although I find that this perception results from New Yorkers retaining a lax vowel in pre-nasal environments, where most Americans would use a tense vowel.”

        I’ve also spotted many a Canadian by listening to that (see blurry map #2). That map, on p. 84 of ANAE is also showing the F1 of /æ/, but this time only before nasals. The blue circles once again represent the lowest F1’s. Notice that the Baltimore-New York corridor once again has many blue circles. Luckily I have access to ANAE online through my college library, so I can see the “unblurry” versions 🙂

        “There also seems to be a gender discrepancy between men’s and women’s ashes: I typically find women’s lower than men’s (at least in GenAm).”

        I’ve noticed this in some places, but in Chicago, for example, females have closer ashes than males (p. 3).

        • Geoff says:

          My previous comment had several links to other web sites in it, so I hope it doesn’t get blocked.

        • Geoff says:

          I said: “The blue circles once again represent the lowest F1′s.”
          I meant: The blue circles once again represent the HIGHEST F1’s.

        • trawicks says:


          I should qualify my earlier statement by saying that I think the gender situation changes when you’re talking about NCS speakers. I might posit that it was not until very recently that the Chicago accent experienced the kind of stigmatization that the NYC or Boston accent did, so it may not have developed the working-class-woman-vs-working-class-men divide you see elsewhere.

        • Geoff says:

          I don’t think the Chicago accent experiences any kind of stigmatization.

    • Amy Stoller says:

      Yes, I think my problem with Harold/Herald is that I preserve the merry-Mary-marry distinction, as most midwesterners do not. Please forget I mentioned it! 🙂

  5. CaitieCat says:

    That “bag/vague” rhyme is also common around this area, Kitchener/Waterloo in Ontario, Canada. Both my kids have it very strongly, as do many of their friends who’ve grown up here. It’s especially strong before voiced consonants; I think this is somehow related to the fairly standard vowel-lengthening, but for some reason it also seems to raise as it lengthens.

    The area has a very strong German influence (Kitchener was Berlin until a fit of patriotic spasm during WWI), and I wonder if it’s a holdover from German ä? Not sure. I’d love to have the support to study the local variety, as it’s showing some neat little changes from the standard “Southern Ontario” variety that’s common from Oshawa to Windsor.

    • AL says:

      The only person I know who rhymes “bag/vague” is from Calgary, so I always assumed it was a Canadian thing.

  6. Peter S. says:

    In bag, bang, and bank, my /æ/ turns into a diphthong /æɪ/. It’s not quite /aɪ/, so dagger doesn’t quite rhyme with tiger, but it’s nearly there. I get this from my mother, who is from the Midlands region (downstate Illinois).

  7. Margaret S. says:

    Hey, I grew up in Kitchener-Waterloo, too. I suspect the “bag/vague” rhyme is a younger generation thing; it sounds completely unfamiliar to me.

    I’ve long been fascinated by the upstate New York “ash” which I know from Buffalo-Tonawanda speakers. I haven’t heard similarity to it in any other American regional accent, but that could be because I don’t know enough about them. Comments from others who know American variations better?

    Examples can be found in ; listen for “have” at 0:54 and “back” at 1:19.

  8. Margaret S. says:

    Link for example of Buffalo-Tonawanta speakers: On Youtube: /watch?v=2ViqiO-860o

  9. Ellen K. says:

    There’s two issues, it seems to me. What is and isn’t part of this phoneme (marry, catch), and how we pronounce the phoneme.

    Alas, the latter is harder to discuss.

    • Thomas says:

      Oh I know. I don’t pronounce every word in the TRAP class the way I pronounce catch. If that were the case, I’d have DRESS-TRAP merger. I just mentioned it because I’ve found that my pronunciation often isn’t listed in dictionaries for some reason, making me think that I’m weird and/or wrong for pronouncing it that way. But I’ll continue to pronounce it that way anyway.

  10. Thomas says:

    *…a DRESS-TRAP merger…

  11. dw says:

    I grew up speaking near-RP, but I also have (or had) a (possibly) phonemic split in the TRAP set. For me, it was phonetically a matter of pure length, with a few members of the TRAP set having a long vowel [æː] as opposed to my usual short vowel [æ].

    Words with the long vowel included “bad”, “man”, “mad” and “family”. It seems to coincide mostly with what Wikipedia reports as the “bad-lad split” (I’m not putting in a link to avoid moderation). However, I think I’m gradually losing this split as I get older, possibly because I now live in the US, with all members of the TRAP set reverting to the short vowel.

    • Sam says:

      That’s interesting, because the American vowel is usually longer than RP vowel, with the exception of the words you mentioned. See Wells p. 246.

      • dw says:

        Yes: I think I can guess what happened. This particular distinction was never very firmly embedded in my personal phonology, and, cut off from regular interactions with others who might have shared the same distinction, it has withered away.

        Since the short vowel has always been the unmarked pronunciation for me, it’s not surprising that the few words with long vowels are being shortened.

    • Sam says:

      I’m referring to actual (phonetic) length though.

  12. Winnen says:

    I mostly use [a], except before nasal and voiced velar consonants where I use [æ], and before voiceless velar consonants where I use [ɛ].

  13. AL says:

    ” Over in California, meanwhile, young people with the most advanced features of the California Vowel Shift may pronounce this with a more ‘cardinal’ /a/ vowel, so that ‘tap’ overlaps with the way many other Americans pronounce ‘top’ (IPA tap).”

    Where in CA do people do this? I grew up in the Bay Area and never heard anyone pronounce tap like top!

    • Evan says:

      They don’t pronounce tap like top, i.e., they don’t merge those 2 vowels. Some of them may pronounce tap in a way that approaches how people from the Great Lakes area say top [tap]. But I don’t know if the majority of Californians do this and I don’t even think the ones who do have this shift use an actual [a]. I think it’s more like [æ̞] or [æ̠] ([æ] with a lowering or retracted diacritic, respectively). But that’s okay.

  14. m.m. says:

    My /æ/ before oral consonants ranges from [æ] to [ä] [with g>d continuity, compared to regular d>g continuity; retracted most before fricatives, so: bag>bad>bath]; before /n/ and /m/ it ranges from [æə] to [eə]; before /ŋ/ ranges from [ɛə] to [eɪ].

    I’m with IVV that the lowering of /æ/ is definitely more of a coastal than inland thing, and with Evan that it’s definitely not currently a majority feature, but lowering isn’t noticed/stigmatized either.