[Update: I made a few slight revisions to this post based on feedback.]
(NOTE: This post uses the International Phonetic Alphabet (IPA). For information about the IPA, please visit my page of IPA Resources.)
What is the #1 thing American actors screw up doing British accents? Four little words: The Trap-Bath Split.
Let me explain. In most American accents, we pronounce trap, cat, and bad with roughly the same vowel as ask, can’t and laugh. This is what we refer to as the short-a sound.
In many British accents, however, these are two separate vowels: the first group of words (i.e. trap) are pronounced with a short-a, (IPA æ) as in America ; but the second group (i.e. bath) is pronounced with a broad-a, IPA ɑ: (i.e. “ah“). Hence the Trap-Bath Split.
To put it in crudely, Americans would say The cat took a bath so that cat and bath are pronounced with the same vowel. For many British people, however, only cat would be pronounced with this vowel; bath is pronounced with the same broad-a vowel as father or palm.
This is tricky for Americans to master. There are no easy rules for which words fall into these two categories. For example, the word chant is pronounced with broad-a in Standard British English (IPA ɑ: or chahnt), but the word ant is pronounced with the short-a.
That is why I have watched many great American actors do flawless British accents until they they let loose an American pronunciation of last or can’t. Ouch.
So, American actor, how can you avoid being a victim of the Trap-Bath Split? Well, first things first, make sure the dialect you are doing actually has the split.
A number of British accents, contrary to popular belief, do not feature the split. And furthermore, the split is used in accents outside the United Kingdom. For handy reference, then, this is a list of accents with the split and those without:
Accents with the Trap-Bath Split:
- Received Pronunciation (Standard British)
- Cockney/London English
- Australian English*
- New Zealand English
- South African English
- Old-fashioned New England Accent (Down East)
Accents without the Trap-Bath Split:
- Nearly all American and Canadian Accents
- Scottish English
- Northern English Accents
- Carribean English (usually)
To make matters more confusing, there are some dialect regions where it’s mixed. Some speakers have the split in these places, some don’t. I would say these are:
- Midlands English (Birmingham, etc.)
- Welsh English
- Irish English (depending on many factors)
- Boston English (still a feature in some very working-class dialects)
My rule of thumb? If you’re playing a character who speaks with Received Pronunciation or of a character from the London area, always use the split. With pretty much any other region in England, do research that is as specific as possible. It can really vary in this day and age.
When you have confirmed that your character indeed has a Trap-Bath Split accent, such as Standard British (RP) or Cockney, here are a few pointers:
1.) Understand which words generally fall into the Bath category. Here are the types of words that usually are pronounced with this broad-a:
- -aff: staff, chaff, etc.
- -aft: daft, after, draft, etc.
- -alf: half, calf, etc.
- -ample: example, sample, etc.
- -ance: dance, lance, etc.
- -anch: ranch, branch, etc.
- -ans: answer, etc.
- -ant: can’t, chant, advantage, etc.
- -aph: graph, etc.
- -as or -ass: ask, trespass, grass, etc
- -ath: bath, path, etc.
- -augh: laugh, etc.
- -aunt: aunt, etc.
Again, there are many individual exceptions to the list above. The important thing is to know where the broad-a tends to appear in British English, not to memorize every single word that is pronounced like this.
2.) If you’re having trouble mastering the split, circle every instance in your script where these kinds of words appear.
3.) Get your hands on a copy of the Longman Pronunciation Dictionary. This is an indispensable volume of British and American pronunciations edited by British linguist John C. Wells. For trap-bath words, it clearly marks whether British English pronounces these words with a æ (trap) or ɑ: (bath). (You can also use dictionary.com, but I prefer the Pronunciation Dictionary because it’s portable, back up by decades of research, and more complete in general).
On a side note, you may be wondering where, exactly, this split comes from? And why do Americans not have the split? Unfortunately, I have no clue. The split seems to have fully developed in Southern England some time in the early 19th-century or thereabouts. That explains why Americans don’t have it: most British immigration had dried up by the time it emerged**. But like a lot of phonological changes, there is unlikely to be a clear, logical explanation.
Trust me, I know it’s hard. As an actor, I still looked up trap-bath words long after I had mastered all the minutiae of certain accents. The rules are so illogical that it requires constant refresher courses.
The bottom line is, the Trap-Bath Split is one of the biggest Achilles heels in learning RP or similar British accents. Without mastering it, it is a dead giveaway that an actor isn’t a Brit!
*As a commenter kindly pointed out, the split is somewhat incomplete in Australia. I would say it’s widespread enough, however, that you can put Oz into the split category.
**Actually, the split began to occur before this. It began with certain vowels being “lengthened” and it was only later that these vowels become the “long-a”. For what it’s worth, it is often suggested that the “tense-lax” split in New York City is a relative of the Trap-Bath split (But that’s a post for another day).
Very informative post; thanks. A quibble, though: how does using the TRAP vowel where the BATH vowel is appropriate qualify as a “nasal” pronunciation? With or without the TRAP/BATH split, the vowel in “can’t” is nasal (assuming the speaker’s nasal passages are clear), and the vowel in “last” is oral (unless it’s a speaker who habitually pronounces vowels nasally).
Neal, you’re totally right. Even I am susceptible to the American linguistic inferiority complex occasionally, and I let slip the old canard about American accents being inherently nasal or whiny. Scratched it out.
With or without the TRAP/BATH split, the vowel in “can’t” is nasal (assuming the speaker’s nasal passages are clear)
It doesn’t have to be. It is quite physiologically possible to keep the velum elevated during the vowel sound and to drop it simultaneously with the alveolar closure of the /n/.
My (roughly near-RP, at least in this respect) speech may not feature perfect synchronization of these movements, but I would claim pretty confidently that my vowel in “can’t” is non-nasal for the majority of its duration.
Stereotypically, RP has a very low degree of nasalization of vowels: far lower than that of much American speech
I’ve thought of a more felicitous way to express what I said in my last comment (unfortunately I can’t seem to delete it).
In my accent (and, to the best of my knowledge, in RP in general), my articulatory targets are always non-nasal vowels. There may sometimes be nasalization adjacent to a nasal consonant because of mechanical failure to achieve those targets. But such nasalization would only be in the durational part of the vowel immediately adjacent to the nasal consonant, not spread through it.
In most American accents that I have heard (I live in California), there appear to me to be nasal targets in many vowels, and very often distinctive pre-nasal allophonic vowel qualities (for example, raised to [e] or thereabouts in words like “can’t”). As far as I know, my accent (and, to the best of my knowledge, RP) lack such distinct allophones.
This can lead to the perception in Britain that American accents are “nasal”, and in this particular instance I believe that the popular perception does rest on a grain of truth.
I’ve been trying to post here but it hasn’t been working. This is a test.
Sorry about that, ‘enry. I use the standard issue wordpress spam-blocking tool (Akismet). It read your earlier comment as spam for some reason. I promise I’m not blocking you!
It’s working! Wonderful! I don’t know what was wrong before.
Hello, I’ve been reading your blog and I really enjoy it. I’m not trying to be a know-it-all here, but AFAIK it isn’t only rural Northern English accents that lack a TRAP-BATH split. It’s really Northern English accents in general. The vast majority of people in the North of England don’t have the split. But there are some very upper-class people who may have the split. You said upper-middle class, but I think it’s mostly upper-class.
Also (and I’m sure you know this) there are words like calf, half, rather, can’t, banana, etc. that have the “broad-a” everywhere in England. So you’ll hear people from Northern England who otherwise don’t have the TRAP-BATH split using the same vowel as PALM in these words. J.C. Wells says of words like these (and I agree with him on this), “if we were considering their pronunciation in England alone, and leaving North America out of account, we should place half, can’t, banana, etc. in the PALM set rather than in BATH.” It may be a good idea for the actor to simply memorize this list of words, because it isn’t that long of a list and it will be useful if they’re doing any sort of English accent.
Once again, I love the blog and I hope you keep it going.
Thanks for the kind words! These are some really good points.
You know, I was debating whether or not to put Urban Northern English accents in the “mixed” category. I decided to put it there based on one or two studies that suggest it’s on the rise, but you’re right, most of those cities don’t have the split. And I’m not going to lie, I sometimes use “upper middle-class” as a gentle euphemism for “rich.”
Actually, if I were to rewrite this post I might recommend focusing on just the areas that feature the complete split, and leave other dialects out of it entirely. Because of course, TRAP-BATH is pretty easy stuff compared to such complex a-splits as Belfast, Scottish Vowel Lengthening and New York’s tense-lax split. Another confusing aspect to this is that within that smaller sub-set of English “broad-a” words you mentioned (“can’t,” “half,” etc.) is one word, “rather,” which is actually broad for some American speakers (although I’ve never quite figured out what the regional distribution of this is, if any).
Not to mention that 95% of the time, when an American needs to do a “British accent,” it is probably going to be Cockney, RP or some other Southern dialect. Sadly, I don’t think I’ve ever seen a play produced in America that takes place in Birmingham or Manchester or really anywhere in the Midlands or North.
I suspect that has its counterpart in that an “American accent” in a UK show or movie is going to be based on “The Sopranos” New “Joysey,” West Coast sit-com Californian, or an amalgam of several “Deep South” drawls. Oh, and “I’ll be back” doesn’t count as American 😉
Part of that is because, while both America and the UK have plenty of dialects, they certainly aren’t spoken by equal amounts of people. GenAm and Estuary are probably spoken by hundreds of times more people than the classic Charleston accent or the Northumberland Burr. Dramatic literature reflects this. (Although there was a popular Northumberland-set play, The Pitmen Painters, which came to Broadway last year, so maybe I shouldn’t speak so soon!)
“Another confusing aspect to this is that within that smaller sub-set of English “broad-a” words you mentioned (“can’t,” “half,” etc.) is one word, “rather,” which is actually broad for some American speakers (although I’ve never quite figured out what the regional distribution of this is, if any).”
Thanks for replying. I don’t really know what the regional distribution is either. My guess would be that it has more to do with education and social class than which region the person is from.
You’ve got a rogue “your” instead of “you’re” in there. (Sorry- I always feel snobby for pointing these things out but it just reads wrong.) Also, the “aff” in BrE “chaff” is definitely always a ‘short a’! And I’d agree that no Northern English dialects have the split. The Welsh thing is interesting as well. I come from North East Wales and in my area people generally don’t have the split. I’d say that it it’s more of a South Wales thing (although I think there are N. Wales areas that have it too but not certain). Anyway, I’ll stop rambling now. Thanks for adding to my geeky hobby of linguistics blogs!
Wells in the LPD actually gives broad A as first choice in “chaff”. But I would agree with you that such a pronuncation would sound odd today. People don’t talk about chaff much these days, so the broad A has been lost.
Yeah, the inclusion of ‘chaff’ in this list brought me up short too. If I heard someone say tʃɑf I’d assume they were trying to mimic the accent and getting it wrong.
Chaff is listed in most dictionaries with the “broad a” given as an alternate pronunciation. However, it’s such an uncommon word that it’s almost a moot point. I don’t think I’ve used chaff even once in my lifetime!
Good stuff, but a couple of important caveats if I may.
First, I know it’s a hackneyed complaint, but why must you use the words accent and dialect interchangeably, when your blog is clearly about accent?
Can I suggest that Americans should be very careful with the word “class” in a British context. If you mean rich, why not just say rich? “Upper-middle class” has a particular meaning, which we might debate the fine points of, but it’s nothing whatsoever to do with “rich”. Class doesn’t have much, if anything nowadays, to do with money, and there is no such thing as “rich” accent in the UK.
Another weird unexplained assumption is that either rich or “upper-middle class” people are more likely to live in the city than in the country, hence, presumably, the fanciful urban/rural split in “Northern England” — that’s just mystifying!
My guess is that most of the time when people say “the British accent” they mean RP. American actors surely aren’t expected to go into any fine detail: RP, London, North of England, Scots and possibly Welsh and Cornish would surely cover 99% of all needs?
I’m sure it’s true that TRAP/BATH is an Achilles heel for Americans trying to speak RP, and of course equally for RP speakers trying to speak American. Another, I’d say, is non-prevocalic R and the linking R that English accents both rhotic and otherwise are so fond of. Perhaps you were coming to that in a future post. Keep up the good work!
Thanks, Harry. You make some really excellent points here which will help me moving forward. I actually just revised the post to remove the bit involving “Urban Northern English” entirely. To be honest, I based that on a single, probably insignificant study I read years ago that I can’t even locate online, so I’m going back to the traditional assumption that there is no TRAP-BATH split in Northern England.
Regarding “accent” and “dialect:” point very much taken. I’ll be a bit more exacting with those in the future.
Actually, you quite often come across an RP speaker from the North of England who doesn’t have the TRAP–BATH Split. Which I suppose means they can’t technically be an RP speaker, but they are in every other respect. There is the term “near RP” but it’s not very helpful here, in that I’m talking about every other feature qualifying as RP except for this one phonemic exception: it’s completely different to having a very slight regional accent which could almost pass for RP. Here’s an example: Matthew Bannister, presenter of the (excellent) BBC Radio 4 obituary programme Last Word.
In practical terms, I would tend to say that an actor could probably get away with NOT bothering to master the TRAP–BATH Split, as long as the quality of the /æ/ phoneme was correct.
I’d also say (speaking off the top of my head) that the nature of the split has changed over the years, and that older styles of RP perhaps use more TRAP and less BATH. So if anyone even noticed the flaw in a perfectly-faked RP minus the Split (and I do wonder how many people would notice that Bannister is not technically an RP speaker unless they stopped to listen carefully), they would probably just perceive it as a little dated or idiosyncratic, rather than unconvincing. Moral: if it’s a problem, focus your energies on other more crucial areas, like rhoticity and vowel quality.
Just a thought.
We actually have an example of this accent in the States, or at least its Midlands variant, in the form of television host Cat Deeley. She is the host of host of “So You Think You Can Dance,” and a native of Birmingham. Her accent could best be described as near-RP with a slight dash of Brummie. Hence when she says the title of her show, “Dance” is notably a short-a, while the rest of her dialect is rather self-consciously non-regional.
I’m sure it’s true that TRAP/BATH is an Achilles heel for Americans trying to speak RP, and of course equally for RP speakers trying to speak American.
Not really. An RP speaker trying to speak American need only learn the membership of the PALM set (far smaller than BATH) and know how to spell. The decision tree for words of the RP-speaker’s merged BATH-PALM-START set is:
* is this word spelled with “r”?. If so it’s in START.
* otherwise, is this word in the PALM set?
* if not, it’s in BATH (so use “flat A” in American accents).
However, an American trying to speak RP would need to know the membership of both the PALM and BATH sets.
Re Australian accents: I wouldn’t say that the split is waning; I would describe it as an incomplete split.
The most widespread accent uses æ for most -ample, -ance and -anch words and with some -ans, -ant and -aph words. The pronunciation of these words varies wildly according to location, age and, most significantly, social class and education. But the rest of those suffixes (and other very frequent words like “can’t”) are solidly ɑ: (or a:) across all social and geographical variations.
There is an interesting chart on Wikipedia’s Australian English Phonology page that shows the distribution of long-a in Australia. Now while there is a citation, we’re talking about Wikipedia here, so proceed with extreme caution. But if it is to believed, it offers a great insight into the trap-bath split. In Adelaide, the great majority of speakers (80%+) feature the complete split. Next comes Syndey, which is mostly split except for -ance words, which are broad for only 30% of the population. Then there’s Melbourne, which is about 60% overall. In Brisbane the split is even less present, and in Hobart, the split is only widespread in -asp and -ast words.
We’re both talking about the table I linked to, right? They cite the Cambridge Encyclopedia of the English Language, which means that we can believe that table. Adelaide certainly has a reputation for long-As (thanks in no small part to Alexɑ:nder Downer). Melbourne’s reputation as a short-A city might be due to the amazingly high rate of “cæstle”, rather than it being true across the board.
But what I wanted to emphasise is that metropolitan area is just one factor. Within Sydney, where I grew up, the division you see in the table is actually variation between suburbs, which correlates strongly with socio-economic status (assume the same goes for Melbourne). This is not something I hear changing significantly across generations (unlike the use of American emphasis, which occurs *much* more amongst younger Australians). Though I’m not sure how you would measure that — as in the UK, you’ll find plenty of rich Australians using working-class accents.
Sorry, not sure why, but I completely missed that link!
Interesting you bring up socioeconomic status in Australia. Here in the States, many of us assume Australia to be a uniformly smiley, middle-class nation, which I’m sure is FAR from reality. Although to be fair, I’ve never considered the socioeconomic dimensions to the Aussie accent. Many of the wealthier Australians I’ve encountered here tend to speak with something akin to near-RP, although I know a rich young man from Brisbane who has an accent so, well, Australian that it sounds almost affected.
Pingback: A Quick Update | Dialect Blog
Nick said: “…unlike the use of American emphasis, which occurs *much* more amongst younger Australians…”
I’m curious to know what you mean by “American emphasis”.
Nick said: “…as in the UK, you’ll find plenty of rich Australians using working-class accents.”
You’ll find plenty of rich Americans with working-class accents too, so I’m not sure how that’s different. Do you mean they’ve been rich their entire lives and they’ve had working-class accents their entire lives despite that or what?
Re American emphasis:
To an Australian ear, “HARRas” sounds antiquated. “reSEARCH” and “comPACT” are still around but “REsearch” and “COMpact” (even as adj.) are far more common, esp. among younger Australians.
But Australians of all ages still say “labOrat’ry” and “LAVat’ry”. (These words are probably not used often enough on TV to influence Australian speakers.) It’s worth noting that words of French origin never seem to use the American final-syllable emphasis either, it’s always “PLATeau”, “BEret” or “SALon”. But I wouldn’t be too surprised if this changed in another generation.
We can’t really do it justice here, but… yes, I would say it’s difficult to tell how rich someone’s parents are from their accent. Of course there’s some correlation but it’s very weak. On the other hand, I have to admit that Australia has nothing as significant as the UK mockney phenomenon (fauxgans not withstanding). I think the strongest correlation you’d find for refined Australian accents (asymptotically approaching the RP) would be education but it’s still very common to meet PhDs with very broad Australian accents (anecdotally, I’d say more common than a regional American accent is in American universities).
The split dates back to the 16th century, not “the beginning of the 19th century” 🙂 Some details are here:
Sort of. I have a footnote about this: the split did, indeed, begin quite early on, but it doesn’t seem to have become what it is today until around 1800, give or take.
You can find a number of early precursors to the split in various accents. The New York/Philadelphia/New Orleans tense-lax split it is one, there’s a similar split in some American Southern accents, and there’s the limited split that you hear among a smaller number of words in Northern and Midlands English. Some early texts from the New England puritans suggest some kind of split as well. But in terms of the full-on split that you hear in RP or Cockney, that seems to have reached its currents state later on.
Of course, it depends on what exactly you mean by “what it is today”: there has always been some variation in exactly which words were in which lexical sets. And the exact phonetic quality of the TRAP and BATH sets has also varied, and indeed still varies today both in time and place.
But there has been a phonemic split betwen “trap” words and “bath” words in the language used by educated people in the London area since the late 17th century. Dobson’s “English Pronunciation 1500-1700”, vol, ii, has more details: it’s cited in the Wikipedia article.
Interesting that you talk about the New England puritans. Since they primarily came from the east of England, beginning with just after the split really got going, one would expect to see it there.
According to John Wells’s “Accents of English”, Caribbean English generally has the split.
Thanks for mentioning that. I think I may just remove Carribean English entirely from the article. The problem is that, while those accents do have the trap-bath split to some degree, they also feature vowel lengthening in other “a” words as well. Jamaican Creole, for example, shows indications of the split, but it also extends the long-a to words we don’t associate with the split, such as “family” and “perhaps.” Likewise, some words that are part of the split in other countries, such as “after” and “advantage” are short-a.
There is a mistake in your diagram in the top left. You’ve described /æ/ as long and /ɑ:/ as short.
I know that it’s a phonetic convention to write /æ/ for TRAP, but those Brits who lack the TRAP-BATH split are very likely to have /a/ rather than /æ/. I think that it would only be in the West Country where you’d get /æ/ for both TRAP and BATH.
The BATH vowel is the most provocative of all vowels. I think that this is because most Brits don’t have the TRAP-BATH split, and they resent how the BBC and other institutions favour the /ɑ:/ pronunciation. John Wells marks the /æ/ pronunciation for “Castleford” as non-RP, but I find this ridiculous when the residents of the town to a person pronounce it with /æ/ and always have done.
Thanks for pointing out that typo. Probably never would have caught that!
^ This convention has infected me. I meant to write that the people of Castleford (in Yorkshire) pronounce it with /a/.
Is the trap-bath split why an Australian who may speak a General Australian Accent could be considered to be British? An Australian who pronounces the splits properly and with the correct sounds on other words.
Although Australia does have the trap-bath split, as you point out, it’s interesting that there’s a sub-class of these words in which there is variation within Australia. Some Australians use the ‘bath’ version in words such as ‘castle’, ‘chant’ and ‘enchant’ whereas many don’t. It’s possible there may be a state division in this regard. It seems to me that many Sydneysiders use the ‘bath’ pronunciation but that in Melbourne it’s more common for these words to be pronounded as in ‘trap’. Either is considered acceptable.
One of the things which I’ve observed is that most
non natives speakers in Europe especially German speakers have something
which I call the France-Francis split in which France is
pronounced with a broad A and Francis with a short a.
Pingback: Dr. Neighbours' Performing Arts Blog » Theatre 2 – British Dialect