General American Revisited

George C. Scott in Patton

I want to share with you a short series of long e-mails between Erik Singer and me on the subject of how to deal with the desire of students to learn some form of “General American.” Erik began the discussion with a thoughtful and thought-provoking e-mail to Phil and me, from which he has given me permission to quote at length:

I’ve been having an interesting year, and wanted to share some thoughts with you. Should either of you have the time and inclination to respond, it goes without saying that I would be fascinated to hear your thoughts in return.

After a year off from institutional/conservatory speech teaching, which I missed terribly, I’ve been teaching speech and phonetics this year in a new one-year conservatory program at HB. The program is in its second year. Ilse is the voice teacher, and has been for been for both years of the program’s existence. We have 18 students this year, half of whom are NNSs [non-native speakers]. Three of the NSs [native speakers] are Aussies, and the rest are North Americans. (Americans really, plus one Canadian.) I have discovered that 18 students is a ridiculously large and unwieldy number for a speech class, and makes the kind of detailed, individual work I like to do very cumbersome and, crucially, time-consuming. More relevantly, however, I have found myself greatly challenged by the NNS/NS division. It’s been wonderful, through the first two-thirds of the year, to have so many different languages in the room (Korean, Japanese, Dutch, Venezuelan Spanish, Mexican Spanish, Iberian Spanish, Catalan—I only wish we’d had more time to delve into the phonetic inventories of all of them). And the K-T approach has been wonderful with this group, because throughout our explorations of anatomy, Omnish, and basic phonetics, the playing field between the NSs and NNSs has been relatively level. As we’ve wended our way into American English, however, I’ve been running smack up against the realization that these two groups have extremely different needs. With the NSs, I am doing what I’ve always done—practicing lots of alternative varieties for individual skills (while continually increasing awareness & sensitivity, articulatory freedom & muscularity, etc.) The NNSs, however, quite legitimately desire something much more prescriptive. What they want, quite desperately, is precisely for me to prescribe a specific speech pattern for them. One that will allow them to sound completely American (and educated, “White,” upper-middle-class, etc.). And, inevitably, not from any particular, identifiable region of the country. In other words, they all want to master an accent that might be described as “General American.” I think that for talented, ambitious actors, this is an entirely legitimate goal. Indeed, if freely chosen rather than imposed, I think I feel this is an entirely legitimate goal for any NNS, or even non-American NS who really aspires to sound as American as possible.. (I wonder whether you both agree, or have reservations?)
Another reason I’m being forced to wrestle very specifically with these questions is that Sam Schwat’s death and the shuttering of his practice seems to be bringing me a fair number of referrals—lots of models and actors from South Africa, Australia, Brazil, Spain, etc., all wanting to perfect their “General” American accents. When they have specific roles that they’re preparing for, of course I coach to that. But most of them just want to find an all-purpose GenAm accent to call their own. And of course, in all circumstances, I have people work from primary source recordings—generally actors they like whose accents fit the bill.
I have been approaching this as an accent coach—as a dialect design job, really—and have been working on and continually expanding a So-Called “General” American breakdown. I have been specifying, in the breakdown and in all my teaching and coaching, which of my recommendations for specific realizations of phonemes are more-or-less absolute (if you don’t do this one, you don’t sound American, or it’s not SCGA because it’s marked and regionally-specific), which are arbitrary, which are whimsical, and which may be subject to change in 5 or 15 years. This feels like the right approach to me, though I realize it is a sort of threading of the needle. It’s been a fascinating exercise, and the breakdown continues to grow alarmingly—it’s now 30 pages and counting. (Indeed I fear it may turn into a book. If it does, it would necessarily include a section on overall philosophy—which would, of course, be very specifically indebted to Dudley’s research and articles and to K-T. And, of course, frequent boldface injunctions that the SCGA accent being described/prescribed is in no way intended to be used as a basis for the prescriptive speech teaching of native AmE speakers!! Would love to hear your thoughts on this. I’m aware that you did away with the Detail Model, Dudley, precisely because of people turning it into a new fixed, standard, pattern. But is there not a legitimate need for the existence of such a pattern? What I’m mainly talking about here is ESL, I suppose, though from a sophisticated, sociolinguistically-informed, accent coach’s perspective. Infinite complexity is simply too much for most NNSs, who have enough to deal with in just getting up to speed with the language.)
This leaves me doing quite a complicated dance in my class at HB, of course. I end up asking for entirely different—indeed, mostly opposite—things from the NSs and the NNSs. When an NNS is totally on top of a particular skill, of course, I will have them try variations. They still want to know which one they should use, though, which one should be the “rule.” They understand completely, I think, at this point in the year, that the notions of what might be “right” or “correct,” especially as regards accent variation, are social constructs built on sand. But they still need to know which choice to make. I am at pains, meanwhile, to make sure that with the NSs, who are in the room for all or most of this, I am underlining the arbitrariness of all of this and the extreme undesirability of “standardizing” their speech in accordance with the prescriptions I am setting out for the NNSs. (The Aussies are a middle group—all three want to work on an American accent, so they’re working with the same breakdown as the NNSs.) It’s quite a juggling act, and I don’t think I am managing it perfectly. I do feel, though, that the whole class is down with the basic project, and clear about what each of them is there for and hopes to get from it.
Whew! Sorry for the length of this. Hope you’re both very very well.

To which, after an unseemly delay, I posted a reply. My comments, and Erik’s responses to them, are below:

Hi Erik,
Well, finally I can get around to at least attempting a very superficial response to your excellent thoughts on the Gen-Am issue. If Phil has a free instant I hope he’ll chime in, to both of us.
I agree with a lot, most really, of what you say and I understand the basic problem that you or any teacher of persons from other dialect areas or languages who want to learn an American accent that will—in a sense—“sell everywhere” will face. Such students want an accent that avoids any stigmatized implications, however stereotypical or contradictory, about education etc.; and one that will also be readily intelligible to any speaker of any accent of English. You’re quite right: if you are working with a non-native speaker who wants to learn an American accent, the obvious question is: which one? Teaching a variety of American accents is far from the first priority for such students.
I have two basic points to make. The first—with which I am sure you will agree—is that the physical articulator skills work that we do in KT-Speechwork is the crucial first step for anyone interested in accents, even the person who wants to travel directly to whatever municipality or farm in the Midwest might conceivably represent that mythical General American.
I certainly do agree. It is the basis for everything I do. Even when I only have an hour or two to teach someone an accent-on-demand, I will include some skill work (often the vowel space and how to get around it). Some corporate clients, alas (I am thinking of one Vietnamese economist, in particular), are resistant to sticking tongues out of mouths, or even touching faces. I do what I can.
If it’s a private non-theatrical client, one can teach the skills work followed by Gen-Am and leave it at that. But if your student is an actor for whom English is not a primary language, the primacy of skills training becomes even greater. The biggest teaching challenge in using this approach is to get everyone to understand that all actors, of whatever ethnicity, need to be able to have the skills to do all accents. An Asian actor not only may need to do a Bronx or other localized American accent at some point in her career, the likelihood is decently good. No actor is served by only learning the Gen-Am and being given the impression that their job is finished. And if they painfully learn Gen-Am first in the curriculum, believe me, they will never go on to the other skills work and will have to cope with new accents with no developed abilities to help them. Actors often want the surprisingly un-instant gratification of going right for the “right American accent,” which is one reason why Skinner is still taught. Un-instant because, as you know, it takes a lot longer to learn it if you don’t have the skills for the job; and teaching it in its un-instant form becomes an indoctrination, not an application of existing skills. They should be dissuaded from this approach.
Yup. Couldn’t agree more.
Practically speaking, however, with limited total instruction time, some sacrifices must be made. So while I will never put the cart in front of the horse (in a class, anyway), and teach SCGA or any other accent before a thorough and lengthy period of skills training, I have realized that with NNSs who are not already advanced speakers of English, I (and they) need to limit my ambitions for them. This means confining their work on American English speech to a single accent pattern, rather than glorying in the wonderful array of possible variations (which is what I do with NSs in the last third of the year). I have never and will never state or imply that the accent pattern they learn is in any way inherently superior to any other accent. I am at pains, in fact, to frequently and explicitly underline the arbitrariness of it all. Nevertheless, there must be a focus. They desperately want and need it.
My second point, which has been a part of our work so far, is that in linguistic reality there is no General American.
Well, you know I agree with this. Part of my whole point, really. If I’ve realized that I need to teach a “GenAm” of some kind, the question I’ve been wrestling with is which “GenAm?”
General American is a term that is often used as a convenience term by linguists, especially UK linguists who don’t really want to focus on American accents too much; but also by American linguists and speech teachers. It’s inaccurate because, as you know, there are still accent gradations throughout the Midwest, with plenty of idiosyncratic urban accent islands, but it’s convenient on a pedagogical level because it can allow us to ignore all those smaller differentiations and focus on a simple amalgam accent that will pass muster to listeners who don’t know the Midwest intimately. (I’m one of these people, but Phil isn’t.) To me, the most honest way to teach a General American would be to pick out a specific accent from somewhere between New York State and Nebraska, and teach that. I will go further: I would recommend to Phil that he start his accent book with several examples of “heartland” accents that—though different from one another—might be sold as General American.
I like this idea very much. I’m embarrassed to admit that it hadn’t occurred to me. Obviously I have some more thinking to do about this. One question, though– wouldn’t you say that I speak a version of “General American?” I don’t come from the Midwest, and have never altered my speech to sound as though I did. I speak more or less the way people from Greenwich speak, but I don’t think my speech is particularly localizable. Which I think means that the accent family known as “GenAm” is broader than that particular island in the Midwest. It is variable in its membership surely, especially if we accept as the definitional the “you-sound-educated-but-I-can’t-tell-where-you’re-from” that is most often meant by the term. What sounds GenAm to Larry may not be so to Moe’s more variation-attuned ears.
I recognize– very clearly!– the danger here. It is precisely the danger Daniel Jones faced in originally defining RP. However descriptive his intentions, in attempting an authoritative description of an accent that was marked by social prestige one risks having that description taken up as a cudgel with which to beat those who do not conform. And I think the “GenAm” I am speaking of here– the SCGA that is demanded by agents, directories & casting directors, mainly– very much is marked, at least in part, by social prestige, even if of a much looser and more variable kind than the old RP.
General American as a catch-all term differs from English RP because it exists on the basis of a shared though imprecise perceptual agreement about what “sounding American” really is. RP, on the other hand, was both a defined and cultivated social accent and an actual regional accent. Regional because it was initially spoken by persons in a defined geographic area. Social because it was expanded geographically and culturally both by inducement—à la Eliza Doolittle wanting to become a lady in a flower shop—or by force: “Ah yes, now that my brain function is returning, I do believe that the chap who was bashing me repeatedly in the head with the butt of his rifle here on the dusty plains near Hyderabad showed a remarkable absence of post-vocalic rhoticity in his curses, along with a slight lip-rounding on the word “wog.”
Yes. And Ha!
Once the skills have been learned, especially those of oral posture, then you can teach anything, including whatever you want to cite as General American. KT Speechwork, despite its emphasis on the descriptive, isn’t a negation of prescription. Rather, it gets an actor ready to receive and embody prescriptive information easily.
Which is why I teach the way I teach. And in case I don’t say it often enough, thank you both.
So Gen-Am away, say I. Just present it in the right context. I’m sure you do and will. The teaching issue is that recurrent student desire to discover what they perceive as the magic accent bullet. The problem is it’s not magic, it’s not only one accent; and they’d better stay on the right end of the muzzle.
This is exquisitely put.
Yes, the context is all. What I’ve been attempting to do, in my written materials, is actually build the context into all the prescriptive choices. So with pre-nasal tensing in TRAP words (the TRAP-TRAM split?), for instance, I point out that it generally sounds old-fashioned not to do it at all, but that excessive raising (past open-mid, or thereabouts) will sound regionally-marked and thus take us out of the orbit of SCGA. So I’m essentially sneaking in 3 allophones, telling them a bit about what each might communicate, and then recommending one for “general” purposes.

I hope that Erik, I and Phil will hear from some of the rest of you on this subject. We’re all had experience in working with students on some version of General American or other, or perhaps some other sanctioned “neutral” accent of English outside of Gen-Am. I hope you will share your views dealing with it both in academia and in private practice.

5 thoughts on “General American Revisited

  1. This is great stuff. I deal with languages where there is no prestige dialect – you have to learn each dialect and pronounce it (sing it) according to who is in the audience. I did notice a blip in the conversation though. RBP is not a dialect. It used to be, but is no longer pronounced by anyone in their home. The people who used to speak in RBP have shifted their pronunciation, especially with the liquids. Most and letters are pronounced as [w]. This is not a speech impediment. Fear of criticizing people with handicaps has prevented a discussion of this. The truth is, that if you ask someone from England who speaks the prestige dialect to remark on a truck carrying blue dye that has overturned in the street to say, “In the the whole world, there’s a load of woad in the road” you would get, “In the hoe wote, theyws a wote of wote in the wote” (unvoiced ending consonants and all). RBP is an artificially maintained pronunciation system that is not spoken by anyone in their homes today. It is therefore not a dialect. This is one reason why I have such a problem with the British Music Conservatory system that teaches it to singers. No one in the audience actually speaks it. They may understand it, but it takes thought to translate it. This slight confusion (requiring a translation to occur in the listener’s mind) drops the audience member out of the suspension of disbelief and plops them back in their seat, wondering when the performance will end so they can grab a beer at the local pub.
    So what I’m saying is what is being stated above, only I’ve moved the discussion to the Commonwealth nations where they teach RBP to performers as how to pronounce all English text (yes, they do this at universities today – I recently heard a negro spiritual being sung using RBP – I almost had a heart attack). Who is in the audience? How do you speak so that every native speaker in English does not have to drop out of the performance a figure out what’s being said on stage? As a singer, I notice this with singers succumbing to their fears at the ends of phrases which tightens their tongues, resulting in palatised ending consonants. This shifts the preceding vowels as well, totally confusing the listener. So “York” becomes “yor-eek’ [yo:Rik’] (Gaelic linguists use an apostrophe after a letter to indicate a palatal position, capital letters to indicate fortis). In the end, you need to be understood. Fear tenses our bodies, resulting in mispronunciation. That’s my slant at least…

  2. Yes, we’ve set something up for next year. I’ll be teaching an intonation class open to regular studio students, and the NNS conservatory students will be required to take it in addition to their conservatory work. I will probably hold a certain number of classes just for them as well.

  3. Erik,

    I didn’t mean to suggest, in talking about targets , and zones of markedness , that these were mutually exclusive. I just wouldn’t want to give targets instead of working on the conceptual framework. And, of course Dudley and I both think that oral posture should be a primary tool in doing this work with NNSs.

    I also think that you should try to get separate time with these students. They need more work to master this skill and the institution should facilitate that. Alternatively, you could set up a private studio and have these students pay to join that class outside of their work at HB.

  4. Phil,

    Well, you’ve more or less said what I was trying to say, just better. And crucially, more succinctly. So it was you who came up with “So-Called General American!” I had thought so, but then begun to doubt it for some reason. Yes, I agree—the term helps frame the whole endeavor in just the right way. I hereby move that we henceforth refer to this family of accents (the ones lacking in markedness) as SCGA.

    I think markedness and salience are absolutely crucial concepts for teaching non-native speakers. And you are quite right that most of the variation one hears in the whole low-back vowel area is remarkably unmarked, as long as CLOTH/THOUGHT don’t get too rounded or acquire offglides, and LOT stays unrounded and not too advanced. Funny you should mention Connecticut, though—I grew up with exactly the same pattern you describe, and it was subtle enough that I could never figure out exactly what I was doing until I studied this stuff.

    However, I do think I need to disagree slightly that giving specific sound targets is unnecessary or inadequate (I think that’s what you’re saying). I wouldn’t have thought so before this last year. But I now believe that very specific prescriptions need to be made, no matter how arbitrary they may be (and I do think that the arbitrariness needs always to be emphasized in the classroom). Perhaps this would not be the case if time were no object, but given the constraints that most teachers and learners are likely to face, I think it is so. I can tell a Korean student, for example, that he may use an open or a closed transition in “What time,” or choose between a glottally-reinforced /t/ or a fully-released one in “It was,” and that the need for linguistic detail in a given circumstance should determine his choice. He may be capable of taking this information in and retaining it. But he still wants to know which of these options he should default to. Which is less marked? I know that this is very much akin to the native-speaking student who simply wants to be told what is correct (the magic accent bullet), but I think that the case of the NNS differs in crucial ways.

  5. This is marvelous!

    I just have a couple of initial comments before I close the computer and go to bed.

    First, I originally proposed the term “So-Called General American” (SCGA) as a way of building into the conversation an appropriate skepticism. You won’t have to look far to find linguists inveighing against the notion of General American, and I have to agree with them. The moment that term was introduced (by George Philip Krapp in his 1925 book The English Language in America) it began to be used to promote a particular “standard” of speech. It also immediately began generating controversy. Hans Kurath in his review of the book wrote:
    “Krapp’s General Type is fictitious at least insofar as pronunciation is concerned, and leads him to assume much more uniformity than actually exists and to make indiscriminate assertions.”

    The claim that GenAm is fictitious has been argued at length in a number of books and articles, and the larger implications: that we could likewise make an argument that all accents are fictitious…I will have to leave for another comment — or perhaps for the book.

    What this means for us is that we can’t make a solid, supportable claim that there is an American accent available for us to teach to NNSs with any authority.

    On the other hand, there are very real pressures to teach something useful to students who come to America looking to participate as actors in the American film and television industry. What they need is something that native speakers of American English often already have: a way of speaking that is unmarked to American audiences.

    How to construct this unmarked accent (so-called American) is the question Erik has raised and it is a question that he’s right to be careful with. the easiest way to work on the problem is to choose a handful of target sounds — and wouldn’t it be convenient if those sounds were the sounds i grew up with in Eastern Iowa?– and to guide students toward those targets.
    But what I’d like to suggest, and I may just be able to articulate this before I fall asleep, is that we identify “zones of markedness” rather than target sounds. If we give specific sound targets we’ll get most of the way to something useful, but unless we identify what the underlying phonological process is, we’ll be leaving the student without robust tools.

    maybe a single example will help me explain. I grew up making a distinction between LOT words and CLOTH/THOUGHT words. I pronounced LOT as /ɑ/ and CLOTH/THOUGHT as /ɒ/. In terms of realization, these were very close, and I daresay someone from Connecticut or California wouldn’t have noticed a difference. But the underlying organizational structure of these sound variations, together with my oral posture makes my accent quintessentially Midwestern.

    All this really means that I agree with what both Dudley and Erik have said, and i think it breaks down into 3 points:
    1) Warn people why they should mistrust the idea of a “general” American accent.

    2) Go into detail about target sounds and variations.

    3) Work skills of listening, describing, embodying speech sounds so that they can be ready to make the subtle changes necessary.

    Good Night

Leave a Reply