Back in 2005 I had a brief post on the Northern Cities Shift; now Rob Mifsud has a nice Slate piece on it:

From Syracuse, N.Y., in the east to Milwaukee in the west, 34 million Americans are revolutionizing the sound of English. Linguists first noted aspects of the change in the late 1960s. In 1972, three linguists, led by William Labov of the University of Pennsylvania, christened the phenomenon the Northern Cities Vowel Shift or, more simply, the Northern Cities Shift (NCS). What they observed may be the most important change in English pronunciation in centuries. …
If news of this radical linguistic shift hasn’t made it to you yet, you are not alone. Even people who speak this way remain mostly unaware of it. Dennis Preston, a professor of perceptual linguistics at Oklahoma State University—he doesn’t merely study how people speak, he studies how people perceive both their own speech and the speech of others—discovered something peculiar about NCS speakers when he was teaching at Michigan State University. “They don’t perceive their dialect at all,” he says. “The awareness of the NCS in NCS territory is zero.”

It’s well worth a read; as John Cowan, who sent me the link, said, “not a single factual error that I could see except for the weirdness of wha instead of aw for /O/.”


  1. Well, I seem to be accent-deaf when it comes to North American. The ‘bosses with antennas on their taps’ (especially ‘bosses’) sticks out a mile, but when I listen to the two ‘radically different’ pronunciations from Detroit and Windsor, I have great difficulty picking out anything ‘Northern City Shift’-ish at all in the Detroit lady’s pronunciation. The only thing I noticed is that the Windsor guy sounds somehow more ‘North American’.

  2. Gown: I have great difficulty picking out anything ‘Northern City Shift’-ish at all in the Detroit lady’s pronunciation.
    So do I. It just sounds Detroitish (we’ve had a Detroiter staying with us this week).
    I love that Speech Accent Archive, where they go meet Stella at the train station. They’ve even got Norwegians saying it, for some reason.

  3. The Slate piece describes the CAUGHT vowel as “short”. I’m not sure how that can be justified.
    In most accents that still distinguish length, the CAUGHT vowel is a long vowel (in my own RP-ish accent, length is the main distinction between COT and CAUGHT). Historically it was a long vowel, deriving from a Middle English diphthong. It may be short in NCS-affected accents, but to what extent to they have a short-long distinction anyway?

  4. In general, American accents obey a variant of the Scottish Vowel Length Rule, so in effect length is non-phonemic, and only vowel quality counts.

  5. I don’t see anything remarkable about people not noticing these phonetic changes, since the changes seem to be subphonemic. Why would they notice? Where’s this revolution?

  6. I love that Speech Accent Archive
    I think it’s neat, but why does the sample contain no examples of the diphthong /ai/? This would be one of the first diphthongs I would listen for.

  7. Not five?

  8. What about St Louis and the Twin Cities?

  9. Charles Perry says

    Here in California, among Valley Girls it’s still all about fronting every possible vowel. Though they still don’t go as far as London dollies (“hello-ee, you-ee”).

  10. There’s an Australian way of saying no that’s like that: “nə-ee”.

  11. John Emerson says

    I grew up in the shift area, more or less, but moved away while it was happening and didn’t catch it. But a friend of mine my age who spent most of his life in Minneapolis did.

  12. Sigh. Thanks MMcM, I’m not sure how I missed it.

  13. There’s an Australian way of saying no that’s like that: “nə-ee”.
    I find it hard to mentally conjure up the pronunciation you’re suggesting. The closest thing I can think of is the pronunciation that goes something like /nay/, where the final vowel is something akin to German ü or French y. I would have trouble giving the exact phonetic symbols, though!

  14. If it’s not happening in NYC or Boston or Seattle, is “Northern Cities” the best name? Would “Rust Belt” be better?

  15. I remember my first encounter with this several years ago, it caused some confusion between this woman and me – I perceive her highly fronted /a/ as [æ].
    Then I had a professor from a NCVS area and while his /a/ wasn’t as fronted, he did pronounce “busses” as “bosses.”
    Now that I recently moved to upstate NY for grad school, I encounter it among the locals all the time, which is quite a trip because it sounds foreign (in the sense of “new”) to me. Also, I thought it was a midwest thing – I had no idea it had spread to a huge area of New York and now to St. Louis, if I remember Dennis Preston saying correctly…

  16. But those short vowels have remained pretty much constant since the eighth century—in other words, for more than a thousand years. Until now.
    I don’t quite understand this claim that the NCS is unique in affecting short vowels. For example, the Wikipedia page on Southern American English lists at least three features (pin-pen merger, drawl, “Southern Shift”) affecting short vowels. Is it just that the NCS is more dramatic and affects more vowels, or is there some other difference?

  17. I don’t see anything remarkable about people not noticing these phonetic changes, since the changes seem to be subphonemic.
    The vowel shifts that separate Australian from RP are mostly sub-phonemic as well, but Australians are very conscious that they don’t speak RP and vice versa.
    I don’t quite understand this claim that the NCS is unique in affecting short vowels.
    Neither do I. In addition to the Southern Shift, there are also the Southern Hemisphere changes, many of which affect the short-vowel system, though less pervasively.

  18. dearieme,
    That usage of “Northern” is in the old North vs. South dichotomy, which gives out as you head west, so that Seattle is equally a part of The North (i.e., not The South) as San Diego. Your point about NYC and Boston stands, though. “Great Lakes Shift” might be better, except that would imply that Toronto participates, when it (oddly) does not.
    Chris S.,
    Western upstate New York can be considered culturally Midwestern, along with western Pennsylvania. And St Louis certainly is.

  19. Not odd. Canadians have their own shift, which is why Windsor and Detroit sound completely different despite being adjacent.

  20. “Your point about NYC and Boston stands, though. “Great Lakes Shift” might be better, except that would imply that Toronto participates, when it (oddly) does not. ”
    Not so odd, really. Those places have never really been in the same speech community ever.

  21. Jim: err, since Southern Ontario (and anglophone Canada more generally, Newfoundland being the main exception) was settled by loyalists from the Northern United States, Toronto was certainly once part of the American speech community, and contact with the United States was never cut off after said loyalists settled in Ontario.
    So: why is the Northern Cities shift not affecting Ontario? A sociolinguist of my acquaintance has claimed to have the answer: he argues that the reason why the Northern Cities Shift leaves Ontario unaffected is because in the Greak Lakes cities in the U.S. this change, which affects white but not black speakers, has as its motivation a desire among white Americans to make their speech as unlike blacks’ as possible.
    From what I recall this claim fits the geographical as well as the chronological facts.

  22. Since I know little about American dialects, I suppose I shouldn’t say this, but I’m not seeing any evidence that a NCS actually exists, in the web references I’ve looked at. There are a bunch of vowel changes in various dialects, but do they have anything to do with one another? In at least one version, I gather, the supposed NCS begins with a change of the low front vowel into something else, then, due to a pull-chain, some other vowel is pulled toward the low front position. But why should such a thing happen? A language can perfectly well do without any low front vowel. And why think that the fronting of a low vowel depends on it being pulled forward to fill some supposed void in the vowel system?

  23. John Cowan,
    The Canadian Shift affects many people on the US side of the border, too (not those with the NCVS, of course), as does Canadian raising.
    Be careful not to overstate the importance of the border, any more than state/province lines. The Great Lakes cities were more tied together than any of them were with their Pacific or Atlantic coasts for the great bulk of their history. I recoil from any description that tries to group the US accents in one pot and the Canada in another, like Detroit has more in common with Savannah, Georgia, than it does with Windsor.

  24. The Canada

    Too bad it’s never been one of those arthrous countries.

  25. I’ve heard the “white flight” explanation for the NCS before and it’s at least very interesting, though I don’t remember being very convinced by the arguments for it. More likely is just that there’s a strong national-identity boundary between Canada and the US, which stops sound changes that are strongly associated with one region from being accepted in the other—Canadians don’t want to sound American.
    Greg Lee, the ontological status of chain shifting in general—whether the movements of the various vowel phonemes are actually *causally* connected to each other—is a hotly debated question, but it’s a totally different question than the existence of the NCS in particular, which is just the empirical fact that all these movements of several vowels do in fact all appear in the same dialect.
    The whole “biggest thing to happen to the short vowels in forever” business is a little bit exaggerated, but not much—the TRAP phoneme has been a low vowel since Proto-Germanic, and the DRESS phoneme has been a front vowel for as long or longer, and the NCS is as far as I know the first dialect to change both of those.

  26. As xyzzyva says, the Canadian shift affects many people on the US side of the border—but that’s probably the result of the same sound change arising independently in multiple regions, rather than, say, Columbus, Ohio being dialectologically connected to Canada.

  27. The guy from Sydney in the Speech Accent Archive is really bunging it on. That is almost a caricature of a Broad Australian accent.

  28. @Etienne, AJD: I’ve long suspected that the NCS was motivated by a desire to distinguish one’s speech not only from AAVE but from Appalachian English, which was the language of many migrants to those cities from WWII to the 1970s. If I’m right, this shift would be even more highly motivated as a matter of distinguishing two speech communities whose faces look the same.
    Northern city joke: “Why don’t n****rs let their kids listen to hillbilly music? They don’t want ’em to grow up too lazy to steal.”

  29. Etienne, AJD & Rodger C – just out of curiosity, why would an accent change in reaction to AAVE happen in the upper midwest, where the cities are highly segregated and the accent sounded nothing like AAVE to begin with, as opposed to the south, where the accents and communities are in closer contact?

  30. @s/o: As for the south, the communities aren’t in close social contact at all, and at any rate white southerners aren’t that concerned with not sounding black, just because the dialects are both closely related and socially isolated from each other. In the case of northern cities, in the first place I prefer my own hypothesis, and in the second, if AAVE is in fact the cause, I suspect the reaction would have been less against AAVE itself than against AAVE-accented northern English.

  31. The “white flight” explanation also fails to account for why the NCS does not occur in the non-Southern urban parts of the US (where whites and blacks live in proximity but with historical differences in dialect) in general, but only in one geographical subset of that area. Surely there must be other historical examples where urban dialects in a particular geographical area changed in tandem with the dialects of the intervening rural areas lagging in their adaptation of the change?

  32. @J. W. Brewer: See my own hypothesis above (avoidance of sounding “hillbilly”). The Appalachian migration was more restricted than the Black migration. I broached this idea on another blog some time ago and was told that the NCS started before the App. migration, but nothing said here seems to suggest that.

  33. Huh. I’ve lived in Chicago my whole life, and I never knew that there was a major migration from Appalachia to the midwestern cities. (Which only means that it wasn’t as noticeable in chicago as, say, Black migration from the south, white flight after the race riots, the floods of Polish immigrants…)

  34. And that I don’t know much about local social history.

  35. @s/o: See the now-classic books Uptown: Poor Whites in Chicago by Todd Gitlin and The Invisible Minority: Urban Appalachians by Bill Philliber and Clyde McCoy. (“Now-classic” is a word I find myself using with dispiriting regularity these days for things I read when they were current …)

  36. Not sure how old s/o is, but I see that the Gitlin book Rodger C references is from 1970. When I arrived in Chicago in ’89 (left again in ’92), various guidebooks I read still referred to Uptown as the “poor Appalachian white” neighborhood, but that was already out of date b/c by then they’d assimilated or moved on.

  37. Rodger C and JW Brewer – Thank you for the recommendations! I’m 28, so I don’t remember the 70’s very well… Uptown is still pretty rough, in part because of the alderwoman’s efforts to keep gentrification from creeping up from lakeview, but it’s a general mishmash of the down-and-out, struggling musicians, and Asian immigrants now.

  38. J.W. Brewer says

    That was an entertaining piece on the “hillbilly problem” and I am grateful to Rodger C for the link. The only country/western bar I can remember patronizing during my now-long-ago time in Chicagoland was the semi-legendary Sundowners Ranch out in Franklin Park, quite a long ways from Uptown.

  39. marie-lucie says

    Canadian raising
    This term refers to the pronunciation of the onset of phonemic diphthongs before voiceless stops compared to voiced stops, as in sight (but not side) or clout (but not cloud). The term is a misnomer since this supposed Canadian innovation of “raising” the initial low vowel of the diphrhong is actually a survival. The change should be identified as “American lowering”, not “Canadian raising”.

  40. That Canadian raising might be a survival of an incomplete version of the GVS, I admit; that it is so, I more than hesitate to assert. In particular, it seems more likely to me that Martha’s Vineyard dialect (which has CR) might actually go back to pre-GVS days than Ontarian English, the ancestor of general Canadian English. Quoth WP:

    Some have hypothesized that Canadian raising may be related historically to a similar phenomenon that exists in Scots and Scottish English. The Scottish Vowel Length Rule lengthens a wide variety of vowel sounds in several environments, and shortens them in others; “long” environments include when the vowel precedes a number of voiced consonant sounds. This rule also conditions /aɪ/ in the long environments and /əɪ/ in the short environments.
    Significantly, though, the Scots Vowel Length Rule applies only before voiced fricatives and /r/, whereas Canadian raising is not limited in this fashion; thus, it may represent a sort of merging of the Scots Vowel Length Rule with the general English rule lengthening vowels before voiced consonants of any sort.

    WP concludes that the evidence is inconclusive.

Speak Your Mind