NZ #9: The Great New Zealand Vowel Shift

By: | Post date: January 4, 2010 | Comments: 4 Comments
Posted in categories: Countries

To the dull-eared outsider, antipodaeans all sound the same. Australians, New Zealanders and South Africans all sound somewhat cockney, because all the dull-eared outsider can pick up is that we drop our r’s. The most dull-eared I’ve ever encountered is a colleague in the States, who was looking to deride my accent, and so started doing his best JFK. Right. Because Australian sounds exactly like Bostonian, and the absence of r’s is the only difference between dialects of English.

To the over-sensitive insider, there is a world of difference between antipodaean dialects. And having only been exposed to one New Zealander at a time in Australia, I’m still adjusting to the fact that they really do all talk like that here.

There’s gradiation of how strong the accent is on TV, although I haven’t worked out the rules yet. News anchors sound close to a prestige variant; to my surprise, the prestige variant does not sound very different from General Australian. Weather readers sound more local, and ads trade in accent stereotype, as they do in Australia. Prestige sells news, folksiness sells kitchenware.

There’s also the added complexity of Maori and Pacific Islander English, which sounds different again: same vowel shifts as in Pakeha English, but the back vowels more back and rounder. I haven’t heard enough to tell whether this is consistently the case, and what the differences are between Maori and the other Polynesians.

So in lieu of actually working out what the social stratification is of New Zealand vowels, I will write on how they shed light on vowel change in Middle English.

English is notorious for its Great Vowel shift. A little after Chaucer, the pronunciation of the long vowels of English rotated, just after English spelling had been fixed. The spelling of long vowels used to make sense: a long <a> was a long version of a short <a>, a long <i> was a long version of a short <i>. Within a century, it all stopped making sense:

  • Long <i> went from [iː] to [əi] (and eventually [ai])
  • Long <e> or <ee> went from [eː] to [iː]
  • <ea> usually went from [ɛː] to [eː] (and eventually [iː] as well). Sometimes it stayed put, which is why weak does not rhyme with steak
  • Long <a> went from [aː] to [ɛː] (and eventually [ei])

And a similar rotation happened with back vowelsː [uː] to [əu] and eventually [au], <oo> from [oː] to [uː], <oa> from [ɔː] to [oː] and eventually [ou].

Now this is a bizarre thing to have happened; that it affected all the long vowels of English tells us that the vowels are an interconnected system, and a disruption to one vowel disrupts all the other vowels in turn.

Two causes have been proposed for this kind of development. The push model says that one vowel changing caused ambiguity with the next vowel, so the next vowel had to change in turn to avoid the ambiguity; but its shift caused ambiguity with the next vowel along, so all the vowels had to change to make way for each other. So once mate stopped being pronounced as [maːt], and started being pronounced as [mɛːt], it became ambiguous with meat. The [ɛː] vowel then had to change in turn, so meat sounded like [meːt], which was now ambiguous with meet. And so on.

That’s a satisfying, logical answer: language is used for communication, ambiguity gets in the way for communication, the vowel shift deals with ambiguity.

The catch is, meet and meat now do sound the same. So it’s not like that strategy paid off.

The alternate account is a pull model. Once a shift in the vowels happened, there was a gap in the vowel system, which made it more odd to learn for children. Children compensated by shifting the next vowel up to fill the gap. But this just created a new gap, and the vowels kept rotating, until the gaps in the system were filled again.

So once [iː] turned into [əi], English turned into a language without an [iː]. A language with long vowels but without a long [iː] is an odd language indeed; and children corrected that oddity by turning [eː] into [iː]. That meant that English now had an [ɛː] and an [iː] but no [eː], and so the vowels kept rotating. (English ended up without an [aː], so there’s still imbalance in American English, although dropping r’s fixed that elsewhere, because [aɹ] turned into [aː]. But the gap filling, like all language change, is by its nature haphazard.)

That account… is not satisfying and logical: it makes of English vowels an algebraic, typological game, and it turns children learning the language into neurotic typologists. The thing is, the shift in New Zealand vowels is its own vowel rotation—this time involving just front short vowels. And the developments in New Zealand support the pull model.

New Zealand English and its development has been thoroughly and meticulously investigated by the linguists here. So the linguists here will be able to tell you precisely when each vowel changed. Because I am not a New Zealand linguist, I will instead reconstruct the sequence based on when Australian ads started making fun of them.

You would have thought a people who some might say sounds like the dregs of the London sewerage system have no standing to deride their neighbours’ accent. You’d be wrong of course, self-criticism is not the point of differentiating yourself from your neighbours, and we so do NOT sound like the dregs of the London sewerage system at all.


The earliest and most stereotypical vowel change was [ɪ] to [ɨ]. To the dull-eared Australian, [ɨ] sounds like [ə], “uh”. Australian jokes about “uh” were already in swing by the early 1980s. “Australia Sux”, the New Zealander is meant to have graffiti’d, to which the Australian daubed underneath “New Zealand Nil”. (Because “sucks” is how New Zealanders are meant to pronounce “six”.) And most Australian jokes still centre around this vowel change, particularly the more halfwitted variety on breakfast radio.

The change of [ɛ] to [ɪ] doesn’t get as much airplay, but it was being mentioned in ads by the 1990s. The first I recollect was for the Mainland Cheese brand.

—Mainland Cheese hilps to keep you fitter.
—Fitta? As in the cheese?
—Not Feta! Fitter!

(I’ve just found out what Mainland means here: it was the South Island literally, when the South Island was where everything was at, and it is the South Island ironically, for the past century.)

The last change to happen, given the distribution of short front vowels in English, is [æ] to [ɛ]. The comedians Across the Ditch [= Tasman] have not yet latched on to this, and I wasn’t sure the change had happened at all; but I have now got confirmation of it. First from a TV ad, extolling the “messive, messive summer sale” at Briscoes. Then a few days later at the winery, with reference to a “desh of gin”. ([mɛsɨv mɛsɨv samə sɛɪl], [dɛʃ əv dʒɨn])

And the sequence means that the change wasn’t motivated by ambiguity. Bid didn’t change to buhd to keep out of the way of bed turning into bid. Bid changing to buhd created a gap, for bed to fill by changing into bid.

I still can’t believe they all actually talk like that here though…


  • Myrtone says:

    I wonder whether the first vowel shift in the sequenc actually originated in Glaswegian where it would have been coused by something in the Scots sound system. Did it appear in Invergargill any ealier than elsewhere under the cloud?

  • John Cowan says:

    PALM may be a confusing word, but if the distinction you are describing is not phonemic (occurs only before /l/), it doesn't matter from a lexical-set perspective. Similarly, CURE is confusing because people often think the /j/ is part of the vowel, but it's not.

    Cool CAPTCHA: "whordism".

  • Marcos says:

    Didn't Wells replace PALM with a different word in his later work? That word is a terrible exemplar for that vowel in my 'lect. The vowel in "palm" is different from that in e.g. "father" – it's rounded and further back, and the /l/ – which is not silent – colors (lateralizes?) it. Thus PALM is in its own category with ALL, FALL, etc, separate from the /a/ words that don't have a following /l/, which would seem to need a different set name. Although, in my case, the /l/-less ones are all merged with CLOTH/LOT/THOUGHT.

  • John Cowan says:

    > The catch is, meet and meat now do sound the same.

    Well, in most places, but not quite everywhere: that merger and the long mid mergers aren't yet universally operative.

    So here are the Revised Wells/Mills/Cowan/Rosta Really Universal, Dammit, This Time, Lexical Sets:


    If you make a stressed-vowel distinction that's not in there, holler.

Leave a Reply

  • January 2010
    M T W T F S S
  • Subscribe to Blog via Email

%d bloggers like this: