Date: Tue, 25 Jul 1995 11:52:33 EDT

From: Douglas Bayer x3701 3NW dbayer[AT SYMBOL GOES HERE]YUKON.HQ.ILEAF.COM

Subject: Re: sherb't

Tim Behrend asked about the pronunciation of "sherbe[r]t."

I'll take a stab:

In this case, the unstressed syllable echoes the vowel

in the stressed syllable. There's simply no other

sound or stress in the word to cause speakers to budge

out of "er" position.

Vowels in unstressed syllables in English centralize--

to "uh" /V/ near back vowels, or to a central "ih" /-I-/

near front vowels and dental consononants.

In fact, the job of an unstressed vowel is not to add or

subtract information or otherwise draw attention to itself.

If you don't centralize such a vowel, you change the stress

pattern of the word.

This "er" is a third variation of "uh" and "ih". To uncurl

the tongue and unround the lips after "sher" and noticibly

move the tongue to "eh" /E/ position is to give "bet"

secondary stress -- it makes "sherb't" a "sure bet."

The er is particularly noticible near the Great Lakes because

of strong vowel raising and the heavily retroflex r there.

So it seems we're gone the next step and reinterpreted

"sherb't" as "sherbert" to better fit our native phonology.

We notice when we're surprised by sherbet's foreign spelling.

--Doug Bayer


I add the following because you work with Asian languages,

Tim, and propably have contact with Asian speakers of English.

(Q: Was Aukland speech born of Northern English and Scots,

like Great Lakes speech, or of Southern English, like

New England?)

Even though "er" is a digraph, the two letters

represent a monothong -- a mid-to-close central vowel.

I grew up in Rochester, NY using noticibly "raised" or

close vowels in "have and "bird" ("hAYIf" and bWRd")

This is characteristic of most "Great Lakes" accents,

(the sales-region of the "Death of a Salesman"...)

Now I live in Eastern Mass, where folks use far broader

vowels ("hahf" and "bI:d"). But the "hard-r"and "r-dropping"

pronunciations are both retroflex. New Englanders may lower

their tongues, but they don't uncurl them.

Another characteristic of the American r is lip-rounding.

We were taught that the initial "w" is silent in "write"

and "wrong," but for us it's alive and well -- as well as

in "read" and "right"." Some English teachers in Japan

have had good success teaching initial r- to their students

by instructing them always to append an invisible w-.

Lip-rounding alone fools most listeners into thinking

that the speaker is curling back her tongue.

Furthermore, the G-L "sh" tends to be lip-rounded, and even

retroflex. This I rarely see noted. But I've had remarkable

and immediate success teaching Japanese students to

pronounce our hard-r by first teaching them my "sh."

You see, a sonogram of the Japanese "sh" /Sy/ shows peak

energy well above 4kHz, near "s," while the American "sh"

peaks as low as 2kHz. This is near the Japanese bilabial

"F" /Hw/. So I instruct them to blend "fu" and "shi"

into "fwsh" for an Amurrican "sh," then tell them to add

voicing and reduce the frictation. Voilar! R!

Finally, Great-Lakers regularly pronounce final -t as a

glottal stop /?/, without changing tongue position.

There's no hint of any t-like gesture toward the ridge

behind the teeth in "ca?" "hur?" "ge?" "visi?"...

"sure be?" or "sherber?"

This leaves "sherb" containing exclusively lip-rounded,

retroflex phonemes /sh/ and /er/, the labial /b/, and a

vowel that is unstressed, so whose job it is not to do


--Doug Bayer