Teenglish from England

[Nathan Sanders]

On Sep 17, 2009, at 1:13 PM, Laurence Horn wrote:

> It's clear that some believe this, and they may be right for their
> own speech.  It is certainly not my impression that raised/tensed [i]
> is the usual (or even that common a) pronunciation for U.S. speakers,
> but unlike some of the "Eenglish" contingent, I am not saying that
> they're necessarily misguided about their *own* pronunciation.  What
> rubs me the wrong way is when I'm told *I* pronounce it "Eenglish" or
> "seeng" and just don't admit it.  But we've been down that road many
> times, and the odds that minds will be changed on this are roughly as
> steep as the odds that the health care debate in Congress will.

It's a shame that TZ hasn't taken a phonetics course, because if he
had, he might very well have learned why [I] sounds like [i] before
[N], or at least learned enough to piece it together on his own.

For the benefit of those who do actually care to learn why this is
true, I'll explain:

The two loudest, most perceptually important resonant frequencies
(vowel formants) that result from a mouth shape for [I] are
approximately F1 = 400 Hz and F2 = 1900 Hz (the actual values for a
given individual will depend on the size of their vocal tract).

Before nasal sounds, vowels in English are often at least partially
nasalized due to anticipatory articulation, as the velum is
prematurely lowered in preparation for the following nasal.  Nasalized
vowels have a very low, extra formant in the 200-300 Hz range due to
the additional resonance of the vowel's sound wave as it travels from
the glottis up into the nasal cavity.

Note that the ordinary F1 value for a plain oral [I] is 400 Hz, which
is very close to the nasal formant value, so close in fact, that
listeners may hear only a single formant rather than two distinct
formants when [I] is nasalized.  When this happens, the perceptual
result is some weighted combination of the nasal formant and F1, which
causes us to perceive a nasalized [I] as having a lower F1 value than
for a plain oral [I].

Since F1 is an acoustic correlate of vowel height and a lower F1 value
corresponds to a higher vowel articulation, a nasalized [I] (which is
found before a nasal) effectively sounds like it is pronounced
slightly higher than it actually is.

Furthermore, a following velar sound causes what is known as a "velar
pinch", in which F2 and F3 values approach each other (i.e., F2
increases while F3 decreases).  Since F2 is an acoustic correlate of
frontness/backness in vowels and a higher F2 value corresponds to a
fronter articulation, [I] before a velar effectively sounds like it is
pronounced slightly fronter than it actually is.

Putting these two facts together: [I] before [N] (which is of course
both nasal and velar) effectively sounds like it is pronounced both
higher and fronter than it actually is.  Since [i] is the only English
vowel that is higher and fronter than [I], [i] is the vowel many
people will hear when [I] sounds higher and fronter (such as when it
is pronounced before [N]).

The next thing that can happen is that some speakers who hear [i] in
this environment will actually produce [i] rather than [I], to mimic
what they think they hear other people saying.

So the end result is that some people say [IN], some people say [iN],
and many people hear both as [iN].

Nathan

--
Nathan Sanders
Linguistics Program
Williams College
http://wso.williams.edu/~nsanders/

------------------------------------------------------------
The American Dialect Society - http://www.americandialect.org