Back to the cot-caught distinction.
I have just looked up some more professional information on it, and
I've found that, according to a Labov study in 1996 and a Harvard
study in 2003, about 60% of Americans retain that distinction, and
among those who display the merger, some of them have merged the
vowels only partially. This means that a only a minority of US
speakers actually shows the full merger.
What, then, would motivate a dictionary publisher, or anyone else, to
completely jettison the distinction in their pronunciation notes, or
in some other way claim the change is more complete than it is?
I have my theory, which is that linguists and linguistics journals
tend to get stuck on certain phenomena and even languages that are
widely and frequently studied, and the phenomena may become
exaggerated in their minds, much as it's been shown that Americans who
watch a lot of TV tend to exaggerate mentally the percentage of the US
population who are policemen.
I think this can happen with individual languages also. Every
linguistics student is exposed to information on Yawelmani, sometimes
again and again, but languages that have far more speakers but do not
display any phenomena of interest to linguists are seldom touched on.
This can create the misimpression in some students that Yawelmani or
some other language that fascinates linguists has more speakers than
it actually does.
Any thoughts, comments or refutations?