cunning, hawk-hock, pop-soda, etc

[Gordon, Matthew J.]

The variation you describe is all subphonemic. The Northern Cities Shift, which gives fronting of the vowel in 'hockey', does not alter phonemic categories, just how they're produced. Shifters and nonshifters have the same number of phonemes.

The back vowel thing is different. The question is whether you (generic) have 2 phonemes or one. Those of us who are blessed with a single low back vowel have the freedom to use a range of forms from [a] to 'open o' in 'hock' as well as 'hawk'. Those of you still saddled with the phonemic distinction don't have such freedom. Sure, there'll be some influence of phonetic context so that you may have a rounder and backer vowel in 'bought' , b/c of the lip rounding of /b/, than in 'taught', but that range of variation should be smaller than for us, mergeratti.

If someone really has a phonemic distinction, it seems unlikely to me that 10% of the time (and I recognize this is not an empirical finding) they would 'mispeak' to the point of confusing the distinction. After all, if this is a performance error (a term variationists despise), it should be random. Do they confuse the distinction between 'bought' and 'boat' also? As you know, phonemic distinctions are phonemic b/c they matter to comprehension, unlike most allophonic differences (e.g., [haki] v. [hAki]).

So, the bottom-line question is what is going on phonemically. You could try a little test of the status of the vowel contrast. Present them with a list of words such as [caught, lot, dog, hawk, box, odd, cough, rough, hot] and ask them which of those has the same vowel as 'bought'.  For people with a clear distinction, this task is pretty easy and they choose 'caught, hawk, cough', and maybe 'dog'. People who've merged tend to choose everything except 'rough'. People who choose 'cough' and 'rough' are likely confused about the task and focussing on spelling. Since the 'open o' phoneme is represented with a number of different spellings, it's very hard for a merged speaker to fake their way through the task - unless they've had training in the history of English.

And, yes Millie, this is the Matt you knew at Michigan.

-----Original Message-----
From:   Millie Webb [mailto:millie-webb at CHARTER.NET]

When I say 90% of the time, Matt, I mean (as I believe I stated in my post)
that they keep the distinction in theory all the time, but in practice,
people misspeak, or their pronunciation is affected by the words surrounding
the lexical item in question, the structure of the phrase or sentence, and
so on.  No one can tell me they "always" say [hahki], [hawki], or [haki]
either, for another example.  No one can tell me their sonogram would look
the same each of ten times they said a given word, or even would be noted
with the same exact "coloring" of the short vowel (influenced by -h, -w,
lengthening, etc.) each of ten times by four different linguists who think
they are coding "the same way".