American Accents/Analysis
It's not a "turnamint", it's a tournament -- toor-na-ment. What the hell is this? I mean, Springsteen doesn't go on "turr." No, he goes on tour. "All right, guys, we're going on turr... One, two, three, furr!"
—Pittsburgh native Craig Shoemaker on the differences between inland-New-England and Texas accents
About the technical details of American Accents.
Although people once thought that the levelling influence of radio and television would decrease the regional accents, broadcasting actually has very little impact on one's accent. In fact, except for the smallest, isolated dialects, dialects have become much stronger than they were even 50 or 100 years ago. The Northern cities vowel shift, the California vowel, and the Canadian vowel shift, which affects part of the Western United States are all recent innovations in the pronunciation of vowels. The merger of the vowels in words such as "cot" and "caught", is also a fairly recent innovation, which affects almost half of the United States, causes speakers with the merger to be unable to produce a distinction in those vowels, or even hear a difference when they hear speakers from non-merging areas. This merger is almost complete in California and the rest of the Western United States. However, it is unlikely to spread to some areas, such as the Inland Northern dialect region, because of its vowel configuration.
One notable feature of the Northern cities vowel shift is that the vowels are being shifted in the opposite direction to that in Western and Central Canada, and the Western United States. This can even lead to the occasional misunderstanding, because in Canada and the Western United States, the vowel in "map" by advanced speakers is the same as the vowel in "mop" by advanced speakers with the Northern cities vowel shift. This also means when one simply crosses the border in NCVS affected regions, one hears a different accent.
Another newly documented feature is the "pin-pen" merger, which merges lax "i" and "e" when they are followed by a nasal consonant (such as "n" or "m"). Speakers with the merger, like the "cot-caught" merger, can neither produce a distinction between "i" and "e" before nasals, nor hear when others make that distinction. Unlike the "cot-caught" merger, however, lax "i" and "e" are distinct vowels when they are not followed by nasal consonants, whereas the "cot-caught" merger completely neutralizes the distinction between those two vowels, thus making words like "hottie" and "haughty" sound the same. The merged vowel can be either of those two vowels or a vowel sound somewhere in between the two. The vowel that is used is determined by many factors, such as the position that the mouth is in when the speaker produces the vowels.
All these vowel shifts are different ways of cleaning up the legacy left by the Renaissance Great Vowel Shift, which left English with 20-odd vowel sounds and its famously chaotic spelling.
There are also some transitional accents such as the Midlands, which is sort of a transitional accent between the North and the South. The Central part of Kansas has a transitional dialect between the accent of western Kansas (which has a Western accent identical to that of Colorado) and the dialect of eastern Kansas, which has a Midlands accent. One of the noticeable features of the Midlands accent is that many people are transitionally "cot-caught" merged, which means that many either produce or perceive "cot" and "caught" as having the same vowel.
All across North America, there is a general trend in pronouncing the "oo" sound (as in "moon") farther to the front of the mouth, and with unrounded lips. This is also a feature of the California vowel shift, and in Western English, "oo" is fronted strongly after coronals such as "t" and "d", making words such as "dude", have an extremely fronted "u" sound, which can sometimes make it sound like it contains a front glide to people with a more back or rounded "oo".
Although there can of course be no truly neutral, or "unaccented" variety of English, when people speak of a "neutral" or General American accent, they are usually referring to a variety that contains no stigmatized features, and is perceived by most North Americans to be fairly unaccented. In the 1950s, newscasters settled on a 'Midwestern' accent for this purpose, modelling their speech on certain parts of the Midwest which had reasonably neutral accents, though they did not exactly imitate any particular Midwestern accent. However, with the advent of the Northern cities vowel shift, the Midwest no longer has a neutral accent. Nor, contrary to common belief, is there only one Midwestern U.S. accent. There are at least three principal Midwestern accents, plus transitional regions.
The North Central accent, described by William Labov in his "Atlas of North American English", is spoken in North Dakota, and the northernmost parts of Minnesota, Wisconsin and the Upper Peninsula of Michigan. Although it lacks the Northern cities vowel shift, it has in general undergone the "cot-caught" merger, and contains a heavy Scandinavian substratum (Finnish in the Upper Peninsula and Norwegian and Swedish most everywhere else).
The Northern dialect, which does not include North Central, is currently being heavily modified by the Northern cities vowel shift particularly in the dialect region defined as the Inland North, which includes Chicago and Detroit. The northern dialect does not exist west of Minnesota. The dialects west of Minnesota are the North Central, and a small transitional accent which finally transitions into the Western accent.
The other part of the Midwest is the Midlands region, which has a very variable pattern. However, about 50 years ago, it had one of the least marked accents and to this day, many people from this region still believe that they have "no accent", or a newscaster's accent. Because of the widespread Northern cities vowel shift, the dialects spoken in the Canada, the Midlands and particularly the Western United States (excluding California vowel-shifted areas) tend to have an accent that is closer to a "neutral accent", even though the West is "cot-caught" merged.
Certain areas in the West are also "pin-pen" merged, particularly in the Southwest. According to William Labov's "Atlas of North American English", page 68, the pin-pen merger is complete in Bakersfield, California. There are also other changes that occur in California, but many also occur in the West and Midwest, such as the marry-merry-Mary merger, in which each of these have the sound as in "error". According to Wikipedia's article on California English, many also pronounce "rang" with the same vowel as in "rain" rather than "ran". Another difficulty of defining a neutral accent is that there is considerable variation in North America on the pronunciation of certain words. For example, the "a" sound in "bag" is pronounced by some to have the same vowel as in "bad", but by others to sound more like the "e" in "beg". In fact some people make no distinction between "bag" and "beg" at all.
Yet another complication comes from pre-rhotic vowels, those immediately before an "r" sound. In Canadian English, and many U.S. areas near the border, "or" in words such as "sorry", "tomorrow" and "horrible", all sound like "ore". Many East Coast accents have "are" instead. In areas such as the Southwest, some words have "ore" (e.g. horrible) and others have "are" (e.g. tomorrow, although even in the Southwest, which is far from the Canadian border, some will pronounce it as "ore"). Usually, in "General American", words such as "horrible" have "ore".