WIXLIB

Figure 5.4 Comparison of DARE map and conventional map of dialect variants (from Cassidy 1985: xxix,figure 7; 1985 by the President and Fellows of Harvard College, reprinted by permission of HarvardUniversity Press)As maps such as the above have become more widely available, they have traveled beyondacademic circles to the general public, who understandably also has an enduring curiosityabout regional dialect variation. Katz’s heat maps have been featured in various nationalmedia outlets, and Vaux and Golder’s survey has even found its way into hundreds ofYouTube videos of people performing their answers to the Harvard Dialect Survey questions.Further evidence of the widespread interest in pinpointing regional dialect variants isevidenced in Katz’s popular dialect quiz, again based on Vaux and Golder’s survey, publishedin the New York Times in December /sunday-review/dialect-quiz-map.html)5.3 The Distribution of Dialect FormsFor some regional items, the distribution of dialect forms shows a GROUP-EXCLUSIVE patternin which communities in one area use one variant while those in another region use a differentone. For example, in the map of pail and bucket displayed in figure 5.3, you can trace a line ofdemarcation that sets apart southern and northern regions of Pennsylvania: South of the lineC05p. 209

bucket is used and north of the line pail is used. When the distribution shows a fairly clear-cutdemarcation, a line indicating the boundaries of the different variants, an ISOGLOSS, may bedrawn. Isoglosses set apart zones of usage in a very discrete way, but not all patterns of usageare as clear as that delimited for the use of pail and bucket in Pennsylvania in the 1930s and1940s. In many cases, variants are more interspersed, making it difficult to draw a meaningfulisogloss. In the South, there are pockets of usage for pail in Virginia, North Carolina, andGeorgia. In addition, there are dialect TRANSITIONAL ZONES, where more than one variantoccurs, with different speakers using different variants, or individual speakers using both. Infact, transitional zones are more typical than the abrupt pattern of distribution implied byisoglosses, especially in more densely populated areas. Boundaries between dialects may beSHARP, where a well-defined, linear boundary exists (e.g., the boundary between Windsor,Canada, and the adjacent city of Detroit, Michigan); FADING, where regional features diminishas one moves from a dialect center to outlying areas (e.g. the areas outlying Pittsburgh to theNorth towards Erie, PA), OVERLAPPING, where linguistic features of different dialect regionsco-exist (Southern and non-Southern traits in Northern Virginia), or NULL, where intermediateareas do have not been found to use of more clearly defined adjacent areas (e.g. regions ofNew Jersey in between New York City and Philadelphia that do not have the complex vowelpronunciation patterns of either city. Isoglosses are certainly useful indicators of theboundaries of regional usage patterns, but they must be viewed with important qualifications.Isoglosses often represent ideal rather than real patterns of delimitation, a “convenient fictionexisting in an abstract moment in time” (Carver 1987: 13).In a microscopic view of regional variation, each boundary line between two differentforms for a given item indicates a different dialect area, but this reduces the definition ofregional dialect to a trivial one. When the overall responses to dialect questionnaires areconsidered, different isoglosses may show similar patterns of delimitation. These clusters, orBUNDLES OF ISOGLOSSES, are usually considered significant in determining regional dialectareas.Predictably, major regional areas are typically determined by having larger bundles ofisoglosses than minor dialect areas. Using this approach, the initial phase of the LinguisticAtlas survey of the Eastern United States ended up proposing several major regional dialectsand some minor dialect areas. For example, Kurath, in his Word Geography of the EasternUnited States (1949), presented a map of major and minor areas that became the standardrepresentation of regional dialects along the Eastern Seaboard for almost a half century. As wediscussed in chapter 4, this map delimits three major regional areas, the North, the Midland,and the South, with a number of subregional dialects for each major area. This map isreprinted as figure 5.6.C05p. 210

Figure 5.6Dialect areas of the Eastern United States, based on Linguistic Atlas isoglosses (from Kurath1949: figure 3; reprinted by permission of the University of Michigan Press)A number of measures have been proposed for determining the relative significance ofisogloss bundles. In Carver’s American Regional Dialects: A Word Geography (1987), whichis based primarily upon lexical data (800 lexical items with regional distribution) taken fromthe files of DARE, the notion of ISOGLOSSAL LAYERING is used to determine major and minorregional varieties. The term LAYER, taken from physical geography, is used to refer to aunique set of areal features, but the importance of this concept lies in the fact that it is used tocapture overlap and divergence in regional dialects by examining levels of layering rather thanindependent sets of isogloss bundles. The most concentrated regional dialect area, where thegreatest number of regionally specific features are present, is the PRIMARY DIALECT AREA. InSECONDARY and TERTIARY dialect areas, there are progressively fewer of these dialectfeatures. For example, the core of the Northeast American English dialect has 20 to 24 wordsfrom Carver’s inventory of regional lexical items, whereas secondary layers have only 15 to19, and so forth. While this approach does not eliminate some of the basic problems withisoglosses we have pointed out, it captures the hierarchical nature of overlap and divergence inregional varieties. As an example of layering, Carver’s (1987) analysis of the extension of theNorthern dialect area into the Northwestern United States is given in figure 5.7. The areaslabeled as primary represent the core areas of the westward extension of Northern and InlandNorthern dialect features, whereas the secondary, tertiary, and quarternary areas of the North,Inland North, and Upper North represent less concentrated layers of these extensions.C05p. 211

Figure 5.7An example of dialect layering in the Northwestern United States (from Carver 1987: 214;reprinted by permission of the University of Michigan Press)Layering can also be represented hierarchically. For example, Carver’s Western dialectlayers can be presented in the form of a hierarchical tree, as in figure 5.8.Figure 5.8An example of dialect layering in the West, represented hierarchically (from Carver 1987: 243;reprinted by permission of the University of Michigan Press)The fact that lexical variation is so often used as a primary basis for regional dialects hasbeen a major source of contention among students of language variation. Some linguists havemaintained that lexical differences are among the most superficial types of linguistic structure,and therefore among the least reliable indicators of dialect areas. However, it should be notedthat Carver’s lexical boundaries correlate well with boundaries arrived at independently incultural geography, including areas delimited by such features as architectural styles, religion,political ideology, and a number of other culturally significant variables. Thus, lexical items,C05p. 212

regardless of their linguistic status, serve as indicators of more broadly based cultural andhistorical foundations upon which regional dialects rest, and they should not be dismissed asinsignificant.As noted in chapter 4, in the late 1990s William Labov, Sharon Ash, and Charles Bobergconducted a comprehensive survey of vowel variation in the United States and Canada using atelephone survey (TELSUR), resulting in the landmark publication, the Atlas of NorthAmerican English (ANAE) (2006). The publication includes an interactive web site allowingone to explore the various dialect regions and their characteristics and the regional distributionof particular features, as well as to listen to samples of vowel differences and to conversationalspeech in each dialect region and community surveyed. At least two speakers from allEnglish-speaking cities in the United States and Canada with a population of more than50,000 people were included in the survey, and all of the vowels for 439 speakers weresubjected to detailed acoustic analysis.The contrasts between Labov, Ash, and Boberg’s ANAE survey and the earlier LinguisticAtlas of the United States and Canada (LAUSC) extend beyond linguistic level (i.e. lexical vs.phonological) and technical sophistication of cartographic plotting. The LAUSC was drivenby an interest in historical dialect features and regions, whereas the ANAE is centered onlanguage change taking place in present-day communities. For LAUSC, ideal intervieweeswere older, rural men with little geographic mobility who preserved older ways of speaking,whereas ideal participants for ANAE were young women, since this population group tends tobe at the forefront of language. The summary map of the major regional varieties andsub-regions from Labov, et al. (2006: 141) is again presented in figure 5.9.Figure 5.9 Summary map from the Atlas of North American English (from Labov, Ash,and Boberg 2006)Though radically different in research design and sampling technique, many phonologicalvariables in Labov, et al. (2006) show patterns of regional variation that parallel those shownfor lexical items. As with lexical variables, phonological features show regional layering. Wemay, for example, expect to find a core Southern or core Northern area, where the highestconcentration of specific phonological features is found, and secondary and tertiary zonessurrounding these primary areas.As we mentioned in chapters 3 and 4, investigations of vowel systems conducted in thepast several decades have revealed that there are several major systematic changes currentlyunder way in the US, delimiting major dialect areas.One pattern of change is called the NORTHERN CITIES VOWEL SHIFT. As Labov describesthis VOWEL ROTATION pattern (Labov 1994: 177–201; see also Labov, et al. 2006: 187-215),the TRAP, LOT, and THOUGHT vowels are moving forward and then upward, and the vowels ofDRESS, STRUT, and KIT are moving backward and then downward. For example, the TRAPvowel is fronting and raising, leaving room for the LOT vowel in a word like lock to moveforward, so that it sounds almost like lack. In turn, this movement leaves room for words inthe THOUGHT class to move toward LOT, so that caught now sounds like what would formerlyhave been pronounced as cot (but now sounds almost like cat). At the same time, the DRESSvowel moves backward towards the STRUT vowel, which is then pushed backward, so that aword like bus might sound more like boss.[INSERT ENHANCEMENT 5.1][INSERT ENHANCEMENT 5.2]C05p. 213

Diagrammatically, the shift may be represented as in figure 5.10. Recall that the vowelsare arranged so that vowels produced with greater tongue height appear at the top of the chart,and those produced with greater fronting of the tongue appear on the left. For convenience,“key words” are given. The arrows indicate the direction in which the vowels are moving,based on Labov’s characterization of this shift pattern.Figure 5.10The Northern Cities Vowel ShiftRegionally, the vowel rotation pattern depicted in figure 5.10 in concentrated in the largermetropolitan areas that start in Western New England and proceeds westward into upstateNew York; the extreme northern portions of Ohio, Indiana, and Illinois; Michigan; andWisconsin. More advanced stages of this change can be found in younger speakers in thelargest metropolitan areas in this Northern region, such as Buffalo, Cleveland, Detroit, andChicago.Exercise 2Identify in the following list of words those items that would be involved in the NorthernCities Vowel ropeAs the Northern Cities Vowel Shift spreads across the northern portion of the US,researchers continue to track its progress, not only in geographic space but also across socialC05p. 214

groups and age groups. In addition, not all researchers are in agreement that the pattern infigure 5.10 tells the whole story, and there may be other vowel movement patterns that arealso an important part of dialect change in the Northern US (e.g. Gordon 2001, Thomas 2001).However, there is no question that sweeping pronunciation changes have been taking place inNorthern metropolitan areas over the course of the past half century, and that these changesare proceeding in quite different directions from the vowel shift patterns affecting otherregions such as the Southern US.In the SOUTHERN VOWEL SHIFT, the PRICE vowel unglides and the nucleus may moveslightly forward, so that a word like time may sound like tom or tam. The FACE vowel lowersbetween the DRESS and PRICE but retains a glide, so that a word like bait may sound somethinglike bite. Concurrently, the DRESS vowel raises toward FACE and takes on a glide, so that aword like set sounds almost like the phrase “say it.” In essence, then, DRESS changes placeswith FACE. In some parts of the South, the KIT vowel may change places with the FLEECEvowel as well, with FLEECE lowering toward FACE and KIT raising and taking on a glide, sothat beat sounds like bait and sit sounds like “see it.” Meanwhile, the back vowels of GOOSEand GOAT are moving forward so that they sound more like gews and gewt, respectively. Therotational patterns that characterize the Southern Vowel Shift are indicated in figure 5.11.Figure 5.11Interactive chart:The Southern Vowel ShiftLabov's Southern Vowel ShiftExercise 3Identify in the following list of words those vowels that would be involved in the SouthernVowel Shift. Is the vowel of the word involved in (1) the gliding of KIT and DRESS vowels(e.g. bid, bed), (2) the lowering of FLEECE and FACE vowels (e.g. beet, late), or (3) the backGOOSE and GOAT vowels moving forward (e.g. boat, boot)?C05p. 215

dBecause the Southern Vowel Shift and Northern Cities Vowel Shift involve very differentrotation patterns, the major varieties characterized by these vowel shifts have over the yearsbecome increasingly different from one another. In fact, this differential rotation is the majorreason why many dialectologists now claim that Southern and Northern speech are currentlydiverging rather than converging. Regionally, the Southern Vowel Shift encompasses most ofthe Southern region indicated on the map in figure 5.11, although the reversal of KIT andFLEECE is confined mostly to the Inland South. In addition, whereas the Southern VowelShift is relatively stable, it is slowly receding while the Northern Cities Vowel Shift continuesto advance. Another important differences between the two shift patterns is the SouthernVowel Shift is more advanced in rural areas of the South than in metropolitan areas, theconverse of what we have seen for the Northern system, in which change radiates outwardfrom, and are most advanced in, urban areas rather than rural locations.Exercise 4Identify whether the vowels in the following words are involved in the Northern Cities Shiftor the Southern Vowel Shift. In some cases, the same vowel may be involved in either theNorthern Cities Shift or the Southern Vowel Shift, but the rotation will be in quite differentdirections. There are three types of answer: (1) Northern Cities Shift, (2) Southern Shift, and(3) both the Northern Cities and the Southern Shift, but rotating in different directions. Incases where the same vowel is subject to both the Northern and Southern Shift, identify thedirection of the rotation for each shift. You might try producing some of these voweldifferences, especially if you know someone who is a good model for the particular aidAlthough these rotational schemes represent two major shifts in American English vowels,they are not the only regional areas involved in vowel shifting. For example, NorthernCalifornia is experiencing a vowel shift indicating its emergence as a distinctive dialectregion. Like the Southern Vowel Shift, the back vowels are moving forward, so the GOOSEvowel becomes more like giws and the GOAT vowel more like gewt. But the front vowels areshifting in quite different directions, so that the KIT vowel is raising towards the FLEECE vowelbefore ng (as in something like “theeng” for thing) but lowering towards the DRESS vowelbefore other consonants (as in something like dead for did). Meanwhile, the DRESS vowel islowering towards the TRAP vowel. The TRAP vowel, in turn, is shifting in two directions; itbecomes a diphthong like stee-and for stand before nasals while shifting towards the LOTvowel elsewhere, as in “bock pock” for backpack. The NORTHERN CALIFORNIA VOWEL SHIFTis shown in figure 5.12.C05p. 216

Figure 5.12The Northern California vowel shift (adapted from Penelope Eckert’s web site:http://web.stanford.edu/ eckert/vowels.html)Interactive chart: Northern CA vowel shiftAudio map: California accentsSome parts of the Northwest such as the Seattle, Washington (Scanlon and Wassink 2010;Wassink forthcoming), and Portland, Oregon (Ward 2003), are currently undergoing vowelshifts that are associated with an emerging regional identity, but other regional dialects aredefined chiefly by their lack of participation in the sweeping rotations of either the NorthernCities or Southern Vowel Shift. In these regions, the TRAP vowel, a pivotal vowel in theNorthern Cities Shift, is relatively stable, and there is a merger of the low back vowel ofTHOUGHT and LOT, as in the same pronunciation for caught and cot or dawn and Don. Theapproximate area encompassed by this LOW BACK MERGER is given in figure 5.13.C05p. 217

Figure 5.13The Distribution of the cot-caught (LOT vowel-THOUGHT vowel) mergerIt appears that the merger in figure 5.13 radiates from three centers. One is in Eastern NewEngland, near the Boston area, which ext

Regional Varieties of English The regional varieties of American English have been a major focus since at least the early part of the twentieth century, when dialectologists began conducting large-scale surveys of regional dialect forms, particularly the Linguisti