Pronunciation Variation in American English
American folklore includes observations about a southern drawl, a Yankee whine, a midwestern twang, and other labels for perceived regional differences in speech. For these differences various fanciful explanations are offered: southerners talk slowly because of the heat; Philadelphians sometimes drop their rs because the humid climate produces such widespread sinus problems that the articulation of that sound is hindered. Popular attitudes toward social and regional versions of AE tend to be stereotyped. A certain kind of Boston speech is perceived as “classy,” whereas the speech of other regions provokes laughter or disapproval. How much of this folklore corresponds to fact? Is there really a southern drawl? A midwestern twang? As with other matters of folk or popular knowledge—medical treatments, for example—there is some truth and some fiction in such regional stereotyping. For example, contrary to the common belief, southern speech is not necessarily slower than that found elsewhere. The use of a rising pitch by some southern speakers where others would use level or falling pitches (making a southern statement sound like a question to outsiders) draws attention. Likewise, the lengthening of vowels at given places in sentences contributes to an impression of slowness. The amount of silence allowed in various types of discourse may also be greater for some southern speakers. But the latter characteristic is not unique to the South, nor do all southerners utilize pauses to the same degree. Although it is clear that different regions do have characteristic mannerisms of speech, which include the characteristics just mentioned and also such things as use of vocal qualifiers (nasality, aspiration, “rough” or “smooth” use of vocal cords), articulation habits, and so on, adequate documentation of the differences is lacking. Many other aspects of social and regional variation of AE have been documented, however, as will be detailed later in this section. But first, there are some general matters whose introduction will aid in the understanding of details.
Dialect. Although this term is popularly applied to speech that is heard as odd or “nonstandard,” most scholars use it to refer to any defined variety of speech, social or regional. From this point of view, every speaker has a dialect, one that reflects social and regional characteristics.
Standard English. Difficult to define, this term usually refers to the forms of English found in edited writing. Pronunciation, however, is much less subject to a recognized norm in American speech. It is normal for standard speakers in Atlanta to differ from their counterparts in Chicago, San Francisco, New Orleans, Boston, and so on. Broadcast (American) English as heard from professional announcers on radio and television has been developing as an informal norm, but it is far from having the force that Received Pronunciation has had in England. In other words, the United States tends to have regional standards of pronunciation.
General American. This much used and abused term most frequently refers to the speech of the western half of the country (excluding Texas and, perhaps, Oklahoma) and most of the territory east of the Mississippi that is north of the Ohio River (excluding the East Coast). Of course there are differences in pronunciation within this large area, but they are often less striking than those of the South or the East. There is some evidence that younger, college-educated speakers in the South and East are adopting some features from the General American area. And, in fact, this term occasionally is used to refer to educated speech across the country. To avoid possible misunderstanding, the term General American will not be used in the following discussion.
Urban vs. Rural. In pronunciation, as in other characteristics of AE, urban areas differ from their surrounding rural areas. Generally speaking, the larger and older cities of the eastern half of the United States are more differentiated than those of the western half, but even in the latter differences exist and seem to be increasing. Smaller, younger cities are more likely to represent an amalgam of features to be found in their rural environs, whereas older, larger cities may contain pronunciation characteristics that are relatively rare in the surrounding territory.
Social Factors. Social pressures to adopt one form of pronunciation or another are real and pervasive. American children very early begin modifying their speech away from that of the home to that of their peers. As their world expands they encounter new varieties of English from which they may choose features. Their choices are based on such influences as frequency of occurrence and their attitudes toward other speakers. The latter influence, sometimes termed prestige, is easily misunderstood. What the culture in general holds to be good is not necessarily what the individual will choose. In fact, there is evidence that many speakers knowingly use nonprestigious forms in order to separate themselves from standard social norms and solidify their identification with subgroups. Indeed, all speakers adjust their speech, including pronunciation, to fit the situation and the audience. These shifts are often so natural that they go unnoticed unless a speaker uses inappropriate forms. Not all pronunciation is affected by social class or situation; only some features, which have become markers within a given region, will vary across social groups and social contexts.
Social forces are constantly shaping the direction and nature of American English pronunciation. Some are national in scope, others regional or local. The degree to which these forces are balanced governs the degree to which speech is homogeneous. It is commonly argued that as such pervasive institutions as education and the media become more national, and as more people travel or change their place of residence, regional and local features of speech are weakened. This contention has some support. For example in the r-less areas of the country (where rs after vowels are not strongly articulated) some younger speakers have more postvocalic /r/ articulation than their counterparts had a generation ago. Contacts among speakers from diverse areas in large regional universities may be encouraging such standardizing trends. Reservoirs of regional and local speech exist in most communities, however, and, with the possible exception of folk pronunciation of individual words, they are likely to remain indefinitely. Furthermore, various appeals to regional, local, ethnic, and other kinds of pride may lead to resurgence and growth of variation. Even though pronunciation possibilities are limited and guided by biological and linguistic forces, degrees of homogenization and variation are also reflections of cultural developments.
Despite its complex patterns of variation, AE pronunciation is remarkably homogeneous over an area of approximately 3 million square miles containing more than 200 million speakers from diverse language backgrounds. Indeed, few sharp regional or social distinctions mark variant pronunciations; more often, gradient transitions and overlapping mixtures characterize what are commonly referred to as the dialects or accents of AE. Popular stereotyping of such features as Brooklyn’s boid and toid for bird and third notwithstanding, very few pronunciation characteristics are unique to an area or social group. Rather, a groups unique pronunciation can usually be defined only by a set of features, any single feature in the set being shared with one or more other areas that may or may not be contiguous with it. It is the combinations of pronunciation features that best reveal the unique character of any given area or social group. In fact, this sharing of individual characteristics extends beyond the boundaries of the United States. Many of the variant pronunciations found in the United States may be found in British English as well, although the proportional number of speakers in the two countries who use a certain pronunciation may differ dramatically, as may the social status of any given feature. Moreover, the combination of features for any individual speaker of British English will differ from those of any given American.
In spite of the high degree of homogeneity of AE, it is clear even to the casual observer that distinctive differences, geographical and social, do exist. Not surprisingly, the earlier-settled East Coast and South have more distinct regional and social differences than do the more recently settled Midwest, Southwest, and Pacific Coast. (We use these regional terms about language with only general boundaries in mind; language differences have little or nothing to do with state boundaries or administrative divisions.) Urban areas, in particular, are internally quite complex and are often distinct from the surrounding rural areas, even in the West where speech is generally more uniform.
Though the members of a group of speakers may sound much the same to an outsider, no area, large or small, achieves complete uniformity of pronunciation: there are likely to be subareal differences (neighborhoods, suburbs, communities) and social differences as well. The social groupings most likely to be reflected in pronunciation differences are age, sex, social class (as seen in education, income, occupation, and community type), and ethnic background. Even informal social groupings, however, may be marked by identifiable differences in pronunciation, many of them so subtle that group members themselves may not be aware of them. Pronunciation becomes even more complex when individual variation is taken into account. Speakers may shift pronunciation to fit the situation and the audience. They may change some pronunciations as they grow older and come into contact with other varieties of the language. In brief, regional, social, and individual variation in pronunciation is not exceptional: it is the norm. Moreover, all of these factors change through time. The reader would do well to keep this complexity in mind when consulting the information on pronunciation given later in this discussion and in the dictionary entries.
Nearly all present-day phonological variations are products of historical processes, recent or past. Many are the result of a sound being influenced by neighboring sounds—a process called assimilation, in which one sound becomes more like another. When such a process has affected one word, the change may then spread to rhymes or near rhymes of that word. This spread may occur rapidly or slowly, sometimes taking generations. Moreover, a process that begins in one regional or social group may spread to other groups. The total set of words involved in the process may not be the same among all groups and may vary over time.
In wash, for example, the /š/ has influenced the way sounds preceding it are articulated by some speakers. One result is an intrusive /r/, that is, an r sound that intervenes between the vowel and the /š/. Thus, many AE speakers pronounce wash as [wɑrš]. Many of these also say [wɑršɪŋ] washing and [wɑršt] washed. But these speakers may or may not pronounce Washington as [wɑršɪŋtǝn]: not all those who say [wɑrš] also have the intrusive sound in all similar words. The same sound pattern, a vowel plus intrusive /r/ before /š/, occurs in other words, such as mush [mɝš], mushroom [mɝšrum], and brush [brɝš], but the vowel that is most likely to invite intrusive /r/ is /ɑ/.
Another example of uneven distribution of phonological processes across regional and social groups is the influence of /š/ and the similar /ž/. Some speakers have only tense or diphthongized vowels before /š/ or /ž/: fish and dish are [fiš], [diš]; bush and push are [buš], [puš]; measure and special are [mežɚ], [spežǝl]; rush and mush are [rʌɪš], [mʌɪš]; ash and mash are [æɪš], [mæɪš]; and wash and Washington are [wɔɪ], [wɔɪšɪŋtǝn]. But many speakers do not have this process for all possible words in a rhyming set; some will say [mežɚ] or [spešǝl], for example, but lack similar variants for other words.
In brief, a phonological process may start with one word for one group (or one individual). It may then spread to other similar words and to other members of the speech community, but the growth of both kinds may be uneven in generalization or in rapidity. Furthermore, if the new variant encounters resistance—for example, if it is made fun of and becomes stigmatized—it may recede, and this shrinking may be as uneven as its spread. The first utterance of the new form may be treated as a peculiarity (good or bad) or may go unnoticed; it will be treated as a variant until it has spread to all (or almost all) members of the word set and throughout the population. (Those words to which it fails to spread become relics.)
Although the types of phonological processes are many and complex, most variations noted in DARE can be accounted for by loss, addition, modification, and change. Loss of a sound (or syllable) may occur in any part of a word and may affect sets of words, clearly revealing a pattern, or perhaps only a single word. One of the outstanding examples of patterned loss is that of postvocalic /r/. In some areas of the United States the r after a vowel in the spelling of many words is not pronounced. Thus farm may be [fɑ:m], rare [rε:ə̯], cord [kɔ:d]. Another patterned loss occurs when consonant clusters are reduced, most often finally in a word but also initially and medially. Some speakers thus may have [wel] for whale, [wil] for wheel, [wɑpɚ] for whopper, while others begin those words with /hw/. Similarly, huge, humor, and Hubert can begin with /j/ rather than /hj/. Some examples of nonpatterned loss are [laɪbεri] for library, [ˈgʌvmǝnt] for government, [ˈklʌmbǝs] for Columbus. For additional examples see sections 3.I.19, 3.I.22, and 3.II.26.
Addition of a sound can be illustrated by the intrusive /r/ in wash [wɑr] and idea [aɪˈdiɚ], the /ǝ/ in athlete [æˈθǝlit] and elm [ˈεlǝm], and the /t/ in across [ǝkrɔst] and once [wʌnst]. For further detail, see section 3.I.23.
Modification results in what native speakers would perceive as variants of the same sound. Modification is very common in AE and provides much of the subtle variation to be heard from one region to another. In some areas, postvocalic /r/, rather than being either lost or fully articulated, has an intermediate nature, producing what others hear as a “weak r.” Similarly, some speakers have a velar (strongly backed) postvocalic /l/ that is heard not as a completely different sound but as a different version of l. These kinds of variation are detailed in section 3.II.
Change or alternation involves the use of one distinctive unit rather than another in a word or set of words. For example, the vowels in fish [fiš] and bush [buš], compared with the more usual [fɪš] and [bᴜš], show the use of different sound units in those words by speakers of different regions or social groups. This kind of variation is examined in section 3.I.
To summarize: phonological variation in the United States is usually the result of sound changes that have spread irregularly through the population, being limited to given regions or social groups and affecting different word sets. The changes have numerous causes, including the influence of neighboring sounds. Of the four types of processes discussed, three (loss, addition, change) work across the sound units of the language, while the fourth (modification) involves slightly different articulation of a given sound unit. The number and kinds of words affected by these processes may produce patterned variation or more limited variation.
Some general comments about systematic approaches to AE pronunciation, discussed in section 1, bear restatement. The number of actual sounds produced by speakers of AE is incalculable, for pronunciation varies with each articulation of every individual. Some abstraction is necessary to deal with this very prolific reality. For the purpose of discussing phonological variation, linguistic scholarship has produced no completely satisfactory system of analysis. The scheme presented here sees AE as having altogether forty sound units or phonemes; all the phonetic elements (articulated sounds) that are taken to be variants of the same sound by native speakers (and that meet certain formal criteria) are considered to be a single sound unit and are represented by a single symbol. (These symbols are enclosed in diagonal slashes.) For example, in bite the vowel is represented by /aɪ/, but the phonetic elements that AE speakers actually produce (for which /aɪ/ stands) can be represented (in square brackets) by [aɪ], [ʌɪ], [ɒɪ], [a:], and others. It is at this phonetic level that much variation in AE takes place, with two exceptions. First, there is one phoneme, /ɔ/, that not all AE speakers have, and second, some processes produce differences of sound-unit occurrence in given words or word-sets. That is, sound units differ in their phonetic manifestations from region to region and group to group and in their incidence in words. No attempt is made in DARE to present all possible or known variants; some, because of the number of words or speakers affected, are of more interest than others. Judicious selection rather than exhaustive listing is the practice followed here.
Much variation in contemporary AE pronunciation is linguistic change working through social and regional variation. This situation is complicated by the fact that AE is, historically, a displaced language. That is, regional and social variations existing in British English were uprooted and mixed in various ways with one another and with other languages in a new context, producing American English. Some pronunciations that were quite limited either socially or regionally in British English became widespread in AE; others remained limited in North America as in Great Britain; still others died out. And, of course, North American innovations have occurred during the three hundred and fifty years since English was brought to the New World. Many, if not most, of the variations found in the United States may also be found in Great Britain; the outstanding difference between British English and AE pronunciation is the mixture of features to be found in any one place and in any one speaker.
The English that was brought to North America in the early seventeenth century was in the latter stages of a period that had already seen great changes in pronunciation. The tense (long) vowels had all raised their points of articulation, a process that simplified the vowel system by eliminating several distinctive sound contrasts but also produced two wide diphthongs, /aɪ/ and /au/. Furthermore, most of the significant changes that were later to appear in Standard London English were probably well under way among some speakers in England. For example, loss of postvocalic /r/ in certain phonological environments is evidenced as early as the sixteenth century in areas within and around London (the source of much early migration to North America). Rural migrants brought it into the speech of lower-class London, where it rose socially, became part of Standard London English, then spread outward again. It was probably indigenous to colonial North American speakers but was later brought more forcefully into AE through those regions which had kept continuous cultural contact with London English.
There is reason to believe that other features that later became standard pronunciations in London English were present in the speech of the lower classes. The seventeenth- and eighteenth-century change in which some words with /ᴜ/ (such as must) became pronounced with /ǝ/ must have begun much earlier, for it is nearly universal in AE, whereas in England some northern dialects retain the older form even today. Likewise most AE speakers have an unround vowel, for example [a] or [ɑ], in “short o” words like hot and stock where most English speakers have a round vowel, [ɔ]. The common AE form, however, does exist in British English, and there is written evidence for its occurrence in Britain as early as the seventeenth century. Again, it must have been brought to North America during early settlement to have spread so thoroughly in AE; at the same time, it became ridiculed as a foppish pronunciation in England and failed to gain general acceptance there. Similarly, the voiced (or flapped) medial t heard as /d/ in words like butter, latter, metal, which has become nearly universal in colloquial AE, has remained quite restricted regionally in England. Other pronunciation changes that took place in England were imported into America with more limited usage. The “broad a,” /ɑ/, of standard British English, rather than /æ/, in words like bath and cant, attained widespread usage only in eastern New England, with much more limited and sporadic use farther south.
It is not impossible that some of these features shared by American and British English are of independent origin. That is, as a result of contact with other languages or for purely phonetic reasons, features parallel with those of British English may have appeared independently in AE. Until more is known about language development, however, it is probably best to assume the simplest hypothesis for the shared feature: importation.
AE pronunciation began in the seventeenth century as a mixture of contemporary British English forms, but not all regional varieties were equally represented. The early colonists came mainly from southeastern England; settlers from other sections came in greater numbers later in the seventeenth century and early in the eighteenth. Nor were all social varieties of British English equally represented. The bulk of the colonial settlers were the more ambitious, dissatisfied, or unlucky members of the lower and middle classes. Thus the composition of the American population favored some social and regional speech varieties over others. This helped to bring about a simplification in AE, compared to the complexity of variation in British English, which remains to the present. The first two or three generations of North Americans probably further reduced variation by permitting some of the less usual forms to lapse. At the same time, some ongoing changes in standard British English were paralleled by those in AE. For example, the diphthongs in words like nice and now, which were probably [ʌɪ] and [ʌᴜ] at the time of settlement, moved (for many speakers) to their present [aɪ] and [aᴜ]. (Some speakers in parts of the East Coast and in Canada retain these earlier forms in such words as nice and about.) The influence of northern British English speakers became significant, especially in the South Midland area of the colonies, with the eighteenth-century immigration of the Ulster Scots, who brought many of the same pronunciations with them.
The eighteenth century was a period of language scrutiny both in England and in the New World. Regularization and formalization according to standards of correctness dictated by logic, Latin structure, and educated usage began to shape popular attitudes toward language. At the same time, the middle class was expanding and efforts toward public education were under way. Late in the century came attempts in the new United States to Americanize the language—especially through the efforts of Noah Webster with his speller and dictionary (1806). Educated use of English, American style, took root. By the late eighteenth century regional distinctions as well as an “American” accent were noticeable. Though the growing regional centers (Boston, New York, Philadelphia, Richmond, Charleston) had developed their own local differences reflecting their unique histories, they were still American. As an educated standard emerged, some peculiarities of pronunciation were probably excluded from the speech of the educated, but regional variations of the standard were tolerated (as they are today).
The American language was also influenced by the great numbers of speakers of other languages with whom the English settlers came into early contact. American Indian, Dutch, Swedish, French, and African languages, among others, were part of the colonial linguistic milieu. The influence of these languages on the lexicon and syntax of regional and standard AE has not been adequately documented. Equally problematic is non-English influence on AE pronunciation. Evidence for any early influence is ambiguous at best, since most of the possibly influential forms also occur in British English. Whether such features as /t/ and /d/ for /θ/ and /ð/ in parts (both north and south) of the East Coast, or lack of postvocalic /r/, are independently derived is open to question. The influence of African speakers, especially in the South where their relative numbers were high, is particularly controversial. Creole English, derived from African and Caribbean languages, was certainly present. How long it lasted and the degree of influence it has had (even on present generations of Blacks) is also open to question. Creole English lacks postvocalic /r/ as well as /θ/ and /ð/; it also lacks /æ/ and the contrast between /ɑ/ and /ɔ/. Neither of the last two characteristics is much evidenced in the current pronunciation of Black or White southerners, whereas the first two are (as also in British English and elsewhere in AE). However, various British English dialects not only have /æ/ and a contrast between /ɑ/ and /ɔ/ but also have speakers who lack postvocalic /r/ and substitute other sounds for /θ/ and /ð/. Thus, no final judgment about non-English influences on early seventeenth- and eighteenth-century AE pronunciation can be made at present.
The question of non-English influence is even more troublesome for the nineteenth and early twentieth centuries. Large numbers of Germanic, Slavic, and Romance-language speakers, Chinese and Japanese, and Protestant and Catholic Irish poured into already settled areas as well as the newly developing agricultural areas. Again the question arises whether identical features in widely separated areas are of independent and coincidental origins or come from a common source. Such cities as Milwaukee, Chicago, Cleveland, Buffalo, New York, Philadelphia, and Baltimore, for example, have numerous speakers who use /t/ and /d/ for /θ/ and /ð/, a common substitution in foreign-accented English. Outside the South, however, these forms are not common in nonurban areas, even those heavily settled by non-English speakers. It is possible that density of population and particular economic and language situations leading to differing patterns of assimilation produced the difference. Other variants such as [ŋg] in the common local pronunciation [lɔŋˈgaɪlǝnd] for Long Island or [ε] rather than [æ] in such words as that and ask in New York City have been attributed to the influence of Eastern European Jews. Yet the same features are found elsewhere in both AE and British English.
Some variations of AE pronunciations do, however, seem likely to be results of non-English settlement. In Wisconsin, Minnesota, and the Dakotas, for example, the tense vowels for many speakers have much less diphthongization than elsewhere, so that /e/ is [e] rather than [eɪ], thus closer to its pronunciation by German, Norwegian, and Swedish speakers who settled the area in great numbers. The same lack of diphthongization exists in parts of eastern Pennsylvania, also heavily settled by German speakers. There are other influences in the Cajun area of Louisiana, in the Pennsylvania German area, the Georgia–South Carolina islands and Low Country, and other communities, urban and rural, where non-English features are still noticeable. In general, however, enough of the evidence for possible non-English influence on majority AE pronunciation is coincidental and ambiguous that prudence dictates withholding judgment on its effect for the present.
The nineteenth-century westward movement of English-speaking Americans and new immigrants established the basic geographical patterning of phonological variation that to a large extent continues today (although developments since World War I have begun to modify it). Except in the South, most of the westward migration began in inland areas, with successive overlapping waves starting from ever more westerly points of departure. Settlers had a tendency to seek out living conditions similar to those they left behind. Thus settlers from western New England moved into New York state, northern Pennsylvania, northern Ohio, Indiana, Illinois, Michigan, Wisconsin, and, more thinly, westward to the Pacific Coast. Those who left northern Maryland, Delaware, and eastern Pennsylvania moved across central Pennsylvania into central Ohio, Illinois, Indiana, and parts of Nebraska, Iowa, Missouri, Arkansas, and Oklahoma. Southern plantation culture moved across the rich Gulf Plains area into eastern Texas. But once these movements reached far enough westward (about the 98th parallel of longitude) their character changed. The lack of navigable rivers in the West made overland travel necessary; moreover, rainfall could not be counted upon to sustain traditional crops. Thus many settlers either doubled back or leaped to the Pacific Coast.
During the mid-nineteenth century, foreign settlements, supplemented by a burgeoning cattle industry moving from northern Texas into the high Plains region, led to a more thorough mixing of settlement, and consequently of speech patterns, through the Rocky Mountains and the length of the Pacific Coast. The penetration of eastern regional pronunciations into the interior, where they came in contact with other varieties as well as with immigrant and American Indian speech, was one of the most important occurrences in the development of AE and its regional patterning. During this period regional consciousness reached a peak from which it is still descending. Inland urban centers with socially complex speech patterns were being formed, a trend greatly accelerated toward the end of the nineteenth century and on into the twentieth.
Another influence affecting pronunciation was that of the schoolmarm and the schoolmaster. Public education for an ever-growing portion of the population was increasing in importance. Along with it came a heightened desire for correctness in language usage. Regional standards of pronunciation were still tolerated, but increasing attention was being paid to better speech, especially in school activities like reading and spelling aloud, and this brought about spelling-pronunciations that in some cases reversed historical trends. Earlier, the -ing endings had been pronounced, in standard as well as nonstandard speech, -n (as in Shakespeare’s normal usage). Except in parts of the South, the -ing pronunciation was reinstated as, at least, the formal pronunciation. Similarly forehead regained a lost h. During the same period the United States came close to having one regional pronunciation become the standard for the country (as has happened in England). The influence of Boston, which enjoyed a cultural flowering in the mid-nineteenth century and which had long been the seat of the highest education, was felt across large parts of the country. New England-bred and -influenced teachers, ministers, and elocutionists spreading across the land left many an [ɑnt] aunt and [ˈrɑɚ] rather behind them.
A combination of sloth, indifference, hostility, and good sense put limitations to this trend, but the influence of public schooling should not be underestimated. Many non-English speakers were to be made American. Many English speakers were rising into the middle class and needed standards to speak by. This became especially important in the developing urban melting pot centers, where class consciousness was very strong. There is some reason to believe that foreign-language communities often adopt a kind of schoolbook English before complete language assimilation takes place. In general, public schooling has done much to reduce variation within regions. American Blacks, who as a group were long denied adequate public schooling, show some marked departures from regional and national norms, retaining older forms derived from British English and, perhaps, early Creole. Likewise, in Appalachia, where until recently public schooling was less constant than elsewhere, many older speakers still retain what are seen today as unusual forms (although the view that some pure Old, Middle, or Elizabethan English is still spoken there is an exaggeration and rests only on the survival of a handful of archaic words and phrases). In addition, the rise of large regional universities, by mixing many young speakers, has done much to level variations of pronunciation not only within but across regions. As a result, many college-educated, socially mobile speakers retain less regional and local pronunciation than their parents do.
Other twentieth-century developments have also modified the regional patterns established during the nineteenth century. Many southerners, Black and White, have moved to large northern cities, coming in sufficient numbers to prevent ready linguistic assimilation. The Detroit area, for example, though a northern city with a majority of northern speakers, also has many speakers with southern pronunciations. Meanwhile many northerners have moved to places like Florida and Arizona, altering the historically expectable patterns there. The development of huge urban centers with their rings of suburbs peopled by a highly mobile, socially selected population also is modifying older patterns. These urban areas, in time, also affect one another, with the result that two widely separated cities can be more like one another than like their closer rural neighbors. And finally, a familiarity with both the artificial standard of national radio and television and the speech of other regions, through the media and through travel, is also having an influence.
In brief, the process that began with London English assimilating many British English regional variations, turning them into urban, socially restricted patterns, some of which became part of the standard pronunciation, has been repeated (short of producing a national standard) in American cities. Moreover, through public education, great mobility, and communication, this process is happening nationally. Even as new variants appear and spread, and as many speakers become aware of the developing national standard, they come to view the older regional and local patterns as social liabilities. One casualty of this process is folk speech, the language of continuing local tradition unaffected by schooling. Formerly widespread, folk pronunciation now occurs less frequently, although conservative usage among some Black and Appalachian speakers gives the appearance of social and regional patterning. Pronunciations like [drin] for drain, [nɑri] for narrow, [ǰaɪn] for join, [ks] for ask, [εlǝm] for elm, all of which date to the eighteenth century or earlier, are now considered rustic even though some of them, like [ǰaɪn], were at one time fashionable, even literary. Meanwhile, even as some changes in regional and social AE pronunciation can be observed, more general changes continue to occur largely beyond conscious comprehension and control, responsive to social movement as well as to internal linguistic pressures.
AE variation is uneven in its density and complexity. The Eastern Seaboard and the South are more diverse than other parts of the United States; greater homogeneity occurs farther west. Not only are local and regional variation greater but urban social diversity is greater in older cities; moreover, western cities tend to reflect the speech of their surrounding regions more directly than do eastern cities. This too is changing, as the older cities extend their influence outward and the western cities begin to differentiate themselves from their rural areas. New York metropolitan speech is pushing into parts of Connecticut and New Jersey and up the Hudson Valley. Greater Los Angeles is beginning to differentiate itself from the rest of southern California (San Francisco has been separate for some time, for other reasons, from northern California). St. Louis is, in a number of ways, different from its surrounding area, whereas the newer and more western Kansas City is still a good reflection of its environs, with only some younger speakers beginning to initiate change. Similarly, between Boston and Richmond, a driving distance of little over five hundred miles, many differences of pronunciation are encountered. In contrast, from Seattle to Los Angeles, more than twice that distance, relative homogeneity exists. If an east-west view is taken, as opposed to a north-south one, a somewhat different picture appears. Starting at the Eastern Seaboard, it is possible to move directly westward finding decreasing variation the farther one proceeds.
The foregoing account needs qualification, however. Some pronunciation features, especially on the East Coast, do run north and south, though with regional pockets and disjunctures. Social features may modify or even override regional ones. Furthermore, lines that separate users of one feature from those of another are seldom clear-cut. If regions and groups are to be defined by their pronunciation, the definitions must be in terms of sets of often overlapping features, and at times they must incorporate quantitative and social qualifications. The more detailed the set of features, the more limited is the social or regional group being described. For example, if the first feature in a set is full articulation of postvocalic /r/, the possible area is very large, excluding only parts of the East Coast and the South. If a fronted pronunciation of /aᴜ/ as [ᴜ] is added, the likely area has been reduced to the South Midland, the western part of the Midwest, and much of the Southwest. If a lack of a contrast between /ɑ/ and /ɔ/ is added and the speaker is middle-aged or old, the area is further reduced to part of the high Plains and Rocky Mountain states. If full articulation of the wide diphthongs is added, it is likely to be an urban area. If alternation of /ε/ for /æ/, and a rounded vowel in “short o” words, are added, the set is typical of, for example, Wichita, Kansas. Again, to be fully specific it is often necessary to build social and quantitative data into the set. Ultimately, of course, each individual speaker has sets that themselves contain variables.
Although overviews of regional variations in AE are somewhat risky, they can be valid and useful if properly qualified. One such overview is that derived from data gathered by a series of projects known collectively as the Linguistic Atlas of the United States and Canada. Based on both lexical and phonological data collected over nearly fifty years to the present, it is particularly revealing of the effects of nineteenth-century migrations. The general areas established by Atlas researchers include four basic regions: (1) the North, representing New England and its westward extensions; (2) the North Midland, representing the Central Atlantic states and their westward extensions; (3) the South Midland, representing the Middle Atlantic states and their westward extensions; and (4) the South, representing the spread of plantation culture. There are, of course, numerous subareas within each of these regions. The data gathered for DARE often corroborate the regions described by Atlas fieldwork, but they also show mixing of features across “boundaries” and suggest the existence of still more subareas.
Two social groups, the young and the Black, are differentiated from those around them sufficiently to warrant special attention. Regional generalizations such as those in the following sections do not always account for pronunciations that are common in Black communities. In the South, the races share many pronunciation features, though the proportions of Black and White speakers having any single feature may vary. For example, /θ/ and /ð/ are frequently pronounced /t/ and /d/ in Black speech, less frequently, though not rarely, among Whites. It is possible that if social and educational backgrounds were held constant a greater similarity of use would be revealed. Another example is the pronunciation of aunt. Although both races have both /ænt/ and /ɑnt/ in the South, Blacks are most likely to have /ɑnt/ and Whites /ænt/. Outside the South, Blacks show varying degrees of assimilation to local and regional patterns. When the Black population is small and dispersed, as in a city like Minneapolis or in many small midwestern towns, assimilation is nearly complete. When numbers are larger or the concentration is great, as in many northern cities, preservation of southern features is common—for example, as in the greater likelihood of dropping of postvocalic /r/, preservation of the contrast between /ɑ/ and /ɔ/ (when Whites in the same community are losing it), and lack of a contrast between pin and pen. Here, too, of course, differing social and educational backgrounds produce marked variation. (Although details of pronunciation are different, these generalizations are also true for Appalachian Whites who have moved to northern and western areas.)
Young speakers, especially socially mobile ones, appear to be breaking with local speech patterns in favor of broader regional and perhaps even newly developing national ones. As the United States becomes more completely urbanized (integrated) in its social structure, the language will necessarily reflect that trend. The movement seems to be toward articulation of vowels farther forward in the mouth; continued loss of marginal contrasts such as /ɑ/–/ɔ/ (cot vs. caught) or /ɔ/–/o/ before /r/ (for vs. four); reinstatement of postvocalic /r/; and lowering of /i/ and /e/ to /ɪ/ and /ε/ before /l/, threatening the contrast between feel and fill, sale and sell. These and other adjustments appear to be spreading (although trends can change quickly and unexpectedly). Thus, many of the features discussed in succeeding sections may be inconsistent in young speakers use, more stable among older speakers.