The Northern Cities Shift

The most significant characteristic of Inland North speech today is the set of pronunciations associated with the Northern Cities Shift (NCS). The NCS describes a series of sound changes affecting six vowel phonemes. These changes are:

– KIT:  is backed and/or lowered to approach  in extreme cases.

– DRESS: /ε/ is backed and/or lowered resulting in forms such as .

– STRUT: /з/ ~ /Λ/ is backed and may also be rounded resulting in .

– TRAP/BATH/DANCE: /æ/ is fronted and raised to a mid or high position and is often produced with an inglide; i.e., [εə] or  . Phonetically these variants resemble those described above for tense /æə​/ in New York and Philadelphia.

– LOT/PALM: /a/ is fronted to near /a/.

– CLOTH/THOUGHT:  is lowered and/or fronted, often with unrounding, to something near [a].

The changes in the NCS are often represented as in figure 1 where the arrows indicate the main trajectories of the shifting vowels.

The NCS appears to be a fairly recent addition to the speech of the Inland North. Linguists first noticed the pattern in the late 1960s though the dialect literature provides evidence that some of the individual changes had been active for at least several decades earlier. For example, the Linguistic Atlas researchers noted the fronting of /a/ as a feature of the Inland North, and studies of college students in the 1930s reported /æ/ raising and /ε/ centralization as characteristics of Upstate New York (Thomas 1935–37). Regardless of when the NCS began, it seems clear that it underwent a great expansion, geographically and phonologically, in the second half of the twentieth century.

The order in which the individual pieces of the NCS appeared is a matter of some debate, but it seems clear that the changes to /æ/, /a/, and  are older than the others. One scenario holds that the shift started with the fronting and raising of /æ/, which drew /a/ forward, which in turn drew  down and forward. The shifting of /ε/ and  began later and their centralizing movement may have sparked the final piece, the backing of /з/ ~ /Λ/. The chronology of these changes is of great theoretical interest because they appear to form a chain shift. Chain shifting describes a series of related changes in which movement of one vowel causes movement in another. Representations like figure 1 make clear the apparent interactions among the shifting vowels. The scenario sketched here for the low vowels describes a “drag chain” where a vowel moves into an empty space vacated by a neighboring vowel. The alternative is a “push chain” where a vowel shifts into another’s space causing the latter to shift to avoid crowding. The interaction between DRESS and STRUT appears to illustrate a push chain.

The changes associated with the NCS operate unconditionally in the sense that the vowels may be shifted in any phonological context. By way of comparison, we might recall that in New York and Philadelphia, for example, the TRAP/BATH/DANCE vowel undergoes raising only in particular environments. In the NCS, by contrast, all instances of this phoneme are potentially subject to raising. Nevertheless, phonological context does play a role in shaping the NCS variation. For each of the shifting vowels, there are some phonological environments that favor the change and others that disfavor the change. Raising of /æ/, for example, is generally favored by following nasals or palatals (e.g., man, cash) and disfavored by following /l/ (Labov, Ash, and Boberg fc.). This does not mean that raised forms do not appear before /l/, only that raising is less common or less advanced (i.e.,  ) in these items. The details of the phonetic conditioning of the NCS can be found in the specialist literature (e.g., Labov, Ash and Boberg fc; Eckert 2000; Gordon 2001). Interestingly, studies of the NCS have not always found consistent patterns of conditioning across various communities. For example, Labov, Yaeger and Steiner (1972) found a following velar stop to be a disfavoring context for /æ/ raising in Detroit and Buffalo whereas it seemed to have the opposite effect in Chicago. More recently, in a study of rural Michiganders Gordon (2001) identified following /l/ as a leading promoter of /æ/ raising, a finding that runs counter to the effects reported by studies of urban speakers.

As the name implies, the NCS is associated with urban speakers from the traditional Northern dialect region. The most advanced forms of the shift are heard in the cities on and near the Great Lakes including Rochester, Buffalo, Cleveland, Detroit, Chicago, and Milwaukee. The national survey conducted by Labov and his colleagues finds evidence of the NCS (or at least some pieces of the Shift) in a vast stretch of the northern U.S. from Vermont, western Massachusetts, and Connecticut, across upstate New York and the Great Lakes region, and westward into Minnesota, northern Iowa and the Dakotas. In Ohio, Indiana, and Illinois the NCS is generally heard only in the northern counties; that is, in those areas included in the traditional Northern dialect region. This pattern is intriguing given that this dialect boundary, which divides the North from the Midlands, was established on the basis of older dialect forms collected over half a century ago. One major exception to the usual geographic restriction is seen in the appearance of NCS pronunciations in the Chicago-to-St. Louis corridor which takes the changes into the traditional Midland region.

The origins of the NCS may lie in the cities, but the changes are certainly no longer limited to urban speech. In Michigan, for example, quite advanced forms of the shift are heard even in small towns and rural areas. The changes appear to follow a pattern of hierarchical diffusion, spreading across large cities, then to smaller cities, and eventually to small towns (Callary 1975).

A number of studies have examined the sociolinguistic distribution of the NCS. This research has often found significant differences across gender lines with women’s speech displaying more advanced forms of the shift. Such a finding is consistent with the common sociolinguistic tendency of women to be in the vanguard of language change. Sociolinguistic studies have also found that the NCS is generally characteristic of white speech; for the most part African Americans and Latinos do not participate in these changes.

Among other sociolinguistic effects, we might also expect to find class-based differences. The results on this score have been variable. Early research along these lines from a survey of Detroit suggested the changes are especially prevalent among the working and lower middle classes, or at least among women of these classes. Men showed very little class differentiation. A similar interaction of class and gender was also found in a later study by Eckert (2000) who conducted ethnographic research in a suburban Detroit high school. Eckert found that some of the changes in the NCS functioned primarily as markers of gender difference while others appeared to have associations with the class-based distinction of the Jocks and the Burnouts, the two main rival groups of students. Today the NCS can be heard in the speech of all social classes and even in the local broadcast media.

As a final sociolinguistic observation, it should be noted that the NCS has acquired very little social awareness in the areas where it has become established. For the most part, speakers with the NCS do not recognize it as a distinctive feature of their region, though the NCS pronunciations are readily noticed by listeners from other areas. The lack of salience of these very distinctive vowel shifts among the native speakers of the Inland North may be related to the traditional position of the dialect as a kind of national norm in the form of “General American”. The belief that their speech is “accentless” remains very common among Northerners (especially Michiganders) today.