There is disagreement in the literature about the earliest stages of the NCS. As mentioned above, Labov and his collaborators (as exemplified in, e.g.,
ANAE) generally describe the raising and tensing of /æ/ as the first stage of the NCS, creating a pull chain in which /o/ is fronted in order to fill the space left in the low front position by the raising of /æ/. McCarthy (2008), on the other hand, argues that the fronting of /o/ was the first stage of the shift; her argument is based on data from several speakers born in Chicago before 1900, who display fronting of /o/ but little to no raising of /æ/.
McCarthy’s contention that /o/-fronting preceded /æ/-raising is consistent with the behavior of /o/ observed in this study. Southwestern New England is the origin of the settlement of the Inland North, and it resembles the
Inland North in that its /o/ is markedly fronter than the /o/ of non–Inland North communities in the Telsur corpus. It does not closely resemble the Inland North with respect to /æ/. This suggests that the fronting of /o/ began early in the history of the NCS, before the present-day Inland North diverged from Southwestern New England speech; thus when the settlers of the Inland North region migrated westward, they already carried with them a somewhat fronted /o/. The Hudson Valley communities that were not settled by New Englanders did not necessarily, under this scenario, already have a fronted /o/, but the fronting of /o/ would have spread to them at a later date from both directions.
The backing of /e/ is a much newer change, as indicated by the fact that it is still in progress in Ogdensburg (and in the Inland North fringe data as a
whole, though not in any of the other individual cities) while raising of /æ/ and fronting of /o/ are not. This change apparently originated in the New York State component of the Inland North after it had already diverged from Southwestern New England, unlike fronting of /o/; for this reason, southwestern New
England’s /e/ is much less backed than New York’s Inland North communities while its /o/ is comparable to at least the Inland North fringe. Like /o/-fronting, /e/-backing appears to have spread from the Inland North to the Hudson
Valley; and then it must have advanced from there to southwestern New
England as well. Thus those two regions have an /e/ that is substantially backer than North American English as a whole, but still not as backed as in the Inland North in New York State.
According to this approach, the raising of /æ would (like /e/-backing) have originated in the Inland North after it had diverged from Southwestern
New England; but then, unlike /e/-backing, /æ/-raising never expanded
substantially into the Hudson Valley or, for the most part, New England beyond it. The raising of /æ/ may also have allowed the Inland North to develop a fronter /o/ than its Western New England predecessor system, by opening up additional phonetic space for /o/ to move forward into. But why should the raising of /æ/ fail to spread while the backing of /e/ and fronting of /o/ apparently spread easily into the Hudson Valley? To attempt to answer this, let us return to ANAE’s account of the origin of the NCS.
To review, ANAE argues that the tensing of /æ/ originated when the construction of the Erie Canal drew settlers from a variety of dialect regions, with a variety of phonological /æ/ patterns, into the same area. This account at face value does not fully account for the distribution of /æ/-tensing in New York State. Several NCS communities in this study are not located near the Erie Canal and did not benefit directly from the increased settlement it drew, but may have benefited from related Canal projects. Of Ogdensburg, for example, Hough (1853) writes that “the Erie canal hindered the growth of this portion of the state,” but the Oswego Canal, which opened five years after the Erie Canal was complete and connected the Erie Canal to Lake Ontario “was the first public work that conferred a benefit upon Ogdensburgh, or St. Lawrence county, as they thus gained a direct avenue to market.” Similarly, Glens Falls is not located on or near the Erie Canal, but it is on the Hudson River, which connects to the Erie Canal, and close to the Champlain Canal that connected the Erie Canal to Lake Champlain. On the other hand, Amsterdam is located along the Erie Canal and was founded and settled in the same general time frame as the NCS
communities in this study; but the presence of the Erie Canal was not sufficient to cause the NCS there.
Combining the Erie Canal explanation with this study’s findings of southwestern New England–origin settlement yields a consistent dialectological picture. Under such a combined explanation, the general raising of /æ/ under the NCS would have been not merely the result of a koineization of multiple incompatible /æ/ systems in one place. Rather, it is the effect of multiple incompatible /æ/ systems coming into contact with the original /æ/ system of Southwestern New England, in communities that were founded by Southwestern New Englanders but subject to increased migration thereafter as part of the Erie Canal population boom. This explains why Southwestern New England itself did not undergo /æ/-tensing—it did not (at least, at the relevant time) have an
inrush of settlers from existing communities with different /æ/ systems. It explains why communities such as Amsterdam that are on the Erie Canal but were not originally settled mostly by southwestern New Englanders did not undergo /æ/-tensing: these communities did not have the same Southwestern New England–derived /æ/ system as the base pattern, and thus did not respond to the phonological pressure of the incoming /æ/ systems in the same way.
Under this analysis, the outcome of the influence of diverse /æ/
phonologies on the underlying Southwestern New England /æ/—i.e., the Inland North general tensing of /æ/—differs in basic phonological structure from the /æ/ found in the Hudson Valley, rather than being merely a different phonetic manifestation of the same phonological features. This hints at why the general tensing of /æ/ did not spread to communities without Southwestern New
England settlement the way /o/-fronting did: although individual surface-level sound changes can diffuse from community to community with relative ease, it is more difficult for a change at the more abstract phonological level to spread directly. The nature of this phonological difference in /æ/, and the nature of the phonological change that would have had to spread into the Hudson Valley in order for full /æ/-tensing to appear there, will be the subject of Chapter 4.
On the other hand, it is still possible for communities in Upstate New York that did not directly experience the Erie Canal population boom but were founded by Southwestern New Englanders to have been affected by /æ/- tensing. These communities would have started with the same Southwestern New England–derived /æ/ system that was the substrate for the development of general tensing in the Erie Canal Inland North cities. By virtue of being in
Upstate New York, many of them along major trade routes that connected to the Erie Canal, they would have been in more or less regular linguistic contact with communities with a greater variety of /æ/ systems, including the Erie Canal communities themselves. Even if these communities did not experience the kind of intense influence from varied /æ/ systems that the cities on the Erie Canal did, their lesser degree of contact could have been sufficient to bring /æ/-tensing to them in a less general and consistent fashion—i.e., in the fashion characteristic of the communities described in this paper as the Inland North fringe.
A possible fault in this speculative scenario is that it requires
Southwestern New England and the Hudson Valley to have had distinct /æ/ systems in the period when the Inland North was beginning to be settled, and there is no direct evidence that they did. Likewise, it does not appear to be the
case that Southwestern New England and the Hudson Valley have distinct /æ/ systems now, nor is there evidence in the data of Kurath & McDavid (1961) that they did in the first half of the 20th century. However, it is far from implausible to suppose that such was the case in the late 18th and early 19th centuries. The settlement of the Hudson Valley, as discussed above, was in large part derived from non–English-speaking populations, either the original Dutch settlers of New Netherland, or more recent Dutch and German immigrants. Indeed, as mentioned above, Dutch was a principal language of Poughkeepsie until only a couple of decades before the migration from southwestern New England into the Inland North began (Platt 1987); and Campbell (1906) writes that in the early years of Oneonta’s settlement, around the turn of the 19th century, “German was the language of common conversation.” So during the decades after the
completion of the Erie Canal in the 1820s, much of the Hudson Valley was at best fairly recently English-speaking. It is highly probable that these second-language English speech communities at that time had substantially different phonologies, influenced by the very recent Dutch and German substrates, from those of the long-standing English-speaking communities of western New England.
The sticking point of this speculative scenario is, as usual, the border between Ogdensburg and Canton. Even if Ogdensburg was settled from
southwestern New England and Canton from Vermont, why should that prevent /æ/-tensing from spreading to Canton as well? Most of Vermont itself was settled from southwestern New England, and Boberg (2001) argues that the present-day dialectological differences between Southwestern and Northwestern
New England are quite recent in origin, so in all probability26 the communities in
Vermont from which Canton’s settlers originated had the same /æ/ phonology as the Southwestern New England origins of the Inland North. This is all the more in need of explanation if we take seriously the speaker in Rutland with a score of 4 as evidence of the NCS extending eastward beyond the Glens Falls area into Vermont, as described above. One possible explanation for the
difference between Ogdensburg and Canton is degree of contact with mixed and varied /æ/ systems at the relevant time. Ogdensburg is located on the St.
Lawrence River, and as mentioned above it commercially benefited from the Oswego Canal connecting the St. Lawrence to trade from the Erie Canal area; Hough (1853) also mentions the intense importance given at the time to the construction of a road from Ogdensburg to the Mohawk Valley. Canton, on the other hand, is not located on any major trade routes connected to more central parts of New York. If this explanation is the correct one, it may not matter whether Ogdensburg was predominantly settled from Northwestern or Southwestern New England. What matters in this case was that, at the time of the Erie Canal boom, Ogdensburg was in enough contact with the trade boom for its /æ/ system to be affected, and Canton wasn’t. Then the raised /æ/ was not able to spread to Canton from Ogdensburg for the same reason it never spread to Amsterdam: underlying phonological changes do not spread as easily as surface-
26 What Hough (1853) says about the settlers of Canton in general is merely that they were from
Vermont; he does not mention what part of Vermont. So it is possible that the preponderance of them came from the eastern part of Vermont, which according to Kurath (1939) was settled more from the Eastern New England dialect region and therefore may have had a different /æ/ phonology than Southwestern New England. (This boundary within Vermont can be seen on Map 22.) This is not a very likely scenario, however, both because the most populous areas of Vermont (and thus the most likely sources of migrants) are in the western part of the state, and because the two Canton pioneers whose towns of origin Hough does identify were both from the
level phonetic changes. This explanation is not entirely satisfactory in that it still does not really account for the speaker in Rutland, but it is perhaps the best than can be done with the limited data available.
This scenario reconciles the two accounts of the “initial stages” of the NCS. As McCarthy (2008) argues, the fronting of /o/ was the first step in the NCS, in the sense that it began earlier than any of the other sound changes thought of as being part of the NCS, before the divergence of the Inland North from Southwestern New England. On the other hand, as ANAE argues, the tensing of /æ/ was the triggering event of the NCS in the sense that that appears to be the change which uniquely distinguishes the NCS and the Inland North from the surrounding regions and their phonological systems.
It also, of course, resolves the conflicts between the accounts of the nature of the relationship between the Inland North and Western New England given by Boberg (2001) and in ANAE. Like ANAE, this chapter contends that the NCS raising of /æ/ is a unique phonological feature that is distinct from the
phonology of Southwestern New England, and would not have happened in an area that did not have the demographic history of New York State. However, the Southwestern New England phonology is also essential to the history of the NCS, to the extent that communities in central and northern New York that were not settled from southwestern New England did not develop it, even if in other respects they resemble the communities that did. Where Boberg’s analysis seems to predict a gradual boundary between the Inland North and Southwestern New England, and the ANAE analysis seems to predict a sharp boundary, this chapter finds a null boundary: the Inland North and Southwestern New England do not
actually meet, but are separated by the Hudson Valley. However, few phonological differences are observed between the Hudson Valley and
Southwestern New England, none of them very large or statistically very robust; from that point of view, the Hudson Valley can be considered to be
dialectologically united with Southwestern New England in the present day.27 In
that respect, the key feature distinguishing the Inland North from the Hudson Valley / Southwestern New England region is the tensing of /æ/. The boundary appears to be a combination of the gradual and sharp types: between the Inland North core and Hudson Valley is the Inland North fringe, where /æ/ is certainly higher than in the Hudson Valley and the other NCS features are more
advanced, but less homogeneously so than in the Inland North core; however, the difference between communities immediately on opposite sides of the boundary can be quite abrupt.