Drawing the Line: Perceptual Dialectology, Borders, and Southernness
University of Kentucky
One might suggest that a border is simply a line. Of course, there are many types of borders: geopolitical ones, like national borders; sociohistorical ones, like residential zoning lines prohibiting groups from living in certain areas; and topographical ones, like those created by rivers. And for each one, there is a convergence of people, practices, cultures, and identities that can be "characterized by conflict and contradiction, material and ideational" (Alvarez 1995: 448).
In the U.S., one of the most prominent lines where people, practices, cultures, and identities are constantly challenged is the line demarcating the American South. For many, the notion of Southernness is inherently dichotomous. It is friendly but stupid; polite but classless; quaint but backwards. Those ascribing to a Southern identity must often contest how they are perceived. And at the edge of the South, this contestation is more pronounced.
In this talk, Louisville, Kentucky is considered as a case study for examining, within a perceptual dialectology framework (e.g., Preston 1989), how being a peripheral member of the South impacts speakers' linguistic acts of identity, especially the perception of such identities. Louisville is one of the northernmost cities to be classified as part of the South. Its location on the Ohio River, on the political and geographic border between Kentucky and Indiana, places Louisville on the isogloss between Southern and Midland dialects as well. It is also conceived of as being located at various sociohistorical, cultural, and psychological borders, all pointing to one question - is Louisville really Southern? This presentation showcases how identity alignments in borderlands are neither simple nor straightforward and examines how perceptual dialectology can help answer questions like this one by providing an appropriate framework for understanding the fluid, complex, and dynamic nature of identities at the border.
Alvarez, Robert R., Jr. 1995. The Mexican-US Border: The Making of an Anthropology of Borderlands. Annual Review of Anthropology 24: 447-470.
Preston, Dennis R. 1989. Perceptual Dialectology: Nonlinguists' views of areal linguistics. Dordrecht, the Netherlands: Foris.
What network structure can (and can't) reveal about linguistic variation: The loss of the Southern Vowel Shift in Raleigh
North Carolina State University
Sociolinguists have recognized the potential of social network analysis to uncover the interactional processes that create patterns of sociolinguistic variation (such as social class patterns) and also promote or inhibit the diffusion of linguistic innovations. Early and recent approaches to network analysis in sociolinguistics have used diverse, often community-specific information about speakers' ego networks, usually generating a single combined index of network scores. Granovetter's (1973) "strength of weak ties" principle - that the diffusion of information hinges on weak rather than strong ties - has been influential: linguistic innovations were, and are, often assumed to travel between groups or communities via weak ties, and sociolinguists commonly believe that early adopters of linguistic innovations are those with many weak ties to speakers outside the local community.
I endeavor to move sociolinguistic network analysis forward both methodologically and theoretically in my investigation of contact-induced vowel shift in Raleigh, North Carolina. Since the mid-20th century, the migration of non-Southerners to the Raleigh area for technology-sector jobs has caused the quick shift away from the Southern vowel system to a (mostly) regionally unmarked system. For example, monophthongal /ai/ as in high has given way to a diphthong. I investigate the role of three interrelated social factors in shaping the trajectory of vowel shift across apparent time: occupation, sex, and social network position. The Raleigh study departs from previous network studies in two ways. First, I model speakers' network characteristics via a bipartite structure in which one set of nodes represents speakers and the other set represents the schools they attended. Second, I assess speakers' vowel systems in relation to their positions in the overall community network, rather than the characteristics of their ego networks.
In this talk I focus on the question as to whether linguistic variation - specifically, any two speakers'a linguistic similarity - is a function of structural equivalence, meaning the extent to which the two members of each dyad have similar network positions. One conclusion is that although linguistic similarity correlates significantly with network similarity, the network effects are stronger for women, and especially women with blue collar jobs. In view of the fact that some previous sociolinguistic network studies, including Milroy (1980) and Labov (2001), also found network effects mostly (or only) for women, this finding suggests that the present bipartite approach reaches the conclusions of earlier, more community-tailored studies in addition to offering new kinds of conclusions.
Continued engagement with contemporary quantitative network methods and research can foster new hypotheses about the relationship between language and social network structure. In particular, I propose a hypothesis that, based on a set of recent simulation-based network studies about the diffusion of cultural contagions, suggests that a "strength of weak tie" hypothesis is inappropriate for many linguistic variables. On the basis of the same simulation-based studies, I further suggest a straightforward hypothesis about the way in which community network structure could underlie a familiar social class pattern in sociolinguistics.