The question of whether socially-conditioned syntactic variation can be modeled in the same way as phonological variation has been debated since very early in the modern study of linguistic variation. While many have called for such an extension (cf. G. Sankoff 1972 and Weiner and Labov 1983), progress in the sociolinguistic study of syntactic variation has been hampered by an adherence to a framework first built for phonological variation, which emphasizes determining semantic equivalence between clearly identifiable variants. As a result, the quantitative studies of syntactic variation that do exist have focused mainly on syntactic variables which behave similarly to phonological variables leaving much syntactic variation unstudied. In this dissertation, I propose a binary distinction between syntactic variables that have clear co-variants (Type 1) and those that do not (Type 2), and I model Type 2 variation as microparametric variation (c.f. McCloskey 1992, Henry 1995, Henry and Cottell 2007, Cornips and Corrigan 2005). I illustrate that there are ways to quantify and understand the social factors affecting Type 2 syntactic variables, without the need to identify a co-variant or establish semantic equivalence, and I utilize an extended case study of the double modal construction of Southern United States English (e.g., I might could go to the store) as a prime example of a Type 2 variable. First, I provide a theoretical account of the syntactic structure of the double modal, showing it to be an example of microparametric variation. Next, I present four studies of the social factors constraining double modal usage and acceptance in Northeast Tennessee, which utilize a variety of methodologies from quantifying acceptability judgments, to corpus linguistics, to language attitudes. These studies show the double modal construction to be associated with lower class and less education while at the same time being associated with politeness. The pragmatic function of the double modal as a way to mitigate direct statements makes this non-standard construction a valuable tool for speakers of all social status levels. The multiple methods utilized in this dissertation highlight that while Type 2 syntactic variables resist many individual sociolinguistic methods, combining multiple sociolinguistic and syntactic methods can be successful. This dissertation, then, can be seen as one step in the direction of quantifying and modeling socially constrained syntactic variation in order to provide a more complete understanding of variation above and beyond phonology
Read
