From Region to Class to Identity: "Show Me How You Speak, and I'll Tell You Who You Are"?

The twentieth century experienced an interesting shift of emphasis, even paradigms, in accounting for grassroots speech patterns in America. Broadly speaking, a speaker's regional background was originally regarded as the prime determinant of his or her speechways, while later social class membership was seen as decisive in shaping one's linguistic behavior. This shifting of emphasis in scholarship went along with a paradigm change from dialect geography to sociolinguistics as the leading subdisciplines, respectively. However, I suggest that neither region nor social class accounts for a person's speech behavior sufficiently; rather, a third factor, identity, has the greatest influence upon how an individual speaks. Clearly, as a research topic this has been an incipient concept, still marginal to linguistic theorizing; nevertheless, I find it significant and indicative of our increased understanding of the complexities of human language.

Perhaps it is necessary first to emphasize a point that to linguists counts as trivial but that popular audiences still find difficult to accept: all real-life speech is rule-governed, systematic and regular, good for all communicative purposes required. Language systems allow for a great deal of variability: we can express the same idea in different ways (carry water in a bucket or a pail, for instance, or drink either soda or pop), choose from alternative but functionally equivalent syntactic patterns (like he gave me a book and he gave a book to me), and utter the same word with varying pronunciations and accents (e.g., say time or tahm. Fundamentally, each of these alternative ways of saying the same thing will fully serve its communicative purpose in its appropriate context, and none is intrinsically "better" or "worse." We need to understand which contexts, or factors of usage, cause the one or the other alternative expression to be used, or, more generally speaking, cause an individual to speak the way he or she does.

Beginning in the 1930s, dialect geographers like Hans Kurath or Raven I. McDavid, Jr., were predominantly interested in regional differences in speech and therefore undertook large-scale regional dialect atlases. These projects yielded a wealth of data on regional speech differences, presumed to have been caused by settlement history and social or topographic patterns that affect communication density, like political boundaries, rivers, forests, mountains, and so on. It is to their credit that, unlike most of their European predecessors, they realized that in the mobile American society social class differences were also important, so they also systematically sampled speech differences between educated, common, and folk speakers from the same region. A later project like Lee Pederson's Linguistic Atlas of the Gulf States (1986-92) aims at a roughly representative sociological sampling of informants and also systematically documents social differences, but essentially a traditional dialect geographer might have said something like "Show me how you speak, and I'll tell you where you're from!"

Since the 1960s, William Labov's sociolinguistic approach, building upon the observation of natural, unmonitored speech, has taught us the importance of structured variability, systematic choices between alternative realizations. Having introduced quantitative tools, he found that speakers with a certain background tend to use certain variants more readily than others: females speak somewhat differently from males, the young prefer other expressions than the old, and African Americans have a dialect somewhat different from that of European Americans. While this approach has also resulted in fundamental insights into principles of language variation and change, essentially the descriptive goal of microsociolinguistic studies in that paradigm has been to document statistical correlations between speech variants, on the one hand, and extralinguistic, social parameters like class, sex, or ethnicity, on the other, with the latter having determined the choice of the former. So, a prototypical sociolinguist, while smilingly tape-recording your speech, might...