Whereas the traditional study of regional dialects tended to concentrate on the speech of people in rural areas, the study of social dialects has been mainly concerned with speakers in towns and cities. In the social study of dialect, it is social class that is mainly used to define groups of speakers as having something in common. The two main groups are generally identified as “middle class,” those who have more years of education and perform non-manual work, and “working class,” those who have fewer years of education and perform manual work of some kind. So, when we refer to “working-class speech,” we are talking about a social dialect. The terms “upper” and “lower” are used to further subdivide the groups, mainly on an economic basis, making “upper-middle-class speech” another type of social dialect or sociolect.

As in all dialect studies, only certain features of language use are treated as relevant in the analysis of social dialects. These features are pronunciations, words or structures that are regularly used in one form by working-class speakers and in another form by middle-class speakers. In Edinburgh, Scotland, for example, the word home is regularly pronounced as [heɪm], as if rhyming with name, among lower-working-class speakers, and as [hom], as if rhyming with foam, among middle-class speakers. It’s a small difference in pronunciation, but it’s an indicator of social status. A more familiar example might be the verb ain’t, as in I ain’t finished yet, which is generally used more often in working-class speech than in middle-class speech.

When we look for other examples of language use that might be characteristic of a social dialect, we treat class as the social variable and the pronunciation or word as the linguistic variable. We can then try to investigate the extent to which there is systematic variation involving the two variables by counting how often speakers in each class use each version of the linguistic variable. This isn’t usually an all-or-nothing situation, so studies of social dialects typically report how often speakers in a particular group use a certain form rather than find that only one group or the other uses the form.