WITHIN the perspective adopted in this book I have claimed that language has no existence apart from the social reality of its users. Although language is a precondition for social life, it does not exist on its own and it does not simply reflect some pre-existing reality. I have tried to show how social and linguistic knowledge are intertwined by looking at some of the various ways in which social differences are encoded in speakers' choices both of variants within what is thought of as one language as well as between languages.
I commented in my Preface that sociolinguistics lacked a convincing theoretical model within which to situate and explain its findings. While sociolinguists have shown the importance of heterogeneity and developed powerful statistical methods for analyzing it, some critics have claimed that they have not really 'explained' it. There has been some confusion in sociolinguistic discussions about what it means to explain something, as well as about cause and effect. This is particularly true in studies making use of quantitative analysis which establishes correlations between certain social and linguistic variables. In fact, it is almost paradoxical that for many this kind of work (discussed in Chapter 3) is synonymous with sociolinguistics because in many cases once a sample of speech data has been obtained from a group of speakers representing particular social categories, the emphasis is subsequently almost entirely on quantifying, formalizing, and analyzing the linguistic variables. The social categories such as class, gender, etc. are taken as given and the social context which motivated the collection of data in the first place is often lost sight of in the final product. Within such statistical studies it is often easy to forget that speakers create and interpret language rather than merely respond passively to variables such as style, social class membership, etc.