Development of a typology for linguistic situations has been a continuing concern of sociolinguistic research. Thus far, however, typological studies have served primarily to demonstrate the difficulties involved in such an undertaking rather than producing results of obvious utility. With few exceptions, notably Greenberg 1956 and Pool 1972, sociolinguistic typology has taken a categorical view of data which is largely scalar. The arbitrary segmentation of scalar attributes leads to undesirable data-loss, but in the absence of a principled measure of significance there is no elegant alternative. Such a measure is to be sought in the study of the ways that sociolinguistic data are applied to practical concerns. In this paper, the general characteristics of sociolinguistic typologies and the particular features of existing models are reviewed and the following major issues are discussed: (1) treatment of scalar attributes and data-loss, and (2) development of objective measures of significance. (Author/AM)