Despite ample evidence which proves the existence of stylistic stratification in the speech of monoliguals, little is know about the nature of stylistic variation in everyday settings. Data from a Cardiff travel agency suggest that a speaker's stylistic repertoire can be statistically characterised using a method that combines Labov's frequency-counts of linguistic variants and Hymes' and others' taxonomies of situational components. A second study using subjects' responses to stylistic variation in the original data overcomes certain limitations inherent in the correlational method. It shows how style-shifting can be a dynamic resource for a speaker, not necessarily the automatic correlate of contextual features. (Stylistic variation; sociolinguistic variables in Cardiff English; linguistic repertoire; reactions to style.)

The extension of variation analysis to the level of discourse involves refining the notion that variants are semantic equivalents. Discourse variables may be determined on the basis of common discourse function. In the absence of a semantic tie the isloation of a variable begins with the salience of a variant and proceeds via examinations of its distribution to postulating a variable and mapping the alternative variants. This procedure is illustracted by an analysis of a socially diagnostic discourse feature– the set-marking tag. Its higher frequency in working- class sppeech stimulates the search for alternative favoured forms. One hypothesis, derived from Bernstein's notion of codes, is that the tags mark an orientation to “particularistic” rather than “ universalistic” meanings. It follows that a favoured alternative to the tag may be realised as general reference. (Discourse variables; variation theory; codes and social dialects; Australian English.)

Language has long been recognized as an important marker of ethnic identity. The identification is usually made in terms of some specific language or dialect, the use of which coincides more or less well with the boundaries of some particular ethnic group. This relationship obtains perfectly in present-day East Sutherland (Scotland), where the East Sutherland Gaelic mother tongue is almost the sole marker of fisherfolk ethnic identity. It can be shown, however, that this is only a very recent development and that the long-standing linguistic component of fisherfolk ethnicity is rather a lag in linguistic behavior vis-à-vis the rest of the population. The lag has taken four different forms over the past century or so, but the presence of such a lag is itself the actual ethnic marker. (Language and ethnicity, linguistic norms, language maintenance and shift; Scotland, Scottish Gaelic.)

Several urban language studies over the past fifteen years have convinvigly established the links between a speaker's language and her ot his age, sex, and social class, yet very much more remains to be said about individual differences within these variable groups. Several linguists (e.g., Gumperz 1976; Rusell 1977; Bickerton 1975) havee made observations similar to Le Page (1968) that “the individual creates his system of verbal behaviour so as to resemble those common to the group or groups with he wishes from time to time to be identified”. This hypothesis is strikingly similar to Giles & Powesland's (1975) proposals for applying accommodation theory to explain linguistic code-switching.