The higher level of achievement is a contribution to the evaluation of the text.
(Halliday 1985: xv)
All branches of linguistics are first and foremost descriptive and thus it is no surprise that text linguistics confines itself to describing what is, in other words to (selections from) already existing and usually published texts. The past thirty years have seen fascinating and lively debate about the nature and boundaries of linguistics, but one tenet has remained unchallenged: that linguistics is concerned solely with making descriptive and not prescriptive statements. While it is universally agreed that evaluating alternative grammars is a proper concern of linguistics, evaluating the comparative communicative success of two alternative sentences generated by any given grammar is not—despite the fact that both pure and applied linguists, in their role as teachers, are daily involved in telling students how to improve their linguistic skills.
There were, of course, in the late 1960s and early 1970s, important sociolinguistic reasons for emphasizing the validity of difference and denying the inherent inferiority of minority dialects. However, this battle has long since been won, following research into West-Indian English in Birmingham by Wight and Sinclair and into Black English in New York by Labov. Now the advances in descriptive linguistics of the last generation should give us the confidence to re-introduce evaluation, to admit what we have always secretly acknowledged, that some texts and some writers are better than others, and to try to account not simply for difference and for how existing texts mean, but also for quality and for why one textualization might mean more or better than another.
It is for this reason that I prefer to see any given text as just one of an indefinite number of possible texts, or rather possible textualizations, of the writer’s message—parts of this chapter, for instance, have passed through more than a dozen drafts, sometimes undergoing minor and sometimes major