ABSTRACT
A second problem with traditional Arabic dialectology is the assumption that the 'most dialectal' forms are to be found amongst the oldest and least educated members of a community, and indeed, the assumption that there is a stable set of 'basic' forms shared by them. This has been shown to be fallacious by many recent studies in other languages. No speaker in any community in which in-depth sociolinguistic research has been done (e.g. Labov 1966 (New York City English), 1972 (Black American English); Trudgill 1974 (Norwich English); Bickerton 1973a (Montreal French); Jahangiri & Hudson 1982 (Tehrani Persian); Russell 1982 (Swahili)) has been shown to be 'single-style'. Whilst it is true that speakers may differ in the extent to which their speech changes as a concomitant of changes in some aspect of speech context, and that they may differ somewhat in the features they change, the major finding is that all speakers in all communities display some degree of variability (Hymes 1967:9). Even the oldest, most uneducated informant, it seems, has a repertoire of variant forms none of which can be considered more basic to him (whatever their historical status) than any other. All speakers, in short, appear to be 'multilectal' (to borrow a term from creolist linguists). If then, multi-lectal grammars are to be accurately described, data will need to be gathered in a variety of contexts which are different from each other, or in a single context in which there is a change in contextual factors.
