New economic and educational opportunities in post-apartheid South African society have seen the rapid growth of a Black middle-class, part of whose making and consolidation is having its children study at elite schools previously reserved for Whites. At the same time the society is experiencing class cleavages as the gap between the rich/new rich versus the poor/very poor has increased (as starkly brought to attention by the student uprisings of 2015). This paper explores the sociolinguistic consequences of these changes in South Africa. One theme that will be discussed at an introductory level is the effect of change on the indigenous African languages, and the scope for their improved status. More salient sociolinguistically is the strengthening of a notably “African” variety of L2 English, labelled “Black South African English”. The features of this variety that show what I call substrate fidelity will be discussed. In contrast is the rise of a “crossover” variety showing little or no such allegiance, as middle-class young Black speakers depart significantly from the traditional L2 of their parents’ generation. This will be the main linguistic focus of the paper, using two related sets of variables: the realisation of schwa as a full vowel, and neutralizations of vowel length. These are studied via the techniques of acoustic sociophonetics, using forced alignment. The first result concerns social class, consequent upon a bifurcation of young Black people according to type of schooling (elite or not). The second result is that there is a consistent gender effect among young people, in contrast to the previous generation’s English. The paper explores the socio-symbolic reasons for this gender differentiation, via commentaries in the press as well as young peoples’ own attitudes. Male speech in the sample appears to be constrained by some degree of African solidarity; middle-class female speech less so.