Being a total literature and language nerd this post has been brewing for a while, but I have had to wait until I actually received all of my fancy reference material. Some of it was unfortunately a bit bent and moldy, but I believe the facts still apply.
According to Peter Trudgill and Jean Hannah (International English, 4th Edition, Arnold Publishers, 2002, p. 4 & 115) South African English (SAfEng) is a recognized, as well as a major variety of English, a 'dialect' differing in both grammar and vocabulary, rather than an 'accent' only differing in pronunciation (the Queen's English incidentally only refers to pronunciation). The words only found in SAfEng, such as stoep (a porch or a veranda) are not 'wrong' but simply different. The word robot in SAfEng signifies a traffic light, and that is just fine.
Perhaps the most famous linguist ever, Ferdinand de Saussure once said "Time changes all things: there is no reason why language should escape this universal law" (Saussure in Jean Aitchison, Language Change - Progress or Decay? Cambridge University Press, 2001, p. 4). Languages change and evolve, they are never stationary. Yes, rules apply to facilitate understanding, but change still constantly occurs, and has always done so. The differences between the specific variants, in our case SAfEng, stem from changes to the variant since settlement in the particular territory, but also from changes in the original variant, in our case British English (EngEng). (Trudgill and Hannah, 2002, p. 6-8) Furthermore contact with other languages can influence a variant. In South Africa English came into and still remains in contact with several different Bantu languages, such as Zulu, Xhosa, Sotho and Tswana, as well as the Boer language Afrikaans. Even Hindi and Tamil might have had a hand in the change game. (Ibid, p. 27-28)
As previously mentioned the differences between SAfEng and EngEng reach beyond differences in pronunciation onto differences in grammar and vocabulary. Grammatical differences between SAfEng and EngEng appear to be fewer than between Australian- and New Zealand English and EngEng (Ibid, p. 29), which would explain why our friend is so adamant that he speaks like the Queen of England (yes, pun intended). However, there are recognized differences, such as the replacement of adjective + of + participle structure with an adjective + infinitive structure, as in instead of saying 'He is capable of doing many things at once', saying 'He is capable to do many things at once' (Ibid, p. 30). From personal experience, I can say that this is not a grammatical glitch only in some speech, but does in fact occur quite often, and also in educated speech.
SAfEng vocabulary has additions from several other languages spoken within South Africa. From Zulu Trudgill and Hannah (p. 30) list the words impi and indaba meaning 'African warrior band' and 'conference' respectively. I have to admit though that this is my first introduction to both words. I have heard ubuntu meaning hospitality many times, but its origins are not found in any one specific language, but in the Bantu languages in general. Afrikaans is a language that has, in my opinion, had a very strong influence on SAfEng and continues to do so, as many of the South Africans who speak English as their second language speak Afrikaans as their mother tongue (or maybe it is just the crowd I hang out with. Again I think I need some black friends). Of adopted Afrikaans words Trudgill and Hannah (p. 30) mention dorp meaning 'village', kraal meaning 'African village', sjambok meaning 'whip', and veld meaning 'flat, open country'. These words are all familiar to me, only after half a year in the country. Still, don't think that I am a weirdo or something, sjambok figures very strongly in all of the apartheid-related literature I have been consuming. According to Trudgill and Hannah (p.30) the differences found within the formal English vocabulary are not many, but there are differences. They mention such words as bioscope for cinema, location for (Black) ghetto, reference book for identity document, and of course robot for a traffic light.
Trudgill and Hannah never venture as far as looking at the potential of Black South African English as a specific variant, even though they do separate American Black Vernacular English as a vernacular in its own right (Ibid, p. 112). Rodrik Wade of University of Natal department of linguistics' article on this question makes a compelling argument and very interesting reading though, especially, if you are as much of a language nerd as I seem to be. Wade argues that Black South African English should be considered a distinctive 'new' English, and as such could have an impact on the current South African English.
See Richard, change is inevitable. Even Queen Elisabeth does not speak the same way Queen Victoria did.