If human language started with some large number phonemes and then lost them over time it would make sense but I would assume the opposite is true. Humans may have started in South Africa and spread from there but that was probably 100K years ago. It wasn't a linear migration but more of back and forth sloshing with the edges expanding outwards. Once humans moved out of South Africa the numbers grew and the populations outside probably exceeded those in South Africa fairly early on. If diversity was the key to more phonemes then I would think areas where a lot of different groups intermingled should be where we see the most phonemes and that would be closer to modern day Israel. I suspect that over the last eight or so thousand years China and India would have held the highest average population density which should give them more phonemes. South Africa has been fairly sparsely populated for a long time and the bush tribes they are describing have had fairly small populations for a long time.

Sorry good theories but they seem to have some large holes.