Africa might be expected to have numerous language isolates, given the antiquity of humanity in the continent. However, there are few undisputed isolates, as well as a number of languages which have been considered isolates but which turn out either to be affiliated to a larger phylum or to be undecidable for lack of adequate data. The paper presents an overview of the main methodological issues which have divided African scholars, in particular discussing the view there may be many more isolates, as some phyla are insecure. The paper focuses on four languages, Laal, Hadza, Bangi Me and Jalaa which are relatively uncontroversial. The longer list, which includes Ongota, Shabo, Meroitic, Mpra and others is discussed in less detail. Finally it considers whether substrates can be detected in the language of foragers who now speak the languages of their settled neighbours.