Kordofanian languages in the context of "Rashad languages"

The Atlantic–Congo languages make up the largest demonstrated family of languages in Africa. They have characteristic noun class systems and form the core of the Niger–Congo family hypothesis. They comprise all of Niger–Congo apart from Mande, Dogon, Ijoid, Siamou, Kru, the Katla and Rashad languages (previously classified as Kordofanian), and perhaps some or all of the Ubangian languages. Hans Gunther Mukanovsky's "Western Nigritic" corresponded roughly to modern Atlantic–Congo.

In the infobox, the languages which appear to be the most divergent are placed at the top. The Atlantic branch is defined in the narrow sense (as Senegambian), while the former Atlantic branches Mel and the isolates Sua, Gola and Limba are split out as primary branches; they are mentioned next to each other because there is no published evidence to move them; Volta–Congo is intact apart from Senufo and Kru.

The Katla languages are two to three closely related languages that form a small language family in the Nuba Hills of Sudan. Part of an erstwhile Kordofanian proposal, they are of uncertain position within the hypothetical Niger–Congo family. They do not share the characteristic morphology of Niger–Congo, such as the noun-class system. Thus Roger Blench classifies them as a divergent branch of Niger–Congo outside the Atlantic–Congo core. A similar situation holds for another Kordofanian family, Rashad; these are not closely related to Katla.

Volta–Congo is a major branch of the Atlantic–Congo family. It includes all Atlantic–Congo except the families of the erstwhile Atlantic and Kordofanian branches and possibly Senufo.

In the infobox at the right, the languages which appear to be the most divergent (including the dubious Senufo) are placed at the top, whereas those closer to the core (the similar "Benue–Kwa" branches of Kwa, Volta–Niger and Benue–Congo) are near the bottom. If the Kwa or Savannas branches prove to be invalid, the tree will be even more crowded.