Here is the name of the supreme being, in one language after another.
Nyame — Akan, Ghana. Nyambe — Koko, Nigeria. Nyambe — Lozi, Zambia. Nzambi — Bakongo. Nzame — Fang, Gabon. Njambe — Herero, Namibia. Nyamien, Niamien — Agni and Baule, Ivory Coast. Amma — Senegal. Amen — among the Berbers, in the north.1
West Africa, Central Africa, Southern Africa, and the north. Different language families. Peoples separated by thousands of miles and, in some cases, by the whole span of the continent.
And the corpus's conclusion, stated flatly: we never stopped worshipping the divinity.2
Most of the comparative claims in this corpus involve two data points — a Kemetic word and an Akan word — and inherit all the problems that come with a pair.3
This one has nine or more, spread across a continent.
That changes the argument's shape considerably, and it's worth being precise about how.
A resemblance between two words can be chance. A resemblance across nine languages in four regions is much harder to attribute to chance, because chance doesn't cluster.
Something is producing the pattern. The question is what.
One: common inheritance. These peoples descend from a population that had this name, and carried it as they dispersed. That's the corpus's reading.
Two: areal diffusion. The name spread through contact, trade, migration and conquest, without common descent — words for high-status concepts travel readily, and divine names travel especially well.
Three: partial family relationship. Several of the languages listed are genuinely related. Bakongo, Herero, Fang and Lozi are all Bantu languages, and Bantu is a well-established family with a documented expansion across central and southern Africa in the last few thousand years.
Three is the one that does most of the work and the corpus never mentions it.
Nzambi, Nzame, Njambe and Nyambe being similar is what you would predict from Bantu relatedness alone. They are not independent data points; they are one data point with four attestations.
Which reduces the list considerably. Strip the Bantu cluster to a single item and you have: a Bantu form, an Akan form, an Agni/Baule form (also Kwa, and closely related to Akan), a Senegalese form, and a Berber form.
Still striking. Much less than nine.
Being fair about this, because the deflation isn't fatal.
Akan and Bantu are not the same family. Akan is Niger-Congo, Kwa branch; Bantu is Niger-Congo, Benue-Congo — related at a very deep level, but far enough apart that a shared divine name isn't automatic.
Berber is Afro-Asiatic, an entirely different family from Niger-Congo, and closer to Egyptian than to any of the others.
So there is a real cross-family observation here: something Amen-shaped appears in an Afro-Asiatic language in the north and something Nyame-shaped appears across Niger-Congo languages in the south and west.
That's the claim worth investigating, and it's a much narrower claim than the list implies.
The corpus supplies a derivation to explain the whole set.
Amen has the title Ariame or Aniame, and that becomes Onyame in Akan.4
And the root senses: amen means to hide, conceal, to be hidden, secret, mysterious. It also means to remain, abide, to continue, to be permanent, stable, fixed.5
Combined: the great hidden, stable, permanent and abiding being.6
The corpus's argument for why those two senses belong together is decent. The supreme being must be hidden — invisible, the source from which visible things come — and must be stable, the primordial thing everything else is built on. Neither alone would do.7
That's an argument about the concept, and it stands independently of whether the phonetic chain holds.
He reads the list and doesn't rush it.
Nyame, Nyambe, Nyambe, Nzambi, Nzame, Njambe, Nyamien, Amma, Amen.
Nine names, nine peoples, and the geographic spread named each time — Ghana, Nigeria, Zambia, Namibia, Gabon, Ivory Coast, Senegal, and the north.
The list is the argument. He doesn't analyse it. He lets the accumulation do the work, and the accumulation is genuinely impressive if you don't know which of them are related to each other.
You look up the language families and the picture changes.
Four of the nine are Bantu. Two more are Kwa and closely related to each other.
The list hasn't stopped being interesting. It's stopped being nine independent things.
You notice how effective the presentation was on you before you checked.
Not because anything false was said — every name on the list is real and every attribution is right.
The selection and ordering did the persuading. Which is worth remembering as a general property of lists.
This claim is unusual in being genuinely testable, so it's worth spelling out how.
What the corpus offers. A list of similar-sounding names for the supreme being across African languages.
What that establishes. That the names are similar. Which nobody disputes.
What would establish common inheritance.
Regular sound correspondence. Not that the names resemble each other, but that the differences between them are systematic — that Akan ny- corresponds to Bantu nz- consistently, across many words, not just this one. That's the comparative method's actual requirement and it's what distinguishes inheritance from resemblance.
Reconstruction. If the name descends from a common ancestor, a proto-form should be reconstructible, and the attested forms should be derivable from it by known sound changes.
Cognates in the same semantic field. If a whole religious vocabulary travelled together — not just the supreme being's name but words for offering, spirit, ancestor, shrine — that's much stronger than one item.
What would establish diffusion instead. A geographic gradient, with forms most similar near a plausible centre and more divergent further out, and no corresponding relatedness in basic vocabulary.
None of these is attempted.
And here's the frustrating part: Niger-Congo comparative linguistics is a real and active field. Proto-Bantu is reconstructed in considerable detail. The work required to test this claim properly is the kind of work that already exists for other vocabulary.
Somebody could check. The corpus doesn't, and neither does anyone arguing against it.
[n≥3 PATTERN] on the resemblance itself — the names really are similar across many languages, and that's a genuine observation. [n=1 OBSERVED] on the inheritance explanation, which is one hypothesis among three and the only one considered.
Strongest support. The resemblance is real, wide, and includes at least one cross-family instance (Berber Amen against Niger-Congo Nyame) that relatedness within Niger-Congo doesn't explain.
The tension. Several of the nine are Bantu and therefore not independent. The corpus presents them as separate confirmations, which materially overstates the evidence.
A second. Diffusion is never considered, and divine names are among the most readily borrowed items in any lexicon.
Open. Does a regular sound correspondence hold between the Akan and Bantu forms across other vocabulary? The comparative material exists. Nobody has run this specific check, and it would either give the corpus its strongest claim by a wide margin or settle the question the other way.
Against Four Words Offered as Proof, this page is the same method with a much better sample, and the comparison is instructive.
That set has four pairs, two of them onomatopoeic, all of them Akan-to-Kemetic. This has nine attestations across a continent.
More data, same missing step. Neither assembles the correspondences into a controlled set or tests for regularity. The demonstration set is too small to test; this one is large enough and isn't tested.
Which is the more frustrating failure, because here the material for a real argument is actually present.
Against Amen-Amenet — Hidden, Stable, Abiding, this page supplies the geographic distribution that page's etymology implies. If amen means hidden-and-abiding, and the name appears from Senegal to Namibia to the Berber north, then the concept travelled as far as the name did — which is the claim, and it's the part that would matter if the linguistics held.
Blumenbach's Three-Race Classification
Both pages are about classification schemes applied to African peoples, and about who gets to draw the boundaries.
The Blumenbach material shows a European taxonomy imposed from outside, drawn on criteria that served the classifier, and then treated as a description of reality rather than as a scheme.
The language-family classifications this page leans on — Niger-Congo, Kwa, Bantu, Afro-Asiatic — have their own history in the same period, built largely by European philologists, and revised repeatedly since.
The insight the pairing produces is a genuine difficulty rather than a resolution.
The corpus's implicit position is that these boundaries are suspect — that a scheme placing Akan and Egyptian in different families is a scheme built by people with an interest in separating them.
And that suspicion is not unreasonable given the history. The classification of African languages was done under colonial conditions by people who were not neutral.
But the response the corpus makes is to ignore the classification rather than to engage it, and that doesn't work. Niger-Congo and Afro-Asiatic are not maintained on nineteenth-century authority; they rest on regular sound correspondences worked out over a century by many hands, and they have been revised where the evidence demanded.
The way to overturn a classification is with the method that built it. Suspicion of the classifier is a reason to check the work, not a substitute for doing so — and the corpus has, in this list, more raw material for that check than anywhere else in its corpus.
Sharpest implication. This is the one claim in the corpus with enough data to be properly tested and an existing comparative literature to test it against — and it is presented as a list to be impressed by rather than a dataset to be analysed. The bottleneck is not evidence.
Generative questions.