Twenty minutes into the first session, before any god has been discussed, the teacher stops.
He says he wants to explain why he uses the Akan language at all.
Then he puts up four pairs of words — one Kemetic, one Akan — and reads the dictionary definitions of each.
That's it. He doesn't build to a conclusion. He shows you four pairs and moves on.
This is the demonstration set: the evidence the entire comparative project is asked to stand on, presented once, early, as a kind of credentials check.
It's worth looking at closely, because everything downstream inherits whatever strength it has.
One. Chattering.
Kemetic k-s-k-s, glossed in the dictionary as to chatter foolishly, to jest. Coptic writes it kas / kes.1
Akan kasa kasa — to talk much, talk noisily, to chatter, to dispute in contending words.2
Two. Roaring.
Kemetic h-m-h-m — to roar, to bellow, a cry, a battle-cry.3
Akan hama hama — boisterous; the sea rages as wind, or agitated by a storm; raging.4
Three. Speaking.
Kemetic ka / kai — to think, think out, meditate, speak, say, cry out, call out, tell out.5
Akan ka — to emit a sound, to utter, speak, say, tell.6
Four. Fixing.
Kemetic sha / shai — to fix, to appoint, to decide, to determine, to destine, to predestinate.7
Akan hye (the H-Y combination sounding as sh) — to fix, to set up, institute, constitute a point, to predestine, ordain beforehand.8
Look at what's actually being matched in each case.
Pairs one and two are onomatopoeic. Chattering and roaring are noises, and languages worldwide make noise-words that sound like the noise. English has chatter, yammer, blah-blah. Roaring gives us rumble, boom, bellow.
That's a real weakness and it's the standard objection. Onomatopoeia is the cheapest kind of cross-linguistic resemblance.
Pair three is better but broad. Ka meaning to speak is two consonants and a very common meaning. Short words with common meanings collide by chance constantly.
Pair four is different in kind.
To predestine, to ordain beforehand, to fix a point in advance is not a noise. It's not a basic concept like speak or water. It's an abstract, specific, slightly unusual semantic bundle.
And both languages attach it to a similar sound.
That's the one that would be hard to get by chance, and the teacher — to his credit — spends the most time on it.
Here's the part that's stronger than people give it credit for, and it isn't about sound at all.
The Akan dictionary was compiled by a Swiss missionary sitting in Akan country in the 1800s, writing down what people said so he could translate the Bible.9
That man had no connection to Egyptology. He wasn't reading hieroglyphic dictionaries. He had no reason to care what sha meant in Kemet and probably no idea the question existed.
He wrote hye means to predestine, to fix, to appoint because that's what people told him it meant.
"A Swiss missionary sitting in the forest in the Akan region... in the so-called 1800s had no connection to the hieroglyphic dictionaries."10
So the two data points were generated independently. Neither could have contaminated the other.
That's a genuine argument. It doesn't establish the conclusion, but it removes one of the most obvious objections — that the resemblance was manufactured by someone who wanted it.
The comparative method in historical linguistics asks for something this set doesn't supply, and the gap is not a technicality.
It wants regular sound correspondence across a controlled cognate set.
Meaning: not four striking pairs, but a systematic pattern where a particular sound in language A corresponds to a particular sound in language B, consistently, across many words, including boring ones.
That's what establishes relatedness, because it's what chance can't produce. Individual resemblances are cheap. Systematic correspondence is not.
Four pairs — two of them onomatopoeic — is not a controlled set.
And there's a further problem the corpus never addresses: Akan is Niger-Congo, Kwa branch. Egyptian is Afro-Asiatic. Those are different families by the standard reckoning, and the standard reckoning was itself built by the comparative method on much larger sets.
To overturn it you'd need to meet or beat that evidentiary standard, not sidestep it.
Two things are true at once and the corpus only says one of them.
The independence argument is good. These weren't cooked. A missionary in the forest and an Egyptologist in a library produced entries that happen to align on an abstract concept.
The set is far too small. Four pairs is an observation worth following up, not a demonstration. [n=2 SUGGESTIVE] at best, and that's being generous by counting only the two non-onomatopoeic pairs.
The teacher presents it in the register of a proof. It's the register that's wrong, not the observation.
Which is the same defect this build keeps finding: the material is more interesting than the confidence is warranted, and stating it modestly would cost nothing and gain credibility.
You're watching the four pairs go up and you notice the pacing.
He doesn't linger on the chattering one. He reads it, points at both entries, moves on.
He slows down on hye. He reads the whole Akan definition out. He repeats predestinate. He tells you about the missionary.
He knows which one is load-bearing. He just never says so.
Later you're looking at a claimed cognate somebody's excited about.
You ask the two questions this set teaches you to ask.
Is the meaning a noise, or a basic concept, or something specific and abstract? Noises and basics are cheap. Specific and abstract is expensive.
Could either source have contaminated the other? If yes, discount heavily. If genuinely independent, that's worth something.
You catch yourself wanting the comparison to work.
Not because you've checked it. Because of what it would mean if it did — a continuity across the continent, a language that never stopped being spoken, a civilisation whose descendants can still read it.
That wanting is doing work on your judgement, and noticing it doesn't stop it working. It just means you'll ask for the sixth pair instead of stopping at the fourth.
Give sha / hye the full treatment, because if any of this survives, it's this one.
The semantic match. Not approximate. To fix, to appoint, to decide, to determine, to destine, to predestinate against to fix, to set up, institute, constitute a point, to predestine, ordain beforehand.11
Those aren't two words that overlap. They're two words with nearly the same range — the same cluster of senses, in the same order of prominence.
That's more unusual than a single shared meaning. Semantic ranges are idiosyncratic; two unrelated words landing on the same range is a stronger coincidence than two words sharing one sense.
The phonetic match. Kemetic sha / shai. Akan hye, where H-Y makes the sh sound.12
Reasonable, and it involves an orthographic explanation — the H-Y digraph — which is a small extra assumption. Not a large one.
The independence. As above. Genuinely strong.
What would have to be true. For this to be inheritance rather than coincidence, you'd want the same sh correspondence showing up across other Akan-Kemetic pairs, systematically. One pair with a striking semantic range is a lead. Ten pairs with a regular sound correspondence is a finding.
What's missing. The corpus never assembles the ten. It has hundreds of comparisons scattered across many courses, but they're presented individually, as illustrations, never gathered into a set and analysed for regularity.
That's the frustrating thing. The material to build a real argument may exist across this teacher's corpus — he's been at this for decades and has published extensively. Nobody has ever assembled it into the form the question requires.
The demonstration set is four pairs pulled out to persuade an audience in twenty minutes. It was never meant to be the evidence, and it's being asked to serve as the evidence because nothing else has been compiled.
Strongest support. The independence of the two lexicographic traditions, and the semantic-range match on pair four. Both are real, and both are better arguments than the corpus's critics usually engage with.
The central tension. Four pairs presented as a demonstration cannot meet a standard built on controlled sets. And two of the four are onomatopoeic, which the corpus never flags as a weakness.
A second. The family-classification problem is never addressed. Niger-Congo versus Afro-Asiatic isn't an opinion someone holds; it's a conclusion from the same method being invoked here, applied at scale.
Open, and genuinely worth doing. Assemble every Akan-Kemetic pair across the whole Odwirafo corpus — hundreds of them — and test for regular sound correspondence. That's a real research project with a real answer. It would either produce the strongest case anyone has made for the continuity thesis, or settle it the other way. Nobody has run it, including the people who most want the answer.
Against Odwirafo's Comparative Linguistic Method, this page supplies the evidence base that page describes. The method is phonetic-and-semantic comparison; this is the sample offered to justify it.
Separating them exposes something. The method is applied hundreds of times across the corpus. The justification for the method is four pairs, presented once, in the first twenty minutes of one course, and never revisited.
Against Asante Dictionary — Primary Source Dating Methodology, there's a direct dependency. That page establishes the dictionary's provenance and date; this page shows what that provenance is doing argumentatively. The dictionary's independence is the strongest card in the corpus's hand, and its 19th-century single-compiler limitation is the same card's other face.
Chi — Unorthodox Etymology and Semantics
Both cases turn on the same question: when does a resemblance between words become evidence of a connection between traditions?
The chi material shows the failure mode in one direction — a term whose meaning drifted through translation until the drift was being taught as the original. Resemblance mistaken for identity.
This page shows the failure mode in the other direction, and it's the more seductive one. Here the resemblance is not manufactured by translation. It arose independently, which removes the obvious contamination route and makes the coincidence feel much more meaningful than it may be.
The insight the pairing produces: independence raises the value of a resemblance but does not convert it into evidence of relatedness. Two unrelated languages can independently attach similar sounds to similar meanings. That's why the comparative method demands regularity rather than independence — regularity is what chance cannot manufacture, and independence merely removes one way of faking it.
Which reframes the whole demonstration set. The teacher's best argument defeats the objection he anticipated — that the match was cooked — while leaving untouched the objection that actually matters. He's won a fight nobody serious was picking.
Sharpest implication. The evidentiary base for a decades-long body of work is four word-pairs offered in twenty minutes, and the material to build something far stronger is scattered, unassembled, across the same author's published corpus. The bottleneck isn't the evidence. It's that nobody has done the compilation.
Generative questions.