n.V4-5.07| WHAT WOULD PROVE US WRONG

01| The method described here is coherent. Coherence is not evidence, and this book part currently rests on more design logic than data. Setting out what would falsify it is the only honest way to end.

02| Five things have to hold.

03| Learn-by-doing has to close a real capability gap, not a small one. Our claim is that people who did not succeed in formal education can reach adequate professional performance through structured real work. If competence on real projects plateaus early — good enough for third-league work and never enough for first-league work — then the method has a ceiling, and a ceiling means the barrio still needs outside professionals for anything complex. That would not kill IkoCiti. It would substantially reduce what we can claim.

04| Peer teaching has to hold quality across generations. This is the one we would bet against ourselves on. If the fifth generation is measurably weaker than the second, the compounding engine is compounding decay, and the tooling and AI monitoring described in V4-5.05 are not sufficient. The test is direct: compare cohort competence by generation, in the same CITI, on comparable work.

05| Contribution analysis has to survive competitive teams. If peer assessment inside a league-ranked team turns out to measure popularity, we have replaced an exam that measured the wrong thing with a social process that measures the wrong thing less visibly. That is worse, not better, because it looks like judgment.

06| The group instruments have to travel. Controversial-topics work, open dissent, drilled disagreement — these are the most culture-dependent components in the entire volume. If they only function in contexts with a particular relationship to public disagreement, then the method is partly local and the standard package shrinks to what actually transfers.

07| The opt-out right has to cost less than it saves. We expect it to preserve people who would otherwise be lost permanently. It could also produce chronic instability in teams and a pipeline that cannot be planned. Return rate after opt-out, and project delivery in teams carrying absences, are the two numbers that decide it.

08| What we do not know is larger than what we do. We do not know time to competence with any precision — the estimates in circulation are guesses with confidence attached. We do not know the ratio of taught modules to project hours that produces the best result. We do not know how much of the method a teaching Maveriq can carry before quality falls, which is the single most important operational number in this book part and the one we have least basis for. And we do not know what appears at volume, which is historically where designs like this meet their real problems.

09| There is also a dependency we cannot design around. Learn-by-doing requires projects, projects require buying power, and buying power belongs to V5-2.00. A CITI whose Exchange runs dry does not slow its training down. It stops having a training system at all, because the classroom was the work.

10| The position we take is this. The subtractions — no classes, no exams, no grades, the right to stop — are well grounded in what V4-3.00 and V4-4.00 establish about stigma and load, and we hold them with some confidence. The additions — contribution analysis, group instruments, the P2P engine — are reasoned but largely untested, and the P2P engine is carrying more weight than any single untested component should.

11| If the engine holds, everything else in IkoCiti scales behind it. If it does not, we have designed an unusually thoughtful training programme, which is a thing the world already has enough of.