A Test That Can Be Convinced

A method for deciding whether two traditions truly converge is only a test if more than one verdict is possible. The translation-strain approach has been built almost entirely as a downgrade engine: it breaks an apparent match into smaller claim units, shows what must be bent, dropped, or added, and the answer is always that the two are less alike than they looked. A device that can only ever lower a score is not measuring convergence; it is performing skepticism, and no comparison can prove it wrong. To become a test, the procedure must be able to reach three honest endings: this much is really shared, this match is false, or we cannot yet tell. It must sometimes raise confidence after decomposition, not only lower it. The discipline that makes it a test is to state, before taking the likeness apart, what a surviving shared role would look like and what would count as confirming it, so the comparison can prove the doubter wrong.

The work on this frontier keeps adding new ways to subtract: check the borrowing history, check whether the repair task survives, check the receiver's wound, check whether a tradition even agrees with itself. Every addition is a reason a likeness is weaker than it seemed. None states what a confirmed convergence would look like, so no case could ever raise the program's verdict, and that one-sidedness is invisible from inside. The distinct move here is a symmetry requirement plus a pre-registered positive verdict: name the upgrade condition before decomposing, and treat a test that never upgrades as a failed instrument rather than a finding about the world. This differs from Smith, who prized difference and the useful manipulation of the gap but never demanded a confirm-or-expose-or-undecidable verdict with a stated upgrade bar, and from general falsifiability talk by tying the demand to a concrete scoring decision.

The asymmetry may be correct rather than a flaw. If genuine independent convergence between historically entangled traditions is genuinely rare, a method that mostly downgrades may simply be tracking reality, and forcing it to produce upgrades could manufacture exactly the perennialist false positives the whole program was built to prevent. Pre-registering an upgrade condition can also be gamed by setting a low bar. The Atman and anatta case is the anomaly that cuts both ways: the de-identification role survives decomposition, which looks like a confirmed convergence, yet the two terms were shaped by centuries of mutual polemic and shared milieu, so the survival may be causal contact rather than independent recurrence, which means both an upgrade verdict and a downgrade verdict mislead. If the three-verdict scheme cannot cleanly separate a thin-but-real shared role from a historically borrowed resemblance, the symmetry requirement is satisfied in name only.