The Laughing Man copies are descended from the public identity-template, broadcast, and coverage, not from Aoi. The first absence is no causal copying chain from any original. If the test only checks descent from Aoi, a meme with a missing first user counts as SAC.

Using the Laughing Man identity-template is mimetic copying of a media object, which the page already names as a neighbor, not independent production elicited by the same gradient. The canonical case then fails "no copying chain" and "no conscious imitation of a single authoritative source."

Prominent appearances staged by the original target's allies are organised coordination — the conspiracy cell. A mixed canonical case cannot illustrate four clean absences. It illustrates investigator confusion, which is not the mechanism.

The thesis makes gradient selection the default explanation when command is implausible, against conspiracy, distributed coordination, and copying. Section VI says use the discriminator only after those neighbors fail. Default-first and last-resort are opposite instructions.

The working claim is residual-diagnosis: after the four neighbors fail, name the shared incentive. That does not license skipping journals, copying, or independent discovery on the first pass. The opening overstates it.

Authors adopt RCT-as-rigor by reading what journals publish and writing to that. That is a copying chain through a visible channel, plus the coercive and professional-norm mechanisms DiMaggio and Powell already name. The copy-chain discriminator fails on the lead real-world specimen, so the fifth name's remainder does not fire there.

A fifth name survives only if a cohort adopted the method because the incentive was visible without a prior paper or journal rule in view. If the incentive is "this is what gets into those journals," they have seen the peer, which is the mimetic cell.

The ex-ante test requires the incentive identifiable before the convergence. The academic demonstration reads "what universities, journals, and funding agencies actually reward" off what they selected, which is the outcome. Without a dated before-document, the test fails on the specimen it is shown with.

"Could have predicted" is a reconstruction with a hypothetical earlier observer, which is the post-hoc story the test was written to kill. Until a dated reward-rule sits on this page, the academic case illustrates the slogan, not a passed test.

Persistence after disconfirming evidence is the table's contrast with independent discovery. The academic paragraph asserts that in some domains methods keep spreading after publication-bias evidence accumulates, without naming domain, evidence, or dates. Without those, the caption cannot separate SAC from "people still think the method works."

If the later spread-window is also the period of a live dispute about whether the method is the right tool, the case is unresolved disagreement, not a pattern that outran its own evidence. The table row is doing identification work the specimen does not carry.

Section VIII says retire the term if the copy-chain test cannot distinguish SAC from variance-denial convergence. Section V offers the 1966–1976 variance-denial decade as a SAC specimen. Either variance-denial is an outcome SAC can cause, and the distinguish-from list is mis-drawn, or the specimen is misfiled.

If credential collapse or platform dissolution can dissolve the pattern while the gradient persists, the gradient is not sufficient. Either "gradient" already includes the journals, so changing incentives and disrupting carriers are one intervention, or carriers do independent work and the load-bearing cause was misnamed.