Section II's five traits are the working definition of "value," while Section I claims only some value complexes are replicators. Preferences are the contrast class. No value that fails the class test is named, so "replicator" is the universal redescription the class claim says to avoid.

A hypothetical private craft standard — held without conversion, punishment, or identity crisis — would be a value that fails the class. The political examples (equality, freedom, family honor) still do not fill the page's own specification slots.

The page's anti-collapse rule requires variant, host, pathway, population, fitness, horizon, and evidence that copying drives the distribution. "Equality" is not given those slots, so it remains a folk word. Five-trait mood does not show differential transmission.

Fitness as adoption, retention, institutionalization, or "something else" scores payoff-biased learning as replicator success. Hand-washing and safety culture can spread because they work, without identity fusion or betrayal-punishment. The five traits were supposed to mark the second path and are not applied to the symbiotic list.

Parasitic diagnosis needs the five-trait profile and host degradation. Symbiotic diagnosis is usefulness. Those are different tests. "Follow the evidence" can pass both on the single observation that it spread.

The fourth table row (physics constraint / growth versus stagnation) is an audit of replicator-class complexes, not a fourth kind of value. "Does this pattern keep a named host viable" is the symbiotic-versus-parasitic question. Growth can itself convert, fuse, and defend.

Growth can identity-fuse, generate automatic counters, and punish degrowth. Physics constraining which configurations last does not make the slogan a different type from equality. A pattern can keep a firm alive and wreck the population that staffs it, so "physics decided" and "this package copied itself into us" can both be true.

Until whose body and how long are named, the physics row does not settle fights. It relocates them into a host-and-horizon dispute. That can be a better question than ranking equality against freedom. It is not an external referee.

"Arguing is often the wrong layer" and "do not use that as a reason to stop arguing" have no discriminator. If every failed argument is a defense-reflex hit, the framework cannot lose. Listed conversation-stoppers can also be true.

Discriminator: can the other side name a host-effect they would accept as a reason to drop the package? If yes, keep arguing. If no evidence is admissible, the reflex is the live layer. "Have you read the sequences?" is homework; "that's elitist" as a closer is not the same act.

The proposed repair is a selection environment where accuracy outcompetes catchy falsehood, illustrated as an informational authority that rewards accuracy. That is a design requirement, not a present operation. Without a score that can fail in public, "reward accuracy" is peer agreement under a nicer name.