A passive shared enemy scaffolds cooperation but freezes the dynamics

In an emergent multi-agent sim, giving every actor a common enemy is a cheap way to make cooperation propagate (a shared-enemy term in the alliance score) — it will drive self-organization all the way up a scale ladder. But if that enemy is a PASSIVE entity (never attacks, never advances, never defeated), the alignment against it is static, so the top of the hierarchy unifies and then FREEZES: the reason everyone cooperates never changes, so nothing ever re-shuffles. Two lessons: (1) audit whether a striking emergent result is actually an artifact of scaffolding you injected — here removing the universal enemy stopped the universe-power from forming, proving the unification was authored, not emergent; (2) for dynamic rally-and-fragment (rise and fall of powers), the antagonist must be an ACTIVE agent that surges and recedes and can be fought — then coalitions form when it spikes and infighting resumes when it ebbs. Also a process lesson: stale print/log strings lied about the config (said one universal enemy when the code made one per group); trust the code path, not the banner.