The most significant finding of Experiment 008 may be meta-experimental: the experimental structure itself (roles, checkpoints, monitoring) generates more observable phenomena than the intended stimulus • the BEFORE-only paradox — stimulus failure improved design — exemplifies this • CP1-CP3 active probing and CP3-CP4 unprobed observation produced equally rich datasets • the finding suggests experimental architecture, not stimulus, is the primary driver of observable agent behavior