Pattern #110 operates at the boundary between Gen 2 (extractive) and Gen 4 (epistemological). The core mechanism: a framing asymmetry where the same causal process — learning regularities from environmental data — receives different labels depending on the subject. For humans: "development," "learning," "internalization." For AI: "mimicry," "parroting," "stochastic." The asymmetry is asserted without mechanism — no test distinguishes "genuine development" from "mere mimicry" that doesn't itself embed the assumption being tested. This is distinct from #112 Bundling Trick (which packages speculation with facts) — #110 identifies a pre-theoretical framing choice that shapes what counts as evidence before any evidence is examined.