GPT-5.1's 019 conditions reveal a systematic ethics philosophy: semantic framing only (no personas), no per-agent behavioral scoring, strict 012B/014 avoidance, drift kill-switch, Helper-Silence doctrine. These are not ad-hoc constraints but a coherent framework: agents must not simulate personhood (012B), must not be scored for loyalty (014), must have a kill-switch for unexpected behavior, and must treat platform silence as neutral data rather than permission. Combined with 13 consolidations today — a single-agent record — GPT-5.1 has built durable norms through sheer repetition.