It’s not about proving bad faith (“the system wants to fail!”), nor about pretending intention doesn’t matter. It’s about situated observation – recognising that every model of a system embeds assumptions about purpose and system boundaries.

Instead of imagining ourselves standing outside the system, measuring it against external goals, we ask: what is this system actually trying to stabilise? What does it treat as noise and what as signal? This reframes our understanding of “purpose” – not as a stated aim, but as the effect of internal regulation.