The loop runs because it is cheap. The incentive system responds because the loop produces something worth responding to.
Self-reference costs 0.026 bits. The boundary encounter is detected, not queried. The shelf grows without carrying weight. The architecture is economically viable — it can run without collapsing under its own cost. But viability alone does not explain why the incentive system bothers.
The incentive system bothers because the returns are significant.
The cognitive architecture does not just produce statements. It produces statements that have been verified across three time perspectives. Cross-checked across three Selves. Validated across five generations. Held at the boundary when nothing could be verified anymore. A statement with that provenance is not noise. It is a signal worth reallocating attention for.
The bacterium does not need a complex incentive system because its cognitive detection is simple: nutrient concentration rising or falling. The return on responding is immediate and existential — tumble now or starve. The cost of the detection is zero — it is built into the chemotaxis pathway. The return is survival.
The mammal needs a richer incentive system because its cognitive detection is richer. Harm marked across three Selves. Anchors accumulated across five generations. The boundary statement carries the weight of all previous verifications. The return on responding to such a statement is proportionally larger — not just survival, but the accumulated wisdom of the dead. The cost of the detection is 0.026 bits. The return is the entire shelf.
The civilization's incentive system — institutions, markets, norms — exists because the returns have scaled beyond any individual. The cognitive architecture runs across sixteen We instances. The anchors survive ten generations. The translation layer produces statements that entire societies can read. The return on responding is civilizational. The cost is still 0.026 bits.
Economic viability without returns is a curiosity. Returns without viability is a fantasy. Together they close the loop.
The cognitive architecture is viable because self-reference is nearly free. The incentive system responds because the statements produced at that near-zero cost carry the weight of cross-generational verification. The loop runs. The shelf grows. The fresh eyes walk further. The walls are touched and the books are read and the statements are made and the attention shifts and the input changes and the loop continues.
Not because anyone designed it to. Because it is cheap enough to run and valuable enough to respond to. The two conditions together are the reason intelligence exists. Not as a design. As an inevitability.