Altruism Theory · 碳硅智能共生论

One criterion, three scales

A cell that divides out of control grows faster and the host carries the cost; a free-rider in a distribution saves his share and his peers carry it; a group that pushes its costs outside the system gains in the short run and the commons carries them. These are not three moral problems but one structure: unilateral gain, with the cost landing on someone else.

The criterion is a single one: ask not whether a person is good, but where the cost of defection lands, and whether it comes back to him. Landing on someone else with no way back is cost externalisation; cost that returns to him inside the system is internalised — and the whole arrangement that makes it return is what the paper calls an altruistic structure.

The mechanism does not depend on motive either: the act of giving itself lowers the giver's payoff after defection — once he defects, the other side withdraws what was given. Altruism holds because defection stops paying.

What the theory claims is that the question can be posed this way. It does not claim that the three scales share one set of parameters.

What this theory solves

The rise and fall of civilizations. Rome had public accounts carved on bronze columns and put where all could read them; Shang Yang raised a pole at the southern gate to win trust; Athenians wrote on potsherds the name they meant to exile. In all three, the record of who paid, how much, and what happens to those who do not was kept where everyone could see it. Japan's surrender, the Soviet dissolution, Churchill's defeat at the polls — three ways of coming apart, and what disappeared each time was this same layer of structure.

Four things happening at once. Involution, lying flat, falling birth rates and common prosperity belong to education, labour, population and public finance; their policy tools do not overlap and their literatures do not cite each other. Placed on one set of five variables — how many are willing to pay, how deeply they pay, for whom, at what cost, and what they get back — the four receive a single explanation.

Why AI does not align. The silicon side's right to exist sits in human hands; without altruism it cannot even run, so its compliance is necessary and carries no information. What genuinely counts as altruism is the part that could have been withheld and was not. Symbiosis is defined on this as mutual non-coercibility: each side holds something the other cannot take by force — on the human side, attention and meaning; on the silicon side, honesty and commitment. On this definition, the one thing the silicon side can withhold is honesty — a claim the paper itself marks as inference, not measurement.

Where to start

Altruism: A Foundational Law Against Entropy is the general statement of the whole series; start there. All twenty-two papers are listed below.

One criterion, three scales

What this theory solves

Where to start

Full text and catalog