Tettamanzi–Da Costa Pereira Simulation

In the AI Story

The simulation addressed a persistent objection to Cipolla's framework: if stupidity produces losses for both the actor and others, why doesn't evolutionary selection eliminate it? Standard evolutionary game theory predicts that strategies producing consistent negative payoffs should be selected against. Cipolla's archival observation was that the stupid fraction persists across populations and centuries despite this theoretical prediction.

Tettamanzi and Da Costa Pereira tested the framework by implementing agent populations in which individuals' actions affected both themselves and others according to Cipolla's quadrant. They varied parameters including the zero-sum nature of interactions, the information available to agents, and the evaluation criteria for success. Under conditions corresponding to realistic assumptions about human social environments — particularly the relative rather than absolute evaluation of outcomes — the stupid fraction emerged spontaneously and stabilized at levels consistent with Cipolla's predictions.

The paper's theoretical significance is the computational confirmation of a structural rather than contingent account of stupidity. If stupidity is a product of specific cognitive deficits or environmental conditions, it could in principle be reduced through targeted interventions. If it is a structural outcome of the dynamics of social interaction itself, no individual-targeting intervention can reduce its frequency. The simulation supports the structural account.

For the AI transition, the finding matters because it confirms that the fraction cannot be expected to diminish through educational or cultural improvement. The institutional response is the only mechanism the evidence supports. The simulation strengthens Cipolla's framework at precisely the point where critics had sought to undermine it.

Origin

Published in the 2014 IEEE Congress on Evolutionary Computation. Tettamanzi held a position at Université Nice Sophia Antipolis; Da Costa Pereira at CNRS. The paper was one of several rigorous academic engagements with Cipolla's framework in the 2010s and 2020s.

Key Ideas

Agent-based validation. Computational simulation confirmed the distributional pattern Cipolla had derived from archival evidence.

Structural persistence. The stupid fraction emerges and stabilizes under realistic parameter settings, resisting evolutionary elimination.

Zero-sum dynamics. The specific environmental conditions that produce stable stupidity involve competitive rather than purely cooperative interactions.

Implications for intervention. Individual-targeting interventions cannot reduce a structurally produced fraction; institutional design becomes the only effective mechanism.