From Monotonic Complementarity to the Maximal Complementarity Zone: A Computational Test of the Relational Deficit Index

A previous paper, Algorithmic Reason on Trial, engaged Daniel Innerarity’s critique of algorithmic reason and proposed the Relational Deficit Index (IDR) together with a testable prediction: humanmachine complementarity gain should increase monotonically with a domain’s IDR. This paper reports a full, reproducible computational test of that prediction, including the model’s assumptions, the synthetic data generated, the statistical results, and the exact source code used to produce them. The result is clear and negative for the original formulation (Pearson r = −0.399, p = 7.25 ×10−13, N = 300), but reveals in its place a robust inverted-U pattern (mean quadratic t R2 = 0.601 versus 0.201 for the linear t, across ten independent seeds; vertex R∗ = 0.410 ± 0.009). This motivates reformulating the construct as the Maximal Complementarity Zone (ZCM). The IDR is retained as a measure of unmet relational demand but is abandoned as a direct predictor of complementarity gain; the ZCM instead defines the region in which that gain is maximized.