I think this should be rephrased as: the disagreement is entirely over what happens when you introduce humans (and, even more specifically, human behavior).

The math is theoretically sound. Anyone disagreeing with that is a fool.

A good example is the argument over network control. The mathematics of blockchain very precisely describe what percentage of accumulated mining power introduces the opportunity to defeat distributed consensus. But it's the fact that there is a human behavioral incentive to try and accumulate that aggregate mining power that might be a real problem, not the math. Granted that's a low likelihood event.

This dichotomy emerges in countless different facets of blockchain, and I think is the simplest way of delineating the line between believers and naysayers.