My regular readers - that ever-rising tide of perspicacious humanity - will recall that I'm always complaining about a certain scientific view of reality as that which can be calculated. A regrettable correlate of this view is the assumption that what hasn't been calculated is unreal. Even more regrettable is the resulting equivalence between what can't be calculated in any sense (e.g., God), and what is merely difficult - or inconvenient - to calculate.
This is the besetting sin of economics in particular, but economics in its quasi-theological mode has shown a worrying ability to infect and cripple other disciplines. If refusing to worry about missing or nonconforming data is convenient for Milton Friedman, it's a moral duty for the rest of us.
Over at Defense Tech, Jeffrey Lewis has a must-read post on nuclear targeting protocols that makes these points brilliantly, along with many others. Starting from the tendency of organizations to "abstract reality in order to manage it," he goes on to discuss the reification of abstract, false, or imaginary data into reality; the logically indefensible exclusion of realities that can't be converted to data; and the episodes of all-too-literal insanity that result.
Only an organization would target 69 nuclear weapons on a single facility (later revealed to be the Sofrino missile defense radar) outside of Moscow in a strike designed to minimize "collateral damage". To take another example, STRATCOM calculates only blast damage from nuclear weapons. STRATCOM does not calculate the damage from any fires that would be ingnited, even though such fires would be far more damaging than any blast effects. Why? Because fire damage is hard to calculate and, therefore, not real.Read the whole post, by all means. Everything Dr. Lewis writes is required reading, but this is one of the best blog posts I've seen in a while.