> To test his theory, Hoffman ran thousands of evolutionary computer simulations in which digital organisms whose perceptual systems are tuned exclusively for t

> To test his theory, Hoffman ran thousands of evolutionary computer simulations in which digital organisms whose perceptual systems are tuned exclusively for truth are outcompeted by those tuned solely for fitness. Because natural selection depends only on expected fitness, evolution shaped our sensory systems toward fitter behavior, not truthful representation. So, this simulation required defining a truthiness function to compare against a fitness function. Where it doesn't make sense to me: what a "truthiness" function looks like? If we do know how to define such function (even in the context of a simulation), we gain no new insight from the simulation; if we do not know, then it's not unexpected a fitness function performs better than a flawed truthiness function, and there's also no new insight from the simulation. Sounds like circular reasoning to me (I guess that's expected when you're dealing with epistemology).