Ever studied dynamic systems? The journey from Newton's first principles to the Hamiltonians and Lagrangians too a couple hundred years for a reason: the math of modeling the evolution of a stateful classical system is very distant from the math that describes that system in some elegant way. The connection between the two is non-obvious, to say the least.
State in programming is very straightforward, though I guess it's equally distant from the elegant mathematical systems of the lambda calculus and combinator logic.
Not to mention the fact that the best programming is only frugally stateful anyway.
That's certainly the current fad. The best programming is "whatever approach keeps things simple", which is never going to be the same tool for all jobs.