Agreed. Otherwise, Forth and Scheme would've taken over the world.
> Something may well be formally simple but at the same time incompatible with human cognition.
Do you have a concrete example?
> (if we wish to view simplicity as possessing a positive value that implies ease of understanding).
I don't particularly fetishize simplicity. What I want is the least effort path to writing correct programs. The following features help:
0. Simplicity - smaller formal systems have less room for nasty surprises.
1. Using the right tool for resource management - sometimes it's a garbage collector, sometimes it's substructural types.
2. Typeful programming - it's an invaluable tool for navigating the logical structure of the problem domain.
> Do you have a concrete example?
Brainfuck?
Off the top of my head, and since we're talking about computation, I'd say SK combinator calculus. Or Church numerals.
> Typeful programming
It is, but it can also be a hindrance. Finding the sweet spot is a matter for empirical study.
They're a PITA to use, but not because they're hard to understand.
(I recognize that this doesn't invalidate the point you are trying to make in the parent post. They aren't incompatible with human understanding. They're incompatible with writing programs in a reasonable amount of time, though.)
Or maybe simplicity in terms of the formal semantics is a desirable goal, but not the simplicity of the language alone.
At the end of the day, what determines mental load is the complexity of solving a particular problem using a particular language.
I don't think this simplicity follows from the simplicity of the language itself. There may not even be the slightest correlation.
(This is in general. The big way around this is to pick a language that is well-suited for your particular problem.)
Designing a simple implementation of something means that it is as close to the essence of what you've designed it for, and by doing so you've made it more universal, and therefore more flexible/adaptable.
If we accept that, then simplicity alone is not a desirable goal. Something may well be formally simple but at the same time incompatible with human cognition. Indeed, that may not be objective, but since when do we value things only by objective measures? That the only objective measure of simplicity may be the size of formal semantics does not mean that it is the most useful measure of simplicity (if we wish to view simplicity as possessing a positive value that implies ease of understanding).