They're also generally much easier to create than constructive proofs, which makes them a convenient go-to for lazy academics.
They're also generally much easier to create than constructive proofs, which makes them a convenient go-to for lazy academics.
def eliminateTripleNot(f: ¬¬¬P): ¬P = {p: P => f({g: ¬P => g(p)})}
For a constructive proof of ¬¬¬P => ¬P? So it's really just ¬¬P => P that causes trouble. It's not clear to me though whether LEM is actually okay in the "fast and loose reasoning is morally correct" sense (i.e. if it's okay as long as you don't cheat and do something like use a non-terminating loop or exception to make your ¬¬P) though? Are there cases where you didn't "obviously" cheat where it's "wrong" to use it? In some sense, I see ¬¬P => P as a "cast" from an opaque function type to specifically be a closure {f => f(p)} for some p: P.> Are there cases where you didn't "obviously" cheat where it's "wrong" to use it?
In particular, it prevents us from adapting the proof to non-binary valued logics full stop.
> I see ¬¬P => P as a "cast" from an opaque function type to specifically be a closure {f => f(p)} for some p: P.
Now this is an interesting thought.
This is generally seen as a good thing by most "lazy academics". I guess your priorities are just different.
Constructivism is also not at all opposed to infinite quantities (that would be finitism). "Real-world domains" deal with infinite spaces (e.g. of all functions R->R) quite regularly across all scientific domains.