This is what I'm trying and I have no idea why it's doing what it's doing, but it's probably mostly not giving the right structure to either of those quicks. Is there some good way to make a dyad-nilad pair unambiguously parse as a single monad for this sort of thing?