1.) Cousin: The infinite and finite are correlatives. The correlation is in thought, and in this sense we obviously can’t know “infinite” without knowing what finite is; but the finite can only be understood relative to another as “this and not-that” and to do this demands reference to what cannot be finite. This is clearest by analogy to the finite things in place: if we are trapped in a room with no windows then we can know the location of all the things in the room but not the location of the room. The finite is known by the infinite in the same way that location is given by what can have no location.
Cousin’s hypothesis is clearest in mathematics but has applications in all discourses.
2.) Parmenides: Every being necessarily exists. What-x-is can’t cease to be x. Whatever sweet is (who knows? a class of molecules? An evolutionary response to calories?) can’t cease to be sweet unless it ceases to be. But whatever is in existence is a being and so can’t cease to be, which rules out the “unless” qualification just given. This is the generalized form of the final immortality argument given in Phaedo.
3.) Revealing nature by experiment. We want to see nature so far as we can control how it acts. But if we had complete control over how it acted it could not reveal anything to us. We would only need to consult our own intentions to know what it will do. The experimenter identifies knowledge with control and yet the locus of such identity could only be in one who thinks the universe into existence, and so would no need experiments to know it.