In this talk, I will share how one undergraduate student used set-based reasoning to reinvent logical principles related to conditional statements and their proofs. This learning occurred in a teaching experiment intended to foster abstraction of these logical relationships by comparing the predicate and inference structures among various proofs (in number theory and geometry). I will portray the progression of Theo’s set-based model from a model-of the truth of statements to a model-for logical relationships. I will connect this to our ongoing design efforts for teaching logic for proving.