Choosing the mix of applicable models is the art of wisdom. Connecting all models pairwise would have O(N^2) complexity.
But I agree, writing down your thoughts on how models interact helps you understand their focal points and limitations.
To extend this line of thinking —one might hope that the synthesis of multiple models would get more elegant — but this would likely come at the expense of interpretability to particular situations. This line of thought is discussed extensively in the philosophy of science and complexity theory.
But nobody needs the full pairwise everything, just a few related concepts, like "If this seems close but doesn't quite fit, have a look at P, Q, R, and S?"
But I agree, writing down your thoughts on how models interact helps you understand their focal points and limitations.
To extend this line of thinking —one might hope that the synthesis of multiple models would get more elegant — but this would likely come at the expense of interpretability to particular situations. This line of thought is discussed extensively in the philosophy of science and complexity theory.