Wikipedia says the following about commutative diagrams:
Commutativity makes sense for a polygon of any finite number of sides (including just 1 or 2), and a diagram is commutative if every polygonal subdiagram is commutative.
However it is not obvious to me that a diagram is commutative if every polygonal sub diagram is commutative. For example consider two commutative squares pasted together with arrows going to the bottom right of each square. I can't find a way to show the commutativity of the big square using the commutativity of the smaller ones.