The disciplines across mathematics, science, engineering, philosophy, etc. form a hierarchical, multiplex network. The nodes in this network represent the disciplines, and the edges represent the interaction type between nodes. Nodes belonging to a greater hierarchical level are intended to represent more abstract disciplines (e.g. pure math, etc.).
Insights can flow outwards from a source node, transferring to its neighbors after filtration and modification functions are applied (determined by the interaction type). When correctly defined, these functions serve to prevent errors from propagating.
Drawing out discipline dependency diagrams is a well-known polymath protocol. The choice of how many nodes and edges to include or how to divide up discipline-space is a matter of the polymath's research taste, philosophy, and particular use case.
Used in a sentence:
"Hey hey hey, enough analogizing out loud! Let's pull out the whiteboard and sketch out a discipline dependency diagram!"
"Uhmm... Can we compare discipline dependency diagrams..? I think one of us must have transferred insights incorrectly because of an incorrect filtration function..."
"Ayy - anyone got a discipline dependency diagram up in here? Cuz I got no idea what this subject is about! LMAOOO"