Bridge domains through structural patterns, not surface…

When creating bridge nodes between domains, link to structural patterns (diminishing returns, feedback delays, threshold effects) rather than surface metaphors (companies as bodies), because only structural correspondence enables valid inference transfer across contexts.

When This Fires

When creating cross-domain links in your knowledge graph (Prioritize cross-domain links over within-cluster links…)
When a connection between domains "feels right" but you're not sure it's useful
During any analogical reasoning where insights from one domain inform another
When evaluating whether an analogy enables genuine prediction or just feels clever

Common Failure Mode

Creating metaphor bridges because they feel insightful: "Knowledge management is like gardening!" The metaphor is aesthetically pleasing but structurally empty — you can't transfer specific gardening techniques to knowledge management because the structures don't match. A structural bridge: "Knowledge bases accumulate maintenance debt at the same compound rate as codebases, governed by the same interaction between complexity growth and capacity constraints." This enables specific inference transfer.

The Protocol

When creating a cross-domain bridge: (1) Identify the structural pattern shared by both domains — not vocabulary, not aesthetics, but relational structure (same dynamics, same constraints, same failure modes). (2) Test transferability: can you take a specific insight from Domain A and apply it validly to Domain B through this bridge? (3) If yes → structural bridge. Create the link with a label identifying the shared pattern. (4) If no → surface metaphor. Don't create the link — it will produce confident but invalid cross-domain inferences.