Prioritize cross-domain links over within-cluster links — weak ties generate better insights
Create cross-domain links between notes from different topic clusters rather than only within-cluster links, because weak ties that bridge disparate domains generate more surprising insights than strong ties that reinforce existing knowledge clusters.
Why This Is a Rule
Granovetter's "strength of weak ties" theory, applied to knowledge graphs: links within a topic cluster (cognitive science ↔ psychology, management ↔ leadership) reinforce what you already know. Links between clusters (cognitive science ↔ software architecture, psychology ↔ organizational design) generate surprises — they connect ideas that your domain-organized thinking would never bring together.
In a knowledge graph, within-cluster links create dense neighborhoods where every node is already related. Cross-cluster links create bridges between neighborhoods, and these bridges are where novel insight lives. The insight that "technical debt accumulates like cognitive load" only emerges if your technical debt notes link to your cognitive load notes — two clusters that topic-based organization keeps separate.
Luhmann's Zettelkasten was deliberately structured to maximize these surprises. His notes were filed by connection, not by topic, which meant a sociology note might sit next to a linguistics note. The juxtaposition was intentional: the filing system itself created cross-domain encounters.
When This Fires
- Creating links for a new note and choosing which existing notes to connect
- Reviewing your knowledge graph and noticing dense clusters with no bridges
- During any linking session where you default to "obvious" same-topic connections
- When your knowledge system feels like it produces confirmations but not surprises
Common Failure Mode
Linking only within topics because those connections are obvious and comfortable. Your "decision-making" cluster has 50 internal links and zero connections to your "systems thinking" or "emotional regulation" clusters — despite deep structural relationships. The within-cluster links are easy to see; the cross-cluster links require you to ask "what does this note remind me of in a completely different domain?"
The Protocol
When linking a note: (1) Create the obvious within-cluster links first. (2) Then ask: "What note from a different domain shares a structural similarity with this one?" Not topic similarity — structural: same dynamic, same tradeoff, same failure mode, different context. (3) Create at least one cross-domain link per note. (4) Label the link with the structural relationship: "same dynamic as," "same failure mode as," "inverse of." Over time, these bridges become the most valuable paths in your knowledge graph.