Modules as Conceptual Bridges
Definition
Modules serve as translation layers between how domain experts conceptualize their subject matter and how data systems actually represent it.
Why It Matters
There’s a persistent tension in knowledge engineering: data structures that optimize for machine processing often don’t match how humans think about the domain. Flat tables, normalized schemas, and formal ontologies can feel alien to experts who hold rich, contextual mental models.
Conceptual modules address this gap by creating intermediate representations that satisfy both sides. They’re formal enough to be computationally useful but bounded in ways that align with expert intuition.
How It Works
A well-designed module captures a “key notion” as domain experts understand it, not as the data happens to be organized. The module for “Physical Sample” in an oceanographic ontology contains everything experts associate with that concept, even if the underlying data is scattered across tables.
This alignment has two benefits:
- For humans: Modules match mental models, making ontologies comprehensible and maintainable
- For LLMs: Conceptually coherent input produces better results than arbitrary chunks
Implications
The bridging function suggests modules aren’t just organizational conveniences, they’re essential infrastructure for human-AI collaboration on knowledge tasks. They translate between human conceptualization (fuzzy, contextual, implicit) and machine representation (precise, explicit, formal).
This echoes a broader pattern: the most effective human-AI workflows often involve intermediate representations that both sides can work with effectively.
Related: 06-molecule—tacit-vs-explicit-knowledge, 06-molecule—seci-framework, 01-atom—human-in-the-loop