Links

Type: index · Status: current

Links are the edges of the knowledge graph. Every link is a decision point for the reader: follow or skip? The quality of surrounding context determines whether that decision is informed or blind.

Foundations

  • title-as-claim-enables-traversal-as-reasoning — claim titles make link traversal read as reasoning; explains why "since [X]" works but "see [X]" is a different link intent, and where the pattern breaks for multi-claim documents

Observations

Decisions

Analysis

  • backlinks — use cases for inbound link visibility: hub identification, source-to-theory bridging, impact assessment, tension surfacing; four design options with trade-offs
  • link-strength-is-encoded-in-position-and-prose — inline premise links carry more weight than footer links; position and prose encode commitment level, creating a weighted graph
  • distilled-artifacts-need-source-tracking-at-the-source — distilled artifacts shouldn't link back to sources (focus), but sources should link forward via "Distilled into:" so source changes trigger staleness review

Reference material

  • link-contracts-framework — framework for systematic, testable linking: link contracts, intent taxonomy, agent implications
  • Toulmin argument — formal argumentation theory behind link semantics: "since [X]" and "because [Y]" links encode Toulmin warrants connecting grounds to claims; the six-part model (claim/grounds/warrant/qualifier/rebuttal/backing) names the structure argumentative links carry
  • Agentic Note-Taking 23: Notes Without Reasons — practitioner validation: an agent inside a curated graph contrasts propositional link semantics ("since [X]") with embedding-based adjacency, arguing the difference is one of kind not degree; strongest external evidence for why link quality (not quantity) determines graph health
  • A-MEM: Agentic Memory for LLM Agents — empirical counterpoint: embedding-based link generation succeeds on QA benchmarks, demonstrating that adjacency-as-linking works for retrieval accuracy even if it lacks propositional semantics; the question is whether the quality gap matters only for navigability and agent reasoning

All notes

  • Agents navigate by deciding what to read next — An agent doing a task navigates by deciding what to read — links, index entries, search tools, and skill descriptions are all pointers with varying amounts of context for that decision
  • Backlinks — use cases and design space — Analysis of where backlinks (inbound link visibility) would concretely help agents working in the KB — use cases, trade-offs, and design options
  • Distilled artifacts need source tracking at the source — Distilled artifacts should not link back to sources (focus), but sources should link forward to distilled targets ("Distilled into:") so that source changes trigger staleness review of downstream artifacts
  • Link contracts framework — source material — Reference framework for systematic, testable linking — link contracts, intent taxonomy, automated checks, agent implications
  • Link strength is encoded in position and prose — Not all links are equal — inline premise links ("since [X]") carry more weight than footer "related" links. Position and prose encode commitment level, creating a weighted graph that affects traversal, scoring, and quality signals.
  • Title as claim enables traversal as reasoning — When note titles are claims rather than topics, following links between them reads as a chain of reasoning — the file tree becomes a scan of arguments, and link semantics (since, because, but) encode relationship types
  • Two kinds of navigation — Link-following is local with context; search is long-range with titles/descriptions; indexes bridge both modes