.me — lineage

me.explain(path)

me.explain("some.path");
// → { value, derivation, meta: { k, recomputed, sourcePath } }

Returns the actual computation that produced the value at path — which write triggered it, which nodes recomputed, and how many (k). Not a self-report generated after the fact: self.explain(path) looks up the derivation record the kernel already keeps and returns it unmodified. No LLM narration in the loop.

Example

me.explain("robots.surgeon.canProceed") →
{
  path: "robots.surgeon.canProceed",
  value: true,
  expr: "canLift && softGripReady && !needsHumanReview && contextAllowsMotion",
  derivation: {
    expression: "canLift && softGripReady && !needsHumanReview && contextAllowsMotion",
    inputs: [
      { label: "canLift", path: "robots.surgeon.canLift", value: true, origin: "public" },
      { label: "softGripReady", path: "robots.surgeon.softGripReady", value: true, origin: "public" },
      // ...
    ]
  },
  meta: {
    dependsOn: [...],
    recomputed: [...],  // what else recalculated in the same wave
    sourcePath: "...",  // which write triggered the recompute
  }
}

If path isn't derived — a plain stored value, not computed from an expression — explain() doesn't invent a cause. It returns expr: null, derivation: null, honestly. Absence of explanation is reported as absence.

How-to

Why recomputed/sourcePath exist without scanning the whole graph on every call: a reverse index from each source path to what depends on it — see Inverted Dependency Indexing.