← Axiom Explorer ← Papers Index
P2.2 - The Parsimony Law
Premise: Occam’s razor is not taste - it’s constraint. Minimal description length wins.
One-sentence version
If multiple explanations fit, the one with less unnecessary machinery is not just “nicer”; it is more stable under extension, prediction, and integration.
The Paper (Narrative)
Parsimony is usually introduced as a rule of thumb: “Don’t multiply entities beyond necessity.” That can sound subjective, like style preference.
But in science and engineering, parsimony behaves more like a survival filter:
- Overfit models are fragile.
- Baroque explanations break when you add new data.
- Unnecessary parameters hide contradictions.
So the project treats parsimony as a constraint on what counts as a defensible ontology.
This matters because it tells you what kind of claims to expect from the rest of the system:
- Not “it could be this, or that, or anything.”
- But “given the constraints, the solution space collapses.”
Later, when the system argues for a terminal observer, a grace operator, or a specific set of boundary conditions, it is not trying to be maximalist. It is trying to be minimal while sufficient.
That is the parsimony law in plain language:
Reality may be deep, but it is not wasteful.
What This Paper Is Not Claiming
- It is not claiming the simplest story is always true in every context.
- It is not claiming we can measure Kolmogorov complexity perfectly (we cannot).
- It is claiming that explanations that require infinite ad-hoc patches are not stable.