Monotonicity Sanity Check¶
Audit — instantiates Additive Measure-Space Design
A cheap consistency check that a containing subset never receives less size than the subset it contains — catching sign errors, overlaps, and broken additivity before they reach a decision.
If one set contains another, it cannot be smaller. Monotonicity Sanity Check is the near-free tripwire built on that fact: for A ⊆ B it verifies μ(A) ≤ μ(B), because B = A ∪ (B∖A) with a nonnegative remainder, so a bigger set's measure can never fall below a part's.[n1] Any inversion — a whole scoring less than one of its pieces — signals a broken nonnegativity, an overlap, or a disjointness error upstream. Its distinctive role is to be the fast, cheap smell-test that runs continuously; it neither builds the measure nor certifies its full additivity, it just catches gross violations early.
Example¶
A warehouse inventory system reports that the "electronics" category holds 4,200 units, while its own sub-bin "laptops" alone reports 4,500. The containing set is smaller than a part it contains — impossible for a coherent count. The monotonicity check flags the containment inversion instantly. The cause turns out to be a pallet scanned twice into the laptop bin, inflating a subset above its parent. The check didn't need the full inventory to be re-audited; a single nested pair was enough to surface that something upstream — here, a double-count — had broken the size rule.
How it works¶
The check exploits containment as a consistency lever:
- Pick nested pairs — subsets A ⊆ B where the relationship is known (a bin inside a category, a region inside a territory).
- Compare their sizes — verify μ(A) ≤ μ(B); an inversion means the remainder μ(B∖A) came out negative, which a nonnegative measure forbids.
- Escalate, don't diagnose — a failure is a signal that nonnegativity or disjointness broke somewhere, handed on to the additivity test or partition table to localise.
It is cheap enough to run everywhere and blunt enough to only catch gross errors — which is exactly its value.
Tuning parameters¶
- Nested pairs checked — how many containment relationships are tested; more coverage catches more, at more cost.
- Strict vs. weak — whether equal sizes for distinct nested sets are tolerated or themselves flagged.
- Cadence — continuous automated checking versus periodic spot checks.
- Tolerance — how large an inversion counts as signal rather than rounding noise.
When it helps, and when it misleads¶
Its strength is cost: it is nearly free, runs anywhere, and catches sign errors, duplicate-driven overlaps, and broken additivity long before a full audit would. Monotonicity is a genuine corollary of a valid measure, so a violation is always a real defect.[n1]
Its failure mode is that it is necessary but not sufficient: passing monotonicity does not prove the measure is additive — a measure can be perfectly monotone and still double-count in ways that preserve ordering. Treating a green tripwire as a certificate of correctness is the classic overreach. The discipline is to read a pass as "no gross error found" and to escalate any failure to Finite or Countable Additivity Test or Partition Sum Table for the real diagnosis.
How it implements the components¶
Monotonicity Sanity Check exercises two axioms by testing their joint consequence:
nonnegative_size_assignment_rule— checks the rule's direct corollary: adding admissible content can never lower a size.disjoint_additivity_rule— leans on μ(B) = μ(A) + μ(B∖A) as the identity a containment inversion violates.
It stress-tests these rules but does not construct the size (Area, Volume, or Counting Template), declare it (Measure-Space Specification), or certify full additivity (Finite or Countable Additivity Test).
Related¶
- Instantiates: Additive Measure-Space Design — this check is the cheap early tripwire on the size rule's coherence.
- Sibling mechanisms: Finite or Countable Additivity Test · Area, Volume, or Counting Template · Partition Sum Table · Measure-Space Specification · Measurable Family Closure Check · Normalization Constant Calibration · Null-Set Policy Register · Probability Measure Construction · Measure Invariance Review
Editorial Notes¶
Form Classification¶
Form family: Assessment, Review & Assurance
Rationale: Monotonicity Sanity Check operates as a bounded evaluation of existing evidence or work that produces a finding or disposition because it a cheap consistency check that a containing subset never receives less size than the subset it contains — catching sign errors, overlaps, and broken additivity before they reach a decision.
Independent corroboration: The frozen evidence defines Monotonicity Sanity Check as 'A cheap consistency check that a containing subset never receives less size than the subset it contains — catching sign errors, overlaps, and broken additivity before they reach a decision', so its operative form is Assessment, Review & Assurance.
Review outcome: Independent reviewer agreement; high confidence.
Origin Attribution¶
Primary origin: Mathematics
Origin pattern: Single lineage
Present-day reach: Multi-domain
Rationale: Monotonicity under set inclusion is a direct consequence and standard sanity condition of mathematical measure theory.
Related originating lineages:
- Statistics & Experimental Design — Applied quantitative work uses the property as a data and probability consistency check.
Review resolution: Both independent reviews agree on primary origin mathematics; reconciliation resolves secondary fields (alternate_origin_disagreement, domain_reach_disagreement, encyclopedia_synthesis_disagreement). Alternate origins retained (statistics_experimental_design) are the union of reviewer-supported formative lineages with explicit rationales, not a list of later application domains. Present-day breadth is represented separately as domain_reach=multi_domain; origin_mode=single_lineage records the historical relationship among lineages. Confidence is conservatively reconciled to high, and encyclopedia_synthesis=true preserves either reviewer's finding that the encyclopedia generalized the mechanism.
Encyclopedia synthesis: The exact catalogued form synthesizes established practice rather than reproducing a single standard historical label.
Review outcome: Reconciled after independent review; high confidence.
Notes¶
The check is a filter, not a proof. Because it is necessary but not sufficient for additivity, a clean run should raise confidence, never close the question — a measure that passes every containment check can still fail the additivity test.
[n1] Monotonicity — A ⊆ B ⟹ μ(A) ≤ μ(B) — follows from nonnegativity and additivity via μ(B) = μ(A) + μ(B∖A) with μ(B∖A) ≥ 0. It is a necessary property of a valid measure but not a sufficient one. ↩a ↩b