CMDG-CM4 — Free Solid Abelian Groups on Profinite Spaces
A human-facing account of GCL's protected CM4 result: what was proved, why the inherited route was difficult, how the Point-functional bypass works, what remains open, and where to follow the live external-review and mathlib-upstream work.
Research surface authority
Canonical campaign tracker: https://github.com/grandchallenge/MATH-PROGRAMME/issues/355
Protected mathematical claims on this page are projections of protected repository records.
Live external-review status is tracked in MATH-PROGRAMME #1222.
Live mathlib-upstream status is tracked in MATH-PROGRAMME #1223.
If this page and an issue disagree about active work, the issue governs the active-work state. If either conflicts with a protected theorem receipt about an already-certified claim, the protected receipt governs the claim.
For readers new to condensed mathematics
Begin with Condensed Mathematics for the Perplexed, a foundational introduction from convergent sequences and profinite probes through sheaf descent and solid abelian groups. This page then explains GCL's restricted CM4 theorem and its exact evidentiary status.
For independent readers and contributors
- Self-contained mathematical note: the exact statement, a finite model, the proof spine, and falsification-oriented review questions.
- Independent reproducibility protocol: immutable source hashes, pinned Lean/mathlib environment, executable verification, expected local evidence and failure gates.
- Specialist review packet and review issue #1222.
- Mathlib integration assessment and upstream issue #1223.
No external expert-review conclusion or upstream mathlib acceptance is implied by the existence of these materials.
1. Status
| Field | Current meaning |
|---|---|
| Result status | Protected formal proof |
| Strongest supported claim | For every profinite space (S), the governed integer-coefficient free solid abelian group is solid at the module level. |
| Formal target | CMDG.CondensedCM4.cm4Target_via_pointFunctional : CM4Target |
| Support route | Formal proof |
| Certification state | Formally checked, independently reviewed, protected on main |
| P3 disposition | PROTECTED_CLOSED_BY_UNDERIVED_BYPASS |
| P4–P6 | Unproved source-route auxiliaries; nonblocking for the selected proof |
| Not claimed | Derived/complex source theorem, arbitrary-ring generalization, C06, CM5, graph/global CMDG completion |
| Live next move | External mathematical review and staged mathlib port; see #1222 and #1223 |
Protected terminal merge: 28aa80edc08833b9945380a848552cd5fb1da363.
2. Plain object
Start with a profinite space (S): a compact, totally disconnected topological space, such as an inverse limit of finite sets.
Condensed mathematics provides a way to form a free condensed abelian group on (S). Solid mathematics then asks for a stronger object, usually written informally as
The CM4 target asks whether this free solid object is actually solid in the precise sense encoded by mathlib's CondensedMod.IsSolid predicate.
For a lay reader: this is a load-bearing interface check. We construct an object intended to live in the solid world, and CM4 proves that it really does.
3. Exact obstruction
Current mathlib still describes the corresponding theorem as a hard TODO:
TODO (hard): prove that ((profiniteSolid ℤ).obj S).IsSolid for S : Profinite.
The inherited source route passes through derived and cohomological machinery. In GCL's dependency decomposition, that created a difficult P3 profinite Ext/cohomology bridge and downstream source-route obligations P4–P6.
The key question became:
Are those obligations necessary for the exact module-level theorem, or only for that particular proof architecture?
Our result answers: they are not necessary for the selected module-level target.
4. Working model
The successful route replaces the difficult global derived obstruction with a detection problem at a single Point fibre.
Very roughly:
- represent a candidate morphism by the scalar information it induces on locally constant integer-valued functions;
- express that information in an integral basis supplied by Nöbeling-type freeness;
- reconstruct arbitrary Point-level measure data from weighted finite Boolean probes;
- prove that if the scalar functional vanishes, the original morphism vanishes;
- use that faithfulness to obtain the required solidity statement.
The intuition is analogous to proving that a complicated linear object is zero by finding a family of coordinates that detects every nonzero element.
5. Restricted claim
The protected result is deliberately narrow:
def CM4Target : Prop :=
∀ S : Profinite.{u},
CondensedMod.IsSolid (ULift.{u + 1} ℤ)
((Condensed.profiniteSolid (ULift.{u + 1} ℤ)).obj S)
and
theorem cm4Target_via_pointFunctional : CM4Target
The ULift is the universe-correct form of the integer coefficient ring used by the fixture. The human mathematical content is the integer-coefficient profinite-solid solidity theorem.
6. Theorem-spine location
Foundations / sheaves / sites
│
CM0 condensed contact
│
CM1 discrete–underlying adjunction [protected closed]
│
CM2 Cartesian-closed condensed sets [protected closed]
│
CM3 abelian / homological structure [protected closed]
│
C05 solid definition boundary [protected closed]
│
CM4 solid mathematics [protected closed: module-level ℤ]
│
CM5 Liquid-Tensor-class benchmark [open frontier]
Inside CM4, the selected route is:
finite Boolean / Nöbeling data
↓
Point-measure reconstruction
↓
kernel product functional
↓
functional faithfulness
↓
coefficient solidity
↓
profinite-solid solidity
↓
CM4 target
The historical source route remains represented separately:
P5 ──→ P4 ──→ P6
↑ ↑
└── P3 ┘
P3 is closed by the underived bypass.
P4–P6 remain unproved source auxiliaries, not requirements of the selected route.
7. Support route
The critical formal interfaces are:
measurePointProjection_zero_reflectsmeasurePointFunctional_zero_reflectsmeasurePointIntegralFunctional_zero_reflectsweightedFiniteBooleanMeasureLimitLift_measurePoint_allTruekernelProductFunctional_eq_zero_of_solidification_kernelcoefficient_eq_zero_of_solidification_kernelcoefficientMappingOutInjectivity_of_pointFunctionalcoefficientObject_isSolid_via_pointFunctionalprofiniteSolid_isSolid_via_pointFunctional
The terminal theorem has axiom readback
[propext, Classical.choice, Quot.sound]
and no sorryAx.
For a reviewer-oriented treatment, see External Mathematical Review Packet.
8. Debt audit and claim boundary
Trust quartet
What is proved?
The governed module-level integer-coefficient CM4 target.
What is checked?
Lean replay, axiom readback, policy/conformance gates, exact-head independent review, protected merge, and terminal protected-main readback.
What remains open?
The broader derived/complex source theorem, arbitrary-ring generalizations, C06, CM5, graph certification, and global CMDG completion.
What requires external verification now?
Independent specialist assessment of the mathematical proof architecture and community review of a mathlib-native upstream port.
Why P4–P6 still appear
They are not deleted because they encode genuine proof debt for the historical source route. They are marked NONBLOCKING_SOURCE_AUXILIARY because the selected underived proof no longer consumes them.
That distinction is part of the result, not administrative cleanup.
9. First executable step
There are now two live bounded actions.
- External mathematical review: independently audit the Point-functional proof and return a disposition on issue #1222.
- Mathlib upstream port: refactor the downstream proof into mathlib-native layers and track the staged port on issue #1223.
The detailed upstream assessment is Mathlib Upstream Readiness Assessment.
Protected evidence
- Terminal CM4 readback: repository record
- Terminal P3 closure: repository record
- Terminal CM4 reconciliation: repository record
- Primary proof source: Point-functional bridge
- Final wrapper: CM4 target fixture