Problem id
derived_solidification_free_CW_homology
What is wrong
At Lean-Eval commit 0e093293afbe96be91504ae6e44947876d3fc84e, Hole 10 asks for
def derivedSolidification_free_CW_derivedNatIso :
derivedSolidificationFreeCWFunctor ≅
singularChainsLightCondAbCWDerivedFunctor := sorry
The fixed category CWTopCat and the pointwise hypotheses in Holes 10–12 only require
[Topology.CWComplex (Set.univ : Set X)]
with no Hausdorff/T2 hypothesis. This admits a non-Hausdorff zero-cell thickening of an
infinite light profinite space. On that object, the requested comparison would make the
non-discrete measure object a retract of a discrete singular-homology object, which is
impossible.
The counterexample has been formalized independently of the unfinished Hole 10:
def counterexample : CWTopCat :=
cwThickening (LightProfinite.toTopCat.obj (ℕ∪{∞}))
theorem no_natural_comparison :
IsEmpty
(CWTopCat.toTopCat ⋙ freeDerivedFunctor ≅
singularChainsLightCondAbCWDerivedFunctor)
theorem no_natural_comparison_of_adjunction
{L : DLightCondAb ⥤ DSolid} (a : L ⊣ derivedInclusion)
{F : CWTopCat ⥤ DLightCondAb}
(spec : F ≅ CWTopCat.toTopCat ⋙ freeLightCondAbOfTopFunctor ⋙
singleAmbient ⋙ L ⋙ derivedInclusion) :
IsEmpty (F ≅ singularChainsLightCondAbCWDerivedFunctor)
There is also a degree-zero obstruction showing that changing the left-adjoint choices
cannot make the pointwise conclusion true:
theorem not_exists_CW_zero_homology_adjunction :
¬ ∃ L : DLightCondAb ⥤ DSolid,
Nonempty (L ⊣ derivedInclusion) ∧
∀ (X : TopCat) [Topology.CWComplex (Set.univ : Set X)],
Nonempty
(isSolid.ι.obj
((DerivedCategory.homologyFunctor Solid 0).obj
(L.obj (singleAmbient.obj (freeLightCondAbOfTop X)))) ≅
singularHomologyLightCondAb X 0)
Verification:
- The formal counterexample was independently rebuilt in a fresh, isolated Lean 4.34.0
workspace with no project/dependency build artifacts.
lake build Submission and an explicit build of all 71 editable helper modules,
including Submission.CWComparisonObstruction, passed.
- The exported axiom audit checked 2,655 declarations and rejected none; the two
negative certificates use only propext, Quot.sound, and Classical.choice.
- The required benchmark API was 9/12: the only unavailable declarations were Holes
10–12, and lake build Solution failed exactly on those three missing identifiers.
- The current Lean-Eval commit
0e093293afbe96be91504ae6e44947876d3fc84e retains the same unrestricted signatures
under Lean v4.35.0-rc3. A direct overlay build of the old proof sources reaches
unrelated 4.34→4.35 compatibility failures in helper modules before compiling the
obstruction; no 4.35 success is claimed here.
Suggested repair: restrict the CW domain used by Hole 10 to the appropriate
Hausdorff/T2 full subcategory (matching the intended mathematical theorem), and add the
corresponding separation hypothesis to Holes 11–12 before regenerating the workspace.
Relevant link (optional)
|
/-- **Hole 10.** The derived comparison theorem, naturally in a CW complex `X`: after applying the |
|
exact derived inclusion from solid light condensed abelian groups to light condensed abelian groups, |
|
the derived solidification of the free light condensed abelian group on `X` is the integral singular |
|
chain complex of `X` as a derived object of light condensed abelian groups. -/ |
|
def derivedSolidification_free_CW_derivedNatIso : |
|
derivedSolidificationFreeCWFunctor ≅ singularChainsLightCondAbCWDerivedFunctor := sorry |
|
|
|
/-- The pointwise component of `derivedSolidification_free_CW_derivedNatIso`. -/ |
|
def derivedSolidification_free_CW_derivedIso |
|
(X : TopCat) [Topology.CWComplex (Set.univ : Set X)] : |
|
derivedInclusion.obj |
|
(derivedSolidification.obj |
|
((DerivedCategory.singleFunctor LightCondAb 0).obj (freeLightCondAbOfTop X))) ≅ |
|
singularChainsLightCondAbDerived X := |
|
(derivedSolidificationFreeCWFunctorSpec.app ⟨X, ⟨inferInstance⟩⟩).symm.trans |
|
(derivedSolidification_free_CW_derivedNatIso.app ⟨X, ⟨inferInstance⟩⟩) |
|
|
|
/-- **Hole 11.** For a CW complex `X`, the homology of the derived solidification of the free |
|
light condensed abelian group on `X` is integral singular homology. Since the derived category |
|
is cohomologically indexed, the `n`-th singular homology group occurs in degree `-n`. -/ |
|
def derivedSolidification_free_CW_homologyIso |
|
(X : TopCat) [Topology.CWComplex (Set.univ : Set X)] (n : ℕ) : |
|
isSolid.ι.obj |
|
((DerivedCategory.homologyFunctor Solid (-(n : ℤ))).obj |
|
(derivedSolidification.obj |
|
((DerivedCategory.singleFunctor LightCondAb 0).obj (freeLightCondAbOfTop X)))) ≅ |
|
singularHomologyLightCondAb X n := sorry |
|
|
|
/-- **Hole 12.** The theorem form of `derivedSolidification_free_CW_homologyIso`: the derived |
|
solidification of the free light condensed abelian group on a CW complex computes integral |
|
singular homology. -/ |
|
theorem derivedSolidification_free_CW_homology |
|
(X : TopCat) [Topology.CWComplex (Set.univ : Set X)] (n : ℕ) : |
|
Nonempty |
|
(isSolid.ι.obj |
|
((DerivedCategory.homologyFunctor Solid (-(n : ℤ))).obj |
|
(derivedSolidification.obj |
|
((DerivedCategory.singleFunctor LightCondAb 0).obj (freeLightCondAbOfTop X)))) ≅ |
|
singularHomologyLightCondAb X n) := sorry |
Problem id
derived_solidification_free_CW_homologyWhat is wrong
At Lean-Eval commit
0e093293afbe96be91504ae6e44947876d3fc84e, Hole 10 asks forThe fixed category
CWTopCatand the pointwise hypotheses in Holes 10–12 only requirewith no Hausdorff/T2 hypothesis. This admits a non-Hausdorff zero-cell thickening of an
infinite light profinite space. On that object, the requested comparison would make the
non-discrete measure object a retract of a discrete singular-homology object, which is
impossible.
The counterexample has been formalized independently of the unfinished Hole 10:
There is also a degree-zero obstruction showing that changing the left-adjoint choices
cannot make the pointwise conclusion true:
Verification:
workspace with no project/dependency build artifacts.
lake build Submissionand an explicit build of all 71 editable helper modules,including
Submission.CWComparisonObstruction, passed.negative certificates use only
propext,Quot.sound, andClassical.choice.10–12, and
lake build Solutionfailed exactly on those three missing identifiers.0e093293afbe96be91504ae6e44947876d3fc84eretains the same unrestricted signaturesunder Lean
v4.35.0-rc3. A direct overlay build of the old proof sources reachesunrelated 4.34→4.35 compatibility failures in helper modules before compiling the
obstruction; no 4.35 success is claimed here.
Suggested repair: restrict the CW domain used by Hole 10 to the appropriate
Hausdorff/T2 full subcategory (matching the intended mathematical theorem), and add the
corresponding separation hypothesis to Holes 11–12 before regenerating the workspace.
Relevant link (optional)
lean-eval/generated/derived_solidification_free_CW_homology/Challenge.lean
Lines 229 to 267 in 0e09329