← Proof comparison · Paper proof graph · Lean proof graph

Definitions: printed form and form used by the proofs

The Lean modules Stafford38.Paper state the definitions of the paper as printed. Some of these are the definitions the Lean proofs work with. Others are stated separately and connected, by a Lean theorem, to a different definition that the proofs use. That definition lives in the working modules, the other modules of the project. This page lists which case applies to each notion: 12 are used as printed, 7 are other names, 4 are Mathlib's, for 16 an equivalence is proved and for 1 an implication.

Notation: k is the ground field of characteristic zero, K an algebraically closed field containing it, over which the geometry of Section 5 takes place. A point is a tuple of coordinates in K. The stand-alone statements are the versions of Theorem 1.1 and Corollary 6.3 written with Mathlib alone, against which the proofs are checked independently (the cards "Challenge comparison" of the proof comparison). The Lean sources are those of commit 1029fdb of the formal repository, with Mathlib at commit c55e6e7.

Under each connecting theorem the page says whether a later proof uses it. Lean stores, for every proof, the declarations it refers to; following these references from Theorem 1.1 (Stafford38.Paper.theorem_1_1) and from Corollary 6.3 gives everything their proofs depend on. A connecting theorem outside this set is still proved; it documents that the printed form agrees, and no later argument rests on it.

Definitions by section

The Weyl algebra and the Bernstein degree

NotionRelationLean form of the printed definitionDefinition the proofs work withHow they are connected
Weyl algebra A_n(k)
Section 1
as printedStafford38Challenge.WeylAlg—The free algebra on x_i, ∂_i modulo the commutation relations. This one definition is used in every statement and in the proofs; PresentedWeyl is a second name for it. The iterated Ore extension A_{n-1}[x][p; ∂] is an auxiliary second construction: the Euler and normal-form arguments of Sections 3 and 4 are carried out there and transported along the isomorphism below.
Bernstein degree deg_B
Definition 2.3, Theorem 1.1
equivalence provedSection2.bernsteinDegreeStafford38FixedSourceChallenge.bernsteinDegree
FixedSource.bernsteinDegree
Three definitions of the same number. Printed: the least j with d ∈ B_j, through the general filtration degree. Stand-alone statement of Theorem 1.1: the least N such that d lies in the span of ordered words of degree at most N. Proofs: the total degree of the normal form of d. Lean proves all three equal.

Section 2: modules, filtrations, characteristic variety

NotionRelationLean form of the printed definitionDefinition the proofs work withHow they are connected
Torsion module
Definition 2.1
another nameSection2.IsTorsionTorsionCyclicity.IsRightTorsionEvery m has md = 0 for some d ≠ 0. The printed name abbreviates the definition used in Corollary 6.3.
Cyclic module
Definition 2.1
as printedSection2.IsCyclic—M = mA, as printed. Corollary 6.3 is stated with this definition; its proof passes to Mathlib's form through the theorem below.
x₁, p = ∂₁, the transverse algebra A_{n-1}
Convention 2.2
another nameSection2.x₁
Section2.p
Section2.transverse
WeylIteratedEquivalence.presentedCoordinate
WeylIteratedEquivalence.presentedMomentum
WeylIteratedEquivalence.previousWeylEmbedding
Names for the first generator pair and for the image of A_{n-1} in A_n. Lean numbers the generator pairs from 0 and writes the rank as n+1 where the paper assumes n ≥ 1: PresentedWeyl k (n+1) has n+1 pairs and is the paper's A_n with n ≥ 1.
C = A_{n-1}[x₁]
Convention 2.2
equivalence provedSection2.C—Printed: the subalgebra generated by the transverse algebra and x₁. The normal form ⊕ p^j C is proved in the Ore extension, whose coefficient ring is shown to be the same subalgebra.
  • Section2.C_eq_range: C = the image in A_n of the coefficient ring of the last Ore step
    used in the proof of Theorem 1.1
Filtration, associated graded, degree, symbol
Definition 2.3
as printedSection2.IsFiltration
Section2.gr
Section2.filtrationDegree
Section2.symbol
—Written as printed for any algebra over k.
Bernstein filtration B and order filtration F of A_n
Definition 2.3
another nameSection2.bernsteinFiltration
Section2.orderFiltration
WeylFiltration.bernsteinPiece
WeylFiltration.orderPiece
Spans of the ordered words x^α ∂^β of bounded weight: weight 1 on every generator for B, weight 0 on x_i and 1 on ∂_i for F.
gr A_n(k) and k[x, ξ]
Definition 2.3
equivalence provedSection2.grCharacteristic.SymbolRing
WeylLeadingSymbol.presentedPrincipalComponent
The proofs take the polynomial ring k[x, ξ] itself as gr A_n, and the symbol of an element is the top homogeneous part of its normal form. The identification with ⊕_j F_j/F_{j-1} is proved as a linear isomorphism together with multiplicativity of symbols. It is not packaged as one ring isomorphism.
Compatible and good filtrations of a module
Definition 2.4
as printedSection2.IsCompatibleFiltration
Section2.IsGoodFiltration
Support.Gabber.GrModule
—As printed, for right A_n-modules and the order filtration. gr M is GrModule, the graded module of Definition 2.3; x_i and ξ_i act on it through right multiplication by x_i and ∂_i, which gives the module structure over k[x, ξ].
Q_m = qF_m and gr Q for Q = A_n/I
Definition 2.4
equivalence provedSection2.cyclicFiltration
Support.Gabber.qFilt
CharacteristicAssociatedGradedModule.OrderAssociatedGradedModuleThe proofs of Sections 3 to 6 work with one fixed module gr Q, built from the images of F_m in A_n/I. The paper modules call it grCyclic. It is isomorphic, over k[x, ξ], to the printed ⊕_m Q_m/Q_{m-1}.
T*𝔸ⁿ_k = Spec k[x, ξ]
Definition 2.6
MathlibSection2.CotangentBundlePrimeSpectrumMathlib's prime spectrum of the polynomial ring k[x, ξ].
Projection π and fibre T*_y X
Section 5.1
as printedSection2.proj
Section2.fibre
—A closed point (y, ξ) is a function on the 2n coordinate indices; π keeps the first n values, and the fibre over y is π⁻¹(y). The fibre is used once, to state Definition 5.6. The projection is used in the last step of Section 6, where W does not meet π⁻¹(H) for the hyperplane H = {x₁ = 0}.
V(J) ⊆ Spec S
Definition 2.5
MathlibSection2.VPrimeSpectrum.zeroLocusMathlib's zero locus in the prime spectrum.
V_K(J), I(S), 𝔪_a
Definition 2.5
equivalence provedSection2.VK
Section2.vanishingIdealOf
Section2.maxIdealAt
MvPolynomial.zeroLocus
MvPolynomial.vanishingIdeal
Classical zero sets and vanishing ideals. I(S) is Mathlib's vanishing ideal by definition. Sections 5 and 6 state their closed sets with Mathlib's zeroLocus and vanishingIdeal.
Char(M) for a good filtration
Definition 2.7
as printedSection2.CharOf—V(Ann gr M), as printed. Independence of the filtration is proved, but no later proof needs it, because Sections 3 to 6 only use the filtration Q_m = qF_m.
Char(A_n/I)
Definition 2.7
equivalence provedSection2.CharCharacteristicAssociatedGradedModule.OrderAssociatedGradedModule
Module.support
Printed: Definition 2.7 applied to the filtration Q_m = qF_m. The proofs use the annihilator and the support of the module gr Q they work with.
Symplectic form, symplectic basis, Gram matrix
Definition 2.8
as printedSection2.IsSymplectic
Section2.IsSymplecticBasis
Section2.gramMatrix
—As printed.
Standard form ω on k^{2n}
Definition 2.8
as printedSection2.omegaCharacteristicSymplecticCompletion.phasePairingω(x_i, ∂_j) = δ_ij as printed. Lean defines it as the negative of the pairing used in the proofs, which has the opposite sign convention.
  • Section2.gramMatrix_omega: the Gram matrix of ω in the generators is -J, J the matrix of commutators [u_i, u_j]
    used in the proof of Theorem 1.1
Poisson bracket
Definition 2.9
equivalence provedSection2.poissonCharacteristic.poissonBracketThe printed formula with partial derivatives. The proofs use a bracket defined in the proof library; the two are equal.
  • Section2.poisson_eq: printed bracket = bracket of the proofs, with the same sign
    used in the proof of Theorem 1.1
Involutive ideal and involutive closed set
Definitions 2.9 and 5.6
equivalence provedSection2.IsInvolutive
Section2.IsInvolutiveSet
Characteristic.PostScalarExtensionPoisson.IsInvolutiveA closed set W is involutive, by definition, if I(W) is an involutive ideal. Theorem 2.10 is stated with the printed definition. Theorem 5.8 is stated with the definition of the proofs, applied to I(W).

Section 3: normal form

NotionRelationLean form of the printed definitionDefinition the proofs work withHow they are connected
Automorphism Φ_g
Section 3
as printedSection3.phiG—The automorphism of A_n extending a linear symplectic map g of the generators; g is given by its matrix in Sp(2n, k).
Monic normal form of d̃
Equation (3.1)
implication provedSection3.IsMonicNormalFormWeylPBWMonicBridge.IsPBWMonicAtPrinted: d̃ = p^N + Σ p^j a_j with a_j ∈ C and deg_B a_j ≤ N - j. The Euler arguments of Section 4 use a monicity condition on the normal form of d̃. Only the direction the proofs need is proved.
Ideal 𝓘 and module Q = A_n/𝓘
Equation (3.2)
another nameSection3.idealI
Section3.Q
Section3.grQ
WeylEulerResidue.canonicalRightIdeal
EulerSurjectivity.RightQuotient
CharacteristicAssociatedGradedModule.OrderAssociatedGradedModule
𝓘 = d̃A_n + x₁^N d̃A_n and its quotient; gr Q is the module of Definition 2.4 for this ideal.
Monic of degree N in ξ₁
Definition 3.4
as printedSection3.IsMonicInXi1—As printed: P, read as a polynomial in ξ₁, has degree at most N and coefficient 1 at ξ₁^N.
Order symbol σ^ord_N(d̃)
Lemma 3.5
equivalence provedSection2.symbolSection3.orderSymbol
WeylLeadingSymbol.presentedPrincipalComponent
Printed: the class of d̃ in F_N/F_{N-1} (Definition 2.3). The proofs use the degree-N part of the normal form, a polynomial in k[x, ξ].
  • Section2.grPieceEquiv: F_j/F_{j-1} ≅ homogeneous polynomials of degree j, sending a class to the degree-j part of the normal form
    used in the proof of Theorem 1.1
  • Section2.grPiece_mk_eq_zero_iff: the class of a in F_j/F_{j-1} vanishes ⇔ the degree-j part of its normal form vanishes
    used in the proof of Theorem 1.1

Section 4: Euler grading and supports

NotionRelationLean form of the printed definitionDefinition the proofs work withHow they are connected
Euler grading and A_{≥0}
Before Lemma 4.1
equivalence provedSection4.eulerDegree
Section4.nonnegativePart
Section4.positivePart
WeylEulerSubring.eulerSubringPrinted: the eigenspaces of the commutator with E = x₁p, and A_{≥0} as the sum of those of degree ≥ 0. The normal-form arguments use the subring generated by the transverse algebra, x₁ and E. Lean proves that the two subrings are equal.
Localization, support, length
Definitions 4.4 to 4.6
MathlibSection4.definition_4_4
Section4.definition_4_5
Section4.definition_4_6
Localization
LocalizedModule
Module.support
Module.length
Mathlib's notions. One theorem per definition states the printed description: equality of fractions, Supp(M) = {𝔭 : M_𝔭 ≠ 0}, and length as the length of any composition series.
M = gr Q, T, f, V = ker f, U = coker f
Before Lemma 4.8
another nameSection4.GrQ
Section4.Tangential
Section4.coordinateMap
Section4.kernelV
Section4.cokernelU
—Names for the objects of the setup paragraph, built on gr Q of Definition 2.4; f is multiplication by the symbol x₁.

Section 5: conormal geometry

NotionRelationLean form of the printed definitionDefinition the proofs work withHow they are connected
Field extensions, Kähler differentials, valuation rings
Definitions 5.1 to 5.3
MathlibSection5.definition_5_1_g_h
Section5.differentialRelations
Section5.definition_5_3
IsTranscendenceBasis
Algebra.trdeg
KaehlerDifferential
ValuationSubring
IsDiscreteValuationRing
Mathlib's notions; the listed theorems state the printed descriptions as their properties. For Ω the printed construction, a free module modulo the Leibniz relations, is proved isomorphic to Mathlib's module of differentials.
  • Section5.definition_5_2_b: the free module modulo the printed relations ≅ Mathlib's Ω, with ⟨a⟩ ↦ da
    used in the proof of Theorem 1.1
Irreducible closed set Y
Definition 5.4(a)
equivalence provedSection5.IsZariskiClosed
Section5.IsIrreducibleClosed
Section5.idealPointPrinted: not a union of two proper closed subsets. The proofs work with the prime ideal 𝔭 = I(Y) as a point of Spec K[x]. The equivalence needs Y nonempty.
Smooth locus Y_sm
Definition 5.4(b)
equivalence provedSection5.JacobianSmoothPointSmoothAffineConormal.SmoothAffinePoint
Algebra.smoothLocus
Printed: the Jacobian rank condition. The proofs use Mathlib's smooth locus of K[x]/𝔭, read at the point y. The equivalence is the Jacobian criterion; it is proved for a prime ideal 𝔭 over an algebraically closed field K of characteristic zero, the setting of Section 5.
Conormal space N*_y Y
Definition 5.4(c)
equivalence provedAffineConormalSpan.equationCovectorSpanAffineConormalSpan.affineConormalSpace
AffineConormalSpan.zariskiTangentSpace
Printed: the span of the differentials df_y (this form is also defined in the geometry modules). The proofs define the tangent space as the common kernel of these differentials and the conormal space as its annihilator. The two agree by duality in finite dimension.
Conormal cone N(Y)
Definition 5.4(d)
equivalence provedSection5.conormalConeProjectiveConormalDirections.smoothConormalFibreProjectionPrinted: the union of the spans of differentials over the Jacobian smooth locus. The proofs use the same union with Mathlib's smooth locus and the annihilator form of N*_y Y.
Direction closure Dir(Y)
Definition 5.5
equivalence provedSection5.conormalDirectionsProjectiveConormalDirections.projectiveHomogeneousClosure
ProjectiveConormalDirections.projectivizedDirectionSet
ProjectiveConormalDirections.smoothConormalDirectionSet
The closure in projective space of the classes [ξ], 0 ≠ ξ ∈ N(Y); closed means cut out by homogeneous polynomials. Dir(Y) occurs only in Theorem 5.7(ii), which neither the paper nor Lean uses afterwards.
  • Section5.definition_5_5: printed Dir(Y) = the closure of directions used for Theorem 5.7(ii)
    not used by a later proof
Fibre-conical set
Definition 5.6
equivalence proved—FibreConicalVanishingIdeal.IsFibreConicalThe printed form, with the fibres T*_y X defined in Section 5.1, has no declaration of its own; it is written out in the statement of the theorem below. The proofs use the condition point by point: (y, ξ) ∈ W implies (y, tξ) ∈ W for t ≠ 0. Theorem 5.8 is stated with the latter.
  • Section5.definition_5_6: the condition of the proofs ⇔ every W ∩ T*_y X is stable under ξ ↦ tξ, t ≠ 0
    used in the proof of Theorem 1.1
Tangent space Θ_y
Subsection 5.2
another nameSection5.tangentSpaceAtAffineConormalSpan.zariskiTangentSpaceThe tangent space of the proofs, taken at the point y = (q_i/q_0) with coordinates in a field L ⊇ K, for the equations with coefficients extended to L.
The map ρ and the cone 𝒞 = ρ⁻¹(Θ_y)
Subsection 5.2
as printedSection5.rho
Section5.tangentCone
—As printed.

Section 6: assembly and Corollary 6.3

NotionRelationLean form of the printed definitionDefinition the proofs work withHow they are connected
V(x₁) and W = V_K(Ann gr Q_K)
Proposition 6.1, proof of Theorem 1.1
another nameSection6.Vx₁
Section6.W
Section2.V
Section2.VK
Names for sets built with Definition 2.5; K is an algebraically closed field containing k.
Stably cyclic module
Corollary 6.3
as printedSection6.IsStablyGenerated
Stafford38Challenge.IsStablyCyclic
—Stafford's condition for l ≥ m generators; stably cyclic is m = 1. The stand-alone statement of Corollary 6.3 writes out the same formula for m = 1.

Which definitions the main statements name

The definitions that occur directly in the Lean statement of each result. Theorem 5.8, for example, is stated with the fibre-conical and involutive conditions of the proofs, which the table above connects to Definitions 5.6 and 2.10.

PaperLean statementDefinitions named in it
Theorem 2.10 (Gabber)Section2.theorem_2_11Characteristic.SymbolRing, Section2.IsGoodFiltration, Section2.IsInvolutive, Support.Gabber.GrModule, Stafford38Challenge.WeylAlg
Theorem 2.10, cyclic case used in Section 6Section2.theorem_2_11_cyclicCharacteristic.SymbolRing, Section2.IsInvolutive
Theorem 2.11 (Nullstellensatz)Section2.theorem_2_12PrimeSpectrum, Section2.V, Section2.VK, Section2.maxIdealAt, Section2.vanishingIdealOf
Lemma 3.5Section3.lemma_3_5Characteristic.SymbolRing, Section2.bernsteinDegree, Section3.IsMonicInXi1, Section3.IsMonicNormalForm, Section3.orderSymbol, WeylFiltration.orderPiece
Proposition 3.6Section3.proposition_3_6Characteristic.SymbolRing, Section2.bernsteinDegree, Section3.IsMonicNormalForm, Section3.grQ, Section3.idealI, Section3.orderSymbol
Theorem 5.7(i) as printedSection5.theorem_5_7_i_printedSection5.IsIrreducibleClosed, Section5.conormalCone, Section5.idealPoint
Theorem 5.7(i) for a prime idealSection5.theorem_5_7_iMvPolynomial.vanishingIdeal, MvPolynomial.zeroLocus, PrimeSpectrum, ProjectiveConormalDirections.smoothConormalFibreProjection
Theorem 5.7(ii) as printedSection5.theorem_5_7_ii_printedProjectivization, Section5.IsIrreducibleClosed, Section5.conormalDirections, Section5.idealPoint
Theorem 5.8Section5.theorem_5_8MvPolynomial.vanishingIdeal, MvPolynomial.zeroLocus, Characteristic.PostScalarExtensionPoisson.IsInvolutive, Characteristic.SymbolRing, FibreConicalVanishingIdeal.IsFibreConical
Proposition 6.1Section6.proposition_6_1Section2.Char, Section2.CotangentBundle, Section3.IsMonicNormalForm, Section3.idealI, Section6.Vx₁
Lemma 6.2Section6.lemma_6_2PrimeSpectrum, Characteristic.SymbolRing, Section2.Char, Section2.CotangentBundle, Section3.Q, Section3.grQ, Section3.idealI
Corollary 6.3corollary_6_3Section2.IsCyclic, Section2.IsTorsion, Stafford38Challenge.WeylAlg
Corollary 6.3, stable cyclicitycorollary_6_3_stablyCyclicSection2.IsTorsion, Section6.IsStablyGenerated, Stafford38Challenge.WeylAlg

Geometry: where the definitions of Section 5 live

Section 5 is the geometric part of the proof. Its Lean definitions sit in three layers. General notions come from Mathlib. The notions specific to this proof are defined in the project's own geometry modules, Stafford38.Geometry. The modules Stafford38.Paper.Section5 restate them as printed and prove the statements of the paper. The separate library AlgebraicAnalysis, on which the project also depends, contributes Ore extensions and filtered modules to Sections 3 and 4 (25 of its modules are used by the proof of Theorem 1.1) and no geometry.

MathlibSpectra, zero sets and vanishing ideals, smooth locus, Kähler differentials, valuation rings, completions, power and Laurent series, projective space.
↓
Stafford38.Geometry236 modules of this project. They define tangent and conormal spaces, the conormal cone, fibre-conical sets and the valuation rings and Laurent expansions of Subsection 5.2, and prove the supporting lemmas. 69 of them are used by the proof of Theorem 1.1.
↓
Stafford38.Paper.Section5Definitions 5.1 to 5.6 in the printed form, the theorems that identify them with the definitions above, and Lemmas 5.9 to 5.13 and Theorems 5.7 and 5.8 proved paragraph by paragraph.
↓
Section 6Theorem 5.8 is applied to W = V_K(Ann gr Q_K) and shows W empty. Then the module Q = A_n/(d̃A_n + x₁^N d̃A_n) is zero, which is the identity of Theorem 1.1.

Core geometric definitions

All are sets or subspaces given by explicit conditions on points y ∈ K^n and covectors ξ ∈ (K^n)*. An irreducible closed set Y enters through its prime ideal 𝔭 = I(Y), a point of Spec K[x]. No schemes or sheaves are involved beyond Mathlib's prime spectrum.

PaperLean definitionMeaning
df_yCoisotropicTranslation.differentialAt
Lean text
/-- Evaluate the differential of a base polynomial at a base point. -/
def differentialAt (y : Fin n → k) (f : MvPolynomial (Fin n) k) (i : Fin n) : k :=
  MvPolynomial.eval y (MvPolynomial.pderiv i f)
The differential of a polynomial f at a point y, as the vector of partial derivatives (∂_1 f(y), …, ∂_n f(y)).
T_y Y, Θ_yAffineConormalSpan.zariskiTangentSpace
Lean text
/-- The embedded Zariski tangent space cut out by the first-order parts of all
equations in `I`. -/
def zariskiTangentSpace (y : Fin n → k)
    (I : Ideal (MvPolynomial (Fin n) k)) :
    Submodule k (AffineTangentVector k n) :=
  (equationCovectorSpan y I).dualCoannihilator
The tangent space of V(I) at y: the vectors killed by df_y for every f ∈ I.
N*_y Y (Definition 5.4(c))AffineConormalSpan.affineConormalSpace
Lean text
/-- The embedded conormal space, defined as the annihilator of the embedded
Zariski tangent space. -/
def affineConormalSpace (y : Fin n → k)
    (I : Ideal (MvPolynomial (Fin n) k)) :
    Submodule k (Module.Dual k (AffineTangentVector k n)) :=
  (zariskiTangentSpace y I).dualAnnihilator
The conormal space: the covectors vanishing on the tangent space. Equal to the span of the df_y.
y ∈ Y_sm (Definition 5.4(b))SmoothAffineConormal.SmoothAffinePoint
Lean text
def SmoothAffinePoint (I : Ideal (MvPolynomial (Fin n) k))
    (y : Fin n → k) : Prop :=
  ∃ e : (MvPolynomial (Fin n) k ⧸ I) →ₐ[k] k,
    (∀ p, e (Ideal.Quotient.mk I p) = MvPolynomial.aeval y p) ∧
    let hprime : (RingHom.ker e.toRingHom).IsPrime := RingHom.ker_isPrime e
    (⟨RingHom.ker e.toRingHom, hprime⟩ :
      PrimeSpectrum (MvPolynomial (Fin n) k ⧸ I)) ∈
      Algebra.smoothLocus k (MvPolynomial (Fin n) k ⧸ I)
y is a point of V(I) at which K[x]/I is smooth over K in Mathlib's sense.
N(Y) (Definition 5.4(d))ProjectiveConormalDirections.smoothConormalFibreProjection
Lean text
def smoothConormalFibreProjection
    (I : Ideal (MvPolynomial (Fin n) k)) : Set (Fin n → k) :=
  {ξ | ∃ y : Fin n → k,
    SmoothAffinePoint I y ∧
      coordinateCovector ξ ∈ affineConormalSpace y I}
All covectors ξ that are conormal to Y at some smooth point. The name refers to the conormal variety {(y, ξ)} ⊆ T*X projected to the fibre coordinate ξ.
Zariski closure in ℙ((K^n)*) (Definition 5.5)ProjectiveConormalDirections.projectiveHomogeneousClosure
Lean text
def projectiveHomogeneousClosure
    (T : Set (Projectivization k (Fin n → k))) :
    Set (Projectivization k (Fin n → k)) :=
  {z | ∀ P : MvPolynomial (Fin n) k, ∀ d : ℕ,
    P.IsHomogeneous d →
    T ⊆ projectiveHomogeneousZeroLocus P →
      z ∈ projectiveHomogeneousZeroLocus P}
The points of projective space at which every homogeneous polynomial vanishing on the given set vanishes.
fibre-conical (Definition 5.6)FibreConicalVanishingIdeal.IsFibreConical
Lean text
/-- Stability under every nonzero scalar dilation in each cotangent fibre. -/
def IsFibreConical (W : Set (PhaseVar n → k)) : Prop :=
  ∀ q ∈ W, ∀ a : k, a ≠ 0 →
    Sum.elim (fun i => q (.inl i)) (fun i => a * q (.inr i)) ∈ W
W ⊆ T*X is stable under (y, ξ) ↦ (y, tξ) for every t ≠ 0.
φ(t) = g(y, t·df_y) (proof of Theorem 5.8)CoisotropicTranslation.fibreLinePolynomial
Lean text
/-- The polynomial line `t ↦ g(y, t v)` in the fibre over `y`, the
zero-fibre specialization of `affineFibreLinePolynomial`. -/
abbrev fibreLinePolynomial
    (y v : Fin n → k) (g : SymbolRing k n) : Polynomial k :=
  affineFibreLinePolynomial y (fun _ => 0) v g
A polynomial g on T*X restricted to the line t ↦ (y, tξ) in the fibre over y, as a polynomial in t.
T*_y X (Section 5.1)Section2.fibre
Lean text
/-- **Definition 2.5 (fibre).** `T^*_y X = π^{-1}(y) ≅ (K^n)^*`. -/
def fibre {K : Type v} {n : ℕ} (y : Fin n → K) : Set (PhaseVar n → K) :=
  proj ⁻¹' {y}
The fibre π⁻¹(y) of the cotangent bundle over a point y, a copy of (K^n)*.

Mathlib notions used by the geometry

MathlibRole in Section 5
PrimeSpectrumSpec; T*X = Spec k[x, ξ] and the points 𝔭 of Spec K[x]used in the proof of Theorem 1.1
MvPolynomial.zeroLocusclassical zero sets V_K(J) ⊆ K^Nused in the proof of Theorem 1.1
MvPolynomial.vanishingIdealvanishing ideals I(S)used in the proof of Theorem 1.1
Algebra.smoothLocusthe smooth locus of K[x]/𝔭used in the proof of Theorem 1.1
KaehlerDifferentialΩ of a ring or field extension, with its universal derivationused in the proof of Theorem 1.1
Algebra.trdegtranscendence degree, dim Y = trdeg of the function fieldused in the proof of Theorem 1.1
ringKrullDimKrull dimension of K[x]/𝔭used in the proof of Theorem 1.1
ValuationSubringvaluation rings of the function fieldused in the proof of Theorem 1.1
IsDiscreteValuationRingthe discrete valuation ring V of Lemma 5.10used in the proof of Theorem 1.1
AdicCompletionthe completion of Vused in the proof of Theorem 1.1
PowerSeriesκ[[t]], the completed valuation ringused in the proof of Theorem 1.1
LaurentSeriesκ((t)), where the generic point is expandedused in the proof of Theorem 1.1
Projectivizationℙ((K^n)*) for Dir(Y)not used for Theorem 1.1 or Corollary 6.3; the earlier proof of the universal statement uses it

Geometry modules behind Theorem 1.1

The repository holds two proofs of the identity 1 = dR + FdS. The proof of Theorem 1.1 follows the paper and uses 69 of the 236 geometry modules. An earlier proof of the same identity (Stafford38.universalStatement) takes a different route through the geometry and uses 191 modules; 60 serve both, and 36 serve neither. The earlier proof is not needed to read the paper. The statements of Section 5 themselves, Lemmas 5.9 to 5.13 and Theorems 5.7 and 5.8, are in Stafford38.Paper.Section5.

The 69 geometry modules used by the proof that follows the paper (internal names)

ActualWitnessConormalAssembly, AffineComponentCoordinateSplit, AffineConormalClosure, AffineConormalSpan, AsymptoticDivisorExistence, CoisotropicTranslation, CompletedDVRCoefficientSection, CompletedDVRPowerSeriesEquiv, ComponentFunctionFieldBoundary, ComponentProjectiveClosure, ComponentProjectiveClosureNormalization, ComponentProjectiveOrder, ConormalAxisContradiction, ConormalPrincipalOpenDensity, ConormalScalarExtensionVanishing, ContinuousPowerSeriesTangentFrame, DivisorTangentLattice, DivisorialVisibleFrameCore, DivisorialVisibleFrameStage4, DivisorialVisibleFrameStage5, DivisorialVisibleFrameStageAssembly, ExactDivisorialVisibleFrameExistence, FibreConicalVanishingIdeal, FiniteGradientResidueExtension, FiniteSeparableDVRChartFoundation, FormalDivisorAxisLift, FormalDivisorLaurentConormal, FormalDivisorTangent, GeneralCoisotropicSets, GeneralConormalContainment, GeneralCoordinateAvoidance, GeneralDivisorialVisibleFrame, GeneralDivisorialVisibleFrameColumn, GeneralDivisorialVisibleFrameCoreData, GeneralDivisorialVisibleFrameWitness, GenericPointKaehlerConormal, GenericSmoothOpen, KaehlerDVRVisibility, KaehlerSpanSeparableAdjoin, KaehlerTranscendenceBasis, KaehlerVisibleDerivationFrame, LaurentConormalDirection, LaurentConormalResidueExtension, NormalizationHeightOne, PaperActualDivisorTangent, PaperHamiltonianFlow, PaperLaurentChartDerivation, PaperResidueDerivationFrame, PaperResidueGroundAlgebraTransport, PaperRetainedChartAssembly, PowerSeriesArcTangency, PowerSeriesTangentLimit, ProjectiveChartCoordinates, ProjectiveConormalDehomogenization, ProjectiveConormalDirections, ProjectiveDivisorOrderGap, ProjectiveValuationNormalization, RelativeCoefficientDVRPlace, RelativeRetainedBoundaryPlace, ResidueMinorSelection, RetainedDVRPlace, RetainedGroundMapIdentification, RetainedPlaceConormalTransport, RetainedProjectiveCompletion, RetractionSpecialization, ScalarExtensionPoints, SmoothAffineConormal, SmoothConormalFibreVanishing, SplitTangentMatrix

Generated by scripts/build-definitions.py from definitions.json, verification/definitions/usage.json and verification/lean-references/references.json. The page is a reading aid for the checked proof.