The valuation spectrum of a ring #
We define the valuation spectrum Spv A following Wedhorn, Adic Spaces
(arXiv:1910.05934v1), Definition 4.1.
Main definitions #
TauCeti.ValuationSpectrum A: The valuation spectrum, the type ofValuativeRelinstances onA.TauCeti.ValuationSpectrum.basicOpen f s: The basic open set{v ∈ Spv A | v(f) ≤ v(s) ≠ 0}.TauCeti.ValuationSpectrum.comap φ: The continuous mapSpv B → Spv Ainduced byφ : A →+* B.TauCeti.ValuationSpectrum.supp v: The support ideal{a ∈ A | v(a) = 0}.TauCeti.ValuationSpectrum.basicOpenFinset_inter: Wedhorn's step (i) in the proof of Lemma 7.5, that the rational opens are stable under finite intersection.TauCeti.ValuationSpectrum.basicOpenFinset_image_mul_right: scaling every numerator and the denominator by a unit gives the same rational open.TauCeti.ValuationSpectrum.comap_preimage_basicOpenFinset: a rational open pulls back to the rational open presented by the images of its numerators and denominator.TauCeti.ValuationSpectrum.isClosed_setOfPred_forall_vlt_one: the sub-unit locus of a set of ring elements is closed — the closedness behind Wedhorn's Corollary 7.12.TauCeti.ValuationSpectrum.quotientLift 𝔞 h: Lift the implicitly inferred pointvwith𝔞 ≤ supp vtoSpv (A ⧸ 𝔞).TauCeti.ValuationSpectrum.localizationComapSection S B v hS: Liftvto a localizationSpv B.TauCeti.ValuationSpectrum.localization_comap_isEmbedding: pullback from the valuation spectrum of a localization is a topological embedding.TauCeti.ValuationSpectrum.suppFun: The continuous support mapSpv A → Spec A.TauCeti.ValuationSpectrum.trivialSection: The continuous section ofsuppFungiven by the trivial valuation attached to a prime ideal; in particularsuppFunis surjective (suppFun_surjective).
References #
- T. Wedhorn, Adic Spaces, arXiv:1910.05934v1, Definition 4.1, Remark 4.3, Remark 4.4, Remark 4.6, Proposition 4.7(2)
Ported from the open Mathlib pull request
leanprover-community/mathlib4#38009
(which supersedes an earlier draft in AINTLIB projects/AdicSpaces); this copy is deleted in
favour of the Mathlib declarations once that pull request reaches the pinned Mathlib.
The localization embedding follows the organization and generated-topology argument of Mathlib's
PrimeSpectrum.localization_comap_isEmbedding, adapted here from prime ideals to valuative
relations.
The valuation spectrum Spv A of a commutative ring A. A point of Spv A is a
ValuativeRel A, but Spv A is kept as a distinct type so it can carry its own topology
without affecting ValuativeRel A.
- ofValuativeRel :: (
- toValuativeRel : ValuativeRel A
The underlying
ValuativeRel Aof a pointv : Spv A. - )
Instances For
The valuation spectrum Spv A of a commutative ring A. A point of Spv A is a
ValuativeRel A, but Spv A is kept as a distinct type so it can carry its own topology
without affecting ValuativeRel A.
Equations
- TauCeti.termSpv = Lean.ParserDescr.node `TauCeti.termSpv 1024 (Lean.ParserDescr.symbol "Spv")
Instances For
Over a subsingleton ring, the valuation spectrum is empty.
Two points of Spv A are equal as soon as their underlying vle relations agree.
Construct a point of Spv A from a valuation v : Valuation A Γ₀.
Equations
- TauCeti.ValuationSpectrum.ofValuation v = { toValuativeRel := ValuativeRel.ofValuation v }
Instances For
Two equivalent valuations define the same point of Spv A.
The basic open subset Spv(A)(f/s) = {v ∈ Spv A | v(f) ≤ v(s) ∧ v(s) ≠ 0}.
Equations
Instances For
The basic open subset for f = s = 1 is the whole spectrum: Spv(A)(1/1) = Spv A.
Not @[simp]: basicOpen_one_right together with ValuativeRel.vle_refl already reduces
the left-hand side, so a simp attribute here would be redundant (simpNF).
The topology on Spv A generated by the basic open sets basicOpen f s
for f, s : A.
Equations
- TauCeti.ValuationSpectrum.instTopologicalSpace = TopologicalSpace.generateFrom {U : Set (TauCeti.ValuationSpectrum A) | ∃ (f : A) (s : A), U = TauCeti.ValuationSpectrum.basicOpen f s}
The topology of Spv A is generated by the basic opens (the defining equation of the
instance).
The valuative relation of a point is determined by its basic opens: v(f) ≤ v(s) holds
iff v lies in basicOpen f s, or s and f both lie in the support — the latter being
detected by the diagonal basic opens basicOpen s s and basicOpen f f.
The sub-unit locus of a set of ring elements is closed: demanding v(a) < 1 at every
a ∈ S cuts out a closed subset of Spv A. The complement is the union over a ∈ S of the
basic opens Spv(A)(1/a) — the condition 1 ≤ v(a) already forces v(a) ≠ 0, so no separate
nonvanishing clause survives.
This is the closedness underlying Wedhorn's Corollary 7.12: Theorem 7.10 describes Cont A
inside Spv (A, IA) by exactly such conditions, so Cont A is the trace of a closed set.
Spv A is T0: inseparable points agree on every basic open, hence carry the same
valuative relation.
The contravariant map Spv B → Spv A induced by φ : A →+* B.
Equations
- TauCeti.ValuationSpectrum.comap φ v = { toValuativeRel := ValuativeRel.comap φ v.toValuativeRel }
Instances For
comap is compatible with ofValuation.
comap φ is continuous.
comap of the identity is the identity.
comap φ is injective when φ is surjective.
Pulling back along two composable morphisms of commutative rings is pulling back along their composite.
The support ideal {a ∈ A | v(a) = 0} of a point v : Spv A.
Equations
- v.supp = ValuativeRel.supp A
Instances For
Membership in the support, as the relation v(x) ≤ v(0).
The support of a point v : Spv A is a prime ideal.
(ofValuation v).vle x y ↔ v x ≤ v y.
The support of ofValuation v equals v.supp.
The canonical valuation associated to a point v : Spv A.
Equations
Instances For
v.valuation is ValuativeRel.valuation for the valuative relation of v.
The two sides are definitionally equal, but the definition of
TauCeti.ValuationSpectrum.valuation is not exposed outside this module, so a downstream module
cannot match v.valuation against Mathlib's ValuativeRel.valuation API without this equation.
It exists to cross that module boundary, not to abbreviate.
Comparison under the canonical valuation of a point is the point's valuative relation.
Strict comparison under the canonical valuation of a point is the point's strict valuative
relation — the strict sibling of valuation_le_iff, and the direct bridge between strict
valuation inequalities and vlt hypotheses.
The support of v : Spv A equals the support of its canonical valuation.
The support of a pullback is the preimage of the support.
The canonical valuation gives back the same point of Spv.
The canonical valuation of the point determined by w is equivalent to w.
𝔞 ≤ supp(comap(mk 𝔞, w)) for all w : Spv (A ⧸ 𝔞).
Lift a point v ∈ Spv A with 𝔞 ≤ supp v to Spv (A ⧸ 𝔞).
Equations
Instances For
comap (mk 𝔞) (quotientLift 𝔞 h) = v.
quotientLift 𝔞 (self_le_supp_comap 𝔞 w) = w.
comap (mk 𝔞) : Spv (A ⧸ 𝔞) → Spv A is a topological embedding.
Send v ∈ Spv A with S ≤ (supp v).primeCompl to the localization Spv B, where B is a
localization of A at the submonoid S.
Equations
Instances For
comap (algebraMap A B) (localizationComapSection S B v hS) = v.
S is disjoint from supp(comap(algebraMap, w)) for w : Spv B.
The range of comap (algebraMap A B) is {v | S ≤ supp(v).primeCompl}.
Pullback of valuative relations along a localization map is injective.
The preimage under localization pullback of the basic open obtained by clearing denominators is the basic open defined by the original fractions.
Pullback of valuative relations along a localization map induces the source topology.
Pullback of valuative relations along a localization map is a topological embedding.
The support map to the prime spectrum #
The support map Spv A → Spec A.
Instances For
The prime ideal underlying suppFun v is the support of v.
suppFun ⁻¹' D(f) = Spv(A)(f/f).
suppFun is continuous.
The trivial-valuation section of the support map #
The trivial-valuation section of the support map (Wedhorn, Remark 4.6): the point of
Spv A given by the trivial valuation attached to a prime ideal.
Equations
Instances For
trivialSection is a section of the support map.
The preimage of a basic open under trivialSection is the corresponding basic open of
the prime spectrum: trivialSection ⁻¹' Spv(A)(f/s) = D(s).
trivialSection is continuous.
The support map is surjective; the trivial valuations provide a section.
Rational opens with a finite numerator set #
Wedhorn's Spv(A)(T/s) for a finite set T: the points where every t ∈ T is
dominated by s, and s is not in the support.
Equations
Instances For
Each numerator gives back an ordinary basic open.
Spv(A)(T/s) is the finite intersection of the basic opens Spv(A)(t/s) for t ranging
over T ∪ {s}; the extra s is what carries the nonvanishing clause when T is empty.
Inserting the denominator among the numerators changes nothing: the extra condition it adds
is v s ≤ v s.
Multiplying a presentation by a unit changes nothing. If u is a unit, then multiplying
every numerator and the denominator of Spv(A)(T/s) by u gives the same rational open.
No injectivity of t ↦ t * u is needed.
The preimage of Spv(A)(T/s) under comap φ is Spv(B)(φ(T)/φ(s)), the finite-numerator
form of comap_preimage_basicOpen.
Wedhorn's step (i) in the proof of Lemma 7.5: the rational opens are stable under finite
intersection. Writing Uᵢ = insert sᵢ Tᵢ for the numerator set augmented by its own denominator,
Spv(A)(T₁/s₁) ∩ Spv(A)(T₂/s₂) = Spv(A)(U₁U₂ / s₁s₂).
The numerator sets on the right carry their own denominators, which
basicOpenFinset_insert_self shows costs nothing — the same absorption IsAdmissible performs
for the admissibility condition.