Bounded subsets of a topological ring #
A subset S of a topological monoid with zero is bounded when every neighbourhood U of zero
absorbs S: there is a neighbourhood V of zero with V * S ⊆ U. This is the notion of
boundedness underlying Huber's theory of adic spaces (Wedhorn, Adic Spaces, Definition 5.27),
where it is the condition cutting out the rings of definition of a Huber ring.
Provenance #
IsBounded and the elementary calculus — IsBounded.subset, isBounded_empty,
isBounded_singleton_zero, isBounded_pair_zero_one, IsBounded.union, IsBounded.mul,
isBounded_singleton and isBounded_finite — follow William Coram's mathlib4#40013, as do
several of their proofs. Everything else here is new: isBounded_iUnion, the nonarchimedean
IsBounded.addSubgroupClosure, IsBounded.add and isBounded_sup, and the image and transport
lemmas. The selection and ordering of results follows AINTLIB's Bounded.lean; its proofs were
not used.
Main definitions #
TauCeti.Huber.IsBounded:Sis bounded, i.e. every neighbourhood of zero absorbsS.
Main results #
TauCeti.Huber.isBounded_iff: unfolding lemma forIsBounded.TauCeti.Huber.isBounded_finsetProd: a finite pointwise product of bounded sets is bounded.TauCeti.Huber.isBounded_finite: finite sets are bounded.TauCeti.Huber.isBounded_of_isLinearTopology: in a linearly topologized commutative ring, such as an adic ring, every subset is bounded.TauCeti.Huber.IsBounded.union,TauCeti.Huber.IsBounded.mul: unions and pointwise products of bounded sets are bounded.TauCeti.Huber.IsBounded.add,TauCeti.Huber.IsBounded.addSubgroupClosure: over a ring with a nonarchimedean additive group, sums and the generated additive subgroup of bounded sets stay bounded.TauCeti.Huber.isBounded_sup: the join of two bounded subrings of a commutative ring is bounded — the boundedness input for joins of rings of definition.TauCeti.Huber.IsBounded.image: a morphism continuous at zero which is open at zero preserves boundedness, andTauCeti.Huber.isBounded_image_ringEquiv_ifftransports boundedness along a topological ring isomorphism.
Implementation notes #
This is not Mathlib's Bornology.IsVonNBounded. That predicate asks that every neighbourhood of
zero absorb S after dilation by a norm-large scalar, so it needs a SeminormedRing of scalars
to have a bornology to be cofinal in; a Huber ring such as ℤ_[p]⟦T⟧ with its (p, T)-adic
topology carries no such norm. The two notions agree over a nontrivially normed field acting on
itself, but neither the statement shape (∃ V ∈ 𝓝 0, V * S ⊆ U against
∀ᶠ a in cobounded, S ⊆ a • U) nor the hypotheses transfer.
Boundedness is not preserved by an arbitrary continuous homomorphism: giving ℚ_[p] the
discrete topology makes every subset bounded, while the identity to the p-adic topology is
continuous and ℚ_[p] is not bounded in itself. IsBounded.image therefore carries the extra
hypothesis that the morphism maps neighbourhoods of zero to neighbourhoods of zero.
References #
- Wedhorn, Adic Spaces, Definition 5.27 and Remark 5.28.
- William Coram, feat: define bounded sets and power bounded elements, mathlib4#40013.
- AINTLIB, branch
dev/adic-spaces,projects/AdicSpaces/Adic spaces/Bounded.lean.
A subset S of a topological monoid with zero is bounded if for every neighbourhood U
of 0 there is a neighbourhood V of 0 with V * S ⊆ U.
Equations
- TauCeti.Huber.IsBounded S = ∀ U ∈ nhds 0, ∃ V ∈ nhds 0, V * S ⊆ U
Instances For
Unfolding lemma for TauCeti.Huber.IsBounded.
Some power of a topologically nilpotent element multiplies a bounded set into any neighbourhood of zero.
Subsets of bounded sets are bounded.
The empty set is bounded.
The singleton {0} is bounded.
The pair {0, 1} is bounded.
A finite union of bounded sets is bounded.
A finite indexed union is bounded exactly when every piece is.
The union of two bounded sets is bounded.
A union is bounded exactly when both parts are.
The pointwise product of two bounded sets is bounded.
Every subset of a discrete monoid with zero is bounded.
Every singleton is bounded.
Every finite subset is bounded.
A finite pointwise product of bounded sets is bounded. Commutativity is
what makes the Finset product of sets available in the first place, and no continuity is
required.
Every subset of a linearly topologized commutative ring is bounded: a neighbourhood of zero
contains an open ideal, which absorbs every subset. This covers the adic rings, such as ℤ_[p]
or W(𝒪_F) with its (p, [ϖ])-adic topology, and in particular the discrete rings.
The additive subgroup generated by a bounded set is bounded in a nonarchimedean ring (the key step of Wedhorn, Proposition 5.30).
The pointwise sum of two bounded sets is bounded in a nonarchimedean ring.
The join of two bounded subrings of a commutative ring is bounded.
A morphism continuous at zero which carries neighbourhoods of zero to neighbourhoods of zero sends bounded sets to bounded sets.
Boundedness is a condition at zero only, so continuity away from zero is irrelevant; but continuity at zero alone is not enough, since refining the topology on the source only creates bounded sets.
An open morphism continuous at zero sends bounded sets to bounded sets.
A topological ring isomorphism sends bounded sets to bounded sets.
Boundedness transports along a topological ring isomorphism.