Negation and topological nilpotence #
Mathlib closes IsTopologicallyNilpotent under the operations that a topology on the ring makes
available — zero, add, mul_left, mul_right, map — but not under negation. This file adds
that, at the hypotheses the statement needs: a MonoidWithZero with a distributive negation, and
continuity of negation.
No uniformity, nonarchimedean neighbourhood basis or commutativity is involved. (-a) ^ n is
a ^ n for even n and -(a ^ n) for odd n, and both of those sequences tend to 0 as soon as
negation is continuous, so every neighbourhood of 0 eventually contains (-a) ^ n whichever
parity n has.
Main results #
IsTopologicallyNilpotent.neg: the negation of a topologically nilpotent element is topologically nilpotent.isTopologicallyNilpotent_neg: the same fact as asimpiff, mirroringisPowerBounded_neg.IsTopologicallyNilpotent.pow: a positive power of a topologically nilpotent element is topologically nilpotent.eventually_mul_pow_mem_of_isTopologicallyNilpotent, and its.existsform: a topologically nilpotent element absorbs any fixed element into any open subring.IsTopologicallyNilpotent.of_pow: an element with a topologically nilpotent positive power is topologically nilpotent.Prod.isTopologicallyNilpotent_iff: an element of a product is topologically nilpotent exactly when both of its components are.IsTopologicallyNilpotent.mem_jacobson_bot,topologicalNilradical_le_jacobson_bot: if an open ideal lies in the Jacobson radical (for instance, the ring is adically complete with respect to it), so does every topologically nilpotent element.IsTopologicallyNilpotent.isUnit_add_left: when the topological nilradical lies in the Jacobson radical, a unit plus a topologically nilpotent element is a unit.IsTopologicallyNilpotent.prod_one_sub_sub_one: for scalarsμfrom a multiset and a topologically nilpotenta,∏_μ (1 - μ a) - 1is topologically nilpotent.
Provenance #
The statement is AINTLIB's IsTopologicallyNilpotent.neg, branch dev/adic-spaces, commit
37bbdaeb, Apache-2.0, Chris Birkbeck, projects/AdicSpaces/Adic spaces/GeometricSeries.lean.
There it sits inside the geometric-series development over a complete nonarchimedean uniform
commutative ring, and the sign on the odd powers is absorbed by an open subgroup. Only the
statement is followed here: the hypotheses are weakened to the ones the result needs, and the
open-subgroup step is replaced by the two-sequence argument above, which needs no nonarchimedean
basis. The iff form has no counterpart in the source.
References #
- C. Birkbeck, AINTLIB, branch
dev/adic-spaces, commit37bbdaeb,projects/AdicSpaces/Adic spaces/GeometricSeries.lean.
The negation of a topologically nilpotent element is topologically nilpotent.
Topological nilpotence is invariant under negation.
A positive power of a topologically nilpotent element is topologically nilpotent: the
powers of a ^ n are a subsequence of the powers of a.
A topologically nilpotent element absorbs any element into any open subring. For s
topologically nilpotent and B open, a * s ^ n lies in B for all large n.
Multiplication by a is continuous, so B pulls back to a neighbourhood of 0, and the powers of
s converge to 0. Neither commutativity nor continuity of addition is used — only
SeparatelyContinuousMul — and no Huber structure enters, which is why this sits here rather than
beside the ring-of-definition form it generalises,
TauCeti.Huber.PairOfDefinition.exists_pow_idealOfDefinition_mul_mem.
The existential form of eventually_mul_pow_mem_of_isTopologicallyNilpotent: some power of a
topologically nilpotent element carries a into an open subring.
An element with a topologically nilpotent positive power is topologically nilpotent.
An element of a product of topological monoids with zero is topologically nilpotent exactly
when both of its components are: the powers of x are taken componentwise, and a sequence in a
product converges exactly when both of its component sequences do.
Topologically nilpotent elements lie in the Jacobson radical as soon as some open ideal
I does, for instance when R is I-adically complete (IsAdicComplete.le_jacobson): a
power of a topologically nilpotent element lies in I, and the Jacobson radical is radical.
In a linearly topologized ring, the topological nilradical lies in the Jacobson radical as soon as some open ideal does.
When the topological nilradical lies in the Jacobson radical (for instance, by
topologicalNilradical_le_jacobson_bot), a unit plus a topologically nilpotent element is a unit.
Compare IsNilpotent.isUnit_add_left_of_commute.
For a multiset s of scalars and a topologically nilpotent a, ∏_{μ ∈ s} (1 - μ a) - 1 is
topologically nilpotent.