Extending a valuation from a subring reached by the powers of an element #
Let R be a subring of a ring A, and let s ∈ R be central in A, with some power of s
carrying each element of A into R: for every a there is an n with sⁿ * a ∈ R. A
valuation w of R that does not vanish at s then has only one possible extension to A,
v a = w (sⁿ * a) * (w s)⁻ⁿ,
and this file shows that the formula is well defined, is a valuation, and is the only valuation
of A restricting to w.
No topology is involved. The hypothesis is met by a ring of definition of a Huber ring — it is
open, so a topologically nilpotent s satisfies it — and
TauCeti.RingTheory.Huber.ExtendValuation is that specialisation.
Why this is not Valuation.extendToLocalization #
Mathlib extends a valuation along a localisation that inverts a set on which the valuation is
nonzero. That does not apply here: the hypothesis does not make s invertible in A, so there
need be no ring map R[1/s] → A at all. Take R = A = ℤ_[p] and s = p, where R[1/s] = ℚ_[p].
What is true, and is all the formula needs, is the one-sided statement that every element of A
is carried into R by a power of s.
Well-definedness #
Independence of n reduces to the case of comparing n with n + j, where
s ^ (n + j) * a = s ^ j * (sⁿ * a) splits off a factor whose w-value is (w s) ^ j, exactly
cancelling the extra (w s)⁻ʲ. Two arbitrary exponents are then compared through their sum.
The two valuation axioms reach a shared exponent differently. For a product, each argument keeps
its own workable exponent — x at m and y at n — and only the product is evaluated at
m + n, because s ^ (m + n) * (x * y) = (sᵐ * x) * (sⁿ * y) already splits that way. A sum has
no such splitting, so there both terms are raised to the common exponent m + n. In each case
the axiom is then inherited from w once the shared factor (w s)⁻⁽ᵐ⁺ⁿ⁾ is divided out.
Main definitions #
Valuation.extendOfPowMulMem: the extension ofwtoA.
Main results #
Valuation.extendOfPowMulMem_apply: the defining formula, at every exponent that works, not just the chosen one. This is the interface; the definition goes throughClassical.chooseand is not meant to be unfolded.Valuation.extendOfPowMulMem_coe: the extension restricts tow.Valuation.eq_extendOfPowMulMem: uniqueness — any valuation ofArestricting towis this one, so the extension is canonical.Valuation.extendOfPowMulMem_congr: consequently the extension does not depend on whichsis used to build it.
References #
- T. Wedhorn, Adic Spaces (arXiv:1910.05934v1), Lemma 7.44(3), which uses this extension for a ring of definition of a Huber ring.
Provenance #
Adapted from C. Birkbeck, AINTLIB, branch
dev/adic-spaces, commit 37bbdaeb9, projects/AdicSpaces/Adic spaces/Lemma745.lean,
declarations vExtFun_step, vExtFun_well_defined, vExtFun_map_mul,
vExtFun_map_add_le_max and exists_valuation_extension. Adapted, not copied: that
development states the result existentially, as ∃ v_ext, …, for a pair of definition of a
Huber ring, and threads the value w s through five separate lemmas as an explicit parameter
with its own defining equation. Here the extension is a def, so it can be named and rewritten
at a call site, the arithmetic is one private lemma rather than four public ones, and the whole
construction is carried out for a subring reached by the powers of s. The uniqueness theorem
and the resulting independence of s have no counterpart there.
The extension of w from R to A, when every element of A is carried into R by
some power of a central element s. For any n with sⁿ * a ∈ R the value is
w (sⁿ * a) * (w s)⁻ⁿ, and extendOfPowMulMem_apply says so at every such n.
Equations
Instances For
The defining formula, at every exponent that carries a into the subring.
The extension restricts to w.
The extension is the only one: a valuation of A restricting to w on R is
extendOfPowMulMem. In particular the extension is canonical.
The extension does not depend on s: two central elements of R whose powers carry
A into R and at which w is nonzero give the same extension.