Pointedness at the place at infinity #
An isogeny of affine Weierstrass curves is a coordinate pullback φ : R(W₂) → K(W₁) carrying the
integrality condition MapsInfinity, the algebraic form of φ(O₁) = O₂. This file reads that
condition off as a statement about places: restricting the place at infinity of W₁ along the
function-field pullback gives the place at infinity of W₂,
((W₁.infinityPlace).comap φ.fieldPullback).IsEquiv W₂.infinityPlace.
The proof is the pole of x. Suppose the pulled-back coordinate φ x₂ had no pole at O₁. The
Weierstrass equation of W₂, pulled back, then forces φ y₂ to have no pole either — a pole of
φ y₂ would make the left-hand side dominate a right-hand side of value at most 1 — so the whole
image φ(R(W₂)) lies in the valuation ring at infinity of W₁. That ring is integrally closed,
so MapsInfinity puts x₁ in it as well, contradicting the double pole v_∞ x₁ = exp 2. Hence
φ x₂ has a pole at O₁, and
WeierstrassCurve.Affine.isEquiv_infinityPlace_of_one_lt identifies the restricted valuation.
Conversely, a coordinate pullback is pointed as soon as the pullback of some function of the target
acquires a pole at the source's infinity, with no injectivity assumed; the coordinate x₂ is the
usual witness, so pointedness of an arbitrary coordinate pullback is exactly a pole of the
pulled-back x₂.
Neither direction uses ellipticity, separability, or the degree of an isogeny.
Main results #
TauCeti.Isogeny.one_lt_infinityPlace_pullback_X: the pulled-back coordinate has a pole at infinity,1 < v_∞ (φ x₂).TauCeti.Isogeny.isEquiv_comap_infinityPlace: the place at infinity restricts to the place at infinity along an isogeny.TauCeti.CoordinatePullback.mapsInfinity_iff_one_lt_infinityPlace: pointedness is exactly a pole ofxat infinity, for any coordinate pullback — the form in which a construction can establish it by one valuation computation.TauCeti.CoordinatePullback.mapsInfinity_map_iff: base change reflects pointedness.TauCeti.CoordinatePullback.mapsInfinity_iff_isEquiv_comap_infinityPlace: the pointedness criterion,MapsInfinity σ ↔ σ_*(O₁) = O₂, for an embeddingσof function fields.
References #
- J. Silverman, The Arithmetic of Elliptic Curves, II.2, III.4.
The pullback of the target's coordinate x has a pole at the source's point at infinity.
If it did not, then neither would the pullback of y — by the Weierstrass equation of the target —
so the whole pulled-back coordinate ring would lie in the valuation ring at infinity of the source;
that ring is integrally closed, so MapsInfinity would put the source coordinate x₁ in it,
against its double pole.
An isogeny carries the place at infinity to the place at infinity: the restriction of the
source's place at infinity along the function-field pullback is equivalent to the target's. This is
the place-level reading of MapsInfinity, that is, of φ(O₁) = O₂.
A coordinate pullback under which some function acquires a pole at infinity maps infinity
to infinity. Any one function of W₂ whose pullback has a pole at the source's point at infinity
suffices; the coordinate x is the usual witness (mapsInfinity_iff_one_lt_infinityPlace).
Injectivity of p is not assumed, in contrast with the criterion for embeddings of function
fields, mapsInfinity_iff_isEquiv_comap_infinityPlace.
Pointedness is exactly a pole of x at infinity. A coordinate pullback maps infinity to
infinity precisely when it sends the target's coordinate x to a function with a pole at the
source's point at infinity, so a construction can establish pointedness by a single valuation
computation.
Changing the coefficient field reflects pointedness. The pole criterion for MapsInfinity
is unchanged because the infinity valuation after base change restricts to the original one.
The pointedness criterion. An embedding σ : F(W₂) → F(W₁) restricts to a coordinate
pullback which maps infinity to infinity exactly when the source's place at infinity restricts
along σ to the target's place at infinity.