23 · Functoriality and descent

11 NAG Tags

The analytic scalar extension XLX_L is constructed by completed tensor products in Section 2.6. That construction determines its points, functions and projection to XX. It does not determine which ground-field calibration is used when integrating upstairs. We therefore specify the calibration before applying a degree or trace formula. The Galois-product example, Example 2.6.3, also shows why a descent argument must use the analytic fibre product rather than the Cartesian product of point sets.

23.1 Field extension, degree, and descent of representatives

Extending the field does not reset the valuation

Fix a complete valued field KK and a finite extension L/KL/K, with the absolute value extended from KK. If πK\pi_K is a uniformizer in a discretely valued example, its value remains −log⁡∣πK∣=λK-\log|\pi_K|=\lambda_K in LL. A uniformizer πL\pi_L has value λK/e\lambda_K/e when the ramification index is ee. Thus skeleton lengths and logarithmic weights are compared using one fixed real scale.

Normalizing the valuation independently in both fields by assigning value one to each uniformizer changes that scale. A formula that compares local heights or slopes across the extension must then include the corresponding conversion factor. The safest convention for an analytic descent problem is to retain the original absolute value throughout.

The finite flat degree formula and the base field

For a finite flat map f:Y→Xf:Y\to X of constant degree dd over one fixed field, the calibrated projection formula gives f∗[Y]=d[X]f_*[Y]=d[X] in the applicable algebraic setting. Scalar extension is normalized differently: with XLX_L integrated over LL, the base-change projection satisfies p∗[XL]L=[X]Kp_*[X_L]_L=[X]_K. Section 23.2 explains and proves the distinction from the cited calibration theorem. A degree factor must therefore be attached to an explicit choice of ground-field calibration, not to the notation XL→XX_L\to X alone.

What averaging can and cannot descend

If an object upstairs belongs to a class stable under the relevant finite trace, one may hope to define a descended object by d−1π∗d^{-1}\pi_* and verify the desired identity by the projection formula. For currents, proper pushforward is often available. For an actual sheaf of delta-forms or another specified representative class, stability under that trace must be proved.

A pushed-forward current satisfying an equation is not automatically an actual form in the required sheaf. The operation may forget the local presentation or introduce singularities outside the admitted class. A descent theorem must show that the trace of the upstairs representative has local expressions downstairs and that those expressions satisfy the correct relations on overlaps.

Finite descent and alteration descent are different

A proper alteration is not a finite flat cover on the entire space. Exceptional loci, changes of relative dimension, and ramification can affect both support and representative regularity. The existence of a proper pushforward current remains useful, but dividing by a generic degree does not automatically reconstruct an actual form on the original complement.

For the strong Green problem, one must in addition control the support: the upstairs construction may require removing the inverse image of the cycle together with an exceptional divisor. A descended current on the original space does not prove that its representative is defined on the exact original complement. This support issue is independent of the normalization by degree.

Compatibility conditions for descent

To use an extension or a cover in a proof, establish in order: the upstairs equation on the full admitted tests; the trace operation on the actual representative class; commutation with d′d' and d′′d''; the projection formula with the correct degree; and preservation of the precise support complement. If the goal is only an equality of currents, say so. If the goal is an actual Green form, the representative step cannot be omitted.

This template is particularly relevant to comparisons between CLD currents, GK delta-forms, and Mihatsch’s analytic delta-forms. Their operations overlap, but the domains of trace and pushforward need not coincide. The correct comparison carries the object and its testing category through every step.

References. The finite-map and current projection formulas in [CLD25] (Chapters 13–16); the representative-sensitive constructions in [Mih24].

23.2 Two different degree formulas

The field of definition of an integral is part of its normalization. A finite map between spaces over one fixed field and a change of the ground field are different operations. The earlier version of these notes did not separate them clearly enough. We correct that ambiguity here.

Theorem 23.2.1. (Scalar extension; external geometric input). Let XX be an analytic space of level nn in the CLD25 framework, let L/KL/K be a complete valued extension with the extended original absolute value, and let p:XL→Xp:X_L\to X be the scalar-extension projection. With the canonical calibrations over LL and KK, respectively, p∗[XL]L=[X]K.p_*[X_L]_L=[X]_K. Equivalently, for an admissible compactly supported top-degree smooth form η\eta on XX, ∫XL,Lp∗η=∫X,Kη.\int_{X_L,L}p^*\eta=\int_{X,K}\eta. There is no factor [L:K][L:K] in this scalar-extension formula, even when the extension is finite.

This is CLD25, Propositions 14.3.6, 15.3.1 and 16.1.1: calibration commutes with scalar extension; the associated measures push forward without changing mass; testing yields the current identity. The calibration theorem is imported here. Once it is given, the displayed integral and current identities follow directly from the definitions.

For a compact morphism f:Y→Xf:Y\to X between analytic spaces over the same field, CLD25 Proposition 16.1.2 instead gives f∗[Y]=(deg⁡cf)[X]f_*[Y]=(\deg_c f)[X], with calibrated degree. In a finite flat algebraic situation where this is the usual constant degree dd, the right side is d[X]d[X]. If L/KL/K is finite, the underlying space XLX_L can also be viewed as a KK-analytic space. Its KK-calibration is then not the LL-calibration in Theorem 23.2.1. One cannot retain the LL-normalized integral and insert the KK-degree factor at the same time.

Example 23.2.2. (Dimension zero). The integration mass of Spec⁡L\operatorname{Spec}L over LL is one. Viewed as a finite zero-cycle over KK, its degree is [L:K][L:K]. Both statements are correct. The discrepancy is entirely the ground-field normalization. It already rules out a universal instruction to divide every pushed-forward base-extended current by [L:K][L:K].

23.3 Pushforward on an actual test space

Proposition 23.3.1. Suppose f:Y→Xf:Y\to X has a continuous pullback f∗f^* on the smooth test spaces of interest, preserves admissible test supports, and commutes with d′,d′′d',d''. For a current TT on YY define f∗Tf_*T by (f∗T)(η)=T(f∗η).(f_*T)(\eta)=T(f^*\eta). Then pushforward commutes with d′,d′′d',d''. If pullback preserves the chosen positive cone, it preserves positivity of currents of the corresponding dual type. If aa is a smooth form, f∗(f∗a∧T)=a∧f∗T,f_*\bigl(f^*a\wedge T\bigr)=a\wedge f_*T, with the same wedge order on both sides.

Proof. Continuity of f∗f^* makes the displayed functional a current. Apply the definition of a distributional derivative and f∗d′=d′f∗f^*d'=d'f^* to a test; the transpose sign is unchanged because the total bidegree shift under pushforward is even. The proof for d′′d'' is identical. A positive test pulls back to a positive test, proving positivity. Finally, f∗(a∧η)=f∗a∧f∗ηf^*(a\wedge\eta)=f^*a\wedge f^*\eta identifies the two pairings in the projection formula. ◻

Proper analytic maps supply the required support control in the standard current formalism; more generally one may work with properness on the support of the current, provided a cutoff argument defines the pairing. A general map does not pull compact supports back to compact supports. Thus the same formula cannot be applied to every map without checking its domain.

Corollary 23.3.2. Under these hypotheses, if Tj⇀TT_j\rightharpoonup T and all pushforwards are defined by the displayed test pullback, then f∗Tj⇀f∗Tf_*T_j\rightharpoonup f_*T. A current identity d′d′′G+δZ=R\mathrm d'\mathrm d''G+\delta_Z=R upstairs pushes forward to d′d′′(f∗G)+f∗δZ=f∗R.\mathrm d'\mathrm d''(f_*G)+f_*\delta_Z=f_*R.

Proof. For each downstairs test, its pullback is one fixed upstairs test, so weak convergence applies. The second assertion follows by linearity and commutation with derivatives. ◻

23.4 The extra step for representatives

The preceding corollary produces a current. Suppose the intended conclusion is an actual form in a specified sheaf Bp,qB^{p,q} on X∖∣Z∣X\setminus|Z|. One additionally needs a trace operation Tr⁡f:Bp,q(Y∖f−1∣Z∣)⟶Bp,q(X∖∣Z∣)\operatorname{Tr}_f: B^{p,q}(Y\setminus f^{-1}|Z|)\longrightarrow B^{p,q}(X\setminus|Z|) with the appropriate degree shift when dimensions differ, compatibility with the current realization, and compatibility with the differentials. This map is additional geometric information. The continuous dual construction of f∗Tf_*T does not provide it.

Lemma 23.4.1. (Exact support in finite descent). Let f:Y→Xf:Y\to X be a finite surjective map of topological spaces. If a representative upstairs is defined only on Y∖(f−1Z∪E)Y\setminus(f^{-1}Z\cup E), then the directly available downstairs domain is at most X∖(Z∪f(E))X\setminus(Z\cup f(E)). To claim a representative on all of X∖ZX\setminus Z, one must remove the obstruction along f(E)∖Zf(E)\setminus Z by an extension theorem or a different construction.

Proof. For x∉Z∪f(E)x\notin Z\cup f(E) every point of f−1(x)f^{-1}(x) lies in the domain upstairs. For x∈f(E)∖Zx\in f(E)\setminus Z at least one point of that fibre is excluded, so a trace requiring data on the whole fibre cannot be evaluated from the given representative. A current extension may still exist, but that is a different assertion. ◻

Averaging over a finite group does not bypass this issue. For continuous functions on a topological quotient X=Y/GX=Y/G, the average ∣G∣−1∑g∈Gg∗u|G|^{-1}\sum_{g\in G}g^*u is invariant and descends continuously by the quotient topology. For differential representatives, invariance must be related to the downstairs sheaf by a separate descent theorem. The function argument alone neither identifies that sheaf nor proves its trace compatibility.

23.5 Finite harmonic maps of metric graphs

Finite maps of curves have a transparent one-dimensional model. A map of metric graphs may be one-to-one on its underlying points and still carry a nontrivial degree, because an edge is traversed with an integral expansion factor. Keeping this factor leads to the correct pullback of measures, trace of functions and Green-kernel normalization.

Let ϕ:Γ′→Γ\phi:\Gamma'\to\Gamma be a surjective map between finite connected compact metric graphs, linear on edges after subdivision, with no contracted edges. For every source edge e′e' mapping into ee, let me′m_{e'} be the positive integral ratio of target length to source length. At a vertex yy above xx, assume that ∑e′↦e at yme′=dy\sum_{e'\mapsto e\text{ at }y}m_{e'}=d_y is independent of the outgoing edge ee at xx. This is the harmonicity condition; dyd_y is the local degree. At an interior point of an edge, the local degree is its expansion factor. We assume both graphs have at least one edge.

Lemma 23.5.1. The sum ∑y∈ϕ−1(x)dy\sum_{y\in\phi^{-1}(x)}d_y is independent of x∈Γx\in\Gamma. Denote it by d=deg⁡ϕd=\deg\phi.

Proof. On the interior of a target edge, the inverse image consists of finitely many edge segments, and the sum of their constant expansion factors is constant. As one approaches an endpoint xx, group these segments by the source vertex yy to which they converge. Harmonicity makes the sum in each group equal to dyd_y, independently of the chosen target edge. The degree therefore has the same value on every adjacent edge. Connectedness finishes the proof. ◻

Trace and pullback

For a continuous function vv on Γ′\Gamma', define its weighted trace by (Tr⁡ϕv)(x)=∑y∈ϕ−1(x)dyv(y).(\operatorname{Tr}_\phi v)(x)=\sum_{y\in\phi^{-1}(x)}d_yv(y).

Proposition 23.5.2. The trace is continuous, positive, and satisfies ∥Tr⁡ϕv∥∞≤d∥v∥∞,Tr⁡ϕ(ϕ∗u)=du.\|\operatorname{Tr}_\phi v\|_\infty\leq d\|v\|_\infty, \qquad \operatorname{Tr}_\phi(\phi^*u)=du. There is consequently a pullback of finite signed Radon measures, uniquely determined by ∫Γ′v d(ϕ∗ν)=∫ΓTr⁡ϕ(v) dν.\int_{\Gamma'}v\,d(\phi^*\nu)=\int_\Gamma\operatorname{Tr}_\phi(v)\,d\nu. It preserves positivity, multiplies total mass by dd, and satisfies ϕ∗ϕ∗ν=dν\phi_*\phi^*\nu=d\nu. In particular ϕ∗δx=∑y↦xdyδy\phi^*\delta_x=\sum_{y\mapsto x}d_y\delta_y.

Proof. Continuity on edge interiors is immediate. At a vertex xx, group the approaching inverse branches by their endpoint yy. Continuity of vv and harmonicity give the limit ∑ydyv(y)\sum_y d_yv(y) from every outgoing direction. The norm estimate and positivity follow from positive local degrees. The representation theorem for continuous linear functionals on C(Γ′)C(\Gamma') defines the measure pullback. Applying its definition to 11 and to u∘ϕu\circ\phi gives the mass and push-pull formulas. ◻

On a single edge of expansion mm, the pullback of a measure with density a(t) dta(t)\,dt has density m2a(ms) dsm^2a(ms)\,ds in a coordinate where t=mst=ms. One factor mm is the local degree and the other is the change of variable. Omitting either factor would break the Laplacian identity below.

Proposition 23.5.3. (Energy identities). For continuous edgewise H1H^1 functions, EΓ′(ϕ∗u,ϕ∗w)=d EΓ(u,w),EΓ(Tr⁡ϕv,u)=EΓ′(v,ϕ∗u),EΓ(Tr⁡ϕv,Tr⁡ϕv)≤d EΓ′(v,v).\begin{align*} \mathcal E_{\Gamma'}(\phi^*u,\phi^*w)&=d\,\mathcal E_\Gamma(u,w),\\ \mathcal E_\Gamma(\operatorname{Tr}_\phi v,u) &=\mathcal E_{\Gamma'}(v,\phi^*u),\\ \mathcal E_\Gamma(\operatorname{Tr}_\phi v,\operatorname{Tr}_\phi v) &\leq d\,\mathcal E_{\Gamma'}(v,v). \end{align*}

Proof. On a source edge, (u∘ϕ)′=me′u′∘ϕ(u\circ\phi)'=m_{e'}u'\circ\phi and dt=me′dsdt=m_{e'}ds. Its contribution to the first energy is therefore me′∫eu′w′ dtm_{e'}\int_eu'w'\,dt. Sum over all inverse edges and use the constant-degree lemma. On the interior of a target edge, (Tr⁡ϕv)′(t)=∑e′↦eve′′(se′(t)).(\operatorname{Tr}_\phi v)'(t)=\sum_{e'\mapsto e}v'_{e'}(s_{e'}(t)). This proves the second identity after change of variable. Weighted Cauchy–Schwarz gives ∣∑e′↦eve′′∣2≤(∑e′↦eme′)(∑e′↦e∣ve′′∣2me′).\left|\sum_{e'\mapsto e}v'_{e'}\right|^2 \leq\left(\sum_{e'\mapsto e}m_{e'}\right) \left(\sum_{e'\mapsto e}\frac{|v'_{e'}|^2}{m_{e'}}\right). The first factor is dd; integrating the second over the target edge recovers the source energies. This proves the inequality and also shows that the continuous trace is edgewise H1H^1. ◻

Laplacians and vertex multiplicities

Theorem 23.5.4. Use the convention Δ=−DD\Delta=-DD of Chapter 20. If vv has a finite measure Laplacian on Γ′\Gamma', then Δ(Tr⁡ϕv)=ϕ∗(Δv).\Delta(\operatorname{Tr}_\phi v)=\phi_*(\Delta v). If uu has a finite measure Laplacian on Γ\Gamma, then Δ(ϕ∗u)=ϕ∗(Δu).\Delta(\phi^*u)=\phi^*(\Delta u). Functions here are continuous and edgewise H1H^1, and the equations hold against all continuous edgewise H1H^1 tests.

Proof. For the trace identity, pair with a test aa on Γ\Gamma and use Proposition 23.5.3: ⟨ΔTr⁡ϕv,a⟩=EΓ′(v,ϕ∗a)=∫Γ′ϕ∗a d(Δv)=∫Γa dϕ∗(Δv).\langle\Delta\operatorname{Tr}_\phi v,a\rangle =\mathcal E_{\Gamma'}(v,\phi^*a) =\int_{\Gamma'}\phi^*a\,d(\Delta v) =\int_\Gamma a\,d\phi_*(\Delta v). For pullback, first let uu be PL. Its Laplacian is concentrated at vertices, with coefficient minus the sum of outgoing slopes. At a source vertex yy above xx, the corresponding sum is ∑e′ at yme′ ∂eu(x)=dy∑e at x∂eu(x),\sum_{e'\text{ at }y}m_{e'}\,\partial_{e}u(x) =d_y\sum_{e\text{ at }x}\partial_eu(x), by harmonicity. This is the asserted weighted pullback of each atom. Subdivision permits the same calculation at any additional break point.

For general uu, let ν=Δu\nu=\Delta u, which has total mass zero. Approximate ν\nu weakly by finite atomic signed measures νj\nu_j of mass zero and uniformly bounded total variation. The graph Green construction gives normalized PL solutions uju_j converging uniformly to u−u(z)u-u(z). In fact they converge in energy: E(uj−u+u(z),uj−u+u(z))≤∥uj−u+u(z)∥∞ ∥νj−ν∥TV⟶0.\mathcal E(u_j-u+u(z),u_j-u+u(z)) \leq\|u_j-u+u(z)\|_\infty\,\|\nu_j-\nu\|_{\rm TV} \longrightarrow0. The pullback energy identity preserves this convergence. The bounded continuous trace map makes ϕ∗νj⇀ϕ∗ν\phi^*\nu_j\rightharpoonup\phi^*\nu. Passing to the PL identity proves the pullback formula on the full energy test class. ◻

Example 23.5.5. (One skeletal point with degree greater than one). The map [0,L]→[0,mL][0,L]\to[0,mL], s↦mss\mapsto ms, has one point in each fibre, but local and global degree mm. Its trace is v↦(t↦mv(t/m))v\mapsto(t\mapsto m v(t/m)), and the pullback of an atom is mδt/mm\delta_{t/m}. This is the skeletal normalization of the monomial map z↦zmz\mapsto z^m between annuli. The algebraic degree is retained by the expansion factor even though the skeletal map is a homeomorphism.

Green-kernel normalization under a finite map

Let μ\mu have mass one on Γ\Gamma, and set μ′=d−1ϕ∗μ\mu'=d^{-1}\phi^*\mu. Both measures have mass one, so the normalized kernels of Theorem 20.15.5 are defined. For x′∈Γ′x'\in\Gamma' and y∈Γy\in\Gamma, gμ(ϕ(x′),y)=∑y′∈ϕ−1(y)dy′gμ′(x′,y′).\begin{equation} g_\mu(\phi(x'),y) =\sum_{y'\in\phi^{-1}(y)}d_{y'}g_{\mu'}(x',y'). \end{equation}

Proof. Both sides, as functions of x′x', have Laplacian ϕ∗δy−ϕ∗μ\phi^*\delta_y-\phi^*\mu. Their difference is constant. The right side has zero μ′\mu'-mean by normalization. For the left side, the measure pullback formula and Tr⁡ϕ(ϕ∗a)=da\operatorname{Tr}_\phi(\phi^*a)=da give ∫Γ′gμ(ϕ(x′),y) dμ′(x′)=∫Γgμ(x,y) dμ(x)=0.\int_{\Gamma'}g_\mu(\phi(x'),y)\,d\mu'(x') =\int_\Gamma g_\mu(x,y)\,d\mu(x)=0. Thus the constant vanishes. ◻

A single lifted point does not generally give the pullback kernel; the weighted sum over its fibre does. These formulas are complete statements in the finite harmonic-graph category. To apply them to a morphism of analytic curves, one must supply compatible skeleta and the theorem identifying local analytic degrees with the graph expansion factors and residual degrees. The existence of such skeleta is a geometric input, not a consequence of the energy identities.

References. The proofs combine the graph Laplacian and Green kernel of [BRg] with the harmonicity condition expressing local degree. They provide a one-dimensional verification of the normalizations used in the analytic projection formulas. The skeletal and slope geometry is treated in [GRW16]; no existence theorem for compatible skeleta is proved by these elementary graph calculations.

23.6 Boundary completions and arithmetic intersection

Local currents and global arithmetic intersection numbers are related through metrics, but have different domains of definition. Cai–Gubler’s abstract divisorial spaces make the completion process itself part of the theory. For an open variety UU and a boundary divisor BB on a compactification, a sequence of metrized divisors is Cauchy when its differences, after passage to common models, are bounded between −ϵB-\epsilon B and ϵB\epsilon B, together with the corresponding inequalities for boundary Green functions. Two consequences matter for applications. First, this topology controls growth at the boundary; uniform convergence on each compact subset of UanU^{\mathrm{an}} is not the defining condition. Second, the completion retains an underlying compactified geometric line bundle, not merely its restriction to UU. Different boundary data can restrict to the same line bundle on UU.

Example 23.6.1. (The open line does not determine its boundary extension). On U=A1⊂P1U=\mathbf A^1\subset\mathbf P^1, every OP1(a∞)\mathcal O_{\mathbf P^1}(a\infty) restricts to the trivial line bundle, but these extensions have degrees aa. Thus the restriction to UU forgets numerical data at infinity. Any comparison theorem requiring the same compactified object has strictly more input than an isomorphism of the restrictions.

Over a proper adelic curve in their sense, [CG26] (Theorem 7.22) extends arithmetic intersection to arithmetically integrable compactified metrized divisors. The pairing is real-valued, symmetric and multilinear, agrees with the proper theory, and has the specified projection, base-change and boundary-continuity properties. The positivity and integrability classes are part of its domain. A positive local CLD current alone is not a membership test for that domain. The relative-energy extension in Theorem 11.3 has a different range: it can take the value −∞-\infty. It requires relatively nef metrics on compactified geometric objects admitting arithmetically nef reference metrics. The reference and the new metric must have the same underlying compactified geometric object, a stronger requirement than having the same line bundle on UU. Mixed relative energy measures their change of arithmetic intersection. In the finite-energy part, this has the form I(h)−I(g)=E(g,h).I(\mathbf h)-I(\mathbf g)=E(\mathbf g,\mathbf h). Writing this as a difference presupposes that the displayed values are finite. The source formulates the extended-valued statement with its admissible energy class, so that no subtraction of two infinities is required. In a proof using the formula, finiteness must be checked rather than inferred from the notation.

Lemma 23.6.2. (Independence of a finite-energy reference). Suppose an energy EE on a class of metrics satisfies the cocycle identity E(a,c)=E(a,b)+E(b,c)E(a,c)=E(a,b)+E(b,c), all terms below are finite, and the known intersection values satisfy I(b)−I(a)=E(a,b)I(b)-I(a)=E(a,b). Then the prescription I(c):=I(a)+E(a,c)I(c):=I(a)+E(a,c) is independent of the reference aa among those known metrics.

Proof. Substitute the two assumed identities to get I(a)+E(a,c)=I(a)+E(a,b)+E(b,c)=I(b)+E(b,c)I(a)+E(a,c)=I(a)+E(a,b)+E(b,c)=I(b)+E(b,c). ◻

The lemma explains the algebra of the extension. It does not prove the analytic construction, finiteness or cocycle property of relative energy; those are substantial inputs from [CG26]. It also separates this global completion theory from the local degree normalization under scalar extension discussed earlier in this chapter.

References. [CG26] (Sections 2, 7, 10 and Theorem 11.3).

23.7 Finite changes of the ground field and local values

The two local calculations below use the original absolute values. The ordinary top-integral calculation uses NAG-002503, NAG-002511, NAG-002514 and NAG-002518 in the scalar-extension source , which accompanies this reconstruction as a separate article. Equality of archimedean local values remains open here.

Proposition 23.7.1. (Conditional place-weight summation). Let F⊂EF\subset E be number fields. Fix unsquared absolute values extending those on Qp\mathbb Q_p and the usual real absolute value, and put cv(F)=[Fv:Qv]/[F:Q]c_v(F)=[F_v:\mathbb Q_v]/[F:\mathbb Q]. Suppose the completion-degree identity ∑w∣v[Ew:Fv]=[E:F]\sum_{w\mid v}[E_w:F_v]=[E:F] is supplied. Suppose real numbers ava_v and awa_w satisfy aw=ava_w=a_v for w∣vw\mid v, and their weighted families are finitely supported or absolutely summable. Then ∑wcw(E)aw=∑vcv(F)av.\sum_w c_w(E)a_w=\sum_v c_v(F)a_v.

Proof. We first establish the completion-degree identity under the stated number-field hypotheses.

Finite places. Let FF be a field with a specified real-valued nonarchimedean absolute value vv. Let H=F^vH=\widehat F_v, let E/FE/F be a finite extension of degree nn, and let w∣vw\mid v mean an absolute value on EE restricting to this very absolute value on FF. Write EwE_w for the resulting completion. Set C=H⊗FEC=H\otimes_F E. There is a canonical indexing of the Artinian local factors CwC_w of CC by w∣vw\mid v, with canonical valued residue field EwE_w. In particular n=∑w∣vℓCw(Cw)[Ew:H].n=\sum_{w\mid v}\ell_{C_w}(C_w)[E_w:H]. For separable E/FE/F all lengths are one.

A basis of E/FE/F gives dim⁡HC=n\dim_H C=n. Its ideals are finite-dimensional HH-subspaces, so descending chains stabilize. By [StacksAlg] (Tag 00JB, Lemma 10.53.6) it is the finite product of its localizations at its maximal ideals. For a maximal ideal PP let D=C/PD=C/P, a finite field extension of HH. NAG-001450 supplies the unique extension of the absolute value of HH, completeness of DD, and equivalence with a finite-dimensional coordinate norm, without separability. The map E→DE\to D, a↦1⊗a mod Pa\mapsto 1\otimes a\bmod P, is a unital field embedding, so defines an extension wPw_P of vv. The image of EE is dense in DD: the images of an FF-basis eie_i of EE span DD over HH, and each ∑ihiei\sum_i h_i e_i is a limit of ∑ifiei\sum_i f_i e_i with fi∈Ff_i\in F. Indeed the error is bounded by max⁡i∣hi−fi∣ ∣ei∣wP\max_i |h_i-f_i|\,|e_i|_{w_P}. Thus DD identifies isometrically with EwPE_{w_P}, fixing EE.

Conversely, for any extending ww, the inclusion of FF into EwE_w extends isometrically to HH. Multiplication gives a map C→EwC\to E_w. Its image is a finite-dimensional HH-domain, hence a field: multiplication by each nonzero element is an injective linear endomorphism and therefore surjective. With its induced norm this field is complete by NAG-001450, hence closed in EwE_w. It contains the dense copy of EE and consequently is all of EwE_w. Its kernel is maximal. Starting from PP, the isometry from the previous paragraph agrees with both structure maps on EE and on HH (on HH by density of FF); thus the resulting kernel is exactly PP. Starting from ww, the norm on the residue field is the induced norm by uniqueness, so this construction recovers ww. This proves the asserted bijection without counting embeddings. The algebra here is formed over the original, possibly incomplete, FF; it is finite-dimensional over HH and already complete for a coordinate norm.

The simple modules in a composition series of CwC_w are its residue field. Each therefore contributes exactly [Ew:H][E_w:H] to the HH-dimension. Adding the contributions of the full local factors gives the formula. To check separability, take a finite tower of simple generators for E/FE/F. At each step the minimal polynomial of the new generator over the intermediate field divides a separable polynomial over FF. It is coprime to its derivative. Starting with HH, if the tensor algebra for the intermediate extension is a product of fields, tensoring with the next simple extension quotients a polynomial ring over each factor by the image of that minimal polynomial. The Bezout identity with its derivative survives in the field. The distinct monic irreducible factors are pairwise coprime, and the Chinese remainder map gives a product of fields again. Induction proves the length-one assertion. In characteristic zero every irreducible nonconstant polynomial is separable, since its derivative is nonzero and of smaller degree.

At a finite place vv of a number field FF over pp, use the absolute value extending the specified one on Qp\mathbb Q_p. The same result for F/QF/\mathbb Q shows that Fv/QpF_v/\mathbb Q_p is finite. For w∣vw\mid v in a finite extension E/FE/F the identification above uses this Fv⊂EwF_v\subset E_w isometrically. Since the number fields have characteristic zero, the length-one case gives ∑w∣v[Ew:Fv]=[E:F]\sum_{w\mid v}[E_w:F_v]=[E:F] at each finite place.

Archimedean places. Let F⊂EF\subset E be number fields, n=[E:F]n=[E:F], and vv an absolute value on FF extending the usual absolute value on Q\mathbb Q. Count each absolute value ww on EE extending this fixed vv once. With the usual unsquared moduli on all archimedean completions, one has Fv⊗FE≃∏w∣vEw,∑w∣v[Ew:Fv]=n.F_v\otimes_F E \simeq \prod_{w\mid v}E_w,\qquad \sum_{w\mid v}[E_w:F_v]=n. If Fv=RF_v=\mathbb R and the factors comprise rr real and ss complex completions, then r+2s=nr+2s=n. If Fv=CF_v=\mathbb C, there are nn complex factors, each of degree one over FvF_v.

Elementary field input. For a nonconstant complex polynomial PP, the leading term implies ∣P(z)∣→∞|P(z)|\to\infty as ∣z∣→∞|z|\to\infty. A sequence approaching its infimum can therefore be taken in a bounded square. Successive subdivision into four closed subsquares, each time retaining one containing infinitely many sequence terms, gives a convergent subsequence by completeness of the real coordinates. Continuity gives a minimizer z0z_0. If P(z0)≠0P(z_0)\ne0, write P(z0+h)/P(z0)=1+chk+O(∣h∣k+1)P(z_0+h)/P(z_0)=1+c h^k+O(|h|^{k+1}) with k≥1k\ge1 and c≠0c\ne0. Choose δ\delta among 1,−1,i,−i1,-1,i,-i with Re⁡(cδ)<0\operatorname{Re}(c\delta)<0. It has a unit-modulus kk-th root uu: writing k=2abk=2^a b with bb odd, the bb-th power permutes those four units, and successive square roots of a unit vv are available since (1+v)/∣1+v∣(1+v)/|1+v| squares to vv for v≠−1v\ne-1, and i2=−1i^2=-1. For small positive tt, ∣P(z0+tu)/P(z0)∣2=1+2Re⁡(cδ)tk+O(tk+1)<1,|P(z_0+tu)/P(z_0)|^2 =1+2\operatorname{Re}(c\delta)t^k+O(t^{k+1})<1, a contradiction. Division by a linear factor and induction show that every complex polynomial splits. Conjugation of roots then shows that each irreducible real polynomial is linear or quadratic with a pair of nonreal conjugate roots. A quotient by such a quadratic is C\mathbb C over R\mathbb R, by evaluation at either root: a nonreal root and 1 already span C\mathbb C over R\mathbb R.

Completion factors. First take any number field DD, with base Q\mathbb Q and H=RH=\mathbb R. More generally let D/BD/B be a finite extension of number fields and fix a completion HH of BB already identified isometrically with R\mathbb R or C\mathbb C. Write A=H⊗BDA=H\otimes_B D. Adjoin a finite generating list for D/BD/B one element at a time. At each step the minimal polynomial over the preceding field is separable: in characteristic zero it is coprime to its derivative. Over each field factor reached so far the Bezout identity with its derivative persists. Factoring the resulting polynomial and applying the Chinese remainder identity for coprime factors gives a product of fields. By the preceding paragraph these fields are real or complex over a real factor, and complex over a complex factor. Starting with HH yields a product decomposition of AA; its HH-dimension is [D:B][D:B] by a basis. In particular all local lengths here are one.

Each factor carries the usual modulus, independent of the choice between conjugate identifications with C\mathbb C. In fact any multiplicative absolute value ν\nu on C\mathbb C extending the usual absolute value on R\mathbb R satisfies ν(i)=1\nu(i)=1, hence ν(z)≤∣Re⁡z∣+∣Im⁡z∣≤2∣z∣\nu(z)\le|\operatorname{Re}z|+|\operatorname{Im}z|\le2|z|. Apply this inequality to zjz^j and let j→∞j\to\infty; applying it also to z−1z^{-1} gives ν(z)=∣z∣\nu(z)=|z|. Thus each factor restricts to the specified absolute value on HH.

The image of DD in a factor spans it over HH; it is dense because the finitely many coefficients of a linear combination can be approximated in BB. The factor is complete, so is the completion of DD for the induced ww. Conversely, for any extending ww, the inclusion of BB in DwD_w extends isometrically to HH. Multiplication defines A→DwA\to D_w. Exactly one primitive factor idempotent maps to 1 in this field, so the image is the corresponding factor. Its norm is its usual one by the preceding uniqueness argument (immediate for a real factor); it is therefore complete and closed. It contains dense DD, hence equals DwD_w. These assignments are inverse: the completion isometry fixing DD fixes the prescribed image of HH by density of BB, hence gives the same kernel in AA. This identifies factors with places themselves, including exactly one factor per complex-conjugate pair over a real place.

Apply the argument first to F/QF/\mathbb Q and R\mathbb R to obtain the claimed alternatives for FvF_v. Apply it again to E/FE/F with H=FvH=F_v. Taking FvF_v-dimensions gives precisely the two counts above.

These two calculations give the asserted degree identity for every place in the proposition. At infinity all moduli are unsquared.

The tower formula gives cw(E)=cv(F)[Ew:Fv]/[E:F]c_w(E)=c_v(F)[E_w:F_v]/[E:F]. Sum first over the finite set of ww above each vv and use the completion-degree identity above. Absolute summability permits regrouping. At the archimedean base place Qv=R\mathbb Q_v=\mathbb R; a complex completion has degree two and its absolute value remains unsquared. ◻

For the non-Archimedean smooth top integrals in the native scalar-extension theorem NAG-002503, the local equality is aw=ava_w=a_v. The extension degree is already in the place weight and must not be inserted a second time into this native local value. The completion-degree identity holds at all places by the preceding proof. Equality of the actual archimedean local height values, the product formula, and summability remain separate inputs.

Finite-place matching of the local value in NAG-002523

The assertion to be applied is at lines 629–632 of the original research_v040/f9_scalar_extension_candidate.tex. Lines 613–614 take equality of the local values and summability as separate hypotheses. Fix a finite extension of number fields E/FE/F, a finite place vv and a place ww above it. Put K=FvK=F_v and L=EwL=E_w, with the isometric inclusion for the original absolute values.

The local assertion.

Let XX be good Hausdorff and pure of analytic dimension dd over KK, let F\mathcal F be coherent on XGX_G, and let α\alpha be a compactly supported ordinary smooth real (d,d)(d,d)-form on XX. Take the entire scalar extension p:XL→Xp:X_L\to X. Use on XLX_L the pulled-back coefficient and the ordinary form presented on inverse images of the presenting domains by the pulled-back units and the same real differential form. These prescriptions define the upstairs data independently of the presentations. Their native local value satisfies aw=∫XL/L,p∗Fp∗α=∫X/K,Fα=av.a_w=\int_{X_L/L,p^*\mathcal F}p^*\alpha =\int_{X/K,\mathcal F}\alpha=a_v. The same conclusion applies to independently specified upstairs data when an LL-isomorphism identifies the space, coefficient and form with this triple.

Checking the object for data from a scheme.

For example, let ZZ be of finite type over FF with coherent module M\mathcal M, and assume X=(ZK)anX=(Z_K)^{\rm an} has the stated analytic hypotheses. The scheme obtained by extending ZEZ_E to EwE_w is canonically ZK×KLZ_K\times_K L, by associativity on affine rings and compatibility with localization. By NAG-001709 its analytification is XLX_L, compatibly with the projection. On an affine Spec(R)\mathrm{Spec}(R) of ZKZ_K let MM represent its coefficient, and let AA be an analytic radius chart. For AL=A⊗^KLA_L=A\widehat\otimes_K L the two coefficient constructions give the same ALA_L-module: AL⊗A(A⊗RM)≃AL⊗RM≃AL⊗R⊗KL(M⊗KL).A_L\otimes_A(A\otimes_R M) \simeq A_L\otimes_R M \simeq A_L\otimes_{R\otimes_K L}(M\otimes_K L). Here NAG-001719 gives the analytic chart description and NAG-001686 gives the coherent pullback. The isomorphisms commute with restriction. For the ordinary test the criterion is equality of its germ on every inverse-image chart with the pulled-back presentation. Scalar extension of the scheme and coefficient alone imposes no such condition on a separately chosen test.

Proof of the integral identity.

NAG-002503, lines 354–384 of the scalar source, constructs the two pullbacks and proves independence on overlaps. The support of the pulled-back test lies in the inverse image of the original compact support. Cover the latter support by the interiors of finitely many affinoids downstairs. Their full inverse images are compact affinoids, and the inverse image of the support is closed in their union, hence compact. The local weight calculation in NAG-002514 restricts to the entire simultaneous norm level of the units at an interior top-cell parameter. It uses the same transverse integral projection and lattice index on both sides. Write GK,GLG_K,G_L for the two complete transverse Gauss fields and QQ for the finite coefficient module on the downstairs transverse Gauss fibre. NAG-002511 identifies the upstairs module with Q⊗GKGLQ\otimes_{G_K}G_L. Each local Artin factor contributes its module length times its full residue-field degree. Summing over all factors computes the vector-space dimension in the respective ground field, so the numerator of the cell weight is unchanged. NAG-002514 keeps the orientation and lattice measure. The compact chart inclusion–exclusion in NAG-002503, lines 386–410, now gives the claimed equality for the signed test.

Fix x∈Xx\in X and an affinoid neighbourhood U=MK(A)U=\mathcal M_K(A) of xx. Put H=H(x)H=\mathcal H(x). By NAG-001616 the entire inverse image of UU has algebra AL=A⊗^KLA_L=A\widehat\otimes_K L. Since L/KL/K is finite isometric and XLX_L is good Hausdorff, NAG-002518 (scalar source, lines 421–469) gives (XL)[K](X_L)_{[K]} with the same points, domains, structure sheaf and completed residue fields. The bounded KK-maps A→ALA\to A_L present a morphism (XL)[K]→X(X_L)_{[K]}\to X over KK. Apply NAG-001617 to this morphism on the entire inverse image of UU. The whole fibre over xx has HH-affinoid algebra AL⊗^AHA_L\widehat\otimes_A H; its points and their completed residue fields are those over xx in XLX_L.

Apply NAG-001689 to K→LK\to L and K→A→HK\to A\to H. Here LL is finite KK-affinoid by NAG-002518, evaluation at xx is bounded, and HH is affinoid over its own complete coefficient field. Tensor symmetry and the bounded iterated pushout give an HH-algebra isomorphism with bounded inverse (A⊗^KL)⊗^AH≃H⊗^KL,(a⊗b)⊗h⟼a(x)h⊗b.(A\widehat\otimes_K L)\widehat\otimes_A H \simeq H\widehat\otimes_K L,\qquad (a\otimes b)\otimes h\longmapsto a(x)h\otimes b . Its inverse sends h⊗bh\otimes b to (1⊗b)⊗h(1\otimes b)\otimes h. A finite KK-basis of LL has bounded coordinate functionals by NAG-001450. Tensoring those functionals and the basis map with HH gives inverse bounded coordinate maps between H⊗KLH\otimes_K L with its projective tensor norm and H[L:K]H^{[L:K]} with its maximum norm. Thus H⊗KLH\otimes_K L is already complete. This finite HH-algebra splits as ∏y∣xBy\prod_{y\mid x} B_y by Stacks Tag 00JB: each local residue field is finite over HH, has its unique complete extending norm by NAG-001450, and gives exactly one analytic point of the factor. Nilpotents vanish under a seminorm. A composition series of ByB_y has successive quotients H(y)\mathcal H(y), whence ∑y∣xlengthBy(By)[H(y):H]=dim⁡H(H⊗KL)=[L:K].\sum_{y\mid x}\mathrm{length}_{B_y}(B_y)[\mathcal H(y):H] =\dim_H(H\otimes_K L)=[L:K]. This is the rank of the scalar fibre. The integral uses the native transverse dimension computed above. These local arguments apply to finite isometric extensions of complete nonarchimedean fields, including inseparable extensions, with the full residue degree and possible nilpotent lengths. For a coherent generic cycle its separate multiplicities are those of NAG-002506 and its finite pushforward has the degree of NAG-002509. Restriction of the constants of integration on the same upstairs space has the factor [L:K][L:K] by NAG-002520.

Remaining premise for arithmetic descent.

The finite-place result discharges the local-value premise for exactly the primitive integral displayed above. For a separately defined local height one still needs a native identification with that primitive on both realizations, or a base-change proof for its own defining operation. At infinity even the real-to-complex case requires a specified primitive and a proof comparing its two normalizations. The reviewed place-weight identity enters only after the local values are matched. The global regrouping in NAG-002523 requires finite support or absolute summability of the weighted families.

Inherited proof dependency.

The integral application uses the F6 full-level argument through NAG-002503, NAG-002514, NAG-002405, NAG-002370 and NAG-002365 in that dependency order. The independent F6 receipt passes the specified NAG-002365 repair subject to its native prerequisites. Full support for this application requires the consuming baseline to include the repaired proof and to discharge its prerequisites. In the consuming English baseline, Theorem 5.41.1 contains that scoped repair.

The finite zero-dimensional calculation

For this subsection, let KK be any field complete for a specified real-valued nonarchimedean absolute value, which may be trivial. Let C≠0C\ne0 be a finite commutative KK-algebra, MM a finite CC-module, and PP an invertible CC-module. Write C=∏pCpC=\prod_p C_p for its Artin local factors, Ep=Cp/mpE_p=C_p/\mathfrak m_p, and wp(M)=length⁡Cp(Mp)[Ep:K]=dim⁡KMp.w_p(M)=\operatorname{length}_{C_p}(M_p)[E_p:K]=\dim_K M_p. The equality follows by a composition series. The degrees are full field degrees. A metric on PP means, at each of these finitely many analytic points, a norm on the one-dimensional EpE_p-space P⊗CEpP\otimes_C E_p compatible with the unique extending absolute value.

Proposition 23.7.2. NAG-002635 (Finite heights, restriction of scalars, and norms). For a generator ss of PP, define hC,M(P,h,s)=−∑pwp(M)log⁡∥s(p)∥h.\begin{equation} h_{C,M}(P,h,s)=-\sum_p w_p(M)\log\|s(p)\|_h . \end{equation} If C→DC\to D is a map of finite commutative KK-algebras, NN is a finite DD-module, and the metric and section are pulled back, then hD,N(D⊗CP,q∗h,1⊗s)=hC,Res⁡CN(P,h,s).h_{D,N}(D\otimes_C P,q^*h,1\otimes s) =h_{C,\operatorname{Res}_C N}(P,h,s). If DD is locally free of constant rank nn over CC and N=D⊗CMN=D\otimes_C M, the right side is nhC,M(P,h,s)n h_{C,M}(P,h,s).

The KK-line NM(P)=det⁡K(P⊗CM)⊗det⁡K(M)∨{\mathcal N}_M(P) =\det_K(P\otimes_C M)\otimes\det_K(M)^\vee has a metric characterized, for every generator ee of PP, by ∥NM(e)∥=∏p∥e(p)∥wp(M).\|\mathcal N_M(e)\| =\prod_p\|e(p)\|^{w_p(M)} . It is well-defined, and hC,M(P,h,s)=−log⁡∥NM(s)∥h_{C,M}(P,h,s)=-\log\|\mathcal N_M(s)\|. For M=CM=C it is the determinant norm line Nm⁡C/K(P)\operatorname{Nm}_{C/K}(P). For h′=e−uhh'=e^{-u}h, hC,M(P,h′,s)−hC,M(P,h,s)=∑pwp(M)u(p),∣Δh∣≤∥u∥∞dim⁡KM.h_{C,M}(P,h',s)-h_{C,M}(P,h,s)=\sum_p w_p(M)u(p),\qquad |\Delta h|\leq\|u\|_\infty\dim_K M . Moreover (23.7.1) equals the frozen KK-analytic integral of the smooth function −log⁡∥s∥-\log\|s\| on M(C)\mathcal M(C) with coefficient module MM.

Proof inputs. NAG-001450 (finite extending absolute values); the already accepted Artin decomposition leaf STACKS-00JB; NAG-002362, NAG-002405, NAG-002424, NAG-002425 (zero-dimensional module weights, chart integration and ordinary germs); elementary finite-dimensional determinants and composition series..

Proof. A multiplicative seminorm on CC kills its nilpotents. Its kernel selects one local factor and its maximal ideal, and the induced norm on the finite field EpE_p is unique by NAG-001450. These are precisely the points of M(C)\mathcal M(C) with its canonical finite topology. The idempotent factors are open and closed. On each factor a function on this finite point set is constant, hence is an ordinary smooth moment function.

Every invertible CC-module has a generator: it is free on each local factor, and the finitely many local generators combine using the idempotents. If a∈C×a\in C^\times, a composition series of MpM_p is stable under multiplication by aa. Each simple quotient is EpE_p with multiplication by its residue a(p)a(p). A basis respecting the series gives block triangular matrices and hence ∣det⁡K(ma∣M)∣=∏p∣Nm⁡Ep/K(a(p))∣length⁡CpMp=∏p∣a(p)∣wp(M).\begin{equation} |\det_K(m_a\mid M)| =\prod_p|\operatorname{Nm}_{E_p/K}(a(p))|^{\operatorname{length}_{C_p}M_p} =\prod_p|a(p)|^{w_p(M)}. \end{equation} For completeness, the last equality does not count separable embeddings. If b∈Ep×b\in E_p^\times has minimal polynomial ff, the characteristic polynomial of multiplication by bb on EpE_p is f[Ep:K(b)]f^{[E_p:K(b)]}, by a K(b)K(b)-basis. A finite splitting extension for ff is obtained by adjoining roots successively. Each root has the same absolute value as bb: the corresponding embedding of K(b)K(b) into that finite extension pulls back an extending absolute value, unique by NAG-001450. Taking the product of all roots with their multiplicities gives ∣Nm⁡Ep/Kb∣=∣b∣[Ep:K]|\operatorname{Nm}_{E_p/K}b|=|b|^{[E_p:K]}, including inseparability.

Replacing a generator ee by aeae multiplies NM(e)\mathcal N_M(e) by det⁡K(ma∣M)\det_K(m_a\mid M) and multiplies the proposed norm by exactly its absolute value, by (23.7.2). Thus the prescription defines a norm on the KK-line and gives the asserted identity for ss. If M=0M=0, both determinant lines have their canonical unit and all sums are zero; the same argument applies.

For restriction of scalars, the pp-part of NN is the direct sum of the DD-local parts lying over pp. Its KK-dimension is the sum of theirs. Uniqueness of extending absolute values makes the pulled-back metric at each such point have exactly the norm of s(p)s(p). Grouping the finite sum proves the projection formula. A locally free rank-nn DD over the Artin semilocal CC is free of that rank, by choosing bases on each factor. Hence D⊗CM≃MnD\otimes_C M\simeq M^n as CC-modules, giving the rank formula. Subtracting the two log norms proves the variation and its bound.

Finally, compute the frozen integral separately on the clopen factors. For the empty moment tuple on CpC_p the rank-zero tangent lattice has index one, the Gauss field is KK, and the full finite algebra is CpC_p. The module weight of NAG-002362 is therefore dim⁡KMp\dim_K M_p. The global-chain construction NAG-002405 supplies this chart and its interior whole-fibre weight; in dimension zero its boundary image is empty. Rank-zero polyhedral integration evaluates the constant at the single point. NAG-002425 identifies that chart sum with the ordinary compact integral. This gives exactly (23.7.1). ◻

Intersection meaning and its boundary. For a specified zero-dimensional intersection represented by a finite KK-algebra CC, its coherent coefficient MM contributes the actual length length⁡CpMp\operatorname{length}_{C_p}M_p at pp, followed by the ordinary residue degree [Ep:K][E_p:K]. In particular C=L[ϵ]/(ϵr)C=L[\epsilon]/(\epsilon^r) has total weight r[L:K]r[L:K]. The determinant on the scalar unit a∈L×a\in L^\times has absolute value ∣a∣r[L:K]|a|^{r[L:K]}. Neither the reduction LL nor a separable point count gives this normalization. This identifies the finite base case; it does not construct a positive-dimensional cycle product.

Finite scalar extension of the Artin local height

The zero-dimensional starting value required at lines 674–691 of the F9 descent supplement is defined in NAG-002635, F9_Local_Height_Interface_v038.tex, lines 197–214. The projection at lines 215–222 uses a single ground field. The following application compares the same primitive in its two native ground fields.

Let KK be complete nonarchimedean, let L/KL/K be finite isometric and let C≠0C\ne0 be a finite commutative KK-algebra with its canonical finite topology. Take a finite CC-module MM, an invertible CC-module PP, a generator ss of PP, and a metric hh as in NAG-002635. Thus at each local factor CpC_p the metric is a norm on the line over kp=Cp/mpk_p=C_p/\mathfrak m_p for its extending absolute value. Write MpM_p for the corresponding summand and s(p)s(p) for the residue of ss. The primitive is hK=−∑pℓCp(Mp)[kp:K]log⁡∥s(p)∥h.h_K=-\sum_p \ell_{C_p}(M_p)[k_p:K]\log\|s(p)\|_h . Form D=C⊗KLD=C\otimes_K L, N=M⊗KLN=M\otimes_K L, and the line and generator PL,sLP_L,s_L obtained by tensoring. All finite tensors carry their canonical complete topologies. Give PLP_L the pullback metric. Then its height hLh_L, defined by the same formula over LL, equals hKh_K.

Proof.

Each idempotent of the factorization C=∏pCpC=\prod_p C_p acts on DD. Decomposing its image further into local Artin algebras gives Cp⊗KL=∏q∣pDqC_p\otimes_K L=\prod_{q\mid p}D_q (Stacks Tag 00JB). Write kq=Dq/mqk_q=D_q/\mathfrak m_q. These are all the analytic lifts of pp: a bounded multiplicative seminorm kills nilpotents and factors through one kqk_q, and each kqk_q gives a point by the unique complete extending absolute value (NAG-001450). Its restriction to kpk_p is the original one by the same uniqueness over KK. More precisely the analytic point fibre over pp has algebra kp⊗KL=(Cp⊗KL)/mp(Cp⊗KL)k_p\otimes_K L=(C_p\otimes_K L)/\mathfrak m_p(C_p\otimes_K L). The quotient map has nilpotent kernel, so the quotient indexes the same lifts. The coefficient at a lift is computed in DqD_q on NqN_q, retaining the original module and the full base-changed algebra.

For a local frame ee of the kpk_p-line the transported norm is ∥aeL∥=∣a∣kq∥e∥h\|a e_L\|=|a|_{k_q}\|e\|_h for a∈kqa\in k_q. Isometry on kpk_p makes this independent of ee; in particular ∥sL(q)∥=∥s(p)∥\|s_L(q)\|=\|s(p)\|. The generator has positive norm at every point. A composition series of NqN_q has quotients kqk_q. The decomposition by idempotents and extension of a KK-basis of MpM_p therefore give, for each pp, ∑q∣pℓDq(Nq)[kq:L]=∑q∣pdim⁡LNq=dim⁡L(Mp⊗KL)=dim⁡KMp=ℓCp(Mp)[kp:K].\sum_{q\mid p}\ell_{D_q}(N_q)[k_q:L] =\sum_{q\mid p}\dim_L N_q =\dim_L(M_p\otimes_K L) =\dim_K M_p =\ell_{C_p}(M_p)[k_p:K]. Multiplying by −log⁡∥s(p)∥-\log\|s(p)\| and summing proves the claim. This argument includes zero coefficient summands and inseparable extensions. When the same upstairs data are measured over KK, the weight at qq uses [kq:K]=[kq:L][L:K][k_q:K]=[k_q:L][L:K]. Thus the restricted value is [L:K]hL[L:K]h_L.

Identification of the integral.

This integral interpretation uses NAG-002635 through NAG-002405 and NAG-002370 to the same scoped NAG-002365 repair in Theorem 5.41.1. The consuming English baseline contains that repair. Its remaining proof prerequisites still require verification. The equality of finite Artin sums above uses the finite field norms and Artin lengths directly.

NAG-002635, lines 242–258 and 299–308, identifies this specific height with an ordinary compact (0,0)(0,0)-integral over either field. The function −log⁡∥s∥-\log\|s\| is constant on each clopen local factor. Its empty moment tuple has lattice index one and coefficient dim⁡KMp\dim_K M_p downstairs, respectively dim⁡LNq\dim_L N_q upstairs. The pullback test on every lift has the same value proved above. Consequently a finite place vv and w∣vw\mid v in NAG-002523 can take av=hFva_v=h_{F_v} and aw=hEwa_w=h_{E_w} for these actual data.

Boundary of the application.

NAG-002523, lines 606–634, assumes equality of the real local numbers and leaves archimedean comparison as a separate input. NAG-002635 assumes a nonarchimedean base at lines 57–63. The recovered passages supply neither an archimedean height primitive on the real and complex realizations nor a convention for its comparison over a real place becoming complex. The all-place degree identity fixes the place weights only. For a higher-dimensional DVR model number, NAG-003211, lines 428–457 of its source, requires a proved downstairs smooth realization. The Artin formula supplies the dimension-zero case only. Finite support or absolute summability of the weighted family is a separate global hypothesis.