Closed

Gauss branches and the finite maximal-rank set

Lemma. (Gauss branches and the finite maximal-rank set). For any analytic map ψ:X→Td\psi:X\to T^d, the underlying set of ψ−1(ηs)\psi^{-1}(\eta_s) is finite. If xx belongs to this fibre and x∈Int⁡(X/K)x\in\operatorname{Int}(X/K), there are arbitrarily small ordinary affinoid neighbourhoods Y=M(B)Y=\mathcal M(B) of xx and N=M(C)N=\mathcal M(C) of ηs\eta_s such that ψ:Y→N\psi:Y\to N is norm-admissible finite and ψ−1(ηs)∩Y={x}\psi^{-1}(\eta_s)\cap Y=\{x\}. One can take CC integral. The ordinary stalk κs:=OTd,ηs\kappa_s:=\mathcal O_{T^d,\eta_s} is a field, and, with E=B⊗^CHs=B⊗CHs,P=M⊗AE,E=B\widehat\otimes_C H_s=B\otimes_C H_s,\qquad P=M\otimes_A E, EE is the actual local Artinian fibre algebra at xx, with residue field H(x)\mathcal H(x). There are natural identifications E=OX,x⊗κsHs,P=Fx⊗κsHs,E=\mathcal O_{X,x}\otimes_{\kappa_s}H_s,\qquad P=\mathcal F_x\otimes_{\kappa_s}H_s, and the finite integer δψ(M,x):=dim⁡HsP=dim⁡κsFx=length⁡E(P)[H(x):Hs]\begin{equation} \delta_\psi(M,x):=\dim_{H_s}P =\dim_{\kappa_s}\mathcal F_x =\operatorname{length}_E(P)[\mathcal H(x):H_s] \end{equation} is independent of the isolating neighbourhoods. After shrinking NN ordinarily and taking the inverse image in YY, the direct images of OY\mathcal O_Y and F∣Y\mathcal F|Y are free, of ranks δψ(A,x)>0\delta_\psi(A,x)>0 and δψ(M,x)\delta_\psi(M,x) respectively. Moreover, writing βp(M)=length⁡ApMp\beta_{\mathfrak p}(M)=\operatorname{length}_{A_{\mathfrak p}} M_{\mathfrak p}, one has δψ(M,x)=∑p∈Min⁡(A)βp(M)dim⁡Hs(E/pE).\begin{equation} \delta_\psi(M,x)= \sum_{\mathfrak p\in\operatorname{Min}(A)} \beta_{\mathfrak p}(M)\dim_{H_s}(E/\mathfrak pE). \end{equation} For every unit tuple ff and every ww, the set SwS_w is finite, without an interior assumption on its points.

Proof

Proof. Choose a compact polyannulus containing ψ(X)\psi(X). The actual fibre over ηs\eta_s is affinoid over HsH_s; 2.33.11 identifies its points and completed residue fields with the points over ηs\eta_s and their original completed residue fields. The Gauss dimension, tower formula, and Abhyankar bound give 0≤dHs(y)=dK(y)−dK(ηs)≤d−d=00\leq d_{H_s}(y)=d_K(y)-d_K(\eta_s)\leq d-d=0 for every such point. Here the dependencies are 3.22.4, 3.22.2, and 3.22.7. The maximum formula therefore makes the fibre zero-dimensional. A zero-dimensional generalized affinoid has finitely many points: strictify it, use dimension invariance and strict finite normalization to make the strictified algebra finite over its field, and project its finite spectrum surjectively back. These are exactly 1.17.6, 3.21.8, 1.12.2, 3.21.1, 1.14.4, 2.35.1, and 2.33.13. No finiteness assertion over the whole Ts\mathbb T_s follows.

We give the field-stalk argument over the original KK. If a generalized affinoid DD has analytic dimension at most dd and z∈M(D)z\in\mathcal M(D) has dK(z)=dd_K(z)=d, then p=ker⁡(D→H(z))\mathfrak p=\ker(D\to\mathcal H(z)) is minimal. Otherwise choose a minimal prime q⊊p\mathfrak q\subsetneq\mathfrak p. Simultaneously strictify D/qD/\mathfrak q and D/pD/\mathfrak p by a finite KK-free extension, making the new field nontrivially valued as well. The domain-preservation and quotient statements of 5.35.1, together with faithful flatness 1.18.8, preserve the strict inclusion of these primes. The strictified domain has Krull dimension at most dd; quotienting by a nonzero prime lowers it to at most d−1d-1, since every prime chain in the quotient can be preceded by the zero prime of the domain. Analytic dimension invariance gives dim⁡(D/p)≤d−1\dim(D/\mathfrak p)\leq d-1, contradicting the Abhyankar bound applied to zz on that quotient. Here equality of strict analytic and Krull dimensions is the strict dimension theory in 3.21.8 and 1.12.2.

An affinoid neighbourhood of ηs\eta_s in TdT^d is an affinoid domain of affine dd-space. It is regular by 2.35.23, and is a finite product of integral affinoid algebras by 2.52.2. Pass to the open-and-closed component containing ηs\eta_s. It is still an ordinary neighbourhood. Its evaluation kernel is minimal by the preceding paragraph, hence zero. If a germ at ηs\eta_s evaluates to zero, represent it on such an integral neighbourhood inside its domain of definition. It is zero there. Evaluation on the ordinary stalk is therefore injective. The description of its maximal ideal in 2.33.16 makes that ideal zero, so this local ring is the asserted field κs\kappa_s. This argument does not identify κs\kappa_s with its completion.

The torus is boundaryless over KK by 2.33.27, and absolute interior at xx implies relative interior for ψ\psi there: test on ordinary affinoid neighbourhoods and use 2.44.6 and the source-factor implication. The finite set already proved makes xx isolated in its analytic fibre. The local finite criterion 2.49.1, applied inside an ordinary neighbourhood disjoint from the other fibre points, gives ordinary open finite neighbourhoods. Restrict the target to an ordinary affinoid neighbourhood, and then to its integral component at ηs\eta_s. Finite base change gives the asserted Y→NY\to N; YY remains an ordinary neighbourhood of xx.

The finite direct-image and stalk identities 2.35.15 and 2.35.17 give B⊗Cκs=OX,x,(M⊗AB)⊗Cκs=Fx.B\otimes_C\kappa_s=\mathcal O_{X,x},\qquad (M\otimes_A B)\otimes_C\kappa_s=\mathcal F_x. There is only one summand because the isolated analytic fibre has only xx. The finite tensor comparison 2.33.19 then gives the displayed identifications with EE and PP, retaining their actual canonical complete norms. A finite algebra over the complete field HsH_s has one analytic point for each local Artinian factor. Thus EE is local, and its residue field is precisely H(x)\mathcal H(x) by 2.33.11. A composition series of PP has copies of this residue field as its simple factors, proving (C.2.7). Since κs\kappa_s is a field, both finite direct-image stalks are free. Apply 2.35.16 to each, and intersect the resulting ordinary neighbourhoods. This derives freeness from the Gauss stalk; no generic-flatness assumption has been made.

For (C.2.8), the fraction-field argument applies: C↪HsC\hookrightarrow H_s factors through Frac⁡(C)\operatorname{Frac}(C), so HsH_s is flat over CC. The domain map A→BA\to B is flat by ordinary single-domain flatness; hence A→EA\to E is flat. The algebraic composition-series argument of the local Artinian calculation now applies verbatim to this actual Artinian algebra: flat going down makes its prime contract to a minimal prime of AA, and tensoring a composition series at that prime gives the displayed sum. Every other minimal-prime quotient of EE is zero. In particular neither nilpotent length nor the full residue-field degree is discarded.

Finally, if rf(x)=dr_f(x)=d, some dd of the nn homogeneous elements fi(x)~\widetilde{f_i(x)} are algebraically independent. For that coordinate subset II, the map ψI=(fi)i∈I\psi_I=(f_i)_{i\in I} sends xx to ηwI\eta_{w_I}: for every Laurent polynomial the nonzero initial terms of maximal norm cannot cancel, which is exactly the weighted Gauss norm. There are finitely many choices of II, and each ψI−1(ηwI)\psi_I^{-1}(\eta_{w_I}) is finite by the first paragraph. Their union contains SwS_w. If n<dn<d, SwS_w is empty. ◻

Citation

Non-Archimedean Geometry Foundations, Tag NAG-002816.

Proof in chapter · Fixed-edition record