12.1 The normalized regularized maximum
A regularized maximum must do more than approximate a nonsmooth function. In a gluing argument it must agree exactly with a branch which dominates the others. For logarithmic branches it must also ignore a branch near its zero set. We prove these properties with one normalization.
Definition 12.1.1. Fix an even nonnegative function with integral one. For and set The subscript is retained when branches are removed. For , the zero mean gives .
NAG-900251 Proof review pending
Theorem 12.1.2. The function is smooth, convex, symmetric in its coordinates, and nondecreasing in each coordinate. It satisfies If , then on a neighbourhood of the function is independent of and equals the regularized maximum of the remaining coordinates. Furthermore,
Proof. Put . After changing variables, is the convolution of the locally integrable function with the smooth compactly supported probability kernel . Differentiating that kernel proves smoothness. For , apply the convexity inequality for to and integrate. This proves convexity. Monotonicity and permutation symmetry follow directly from the integral.
Translation by adds inside the maximum, and the kernel has mass one. Jensen’s inequality gives the lower bound in (12.1.2), since . The upper bound follows from . The pointwise estimate proves (12.1.3).
For the deletion assertion choose realizing the other maximum. Then throughout the kernel support. The maximum does not use the th coordinate; integrating out contributes the factor one. The strict gap persists near . Finally, monotonicity and convexity give the derivative inequalities. Differentiate (12.1.1) once in , then once in , to obtain the two remaining identities. ◻
The factor is essential if the kernel is written as . Without it the mass is , and translation adds . This is the normalization error in the displayed regularization in CLD25, Proposition 16.9.1, p. 466. Formula (12.1.1) supplies a direct check independent of any positivity argument.
12.2 Composition and gluing
The smoothness just proved is an ordinary real statement. To use it in CLD geometry, all functions involved must first be represented on a common refinement of their charts. The following chain rule then descends to actual forms.
NAG-900252 Proof review pending
Proposition 12.2.1. Let be original smooth functions on an analytic open set, and let be smooth near their joint image. Then If every is positive, is convex, and , then is positive. In particular is smooth psh.
Proof. Apply to and use the product rule. At a point, factor the positive semidefinite Hessian as . The second sum becomes , where . This is a positive -form. The first sum is positive because its scalar coefficients are nonnegative. These identities commute with the pullbacks defining the sheaf. ◻
NAG-900253 Proof review pending
Corollary 12.2.2. (Gluing with a collar). Let be open, let be smooth psh on , and let be smooth psh on . Suppose an open neighbourhood of the relative boundary of in satisfies on . Then is smooth psh on , and equals wherever .
Proof. Deletion and show equality on the collar, so the definitions agree on an open neighbourhood of every interface point. Composition proves positivity on , and the old function supplies it elsewhere. The last assertion follows by deleting the other branch. ◻
A partition of unity does not replace this argument. Although it glues unrestricted smooth functions, derivatives of its factors introduce mixed terms of no fixed sign. The collar inequality is the geometric input that makes positive gluing possible.
12.3 Logarithmic branches and finite barriers
NAG-900254 Proof review pending
Proposition 12.3.1. Let be analytic functions on an analytic open set , let and be real constants, and put Interpret . Applying to the finite branches and deleting the branches equal to defines a global original smooth psh function satisfying .
Proof. Fix . If , continuity of gives a neighbourhood on which . Delete that branch throughout this neighbourhood. For every remaining branch, shrink the neighbourhood so that is a unit. Its logarithmic absolute value is then an original smooth function with zero . The constant branch remains finite everywhere. Proposition 12.2.1 therefore proves smoothness and positivity locally. On overlaps the deletion identity shows that the local formulas coincide. The estimates follow from (12.1.2). ◻
NAG-900255 Proof review pending
Theorem 12.3.2. (A barrier from nested rational domains). Let be an affinoid domain in a good Hausdorff analytic space . Let have no common zero and let . For real , define Suppose , and let satisfy all the inequalities for strictly. There exist an open neighbourhood of , an original smooth psh function on , and an open neighbourhood of such that
Proof. The no-common-zero assumption implies on every : otherwise every would vanish there. These sets are compact rational domains. Choose and set on the locus in . It is finite and continuous. Choose so that , and put . This subset is contained in , hence in the ordinary interior of ; it is therefore open in . It contains . For , its sublevel is exactly .
Choose and a real with . The preceding proposition applies to the analytic functions on . Put . Then . Its nonpositive sublevel is closed in and contained in . As a closed subset of the compact set , it is compact. Continuity gives the required . ◻
The theorem is deliberately stated with the nested-domain input visible. Establishing such rational data near every point uses the local domain theorems of analytic geometry. The analytic construction above proves what follows once those data have been supplied. It also distinguishes the larger domain from the smaller set on which is negative; these two sets must not be identified.
12.4 Uniform approximation and decreasing sequences
The domain of approximation is part of the assertion. The useful elementary conversion below does not change that domain with the accuracy parameter.
NAG-900256 Proof review pending
Lemma 12.4.1. Let be continuous on and let be stable under addition of real constants. Suppose is a locally uniform limit of functions in . For every relatively compact , there is a sequence in decreasing to .
Proof. Choose with . Set . Then The squeeze proves the limit. Every function remains defined on the common set . ◻
Conversely, decreasing continuous functions with a continuous limit converge uniformly on every compact subset, by Dini’s theorem. Neither direction proves that a continuous section at an arbitrary transfinite decreasing-limit stage admits continuous approximants in the preceding stage. That is the missing replacement step discussed in the source concordance, not a consequence of Dini’s theorem.