Contents of this source
The protected background transpose and its critical tangent currents
The positive background in the remaining threefold Dirac problem has degree one. On a protected finite moment domain its exact action on pulled-back smooth tests is a positive closed real supercurrent of bidimension two. Its trace has quadratic growth. This supplies a uniform critical bound, without assuming a local potential for the background or discarding any part of a moment fiber.
An exact transpose on a complete proper moment domain
Let be an original ordinary domain of pure dimension three, and let , , be one finite actual moment map on a GP-protected domain. Thus is open, and The inverse image in MT1 is the entire ordinary inverse image. It is not an old-source inverse image. Let be any original symmetric middle-positive separately - and -closed current of degree on . Define, for every compactly supported smooth real -superform on ,
Proposition (Middle positivity and both closedness equations). Formula MT2 defines a symmetric middle-positive real supercurrent of bidimension , hence bidegree , with For , its positive trace measure is for relatively compact Borel subsets of . Every coefficient measure of has total variation bounded by a fixed dimensional multiple of .
Proof. MT1 makes an original compact smooth CLD test. On each fixed compact support the finite pullback is continuous in the usual smooth test topology, proving that MT2 is a current. Column exchange commutes with this pullback, giving symmetry. If is a smooth real -form, the standard signed form is middle-positive and so is its pullback. The assumed original middle positivity of therefore gives exactly middle positivity of . No equivalence with weak or strong positivity in the ambient dimension has been used.
Pullback commutes with both first derivatives. Compact support remains inside by MT1, so applying either closedness equation of to a pulled-back test proves MT3. In particular no source cutoff has been inserted: such a cutoff could have introduced an additional boundary derivative, whereas MT2 uses the whole protected inverse image.
The scalar functional obtained by testing against is a positive Radon measure. The equality for smooth in MT2 identifies it with the proper measure transpose in MT4, and hence gives that equality for Borel sets. In a constant basis the cocoefficient measures form a positive-semidefinite matrix. Its diagonal sum is the trace, and polarization bounds each off-diagonal variation by one half the sum of the two corresponding diagonal measures. This gives the asserted coefficient control, with the fixed orientation and basis constants. ◻
This construction is an exact finite-presentation test transpose. It does not assert a higher-degree analytic action, a pushforward formula for an arbitrary cap carrier, or that tests in this one target replace all original tests. The current still has its original complete carrier. In particular MT4 includes background mass at every point of the relevant moment fiber, regardless of its old rank.
The exact quadratic trace estimate
Translate to zero and fix . Let . The following is the bidimension-two case of the real Lelong monotonicity calculation; the normalization is recorded explicitly.
Theorem (Quadratic growth and finite critical density). The function is nondecreasing for . Consequently No assertion that is part of the theorem. In particular , while the density in MT6 can be positive.
Proof. First suppose is smooth. On an annulus write and use On the sphere of radius , the first differential pulls back to zero. Stokes, using , therefore gives for To check the constant, apply Stokes first to , and note on a sphere that and after tangential pullback in the Stokes boundary expression. The resulting boundary identity is Subtract this identity at from that at , and apply Stokes on the annulus, where is smooth. This proves MT8 without introducing any product of at the origin. The integrand on the annulus is positive: is semipositive and its square is strongly positive, so it may be paired with the middle-positive .
For a general coefficient-Radon current convolve its coefficients by a nonnegative compact smoothing kernel, inside a slightly smaller ball. The resulting currents are smooth, middle-positive and separately closed. Their trace measures converge weakly to . At radii whose spheres have zero trace mass, MT8 passes to the limit. There are at most countably many exceptional radii. Approximation of arbitrary open balls from below by nonexceptional radii gives the monotonicity in MT5 for all . The finite measure on the fixed larger compact ball gives MT6; letting also gives the absence of an atom of the trace itself. ◻
The printed statement of Lagerberg’s Proposition 3.4 omits the word closed, but its Stokes proof and the following corollary use closedness. MT3 is imposed and proved here. Middle positivity is also essential for coefficient control: Lagerberg’s Remark 3.3 exhibits a weakly positive nonzero tensor with zero trace. No such weaker cone is substituted in MT4.
The exponent two is sharp. A positive integration current of a complete real two-plane has , with . Thus MT6 supplies the critical bound required for a first-jet replacement in CR; it does not supply CR5’s little-o bound for the leading homogeneous term.
Compactness and conicity of critical tangent currents
For , let and define on The pushforward here is only the ordinary real dilation of a real supercurrent. Equivalently, each coefficient measure transforms as The factor in MT9 accounts for both columns of a test; it is not a change of the original non-Archimedean absolute value.
Corollary (Critical tangents are positive closed cones). For every sequence , a subsequence of converges on all compact smooth real tests to a global middle-positive separately closed supercurrent of bidimension two. Every such limit satisfies When the original complete moment-image germ is contained in a closed cone , each limit is supported in . The tangent limit need not be unique or polyhedral.
Proof. For any fixed , MT6 gives for small . MT4 bounds every coefficient measure on each compact ball. Weak compactness of Radon measures and a diagonal selection on increasing balls give a subsequential current limit. Positivity, symmetry and both closedness equations pass to smooth tests. Cone support is preserved under dilation and under this limit.
The left side of MT12 converges numerically to . The open- and closed-ball inequalities for weak measure convergence, followed by and , show that the limit measure has exactly that ball mass. In particular it charges no sphere and no atom at the origin.
Apply the annular identity MT8 to , by the same local regularization argument. Its constant normalized ball mass gives In an orthonormal frame whose first vector is radial, is times the identity on the orthogonal hyperplane and is zero on the radial vector. The middle-positive cocoefficient matrix of is positive semidefinite on two-vectors. MT13 makes its whole tangential two-vector block zero; positivity then makes every row and column meeting that block zero. Only two-vectors containing the radial direction remain. In current-form notation this is The origin has zero coefficient mass, so these identities hold globally. The one-column Cartan identity and now imply, for every distributional coefficient of bidegree , Equivalently for every compact smooth scalar . Differentiating therefore gives zero. This is exactly the dilation identity in MT11, by MT10. ◻
The conicity conclusion constrains the possible critical backgrounds, but it does not put them in the retained analytic weak delta-background class . It also does not prove that their full paired action on the leading finite-PS tensor vanishes. Such a density or annihilation assertion is a further theorem.
Primary sources and original inputs.
The real trace and monotonicity formulas were checked directly in A. Lagerberg, Super currents and tropical geometry, arXiv:1008.2856v1, Definition 3.1, Propositions 3.2 and 3.4, and Remark 3.3, pp. 19–20, https://arxiv.org/pdf/1008.2856. The closed-current scope and exponent are also explicit in F. Elkhadhra and K. Zahmoul, Lelong–Jensen formula, Demailly–Lelong numbers and weighted degree of positive supercurrents, arXiv:1809.00992v2, Theorem 1 and Corollary 2, p. 11, https://arxiv.org/pdf/1809.00992. MT2–MT4 use the retained original GP compactness and middle-positive coefficient-Radon interface. The critical scaling and conicity deduction are recorded above with their full normalizations.