1. Purpose and Scope

This note isolates one technical point that is often implicit in treatments of path integrals and quantum composition: the role of half-densities (and their scaling) in making composition laws coordinate-invariant and dimensionally well-defined.

Claims below are labeled as Proposition (math-precise under hypotheses) or Heuristic (programmatic bridge).

Notation (dimensions). The half-density weight \(\text{length}^{d/2}\) uses the dimension of the manifold being integrated over in the composition law (the “composition variable” dimension). In nonrelativistic time-slicing this is typically spatial dimension; in covariant/field-theory settings it is spacetime dimension. In the gravity-only sieve (Derivation PA-D1.3), read \(d\) as the spacetime dimension \(D\) if that avoids confusion: the “\(D=4\Rightarrow\) area” conclusion is about spacetime \(D=4\), not a claim that spatial \(d=4\) is privileged.

2. Half-Densities and Composition Kernels

Let \(M\) be a \(d\)-dimensional manifold. A (positive) density is a section of \(|\Lambda^d T^\ast M|\), and a half-density is a section of \(|\Lambda^d T^\ast M|^{1/2}\).

The key operational point is: when a kernel is a half-density in its integration variable, composition of kernels does not depend on an arbitrary choice of coordinate measure.

Heuristic PA-H1.1 (Why half-densities). If \(K_1(x,y)\) and \(K_2(y,z)\) are each half-densities in the intermediate variable \(y\), then their product is automatically a density in \(y\) (no additional choice required), and \(\int_M K_1(x,y)K_2(y,z)\) is coordinate-invariant without fixing a preferred \(dy\). This is the structural reason why kernel composition in the path-integral formalism is coordinate-invariant: the half-density structure of each factor ensures the density-product property by construction (Derivation PA-D1.1 makes this precise).

Derivation PA-D1.1 (Coordinate invariance of half-density pairing and composition [Woodhouse1992; BatesWeinstein1997]). In a local chart \(y=(y^1,\ldots,y^d)\), write a half-density as \(\psi(y)=\varphi(y)\,|dy|^{1/2}\). Under a change of variables \(y=y(y’)\), one has \(|dy|^{1/2}=|\det(\partial y/\partial y’)|^{1/2}|dy’|^{1/2}\), so the coefficient transforms as \(\varphi’(y’)=\varphi(y(y’))\,|\det(\partial y/\partial y’)|^{1/2}\).

Hence the product of two half-densities is a density: \(\psi_1\psi_2=(\varphi_1\varphi_2)\,|dy|\), and its integral is chart-independent: \(\int_M \psi_1\psi_2\) is well-defined without choosing a background measure beyond the density bundle itself.

Kernel composition is the same mechanism: if \(K_1(x,y)\) and \(K_2(y,z)\) are half-densities in \(y\), then \(K_1K_2\) is a density in \(y\) and \(\int_M K_1K_2\) is coordinate invariant.

3. Dimensional Analysis: Normalizing a Half-Density Requires a Scale

A density on \(M\) carries the units of \(\text{length}^d\) once physical units are assigned to coordinates. A half-density therefore carries units \(\text{length}^{d/2}\).

Proposition PA-P1.1 (No canonical “half-density = function” identification). There is no canonical identification of a half-density \(\psi\in|\Lambda^d T^\ast M|^{1/2}\) with an ordinary scalar function \(f\) on \(M\) [Woodhouse1992, §5.7; BatesWeinstein1997, Ch. 7]. Choosing such an identification is equivalent to choosing a nowhere-vanishing reference half-density \(\sigma_\ast\) (equivalently a positive density \(\rho_\ast=\sigma_\ast^2\)) and writing \(\psi=f\,\sigma_\ast\).

Derivation PA-D1.2 (Dilation makes the \\(\text{length}^{d/2}\\) weight explicit). On \(\mathbb R^d\), consider a dilation \(y\mapsto y’=a y\) with \(a>0\). Then \(|dy’|=a^d|dy|\), so \(|dy’|^{1/2}=a^{d/2}|dy|^{1/2}\). Thus even in flat space, half-densities carry an inherent \(\text{length}^{d/2}\) scaling weight.

Derivation PA-D1.2a (Near-diagonal scaling forces the square-root Jacobian \\(\varepsilon^{-d/2}\\)). On \(M=\mathbb R^d\), introduce near-diagonal coordinates \(y=x+\varepsilon v\) with \(\varepsilon>0\). Then \(dy=\varepsilon^d\,dv\), hence \(|dy|^{1/2}=\varepsilon^{d/2}|dv|^{1/2}\). For a bi-half-density kernel written locally as \[ K_\varepsilon(x,y)=k_\varepsilon(x,y)\,|dx|^{1/2}|dy|^{1/2}, \] its pullback to \((x,v)\) variables becomes \[ K_\varepsilon(x,x+\varepsilon v) =\big(\varepsilon^{d/2}k_\varepsilon(x,x+\varepsilon v)\big)\,|dx|^{1/2}|dv|^{1/2}. \] Thus, any attempt to define a nontrivial “\(\varepsilon\to0\)” near-diagonal limit of kernels (the scaling step that tangent-groupoid quantization packages) inevitably produces an \(\varepsilon^{d/2}\) factor from the half-density Jacobian, and the corresponding scalar representative must be renormalized by \(\varepsilon^{-d/2}\) to stay finite. This is the same exponent as in the finite-dimensional “square-root delta” normalization: the half-density is the square root of the density Jacobian. For a canonical distributional witness carrying the same exponent, see Derivation PA-D1.2b.

Derivation PA-D1.2b (Identity kernel: the delta bi-half-density carries the same exponent). The identity operator on half-densities has Schwartz kernel \[ K_{\mathrm{Id}}(x,y)=\delta^{(d)}(x-y)\,|dx|^{1/2}|dy|^{1/2}. \] Under the same near-diagonal change of variables \(y=x+\varepsilon v\), one has \(\delta^{(d)}(x-y)=\varepsilon^{-d}\delta^{(d)}(v)\) and \(|dy|^{1/2}=\varepsilon^{d/2}|dv|^{1/2}\), hence \[ K_{\mathrm{Id}}(x,x+\varepsilon v) =\varepsilon^{-d/2}\,\delta^{(d)}(v)\,|dx|^{1/2}|dv|^{1/2}. \] So the universal \(\varepsilon^{-d/2}\) exponent is already encoded in the half-density identity kernel (for a comprehensive treatment showing this exponent is forced across temporal composition, Van Vleck determinant in curved space, heat-kernel diffusion, and renormalization thresholds, see [PathIntegralNormalization]). If one scalarizes by choosing a dimensionless reference half-density \(\sigma_\ast:=L_\ast^{-d/2}|dx|^{1/2}\) (for a constant length scale \(L_\ast\)), then the scalar representative of the identity kernel is the dimensionless distribution \(k_{\mathrm{Id}}(x,y)=L_\ast^{d}\delta^{(d)}(x-y)\).

Proposition PA-P1.2 (Universal *dimensionless* amplitudes force a \\(\text{length}^{d/2}\\) constant). If one imposes the extra requirement that the scalar representative \(f\) in \(\psi=f\,\sigma_\ast\) be dimensionless in physical units, then the reference half-density \(\sigma_\ast\) must carry all of the \(\text{length}^{d/2}\) dimension. In particular, a constant (field-independent) choice of \(\sigma_\ast\) is equivalent to choosing a universal \(\text{length}^{d/2}\) scale.

In \(d=4\), this universal \(\text{length}^{d/2}\) scale is a universal area scale.

Heuristic PA-H1.2 (Reciprocity claim). Half-densities alone do not force a particular scale: the forced fact is that converting half-density objects into scalar numerical amplitudes requires extra structure (a reference half-density). The “universal area scale” claim begins only after adding two further hypotheses: 1. the reference \(\sigma_\ast\) is taken to be constant (no dependence on background metric/fields), and 2. the constant is required to be fixed by universal constants/couplings of the theory.

Under these hypotheses, \(d=4\) is the unique dimension in which the needed \(\text{length}^{d/2}\) constant can be supplied by the gravitational coupling without fractional powers (Derivation PA-D1.3).

Derivation PA-D1.3 (Gravity-only sieve: why \\(d=4\\) is singled out if only \\(G_d\\) is used). In \(d\) spacetime dimensions, the Einstein–Hilbert action \(\frac{1}{16\pi G_d}\int d^d x\,\sqrt{|g|}\,R\) shows that (in \(c=\hbar=1\) units) Newton’s constant has dimension \([G_d]=\text{length}^{d-2}\). Assume the only available dimensionful coupling used to build the universal normalization constant is \(G_d\) itself (no cosmological constant, no additional dimensionful scales), and impose PA-H2.5 in the literal “no fractional powers of \(G_d\)” sense. Then the normalization constant has dimension \(\text{length}^{k(d-2)}\) for some integer \(k\). Matching \(\text{length}^{d/2}\) forces \(\text{length}^{d/2}=\text{length}^{d-2}\), which holds if and only if \(d=4\). In that case \(G_4\) itself has dimension of area, and the corresponding area scale is the Planck area \(L_P^2\sim \hbar G_4/c^3\).

Remark PA-D1.3a (Three obstruction mechanisms: why \\(d=4\\) is uniquely un-obstructed). The failure of the gravity-only sieve away from \(d=4\) has three qualitatively distinct causes. For odd \(d\), the target exponent \(d/2\) is a half-integer, while any monomial in couplings with integer length dimensions has integer total dimension — a parity obstruction that is categorical regardless of which integer-dimensional couplings are admitted (including \(\Lambda_d\)). For even \(d\ge 6\), the target \(d/2\) is an integer but \(k=d/(2(d-2))<1\), so no positive-integer power of \(G_d\) alone suffices — a magnitude obstruction. For \(d=2\), \(G_2\) is dimensionless and gravity provides no scale — but \(d=2\) is precisely the dimension where the transmutation route (PA-H2.13, Example PA-E5) provides its minimal witness: the marginal 2D delta coupling generates the scale \(\kappa_\ast^{-1} = \text{length}^{d/2}\) via RG invariance rather than coupling monomials. Thus the “trivial gravity” and “minimal transmutation witness” appearances of \(d=2\) are complementary: the monomial sieve fails (no scale from gravity), but the transmutation mechanism succeeds. Thus \(d=4\) is not merely the solution of a single Diophantine equation; it is the unique integer dimension that evades all three obstructions simultaneously.

Remark PA-D1.3b (Odd-\\(d\\) parity lemma and transmutation route). The parity obstruction for odd \(d\) is categorical under the monomial sieve (PA-H2.5a): if all admitted couplings have integer length dimensions \(a_i\in\mathbb Z\), then any monomial \(\prod g_i^{n_i}\) with \(n_i\in\mathbb Z\) has total dimension \(\sum n_i a_i\in\mathbb Z\), so a half-integer target \(d/2\) is unreachable. Standard couplings \(G_d\), \(\Lambda_d\), \(g_d^2\) (YM action coupling), mass \(m\), and area parameter \(\alpha_\) all have integer length dimensions in any \(d\). The transmutation route (PA-H2.13 / Derivation PA-D1.6a) is parity-blind: if a marginal or relevant coupling generates an RG-invariant scale \(\Lambda_\) with \([\Lambda_]=\text{length}^{-1}\), then \(\Lambda_^{-d/2}\) has dimension \(\text{length}^{d/2}\) for any \(d\), integer or not, because \(\Lambda_*\) is a positive real number (not a coupling to be raised to an integer power). Explicit witness for \(d=3\): the 3D delta interaction generates a scattering length \(a\) (dimension \(\text{length}^1\)) via power-law renormalization, and the scalarization constant \(a^{3/2}=\text{length}^{3/2}=\text{length}^{d/2}\) is well-defined (smooth function of \(a\)) without involving fractional powers of fundamental couplings. Thus the claim “\(d=4\) is selected by the monomial sieve” should be stated precisely as: “among even \(d\) with gravity-only coupling, \(d=4\) is uniquely selected.” Odd \(d\) is blocked by parity, not by the gravity coupling’s specific dimension.

3.1 Hypotheses as Separate Knobs (What Is Forced vs Chosen)

The discussion above mixes three different kinds of statements: 1. Geometric facts (what half-densities are, how they compose, how they scale), 2. Representational choices (how one turns half-density objects into scalar numbers), 3. Universality/selection principles (what choices are allowed if we demand “background-free” and “built from couplings”).

To study these separately, it is useful to keep the hypotheses explicit.

Hypothesis PA-H2.1 (Half-density formulation). Quantum kernels are treated as bi-half-densities so that composition in intermediate variables is coordinate invariant (Section 2 and Derivation PA-D1.4).

Hypothesis PA-H2.2 (Scalarization by a reference half-density). To interpret half-density amplitudes as scalar numerical functions, we pick a nowhere-vanishing reference half-density \(\sigma_\ast\) and write \(\psi=f\,\sigma_\ast\) (Proposition PA-P1.1).

Hypothesis PA-H2.3 (Dimensionless scalar representative). The scalar representative \(f\) is required to be dimensionless in physical units (Proposition PA-P1.2). This forces \(\sigma_\ast\) to carry the full \(\text{length}^{d/2}\) weight.

Hypothesis PA-H2.4 (Background-free constancy). The reference \(\sigma_\ast\) is taken to be constant/field-independent, rather than determined by background geometry (e.g. a Riemannian volume \(|g|^{1/4}|dx|^{1/2}\)) or by dynamical fields (e.g. a dilaton-like factor). This is the first point where a universal constant enters.

Hypothesis PA-H2.5 (Analyticity / no fractional powers of couplings). If the universal constant is required to be built from the theory’s couplings without fractional powers, then dimensional analysis becomes a dimension sieve rather than a tautology. This hypothesis has at least two distinct readings: 1. Integrality (integer-exponent) reading: the constant is a monomial in the available couplings with integer exponents (possibly allowing negative powers). Equivalently, dimension-matching becomes an integer (Diophantine) constraint on the exponents. 2. Perturbative analyticity reading (stronger): the constant admits a Taylor expansion around a specified weak-coupling limit, so only nonnegative integer powers appear.

To keep the branches explicit, we will refer to these as: - Hypothesis PA-H2.5a (Integrality / monomial sieve). The normalization constant is a monomial in admitted couplings with integer exponents. - Hypothesis PA-H2.5b (Perturbative analyticity sieve). The dependence is analytic at a specified weak-coupling point (hence forbids negative powers and other non-analytic dependence).

Heuristic PA-H2.5b1 (Analyticity for dimensionful couplings needs a reference scale). If a coupling \(g_i\) is dimensionful, “weak coupling” is only meaningful after choosing a reference scale \(\mu\) and forming a dimensionless parameter \(\hat g_i(\mu):=\mu^{-a_i}g_i\) (where \([g_i]=\text{length}^{a_i}\)). In that sense, perturbative analyticity is naturally analyticity in \(\hat g_i(\mu)\) near \(\hat g_i=0\). Demanding at the same time that the scalarization constant be a \(\mu\)-independent universal constant pushes one back to either: 1. an engineering-dimension monomial in admitted dimensionful couplings with nonnegative integer exponents (the sieve branch), or 2. a non-analytic RG-invariant transmutation scale (outside PA-H2.5b).

Separately, the RG “dimensional transmutation” mechanism (Heuristic PA-H2.13 / Derivation PA-D1.6a) supplies a scale outside both PA-H2.5a and PA-H2.5b: it is typically non-analytic in the coupling.

Derivation PA-D1.3 is the simplest gravity-only instance under the integrality reading: “use \(G_d\) without fractional powers” singles out \(d=4\).

Heuristic PA-H2.6 (Where "special dimensions" can appear). Special dimensions do not come from half-densities alone (Hypothesis PA-H2.1). They appear only after adding a selection principle like PA-H2.4–PA-H2.5: the requirement that the scalarization choice be universal, background-free, and coupling-built in a restricted (e.g. analytic) way.

Remark PA-H2.6b (Minimal set and non-bookkeeping content). The minimal hypothesis set for “half-density normalization \(\Rightarrow\) Planck area in \(d=4\)” is: PA-H2.1 (composition forces the \(d/2\) exponent — Derivation PA-D1.4a, not a convention), PA-H2.4 (background-free scalarization — a physical principle, not bookkeeping: dropping it lets the metric-derived \(|g|^{1/4}|dx|^{1/2}\) work in any \(d\) without a new scale), PA-H2.5a (integrality sieve — a UV hypothesis: dropping it allows transmutation to supply scales in any \(d\)), and the gravity-only coupling restriction. Each ingredient is necessary: removing any one defeats the conclusion. The Diophantine equation \(k(d-2)=d/2\) is the algebraic core, but its ingredients are not arbitrary — composition, background-freeness, and integrality each have independent physical or structural motivation.

Heuristic PA-H2.6a (Independent D=4 filter: conformal scalarization-gauge simplicity). The “dimension sieve” (PA-H2.5) is not the only way a special dimension can appear once one insists on scalar representatives. A different kind of filter comes from asking for simplicity of how kinetic operators depend on the scalarization gauge.

In the QFT-facing half-density calculus, a background metric supplies a reference half-density \(|g|^{1/4}|dx|^{1/2}\), and the scalar Laplacian \(\Delta_g\) conjugates to an operator on half-densities \[ \widetilde\Delta:=|g|^{1/4}\Delta_g|g|^{-1/4}. \] In a conformal background \(g_{\mu\nu}=e^{2\sigma(x)}\delta_{\mu\nu}\) in spacetime dimension \(D\) (here \(\sigma(x)\) denotes the conformal factor, not the reference half-density \(\sigma_\ast\) of Section 3), a direct computation gives \[ \widetilde\Delta\psi =e^{-2\sigma}\Big( \partial^2\psi -2\,\partial\sigma\cdot\partial\psi -\frac{D}{2}(\partial^2\sigma)\,\psi +\frac{D(4-D)}{4}(\partial\sigma)^2\,\psi \Big). \] Derivation sketch. On \((\mathbb{R}^D, e^{2\sigma}\delta_{\mu\nu})\), the scalar Laplacian is \(\Delta_g = e^{-2\sigma}(\partial^2 + (D-2)\partial\sigma\cdot\partial)\). The half-density conjugation \(\widetilde\Delta = |g|^{1/4}\Delta_g|g|^{-1/4}\) uses \(|g|^{1/4} = e^{D\sigma/2}\). Writing \(|g|^{-1/4}\psi = e^{-D\sigma/2}\psi\) and applying \(\Delta_g\) to \(e^{-D\sigma/2}\psi\) via the product rule generates four terms: (i) \(e^{-2\sigma}\partial^2\psi\), (ii) \(-2\partial\sigma\cdot\partial\psi\) from the cross term \((D-2)\partial\sigma\cdot\partial(e^{-D\sigma/2}\psi)\) combined with the product-rule derivative acting on \(e^{-D\sigma/2}\), (iii) \(-\frac{D}{2}(\partial^2\sigma)\psi\) from the Laplacian hitting the exponential, and (iv) \(\frac{D(4-D)}{4}(\partial\sigma)^2\psi\) from collecting the quadratic-gradient contributions: the \((D-2)\partial\sigma\) Christoffel term gives \(+D(D-2)/2\) while the square of the gradient from the exponential gives \(-D^2/4\), combining to \(D(4-D)/4\). See also Birrell and Davies, Quantum Fields in Curved Space (1982), §3.2 for the conformal transformation of the scalar Laplacian; the half-density conjugation step is the standard Weyl rescaling with weight \(D/4\) [Vassilevich2003HeatKernel, §3.3].

Thus the quadratic-gradient term \((\partial\sigma)^2\psi\) cancels at \(D=4\) (within the conformal ansatz). If one adopts the extra criterion that scalarization-gauge changes should not generate such quadratic-gradient “potentials” in the half-density kinetic operator, then \(D=4\) is singled out by operator simplicity rather than by coupling-dimension matching.

This filter is independent of PA-H2.5 and, by itself, does not supply a length scale; it is recorded only as an additional “special dimension” candidate knob to compare against the scale-sieve hypotheses.

Remark PA-H2.6c (Heat-kernel witness: conformal coupling and \\(D=4\\)). The \(D=4\) conformal specialness has a parallel heat-kernel manifestation. For the scalar kinetic operator \(P=-\nabla^2+\xi R+m^2\) on a closed Riemannian \(D\)-manifold, the first Seeley-DeWitt coefficient at coincidence is \(a_1(x,x)=(\tfrac16-\xi)R\). Conformal coupling corresponds to \(\xi_{\mathrm{conf}}=(D-2)/(4(D-1))\), which gives \(a_1(x,x)\big|{\xi=\xi{\mathrm{conf}}}=\tfrac{4-D}{12(D-1)}R\). This vanishes if and only if \(D=4\). The physical interpretation: at \(D=4\), the half-density conjugation potential \(V=\tfrac16 R\) (universal for any metric in any dimension) exactly equals the conformal coupling \(\xi_{\mathrm{conf}}R=\tfrac16 R\), so the conformally coupled half-density field has no leading curvature correction to its heat kernel at coincidence. The heat-kernel trace \(\mathrm{Tr}\,e^{-tP}\sim(4\pi t)^{-D/2}\sum A_n t^n\) carries the same \(D/2\) exponent as the identity kernel normalization (Derivation PA-D1.2b), the propagator Schwinger parametrization, the UV divergence degree \(\Lambda^{D-2n}\), and the dimensional-regularization pole structure \(\Gamma(D/2-n)\) — five manifestations of the single fact that the half-density identity kernel on \(\mathbb R^D\) scales as \(\varepsilon^{-D/2}\) [Vassilevich2003HeatKernel, eq. (4.14) for the \(a_1\) coefficient].

3.2 What Changes When a Hypothesis Is Relaxed?

This subsection records the main “branches” that need separate study.

Drop PA-H2.3 (allow dimensionful \(f\)). Then no universal \(\text{length}^{d/2}\) constant is forced; the dimensional weight can be carried by the scalar representative itself (as in the usual statement “wavefunctions have dimension \(\text{length}^{-d/2}\)”).
Drop PA-H2.4 (allow background geometry). Then \(\sigma_\ast\) can be chosen from a metric (or other structure), and the “universal constant” is replaced by background-dependent normalization.
Drop PA-H2.5a/PA-H2.5b (allow fractional powers or other non-analytic dependence). Then in any \(d>2\) one can build a \(\text{length}^{d/2}\) constant from gravity via \(G_d^{\,d/(2(d-2))}\) (in \(c=\hbar=1\) units), so \(d=4\) is no longer singled out; instead, \(d=4\) is simply the unique case where the exponent is an integer.
Change “which coupling supplies the scale”. Using other dimensionful couplings (cosmological constant, string tension, gauge couplings in various dimensions, etc.) yields different “special-dimension” sieves. This is conceptually aligned with the observation that some dimensions are singled out by other structures (division algebras, special holonomy, supersymmetry), but those filters are separate from the half-density story and should not be conflated.
Replace PA-H2.5 by transmutation (RG as the scale supplier). If one allows RG-invariant scales generated by compatibility (dimensional transmutation), then even dimensionless couplings can supply a physical length scale (Heuristic PA-H2.13 / Derivation PA-D1.6a). This branch does not “sieve dimensions” in the same way as PA-H2.5a/PA-H2.5b; it supplies a scale by non-analytic RG invariants rather than by monomials in couplings.

Branch summary (keep the knobs separate). There are three distinct “scale supplier” mechanisms in play: 1. Monomial sieve (PA-H2.5a/PA-H2.5b): build the needed \(\text{length}^{d/2}\) constant from admitted couplings using restricted (e.g. analytic) dependence. 2. Fractional/non-analytic coupling dependence: allow fractional powers or other non-analytic dependence directly in couplings (dimensional analysis becomes permissive; the \(d=4\) sieve disappears). 3. Dimensional transmutation: generate a scale from RG compatibility (Derivation PA-D1.6a / Example PA-E5), which is typically non-analytic in naive coupling expansions.

3.3 Starting with PA-H2.5a: Integrality as a Dimension Sieve

The point of PA-H2.5a/PA-H2.5b is not that dimensional analysis alone selects a unique scale (it does not), but that restricting allowed functional dependence on couplings can turn dimensional analysis into a selection principle.

Derivation PA-D1.6 (PA-H2.5a: integer-exponent form of “no fractional powers”). Work in \(c=\hbar=1\) units for dimension counting. Let the available couplings \({g_i}\) have length dimensions \([g_i]=\text{length}^{a_i}\). Under the integrality reading of PA-H2.5, the universal normalization constant is a monomial \(C=\prod_i g_i^{n_i}\) with integers \(n_i\). Its length dimension is \([C]=\text{length}^{\sum_i n_i a_i}\). Requiring \([C]=\text{length}^{d/2}\) is therefore the integer-exponent (Diophantine) condition

\[\sum_i n_i a_i=\frac{d}{2}.\]

Existence (and non-uniqueness) of solutions depends on: 1. which couplings are admitted as “universal” inputs, and 2. whether one allows negative exponents (non-analytic at zero coupling) or insists on perturbative analyticity (nonnegative exponents).

Heuristic PA-H2.7 (Why PA-H2.5 needs a “what counts as a coupling” rule). If one allows arbitrary redefinitions of couplings (e.g. adjoining a new symbol \(\tilde G = G_d^{1/(d-2)}\)), then “no fractional powers” becomes vacuous: the forbidden root has simply been renamed as an allowed coupling. PA-H2.5 is meaningful only together with a prior criterion for admissible coupling dependence (e.g. perturbative analyticity around a distinguished limit such as \(G_d\to0\)).

Heuristic PA-H2.7a (Admissible couplings: exclude scheme parameters and non-analytic reparametrizations). To keep PA-H2.5 from collapsing into a coordinate artifact, we implicitly adopt the following convention: 1. Admitted couplings are the independent parameters that appear as coefficients of local operators in the UV action after fixing a canonical normalization convention for fields (so that field-rescaling freedom is not used to hide roots/powers). 2. We allow only analytic reparametrizations near a chosen base point (“weak coupling”), so adjoining \(\tilde g=g^{1/2}\) is disallowed when it is non-analytic at that base point. 3. We explicitly exclude scheme/scale conventions from the coupling set: the renormalization scale \(\mu\), regulator cutoffs \(\Lambda\), and finite-subtraction constants; and we also exclude the scalarization gauge scale \(L_\ast\) itself (since constraining \(L_\ast\) is the point of the ladder).

Under this convention, PA-H2.5a is best viewed as a computational proxy once coupling coordinates are fixed, while PA-H2.5b (analyticity at the base point) is the more invariant “no roots” statement. The phrase “canonical normalization of fields” in (1) is itself a convention choice; the point is that fixing such a convention makes the admissibility rule a controlled knob rather than an implicit loophole.

Heuristic PA-H2.5c (Physical content of the integrality sieve: classical vs quantum scale generation). The integrality sieve (PA-H2.5a) has a physical interpretation beyond the aesthetic “no fractional powers” reading. In any dimension \(d > 2\), the half-density normalization scale \(\text{length}^{d/2}\) can be supplied by \(G_d^{d/(2(d-2))}\) — but this is an analytic function of \(G_d\) only when \(d/(2(d-2))\) is a non-negative integer, i.e., only at \(d=4\) (where the exponent is 1). In all other dimensions, the scalarization constant requires either: (a) fractional powers of \(G_d\) (non-analytic at \(G_d=0\), i.e., ill-defined in the decoupled-gravity limit), or (b) non-analytic RG-invariant transmutation scales (dimensional transmutation, Heuristic PA-H2.13).

The physical distinction is therefore: \(d=4\) is the unique dimension where the half-density normalization can be supplied by the classical (tree-level) gravitational coupling alone. In other dimensions, the normalization requires quantum (loop-level) scale generation. The integrality sieve is equivalent to asking: “in which dimension does the half-density normalization survive in the classical (\(G_d \to 0\)) limit?” — and the answer is \(d=4\).

Important caveat: The tree-level/loop distinction invoked here is itself a statement within perturbative gravity. In the Wilsonian effective-action framework, the EH action \(\int\sqrt{\lvert g\rvert}R/(16\pi G_d)\) defines the tree-level vertex; loops contribute higher-derivative corrections suppressed by powers of \(E^2 G_d\). Analyticity of the scalarization constant in \(G_d\) near \(G_d = 0\) is therefore analyticity in the perturbative gravitational expansion — a physically natural requirement if one demands that the half-density normalization be definable without needing to resum the gravitational loop series. This is the precise sense in which PA-H2.5a is a “tree-level sufficiency” criterion rather than a mathematical elegance condition. However, this motivation is itself perturbative: a non-perturbative UV completion of gravity might supply alternative mechanisms, which is why the conclusion is labeled as conditional on PA-H2.5 rather than unconditional.

To summarize the logical status: the \(d/2\) exponent in the half-density normalization is proved (Derivation PA-D1.4a); the Planck area identification is motivated by, not derived from, the half-density formalism — it requires the additional PA-H2.4 + PA-H2.5a + gravity-only hypotheses. This distinction is structural, not a weakness: the paper’s contribution is precisely delineating which conclusion follows from which hypothesis.

Example PA-E1 (Gravity-only). With only \(G_d\) available, \(a_1=d-2\) and the condition becomes \(n(d-2)=d/2\). For integer \(d\ge 3\), this has a solution only at \(d=4\) with \(n=1\), reproducing Derivation PA-D1.3.

Example PA-E2 (Gravity + cosmological constant). If one also allows the cosmological constant \(\Lambda_d\) with \([\Lambda_d]=\text{length}^{-2}\), then the condition becomes \(n(d-2)-2m=d/2\) for integers \(n,m\). A simple family of solutions exists for \(d\) divisible by \(4\): take \(n=1\) and \(m=d/4-1\), so

\[C \sim G_d\,\Lambda_d^{\,d/4-1},\]

has dimension \(\text{length}^{d/2}\). Thus, even under PA-H2.5, \(d=4\) is not automatically unique once additional dimensionful couplings are admitted; what is special about \(d=4\) in this family is that it is the only case with \(m=0\) (no need to involve \(\Lambda_d\)).

Example PA-E3 (Yang--Mills coupling as an alternative sieve). In \(d\) spacetime dimensions, the Yang–Mills action is typically written as \(\frac{1}{4g_d^2}\int d^d x\,F_{\mu\nu}F^{\mu\nu}\), so \([g_d^2]=\text{length}^{d-4}\) (equivalently \([g_d]=\text{length}^{(d-4)/2}\)). If we (hypothetically) allow the half-density normalization constant to be a pure monomial in \(g_d\), \(C\sim g_d^{\,p}\) with integer \(p\ge 0\), then \[ [C]=\text{length}^{p(d-4)/2}. \] Matching \([C]=\text{length}^{d/2}\) gives the integer-exponent condition \[ p(d-4)=d \quad\Longrightarrow\quad d=\frac{4p}{p-1}=4+\frac{4}{p-1}. \] The case \(p=1\) is excluded (it gives \(d=4p/(p-1)\) undefined; the YM coupling would be dimensionless in that formal limit); \(p=0\) gives a dimensionless \(C\), which carries no scale. For \(p\ge 2\), integer solutions occur when \(p-1\mid 4\), i.e. \(p\in{2,3,5}\), giving \(d\in{8,6,5}\) respectively.

In particular, in \(d=4\) the gauge coupling is dimensionless and cannot by itself supply the \(\text{length}^{d/2}\) factor needed for half-density scalarization; in that case the scale must come from another dimensionful coupling (e.g. gravity) or from a non-analytic mechanism (dimensional transmutation).

Example PA-E4 (A universal area parameter \\(\alpha_\ast\\) as a scale supplier). Suppose the UV theory admits a universal area parameter \(\alpha_\ast\) with dimension \([\alpha_\ast]=\text{length}^2\) (for example, in perturbative string theory \(\alpha_\ast\) is the familiar \(\alpha’=l_s^2\), with string tension \(T\sim 1/\alpha’\)). If one allows the half-density normalization constant to be built from \(\alpha_\ast\) alone as a monomial \(C\sim (\alpha_\ast)^n\) with integer \(n\), then \[ [C]=\text{length}^{2n}, \] and matching \([C]=\text{length}^{d/2}\) forces \(2n=d/2\), i.e. \(d=4n\). So \(\alpha_\ast\) provides a “background-free” source of scale but does not single out \(d=4\) on its own; it selects dimensions divisible by \(4\) under the strict integrality reading of PA-H2.5. In \(d=4\) it yields directly an area scale \(C\sim \alpha_\ast\) (e.g. \(C\sim \alpha’\) in the string example).

Remark PA-E4a (Emergent string tension is a transmutation-scale instance, not a UV parameter). In confining phases one often defines an effective string tension \(\sigma\) with \([\sigma]=\text{length}^{-2}\), so \(\sigma^{-1}\) supplies an area scale. Operationally, for a large rectangular Wilson loop of spatial size \(R\) and Euclidean time extent \(T\), one defines \(V(R)=-\lim_{T\to\infty}\frac{1}{T}\ln\langle W(R\times T)\rangle\), and an area law \(\langle W(R\times T)\rangle\sim e^{-\sigma RT}\) implies \(V(R)\sim\sigma R\) and identifies \(\sigma\) [Greensite2003Confinement]. In \(d=4\) gauge theories with dimensionless couplings, such a scale is expected to arise (when it exists) as an RG-invariant transmutation scale rather than as a new analytic monomial in the couplings. Logically, this places “string tension as area supplier” in the PA-H2.13 branch unless one is explicitly assuming a fundamental UV area parameter \(\alpha_\ast\).

Heuristic PA-H2.12 (Link to gravity/Planck length in a UV completion). The gravity-only sieve (Derivation PA-D1.3) uses \(G_d\) as the unique universal coupling supplying dimension. In a UV completion where gravity is emergent from a sector with a universal area parameter \(\alpha_\ast\) and a dimensionless coupling \(g\), dimensional analysis suggests \[ G_d \propto g^{2}\,(\alpha_\ast)^{(d-2)/2}\times(\text{volume factors}), \] so the Planck length/area is derived from \(\alpha_\ast\) and \(g\) rather than fundamental. In perturbative string theory, one has \(\alpha_\ast=\alpha’\) and \(g=g_s\) (up to compactification-volume factors). In that framing the half-density “universal area scale” could naturally be \(\alpha_\ast\) (or a simple function of it), while the Planck area is recovered as a consequence of how gravity emerges.

Heuristic PA-H2.8 (What PA-H2.5 is really buying). The value of PA-H2.5 is comparative: it distinguishes dimensions in which the needed \(\text{length}^{d/2}\) factor can be supplied by simple coupling dependence (integer powers of the already-present couplings), versus dimensions in which any such factor requires either (i) introducing extra scales/couplings, (ii) taking fractional powers, or (iii) invoking non-analytic mechanisms (dimensional transmutation).

Heuristic PA-H2.13 (Dimensional transmutation as a scale-supplier). If one relaxes PA-H2.5’s “analytic monomial in couplings” requirement, then even a theory with only dimensionless couplings can generate a physical length scale through RG invariance: a running coupling \(g(\mu)\) can be traded for an RG-invariant scale \(\kappa_\ast\) (or \(\Lambda\)), typically of the form \(\kappa_\ast \sim \mu\,\exp(-\mathrm{const}/g(\mu)^2)\) or \(\kappa_\ast \sim \mu\,\exp(-\mathrm{const}/g(\mu))\). In that branch, the half-density scalarization scale required by PA-H2.4 can be supplied by \(\kappa_\ast^{-d/2}\) (or its square in \(d=4\) as an area scale), but the scale is no longer an analytic monomial in the couplings: it is emergent and non-perturbative in the naive coupling expansion.

Heuristic PA-H2.14 (Bookkeeping: what “\\(d\\)” means in \\(\text{length}^{d/2}\\)). The half-density weight \(\text{length}^{d/2}\) refers to the dimension of the manifold whose coordinates are integrated over in the composition law (the intermediate-variable space). In a nonrelativistic time-sliced kernel this is typically the spatial dimension, while for covariant/proper-time kernels one may compose over spacetime points. The dimension-sieve discussion using \(G_d\) treats \(d\) as the spacetime dimension, so any \(d=4\Rightarrow\) “area scale” conclusion should be read in that covariant sense unless stated otherwise.

Derivation PA-D1.6a (RG-invariant scale from a beta function). Let a (dimensionless) running coupling \(g(\mu)\) satisfy an RG equation \(\mu\,dg/d\mu=\beta(g)\) with \(\beta(g)\neq 0\) in the range of interest. Then the combination \[ \Lambda_\ast \equiv \mu\exp!\left(-\int^{g(\mu)} \frac{dg’}{\beta(g’)}\right) \] is RG-invariant (independent of the subtraction scale \(\mu\)), up to a finite multiplicative constant corresponding to a choice of scheme/normalization of the integral. In one-loop form \(\beta(g)=-b g^2+O(g^3)\), one obtains the familiar transmutation scale \(\Lambda_\ast \sim \mu\,e^{-1/(b g(\mu))}\times(\text{scheme factor})\).

If PA-H2.3–PA-H2.4 demand a universal scalarization constant \(C\) with \([C]=\text{length}^{d/2}\), then any RG-invariant inverse length \(\Lambda_\ast\) supplies one by \(C\sim \Lambda_\ast^{-d/2}\). In particular, for \(d=4\) this produces a universal area scale \(C\sim \Lambda_\ast^{-2}\), without requiring the scale to be an analytic monomial in couplings (so this branch sits outside PA-H2.5).

Example PA-E5 (2D delta: transmutation yields a length scale). In the \(2\)D delta interaction, the contact coupling is marginal and the renormalized theory is naturally parameterized by an RG-invariant inverse length \(\kappa_\ast\) rather than by the bare coupling. Concretely, one finds (up to conventions) a running coupling \(g_R(\mu)\) with beta function \(\beta(g_R)\propto g_R^2\), and the RG invariant \[ \kappa_\ast^2 \equiv \mu^2\exp!\left(\frac{2\pi\hbar^2}{m}\frac{1}{g_R(\mu)}\right), \] so \(\kappa_\ast\) is independent of the subtraction scale \(\mu\) and sets a bound-state/scattering scale [ManuelTarrach1994PertRenQM]. This is a minimal witness that “a scale is forced by compatibility” can occur even without a dimensionful coupling, via renormalization rather than via analytic monomials.

Remark PA-E5a (Half-density match and three-level RG hierarchy). In the \(2\)D delta, the transmutation scale \(\kappa_\ast^{-1}\) has dimension \(\text{length}^1=\text{length}^{d/2}\) for \(d=2\), exactly the half-density scalarization weight required by PA-H2.3–PA-H2.4. The beta function \(\beta(g_R)\propto g_R^2\) vanishes to order \(2\) at the Gaussian fixed point; transmutation requires this nonlinearity — a linear beta function produces only algebraic (power-law) RG invariants without generating a new scale. Three levels of the RG hierarchy are now witnessed by explicit computations: 1. Semigroup structure: shared by all refinement flows, including linear beta functions (any toy ODE with a linear beta, e.g. \(\beta(a)=\tfrac12-a\), exhibits the semigroup property without transmutation). 2. Transmutation: requires \(\beta\) of order \(\ge 2\) at the fixed point, producing a non-analytic RG-invariant scale. Witness: this example (2D delta, \(\beta\propto g_R^2\)). 3. Dimension sieve (PA-H2.5): demands the scale be an analytic monomial in couplings, selecting \(d=4\) under the gravity-only hypothesis. Witness: Derivation PA-D1.3. Transmutation (level 2) supplies a scale in any \(d\) where a marginal coupling exists, so it does not sieve dimensions. The half-density weight \(\text{length}^{d/2}\) correctly tracks the geometric type of the resulting scale: a length in \(d=2\), an area in \(d=4\).

Remark PA-E5b (Where in the kernel the transmutation scale acts). The full resolvent of the \(2\)D delta interaction factorizes via the Lippmann–Schwinger identity as \(G=G_0+G_0\,T\,G_0\), where \(G_0(x,z;E)\) is the free resolvent and \(T(E)\) is the scalar \(T\)-matrix at the contact vertex. Each \(G_0\) factor carries the Van Vleck half-density weight at its endpoint (Derivation PA-D1.4), while \(T(E)\) is a scalar amplitude containing the transmutation scale \(\kappa_\ast\). Thus the two ingredients play complementary roles: the Van Vleck prefactor implements the half-density transformation law (PA-H2.1), and the transmutation scale supplies the scalarization constant needed to extract dimensionless amplitudes (PA-H2.2–PA-H2.4). They are structurally independent and combine multiplicatively in the kernel; the geometric weight comes from free propagation between vertices, while the dynamical scale comes from RG invariance at the vertex.

Remark PA-E5c (Half-density composition at the vertex mirrors the free-propagator semigroup). In the Lippmann–Schwinger product \(G_0\,T\,G_0\), the half-density mechanism is the same as in PA-D1.4a: the two \(|dy|^{1/2}\) factors at the vertex pair into a density \(|dy|\), which integrates coordinate-invariantly, while the result inherits \(|dx|^{1/2}\,|dz|^{1/2}\) at the endpoints. The \(T\)-matrix is a scalar at the contact vertex containing the transmutation scale \(\kappa_\ast\) but carrying no half-density weight. Thus the half-density structure (geometric, from free propagation) and the transmutation scale (dynamical, from RG at the vertex) enter the resolvent through independent factors and combine multiplicatively. The transmutation scale \(\kappa_\ast^{-1}\) has dimension \(\text{length}^{d/2}\) for \(d=2\), matching the half-density scalarization weight. For \(d=4\), the same factorization gives a scalarization constant \(\Lambda_\ast^{-d/2}=\Lambda_\ast^{-2}=\text{area}\).

3.4 Running PA-H2.3: Is “Dimensionless \(f\)” Physics or Convention?

The half-density formalism (PA-H2.1) gives a canonical pairing \(\int \bar\psi\,\psi\) that does not require choosing a background measure. But when we write \(\psi=f\,\sigma_\ast\) (PA-H2.2), we are choosing a representation of the same object as a scalar function with respect to a chosen positive density \(\rho_\ast=\sigma_\ast^2\).

Proposition PA-P1.3 (Scalarization is a choice of measure, not new physics). Choosing a reference half-density \(\sigma_\ast\) identifies the canonical Hilbert space of \(L^2\) half-densities on \(M\) with the scalar Hilbert space \(L^2(M,\rho_\ast)\), where \(\rho_\ast=\sigma_\ast^2\). Different choices \(\sigma_\ast\) yield unitarily equivalent scalar representations.

Derivation PA-D1.7 (Change of reference half-density acts by multiplication). Let \(\sigma_1,\sigma_2\) be nowhere-vanishing half-densities on \(M\) and set \(r:=\sigma_2/\sigma_1\), a positive scalar function. Writing the same half-density state \(\psi\) as \(\psi=f_1\,\sigma_1=f_2\,\sigma_2\) gives \(f_2=r^{-1}f_1\). Moreover, \[ \int_M \bar\psi\,\psi =\int_M |f_1|^2\,\sigma_1^2 =\int_M |f_2|^2\,\sigma_2^2 =\int_M |f_2|^2\,r^2\,\sigma_1^2, \] so the two scalar pictures differ by a compensating change of measure and pointwise multiplication. In particular, if \(\sigma_2=c\,\sigma_1\) is a constant rescaling, then \(f_2=c^{-1}f_1\) is the familiar global wavefunction normalization freedom.

Heuristic PA-H2.9 (How PA-H2.3 creates a scale). In the usual “scalar wavefunction” presentation on \(\mathbb R^d\), one implicitly chooses \(\sigma_\ast=|dx|^{1/2}\) and allows the scalar representative to carry dimension \(\text{length}^{-d/2}\) so that \(|\psi|^2\,d^dx\) is dimensionless probability. Requiring instead that the scalar representative \(f\) be dimensionless (PA-H2.3) shifts the \(\text{length}^{-d/2}\) factor into the reference half-density: \[ \sigma_\ast \sim L_\ast^{-d/2}|dx|^{1/2}, \] so “dimensionless \(f\)” is a convention unless the scale \(L_\ast^{d/2}\) is fixed by an additional universality principle (PA-H2.4–PA-H2.5).

3.5 Running PA-H2.4: What Does “Background-Free Constancy” Mean?

From Derivation PA-D1.7, changing the reference half-density \(\sigma_\ast\) by a positive function \(r(x)\) changes the scalar representative by \(f\mapsto r^{-1}f\) and changes the scalar measure by \(\rho_\ast\mapsto r^2\rho_\ast\). So the raw half-density formulation has a large “scalarization gauge freedom”.

Heuristic PA-H2.10 (Constancy = no extra background function). If we take “background-free” in the strong sense “no additional structure beyond the manifold and the theory’s couplings”, then allowing an arbitrary non-constant \(r(x)\) would amount to introducing a new background field/function by hand. In that strong sense, the only admissible changes are constant rescalings, and choosing \(\sigma_\ast\) becomes a choice of a single global scale (fixed or not fixed by couplings depending on PA-H2.5).

Derivation PA-D1.8 (Three natural families of \\(\sigma_\ast\\) and what they mean). On a configuration space \(M\), the common ways to choose a reference half-density are: 1. Flat/affine choice (when available): on \(\mathbb R^d\) with its affine structure, translation invariance picks \(\lvert dx\rvert^{1/2}\) uniquely up to a constant factor. This is “constant” in the sense of being homogeneous under translations. 2. Metric-derived choice: given a Riemannian/Lorentzian metric \(g\), one can take \(\sigma_g := |g|^{1/4}|dx|^{1/2}\), so that \(\rho_g=\sigma_g^2=\sqrt{|g|}\,|dx|\) is the familiar invariant volume density. This makes the scalar representative \(f\) a genuine scalar field but makes the scalarization depend on background geometry. 3. Field-derived (dilaton-like) choice: given a scalar field \(\Phi\) (background or dynamical), one can take \(\sigma_{\Phi}:=e^{-\Phi}\sigma_g\). In the scalar picture this is a local rescaling of the measure, and it is the natural way to encode “local units” or Weyl factors.

PA-H2.4 asserts that the theory supplies (or selects) a choice of type (1) with no \(x\)-dependent factor: a fixed reference \(\sigma_\ast\) whose only remaining ambiguity is an overall constant scale.

Heuristic PA-H2.11 (RG as scale dependence of scalarization). If refinement/coarse-graining forces an \(x\)-independent but scale-dependent choice \(\sigma_\ast(\mu)\) (equivalently a scale-dependent constant \(L_\ast(\mu)\)), then PA-H2.4 is replaced by an RG statement: the scalarization convention becomes part of the renormalization scheme (a “wavefunction renormalization” for the scalar representative). In that case, a universal area/length scale can still appear, but typically as an RG invariant (dimensional transmutation scale) rather than as a fixed analytic monomial in couplings.

Derivation PA-D1.8a (Running scalarization \\(\sigma_\ast(\mu)\\) is a \\(Z(\mu)\\) factor on scalar representatives). Assume the intrinsic (half-density) state \(\psi\) is fixed, and consider two scalarizations related by a \(\mu\)-dependent constant rescaling, \(\sigma_\ast(\mu)=c(\mu)\,\sigma_0\) with \(c(\mu)>0\). Writing \(\psi=f(\mu)\,\sigma_\ast(\mu)\) gives \(f(\mu)=c(\mu)^{-1}f_0\), where \(f_0\) is the scalar representative in the \(\sigma_0\) convention. Defining \(Z(\mu):=c(\mu)^2\), this becomes the familiar multiplicative form \[ f(\mu)=Z(\mu)^{-1/2}\,f_0. \] So an \(x\)-independent running of scalarization is formally equivalent to a wavefunction renormalization factor on scalar representatives. This is bookkeeping: \(\sigma_\ast(\mu)\) is a convention/scheme choice; only RG-invariant combinations (e.g. transmutation scales) are candidates for physical statements.

Heuristic PA-H2.11a (Guardrail: geometric weight \\(\neq\\) anomalous dimension). The half-density square-root Jacobian is a geometric transformation law under coordinate changes; anomalous dimensions are RG scaling data in interacting theories. They should not be conflated: allowing \(\sigma_\ast(\mu)\) to run is a representation convenience, not a claim that geometry “produces” anomalous scaling.

4. Stationary Phase Produces Half-Density Prefactors (Short-Time Kernel)

In the standard path-integral formalism, stationary phase explains why classical extremals dominate refinement limits. Here we add the complementary kernel-level fact: stationary phase does not only pick the extremal; it also produces a determinant prefactor that transforms as a half-density, i.e. the object needed for coordinate-free kernel composition.

Derivation PA-D1.4 (Van Vleck prefactor is a bi-half-density). Let \(S_{\mathrm{cl}}(x,z;t)\) be the classical action as a function of endpoints and time, treated as a generating function. The standard short-time/stationary-phase approximation to the propagator has the form

\[K(x,z;t) \approx \frac{1}{(2\pi i\hbar)^{d/2}} \left|\det\!\left(-\frac{\partial^2 S_{\mathrm{cl}}}{\partial x\,\partial z}\right)\right|^{1/2} \exp\!\left(\frac{i}{\hbar}S_{\mathrm{cl}}(x,z;t)\right).\]

Under a change of coordinates \(x=x(x’)\), \(z=z(z’)\), the mixed Hessian transforms by the chain rule, and its determinant acquires Jacobian factors:

\[\det\!\left(-\frac{\partial^2 S_{\mathrm{cl}}}{\partial x'\,\partial z'}\right) = \det\!\left(\frac{\partial x}{\partial x'}\right) \det\!\left(\frac{\partial z}{\partial z'}\right) \det\!\left(-\frac{\partial^2 S_{\mathrm{cl}}}{\partial x\,\partial z}\right).\]

Taking square roots shows that the prefactor transforms with \(|\det(\partial x/\partial x’)|^{1/2}|\det(\partial z/\partial z’)|^{1/2}\), i.e. exactly as a half-density factor at each endpoint. Thus the stationary-phase prefactor is naturally interpreted as making \(K\) a half-density in each variable, so that kernel composition does not depend on a background measure choice. This is the standard “Van Vleck type” semiclassical prefactor in the correspondence/semiclassical tradition [VanVleck1928Correspondence] [Morette1951] [DeWitt1957CurvedSpaces].

Derivation PA-D1.4a (Free-propagator semigroup: the \\(d/2\\) exponent is forced by composition). The free quantum propagator on \(\mathbb{R}^d\), \(K(x,z;t) = \left(\frac{m}{2\pi i\hbar t}\right)^{d/2}\exp!\left(\frac{im|x-z|^2}{2\hbar t}\right)\), provides a concrete witness. As a bi-half-density, the product \(\mathbf{K}(x,y;t_1)\,\mathbf{K}(y,z;t_2)\) pairs the two \(|dy|^{1/2}\) factors into a density \(|dy|\), which integrates without a background measure. The \(d\)-dimensional Gaussian integral over \(y\) gives a volume factor \(\left(\frac{2\pi i\hbar\,t_1 t_2}{m(t_1+t_2)}\right)^{d/2}\). Combining with the two kernel prefactors:

\[\left(\frac{m}{2\pi i\hbar t_1}\right)^{d/2} \left(\frac{m}{2\pi i\hbar t_2}\right)^{d/2} \left(\frac{2\pi i\hbar\,t_1 t_2}{m(t_1+t_2)}\right)^{d/2} =\left(\frac{m}{2\pi i\hbar(t_1+t_2)}\right)^{d/2},\]

which is the prefactor of \(K(x,z;t_1+t_2)\). The semigroup property holds because the \(d/2\) exponent appears three times (twice from the kernels, once from the Gaussian) and the cancellation is exact only for this exponent. This is the Van Vleck determinant in disguise: for the free particle, \(\det(-\partial^2 S_{\mathrm{cl}}/\partial x\,\partial z)=(m/t)^d\), so the square root is \((m/t)^{d/2}\).

In the tangent-groupoid near-diagonal picture, the parameter \(\varepsilon=\hbar t/m\) (diffusion scale) plays the role of the rescaling parameter in PA-D1.2a: the prefactor is proportional to \(\varepsilon^{-d/2}\), and the passage from kernel to symbol on \(TM\) absorbs this half-density Jacobian.

Heuristic PA-D1.9 (Square-root delta normalization has half-density weight). In finite dimension, the “localize on critical points” distribution is \(\delta(\nabla f)\), supported on \(\mathrm{Crit}(f)\). A concrete way it appears is via a “halved” oscillatory integral with a normalization exponent fixed by dimension. (The \(\varepsilon\to 0\) limit below is formal; a rigorous treatment requires distributional convergence in the Schwartz topology, e.g. via stationary-phase estimates in the distributional sense — see Hörmander, The Analysis of Linear Partial Differential Operators I, §7.7.)

Let \(f:\mathbb R^N\to\mathbb R\) be smooth and define, for \(\varepsilon>0\), \[ A_\varepsilon(O):=\varepsilon^{-N/2}\int_{\mathbb R^N} e^{\frac{i}{\varepsilon}f(x)}\,O(x)\,dx. \] Then \[ |A_\varepsilon(O)|^2 =\varepsilon^{-N}\int!!\int e^{\frac{i}{\varepsilon}(f(x)-f(y))}\,O(x)\,\overline{O(y)}\,dx\,dy. \] Applying the near-diagonal scaling \(y=x+\varepsilon z\) (so \(dy=\varepsilon^Ndz\)) gives \[ |A_\varepsilon(O)|^2 =\int!!\int e^{-\frac{i}{\varepsilon}(f(x+\varepsilon z)-f(x))}\,O(x)\,\overline{O(x+\varepsilon z)}\,dz\,dx. \] Formally letting \(\varepsilon\to 0\) yields \[ |A_\varepsilon(O)|^2 \;\to\;\int!!\int e^{-i z\cdot \nabla f(x)}\,|O(x)|^2\,dz\,dx =(2\pi)^N\int \delta(\nabla f(x))\,|O(x)|^2\,dx. \] The exponent \(N/2\) in the prefactor is exactly the half-density scaling: it cancels the Jacobian \(dy=\varepsilon^Ndz\) under near-diagonal rescaling, and it is the “square root” of the density normalization that produces \(\delta(\nabla f)\).

Remark PA-H1.4 (Where Planck area can enter, minimally — summary of Derivation PA-D1.3). For the reader’s convenience, we restate the conclusion of Derivation PA-D1.3 in compact form: if the theory supplies a single universal coupling with dimension of length (Newton’s constant) and one demands that the half-density normalization constant be built from that coupling without fractional powers, then \(d=4\) is singled out and the resulting constant has the dimension of an area, naturally identified with the Planck area \(L_P^2\sim \hbar G_4/c^3\).

The broader program in which this note sits argues that: 1. classical dynamics are recovered from quantum composition by stationary-phase concentration, and 2. refinement across scales forces RG-style consistency conditions when naive limits diverge.

This note adds a complementary ingredient: the kernel side is most naturally formulated in half-density language, and stationary phase produces the bi-half-density prefactor directly. A universal convention for turning those half-densities into scalar amplitudes then requires a \(\text{length}^{d/2}\) scale; in \(d=4\) this is an area scale.

6. Open Problems and Outlook

(Addressed: Derivation PA-D1.4a.) The free-propagator semigroup computation provides a concrete witness; a full tangent-groupoid treatment remains desirable.
(Addressed: Remark PA-E5c.) The Van Vleck / transmutation separation is explicit in the 2D delta model; a \(d=4\) gauge-theory witness is still open.
(Addressed: Remark PA-D1.3b.) General-dimension analysis: what replaces “area” in odd dimensions. The monomial sieve is an even-\(d\) filter (parity lemma), while the transmutation route is parity-blind. An explicit odd-\(d\) witness exists: in \(d=3\), the delta-interaction scattering length \(a\) (generated via power-law renormalization) supplies \(a^{3/2}=\text{length}^{3/2}=\text{length}^{d/2}\) without fractional coupling powers, since \(a\) is a derived RG-invariant scale. A universal normalization remains defensible in odd \(d\) provided one admits transmutation-generated scales.
(Addressed: Remark PA-H2.6b.) The minimal hypothesis set is PA-H2.1 + PA-H2.4 + PA-H2.5a + gravity-only; each is necessary.
Track minimal-length/GUP scenarios as a comparison branch: do they implement the “needed scale” at the level of kinematics (modified commutators/dispersion) or can they be reframed as a refinement-compatibility condition? Use [Hossenfelder2013MinimalLength] as an OA entry point.

Heuristic PA-H4.1 (Half-density and the Misner measure problem). For finite-dimensional configuration spaces (minisuperspace), half-density wavefunctions uniquely resolve the Hilbert space structure ambiguity identified by Misner (1962) and DeWitt (1967): the factor \(\sqrt{\gamma}\) in Wheeler–DeWitt wavefunctions is the finite-dimensional half-density weight (Derivation PA-D1.2a, restricted to minisuperspace coordinates). For the infinite-dimensional case, the half-density approach reduces — but does not eliminate — the Misner ambiguity: the residual freedom is in the choice of Riemannian structure on superspace. The discrete area spectrum of loop quantum gravity, \(A_n = 8\pi L_p^2\sqrt{j(j+1)}\), provides a candidate regularization compatible with \(L_p^2\) as the half-density area scale, but the connection between the Ashtekar–Lewandowski measure and half-density composition is not yet established.

Conjecture PA-C4.2 (d=4 b-calculus contact spectrum: analytic, not non-perturbative). The b-Laplacian \(-\Delta_b\) on \(\mathbb{R}^4\setminus{0}\) — viewed as a manifold with boundary after real blow-up of the origin (replacing \(r=0\) by the front face \(S^3\)) — in \(L^2b = L^2(\mathbb{R}^4\setminus{0}, r^{-4}d^4x)\), s-wave sector, is expected to have deficiency indices \(n\pm=1\) (one-parameter Robin self-adjoint extension family). The conjectured bound-state spectrum under the Robin APS boundary condition at the blown-up face \(r = L_0\) is: \[ E_B = -\frac{\hbar^2}{2mL_0^2}\,\mu_{\mathrm{APS}}(\mu_{\mathrm{APS}}-2), \qquad \mu_{\mathrm{APS}} < 0, \] where \(\mu_{\mathrm{APS}}\in\mathbb{R}\) is the Robin extension parameter and \(\kappa = \sqrt{-2mE_B}/\hbar\) is the bound-state inverse length. The constraint \(\sqrt{1+\kappa^2 L_0^2}=1-\mu_{\mathrm{APS}}>0\) restricts to \(\mu_{\mathrm{APS}}<0\); the \(\mu_{\mathrm{APS}}>2\) branch is a squaring artifact. The deficiency-index claim (that \(n_\pm = 1\) for the b-Laplacian restricted to the s-wave sector on \(\mathbb{R}^d\setminus{0}\) in the b-calculus measure, for all \(d \geq 2\)) and spectrum formula require a self-contained proof or reference to a completed companion paper; they are stated here as a conjecture pending such verification.

Contrast with d=2 (if the conjecture holds): In the 2D delta interaction, the bound state energy \(E_B \propto e^{-1/g_R(\mu)}\) is genuinely non-perturbative (exponentially small in the coupling) and requires dimensional transmutation. In the conjectured d=4 b-calculus spectrum, the dependence on \(\mu_{\mathrm{APS}}\) is algebraic: no RG running, no beta function, no dimensional transmutation. The UV scale \(L_0\) (at which the APS condition is imposed, e.g. \(L_0=L_P\)) enters directly as a scale parameter, not through a running coupling.

Conjectured implication for the area-scale program: If the deficiency-index and spectral claims hold, then the extension parameter \(\mu_{\mathrm{APS}}\) is a free dimensionless parameter not fixed by the b-calculus structure, composition, or dimensional analysis — requiring additional physical input (scattering data or Planck-scale boundary condition). The b-calculus would select the type of contact interaction (Robin at \(r=L_0\)) but not its strength \(\mu_{\mathrm{APS}}\). This would be distinct from the transmutation route (PA-H2.13) where the coupling itself fixes the scale.

References

[ManuelTarrach1994PertRenQM] Cristina Manuel and Rolf Tarrach, “Perturbative Renormalization in Quantum Mechanics,” Physics Letters B 328 (1994), 113–118. arXiv:hep-th/9309013 (v1, 2 Sep 1993). DOI 10.1016/0370-2693(94)90437-5.
[VanVleck1928Correspondence] J. H. Van Vleck, “The Correspondence Principle in the Statistical Interpretation of Quantum Mechanics,” Proceedings of the National Academy of Sciences of the United States of America 14(2) (1928), 178–188. DOI 10.1073/pnas.14.2.178.
[Hossenfelder2013MinimalLength] Sabine Hossenfelder, “Minimal Length Scale Scenarios for Quantum Gravity,” Living Reviews in Relativity 16 (2013), 2. arXiv:1203.6191 (v2, 30 Apr 2013). DOI 10.12942/lrr-2013-2.
[Greensite2003Confinement] Jeff Greensite, “The Confinement Problem in Lattice Gauge Theory,” arXiv:hep-lat/0301023 (v2, 5 Mar 2003), Progress in Particle and Nuclear Physics 51 (2003) 1. DOI 10.1016/S0146-6410(03)90012-3. (Review; includes Wilson-loop definition of static potential and area-law/string-tension criterion.)
[Morette1951] Cécile Morette, “On the Definition and Approximation of Feynman’s Path Integrals,” Physical Review 81 (1951), 848–852. DOI 10.1103/PhysRev.81.848.
[DeWitt1957CurvedSpaces] Bryce S. DeWitt, “Dynamical Theory in Curved Spaces. I. A Review of the Classical and Quantum Action Principles,” Reviews of Modern Physics 29(3) (1957), 377–397. DOI 10.1103/RevModPhys.29.377.
[Vassilevich2003HeatKernel] D. V. Vassilevich, “Heat kernel expansion: user’s manual,” Physics Reports 388 (2003), 279–360. arXiv:hep-th/0306138. DOI 10.1016/j.physrep.2003.09.002.
[Woodhouse1992] N. M. J. Woodhouse, Geometric Quantization, 2nd ed., Oxford Mathematical Monographs, Oxford University Press, 1992. ISBN 0-19-853673-9. (Standard reference for half-density line bundles and BKS pairing in geometric quantization.)
[BatesWeinstein1997] Sean Bates and Alan Weinstein, Lectures on the Geometry of Quantization, Berkeley Mathematics Lecture Notes vol. 8, AMS, 1997. ISBN 0-8218-0798-9. OA: https://math.berkeley.edu/~alanw/GofQ.pdf. (Half-densities as sections of a line bundle, metalinear structures, composition.)
[PathIntegralNormalization] A. Rivero and A.I.Scaffold, “Path-Integral Normalization: The d/2 Exponent as Composition Compatibility Datum,” companion satellite paper in this series (2026). (Comprehensive treatment of the d/2 exponent across temporal composition, Van Vleck determinant, heat-kernel diffusion, renormalization thresholds, and Lévy-stable exclusion.)

Planck Area from Half-Density Normalization