Schoenberg Type Inequalities

• Schoenberg type inequalities are a collection of analytic estimates comparing objects to their transforms using positivity and smoothness conditions across polynomial geometry, approximation, and harmonic analysis.
• They provide sharp bounds that relate the magnitudes of polynomial zeros and critical points, error measures in spline approximation, and spectral characteristics in commutative as well as noncommutative frameworks.
• Unified via operator interpolation and functional calculus, these inequalities bridge classical moment identities with modern positivity-preserving methodologies, impacting diverse areas of analysis.

Schoenberg type inequalities are a family of estimates and correspondence principles associated with the work of I. J. Schoenberg that occur in several distinct parts of analysis. In the literature, the expression refers not to a single theorem but to a cluster of results with a common structural theme: a quantitative comparison between an object and a transformed object, typically mediated by positivity, smoothness, or spectral structure. In the sources considered here, the term covers inequalities for zeros and critical points of polynomials, lower and upper error bounds for Schoenberg operators in spline approximation, Bochner-Schoenberg-Eberlein bounds on spectra of commutative topological algebras, and positivity-preserving correspondences for kernels, semigroups, and noncommutative products (Tang, 14 Aug 2025, Nagler et al., 2013, Amiri et al., 2020, Schürmann et al., 2012).

1. Scope of the term

A central misconception is that “Schoenberg type inequality” denotes only the classical quadratic inequality for polynomial critical points. The current literature uses the phrase in a broader sense. In polynomial geometry, it means inequalities relating the moduli of critical points to the moduli of zeros. In approximation theory, it means direct and inverse estimates comparing the error of the Schoenberg operator with a modulus of smoothness. In harmonic and functional analysis, it appears as the Bochner-Schoenberg-Eberlein inequality on the spectrum of a commutative Banach or Fréchet algebra. In the theory of positive definite kernels and noncommutative probability, it appears through Schoenberg correspondences, coefficient relations, and positivity-preserving functional calculi (Tang, 14 Apr 2025, Nagler et al., 2013, Amiri et al., 2020, Pascoe, 2019).

Setting	Typical objects	Representative statement
Polynomial geometry	zeros $z_j$, critical points $w_k$	bounds for $\sum |w_k|^p$ in terms of $\sum |z_j|^p$
Spline approximation	Schoenberg operators $S_{n,k}$	equivalence of approximation error and $\omega_2(f,t)$
Commutative topological algebras	BSE-functions on $\Delta(A)$	$$\left|\sum c_i \sigma(\varphi_i)\right|\le \beta_M P_M(\sum c_i\varphi_i)$$
Positive definiteness and correspondence	kernels, semigroups, generalized Schur products	conditional positivity implies positive semigroups or positive functional calculus

This multiplicity is not merely terminological. A plausible implication is that the modern usage identifies a recurrent Schoenbergian pattern: an analytic or algebraic generator is controlled through a positivity condition, and that positivity propagates to a transformed family.

2. Polynomial geometry: from quadratic to all orders

For a monic polynomial

$p(z)=\prod_{j=1}^n (z-z_j),$

with critical points $w_1,\dots,w_{n-1}$, the classical Schoenberg inequality is the order-$w_k$0 case under the centroid condition

$w_k$1

In the 2025 interpolation framework, this becomes the sharp estimate

$w_k$2

and more generally, for all $w_k$3,

$w_k$4

while for $w_k$5,

$w_k$6

These inequalities are stated as sharp for all $w_k$7, and the method uses complex interpolation between $w_k$8 spaces and Schatten $w_k$9-classes, with the compression

$\sum |w_k|^p$0

for which

$\sum |w_k|^p$1

on the centroid-zero subspace (Tang, 14 Aug 2025).

Before that interpolation result, higher-order progress was uneven. The order-$\sum |w_k|^p$2 case had been established earlier, and a 2025 paper produced an explicit sharp order-$\sum |w_k|^p$3 inequality: $\sum |w_k|^p$4 with equality if and only if the zeros lie on a straight line. The same paper also gave an order-$\sum |w_k|^p$5 inequality and described it as the first known odd-order result, thereby partially answering two open problems posed by Kushel and Tyaglov and connecting the general framework to Sendov’s conjecture (Tang, 14 Apr 2025).

The methodological shift is significant. Earlier even-order results depended on explicit trace calculations, companion-matrix techniques, and majorization. The interpolation approach replaces this case-by-case structure with a continuous $\sum |w_k|^p$6-scale. This suggests that the polynomial branch of Schoenberg type inequalities has moved from isolated moment identities to operator-space methods (Tang, 14 Aug 2025).

3. Dual and noncommutative polynomial variants

A different direction asks for a reverse estimate: can the zeros be bounded in terms of the critical points? The dual Schoenberg type inequality answers this in the affirmative, but with an additional correction term. For a degree-$\sum |w_k|^p$7 polynomial $\sum |w_k|^p$8 with simple critical points $\sum |w_k|^p$9 and mean

$\sum |z_j|^p$0

the bound is

$\sum |z_j|^p$1

The proof uses matrix differentiation and integration, a characterization of integrability for diagonalizable matrices, and Schur’s inequality

$\sum |z_j|^p$2

In this setting, Schoenberg type inequality no longer means a one-way contraction from zeros to critical points; it means a two-sided geometric relation, although the reverse direction requires extra data (Danielyan et al., 2020).

The noncommutative analogue is the C$\sum |z_j|^p$3-algebraic Schoenberg conjecture. Let

$\sum |z_j|^p$4

over a C$\sum |z_j|^p$5-algebra $\sum |z_j|^p$6, and suppose

$\sum |z_j|^p$7

The conjecture asserts the operator inequalities

$\sum |z_j|^p$8

and

$\sum |z_j|^p$9

The scalar case reduces to the Malamud-Pereira theorem, while the paper proves the conjecture for degree $S_{n,k}$0, with equality. For $S_{n,k}$1, it remains open (Krishna, 2022).

These developments show that the polynomial meaning of Schoenberg type inequalities now has three layers: the classical direct inequality, dual inequalities, and noncommutative operator-order analogues.

4. Approximation theory and the Schoenberg operator

In spline approximation, Schoenberg type inequalities compare the error of a Schoenberg operator with a modulus of smoothness. For uniform knots and splines of degree $S_{n,k}$2, the uniform Schoenberg operator $S_{n,k}$3 satisfies the lower bound

$S_{n,k}$4

for all continuous $S_{n,k}$5 on $S_{n,k}$6, where

$S_{n,k}$7

The same work establishes the two-sided equivalence

$S_{n,k}$8

with constants depending on $S_{n,k}$9 and $\omega_2(f,t)$0 but not on $\omega_2(f,t)$1. A key technical inequality is

$\omega_2(f,t)$2

combined with the spline-space estimate

$\omega_2(f,t)$3

The paper presents this as the first direct lower bound for the uniform Schoenberg operator in terms of the classical second order modulus of smoothness and as a confirmation of a conjecture posed by Beutel et al. for fixed degree and vanishing mesh size (Nagler et al., 2013).

A broader spline setting appears in quasi-collocation with CCC-Schoenberg operators. There the operator is

$\omega_2(f,t)$4

where $\omega_2(f,t)$5 is a CCC B-spline basis and $\omega_2(f,t)$6 are typically CCC-Greville abscissas. The principal approximation estimate is

$\omega_2(f,t)$7

with

$\omega_2(f,t)$8

When a second-order boundary value problem is solved by quasi-collocation through

$\omega_2(f,t)$9

the solution error satisfies

$\Delta(A)$0

In this branch of the theory, Schoenberg type inequalities are approximation-order statements: they certify that the operator is a second-order approximation process and transmit that order to numerical schemes (Bosner, 2021).

5. Positive definite kernels, spheres, and Schoenberg coefficients

A different Schoenbergian tradition concerns positive definite functions on spheres and their coefficient sequences. For a real sphere $\Delta(A)$1, a continuous isotropic positive definite function has a Gegenbauer expansion with coefficients $\Delta(A)$2, the $\Delta(A)$3-Schoenberg sequence. For a complex sphere $\Delta(A)$4, the analogous coefficients are $\Delta(A)$5 in disk-polynomial expansions. Explicit dimension-walk relations are known in both directions. In the real case, one forward recursion is

$\Delta(A)$6

and the inverse formula is

$\Delta(A)$7

In the complex case,

$\Delta(A)$8

These relations are applied to strict positive definiteness and show that coefficient data can be transferred across dimensions (Bissiri et al., 2018).

The extension from coefficients to functions is given on product spaces $\Delta(A)$9, with $$\left|\sum c_i \sigma(\varphi_i)\right|\le \beta_M P_M(\sum c_i\varphi_i)$$0 a locally compact group. The class $$\left|\sum c_i \sigma(\varphi_i)\right|\le \beta_M P_M(\sum c_i\varphi_i)$$1 consists of continuous functions $$\left|\sum c_i \sigma(\varphi_i)\right|\le \beta_M P_M(\sum c_i\varphi_i)$$2 such that

$$\left|\sum c_i \sigma(\varphi_i)\right|\le \beta_M P_M(\sum c_i\varphi_i)$$3

is positive definite. The generalized Schoenberg expansion is

$$\left|\sum c_i \sigma(\varphi_i)\right|\le \beta_M P_M(\sum c_i\varphi_i)$$4

with uniform convergence, $$\left|\sum c_i \sigma(\varphi_i)\right|\le \beta_M P_M(\sum c_i\varphi_i)$$5, and $$\left|\sum c_i \sigma(\varphi_i)\right|\le \beta_M P_M(\sum c_i\varphi_i)$$6. For abelian $$\left|\sum c_i \sigma(\varphi_i)\right|\le \beta_M P_M(\sum c_i\varphi_i)$$7, each $$\left|\sum c_i \sigma(\varphi_i)\right|\le \beta_M P_M(\sum c_i\varphi_i)$$8 admits a Fourier-type representation by a positive finite measure on $$\left|\sum c_i \sigma(\varphi_i)\right|\le \beta_M P_M(\sum c_i\varphi_i)$$9. This extends Schoenberg’s theorem from scalar coefficients to function-valued Schoenberg coefficients and is explicitly motivated by space-time covariance models on the sphere (Berg et al., 2015).

The dimensional continuation problem for radial positive definite functions on $p(z)=\prod_{j=1}^n (z-z_j),$0 introduces a further inequality-like phenomenon. If $p(z)=\prod_{j=1}^n (z-z_j),$1 denotes the class of radial positive definite functions on $p(z)=\prod_{j=1}^n (z-z_j),$2, then $p(z)=\prod_{j=1}^n (z-z_j),$3. A transition formula characterizes when $p(z)=\prod_{j=1}^n (z-z_j),$4 actually belongs to $p(z)=\prod_{j=1}^n (z-z_j),$5 in terms of the Schoenberg representation measure $p(z)=\prod_{j=1}^n (z-z_j),$6. The paper also proves directly that the Schoenberg kernel $p(z)=\prod_{j=1}^n (z-z_j),$7 lies in $p(z)=\prod_{j=1}^n (z-z_j),$8, that

$p(z)=\prod_{j=1}^n (z-z_j),$9

and that

$w_1,\dots,w_{n-1}$0

Moreover, if $w_1,\dots,w_{n-1}$1 with $w_1,\dots,w_{n-1}$2, then for every $w_1,\dots,w_{n-1}$3 there exists a finite Schoenberg matrix

$w_1,\dots,w_{n-1}$4

with at least $w_1,\dots,w_{n-1}$5 negative eigenvalues. This gives a strong form of failure of positive definite continuation (Golinskii et al., 2015).

6. Bochner-Schoenberg-Eberlein inequalities and noncommutative correspondences

In commutative Banach and Fréchet algebras, the Bochner-Schoenberg-Eberlein property replaces metric or polynomial data by spectral data on the Gelfand spectrum. For a commutative semisimple Fréchet algebra $w_1,\dots,w_{n-1}$6, a bounded continuous function $w_1,\dots,w_{n-1}$7 is a BSE-function if there exist a bounded set $w_1,\dots,w_{n-1}$8 and $w_1,\dots,w_{n-1}$9 such that

$w_k$00

for all finite scalar sequences and characters. The resulting algebra $w_k$01 is itself a commutative semisimple Fréchet algebra. One of the main structural statements is

$w_k$02

There is also a characterization of unitality in terms of

$w_k$03

which is described as a strict Schoenberg-type inequality for the constant function. The direct sum $w_k$04 is a BSE-algebra if and only if $w_k$05 and $w_k$06 are BSE-algebras, and commutative semisimple Fréchet C$w_k$07-algebras and commutative semisimple uniform Fréchet algebras are BSE-algebras via inverse-limit arguments (Amiri et al., 2020, Amiri et al., 2020, Amiri et al., 2020).

In noncommutative probability, Schoenberg correspondence generalizes the classical principle that conditional positivity exponentiates to positivity. On a dual semigroup $w_k$08, if $w_k$09 is conditionally positive, then the convolution exponential

$w_k$10

defines a convolution semigroup $w_k$11 of states. The proof uses approximations of the form

$w_k$12

together with an estimate

$w_k$13

A related 2023 result extends this to non-unital semigroups on $w_k$14: for a closed convex cone $w_k$15 and an idempotent $w_k$16, one has

$w_k$17

if and only if $w_k$18 is $w_k$19-conditionally positive on $w_k$20. This characterizes generators of semigroups that are $w_k$21-positive, $w_k$22-superpositive, or $w_k$23-entanglement breaking, and recovers the Lindblad-Gorini-Kossakowski-Sudarshan theorem as a corollary (Schürmann et al., 2012, Bhat et al., 2023).

A further noncommutative extension appears in generalized Schur products. Products on matrix spaces that preserve rank one and positive semidefiniteness are classified as generalized Schur products, and the corresponding Schoenberg theorem states that a positivity-preserving noncommutative function has an expansion

$w_k$24

with all $w_k$25, where powers are taken with respect to the generalized product. This places Schoenberg type inequalities and correspondences within a broad functional-calculus framework (Pascoe, 2019).

Taken together, these developments show that Schoenberg type inequalities now designate a mathematically diverse but structurally coherent body of results. The common thread is the passage from an underlying positivity or smoothness condition to a transformed object whose size, spectrum, or approximation error is sharply controlled.