---
title: Skew von Neumann Constant
url: https://www.emergentmind.com/topics/skew-von-neumann-constant
type: topic
---

# Skew von Neumann Constant

The expression **skew von Neumann constant** does not designate a single universally fixed invariant in the current literature. In the materials considered here, it appears in three principal senses. In Banach-space geometry it denotes a family of weighted or skewed von Neumann–Jordan-type constants built from the pair of combinations \(\lambda x+\mu y\) and \(\mu x-\lambda y\). In operator algebra it denotes, in an interpretive sense, the central self-adjoint unitary \(Z\) that classifies strong skew commutativity preserving maps on a von Neumann algebra. In a separate entropy-theoretic setting it can be interpreted as the limiting skewness of the von Neumann entanglement entropy distribution, and in that natural asymptotic regime the limiting value is \(0\) [2411.10094, 1204.1841, 1910.01199]. Earlier work on upper, lower, and modified \(n\)-th von Neumann–Jordan constants did not use the phrase as a formal term, but explicitly treated such one-sided or sphere-restricted modifications as conceptually “skew” variants [1811.01652]. The phrase should also be distinguished from the **Murray–von Neumann coupling constant** \(y(M)\) attached to a representation of the universal von Neumann algebra of a smooth four-manifold [1712.01828].

## 1. Terminological scope and principal usages

Within Banach-space geometry, the term belongs to the theory of von Neumann–Jordan constants and their weighted generalizations. The basic idea is to replace the symmetric pair \(x+y\), \(x-y\) by the skewed pair \(\lambda x+\mu y\), \(\mu x-\lambda y\), and then normalize by a power of \(\lambda^p+\mu^p\) and the norms of \(x\) and \(y\). This produces a two-parameter family of geometric invariants sensitive to nonsquareness, uniform convexity, uniform smoothness, normal structure, and related properties [2411.10094, 2505.17628, 2506.13142].

Within von Neumann algebra theory, the phrase has a different status. The paper on strong skew commutativity preserving maps studies the skew Lie product \([A,B]^*:=AB-BA^*\) and shows that every surjective strong skew commutativity preserving map on a von Neumann algebra without central summands of type \(I_1\) is multiplication by a central self-adjoint unitary \(Z\) with \(Z^2=I\). The exposition explicitly suggests that this central symmetry can be thought of as a “skew von Neumann constant” because it is fixed and globally twists the skew commutator structure [1204.1841].

A third usage is only interpretive. In the study of random bipartite pure states, the skewness of the von Neumann entanglement entropy is computed from the third cumulant of the entropy distribution. The same source states that if one defines a “Skew von Neumann Constant” as the high-dimensional limiting skewness of the standardized entropy, then the constant is \(0\) [1910.01199].

| Context | Defining object | Representative source |
|---|---|---|
| Banach-space geometry | Weighted von Neumann–Jordan-type constant | [2411.10094] |
| Operator algebra | Central symmetry \(Z\in Z_s(M)\), \(Z^2=I\) | [1204.1841] |
| Entropy theory | Limiting skewness of von Neumann entropy | [1910.01199] |

## 2. Operator-algebraic meaning: the central symmetry \(Z\)

In the operator-algebraic setting, the relevant structure is the **skew Lie product**
\[
[A,B]^*:=AB-BA^*.
\]
A map \(\Phi:R\to R\) is **strong skew commutativity preserving** if
\[
[\Phi(A),\Phi(B)]^*=[A,B]^*
\]
for all \(A,B\). For a von Neumann algebra \(M\) without central summands of type \(I_1\), the principal theorem states that a surjective map \(\Phi:M\to M\) is strong skew commutativity preserving if and only if there exists an element \(Z\in Z_s(M)\) such that
\[
Z=Z^*,\qquad Z^2=I,\qquad \Phi(A)=ZA\quad\text{for all }A\in M.
\]
Here \(Z_s(M)=\{Z\in Z(M):Z=Z^*\}\), so \(Z\) is a **central symmetry**, equivalently a central self-adjoint unitary [1204.1841].

This theorem makes the “constant” interpretation precise. The map is completely determined by a fixed central element independent of \(A\) and \(B\), and the skew commutator identity reduces to the relations \(Z\in Z(M)\) and \(Z^2=I\). In a factor, where \(Z(M)=\mathbb CI\), the only possibilities are \(Z=\pm I\), so the only surjective strong skew commutativity preserving maps are \(\Phi(A)=A\) and \(\Phi(A)=-A\). In a direct sum decomposition into central summands, \(Z\) decomposes as a direct sum of signs on each summand. The same paper also shows that in prime \(*\)-algebras with involution of the second kind, the only strong skew commutativity preserving surjective maps are again \(\pm\mathrm{id}\), so the “skew constant” reduces to a global sign [1204.1841].

The significance of this usage is structural rather than metric. The object called a constant is not a scalar geometric modulus but a central symmetry that rigidly classifies all such preservers. In that sense, it is a classification parameter for skew commutator invariance.

## 3. Banach-space geometric constants: weighted and spherical forms

The main geometric usage begins from weighted analogues of the von Neumann–Jordan constant. One sphere-based formulation is the **skew generalized von Neumann–Jordan constant on the unit sphere**
\[
\widetilde C_\mathrm{NJ}^p(\xi,\nu,X)
=\sup \left\{
\frac{\|\xi x+\nu y\|^p+\|\nu x-\xi y\|^p}
{2^{p-1}(\xi^p+\nu^p)}
: x,y\in S_X
\right\},
\]
where \(X\) is a Banach space, \(\xi,\nu>0\), and \(1\le p<\infty\) [2411.10094]. This is the sphere version of earlier skew constants \(L_{YJ}(\xi,\nu,X)\) and \(C_\mathrm{NJ}^p(\xi,\nu,X)\), and it extends the classical symmetric case obtained when \(\xi=\nu\).

Its first universal estimates are
\[
\frac{(\xi+\nu)^p+|\nu-\xi|^p}{2^{p-1}(\xi^p+\nu^p)}
\le
\widetilde C_\mathrm{NJ}^p(\xi,\nu,X)
\le
\frac{(\xi+\nu)^p}{2^{p-2}(\xi^p+\nu^p)}.
\]
The upper bound is global over all Banach spaces, and the lower bound is attained by taking \(y=x\in S_X\). In the two-dimensional spaces \((\mathbb R^2,\|\cdot\|_1)\) and \((\mathbb R^2,\|\cdot\|_\infty)\), the constant attains the upper bound exactly:
\[
\widetilde C_\mathrm{NJ}^p(\xi,\nu,X)
=
\frac{(\xi+\nu)^p}{2^{p-2}(\xi^p+\nu^p)}.
\]
The same source also computes explicit values for a mixed \(l_1\)-\(l_\infty\) octagon-type norm, for \(l_r\) and \(L_r\) with \(r\ge2\), and for a modified \(c_0\)-type norm, thereby showing that the skew constant can interpolate between extremal nonsmooth behavior and more Hilbert-like behavior [2411.10094].

The geometric role of the constant is explicit. It is related to the classical sphere constant \(\widetilde C_{\mathrm{NJ}^{(p)}}(X)\), to the James constant \(J(X)\), and to the modulus of convexity \(\delta_X(\epsilon)\). In particular, the paper proves that the following are equivalent: \(\widetilde C_{\mathrm{NJ}^{(p)}}(X)=2\); for all \(\xi,\nu>0\),
\[
\widetilde C_\mathrm{NJ}^{p}(\xi,\nu, X)
=
\frac{(\xi+\nu)^{p}}{2^{p-2}(\xi^{p}+\nu^{p})};
\]
and the same equality for some \(\xi_0,\nu_0>0\). It also proves that \(X\) is uniformly non-square if and only if
\[
\widetilde C_\mathrm{NJ}^{p}(\xi,\nu,X)
<
\frac{(\xi+\nu)^{p}}{2^{p-2}(\xi^{p}+\nu^{p})}.
\]
Thus the skew constant functions as a quantitative detector of extremal nonsquareness [2411.10094].

## 4. Type and orthogonality variants

A distinct but closely related line replaces the sum of the two skew norms by their minimum. The **skew generalized von Neumann–Jordan type constant** is
\[
C_{L_-}(\lambda,\mu,X)
=
\sup\left\{
\frac{\min\{\|\lambda x_1+\mu x_2\|^p,\ \|\mu x_1-\lambda x_2\|^p\}}
{2^{p-3}(\lambda^p+\mu^p)(\|x_1\|^p+\|x_2\|^p)}
:
x_1,x_2\in X,\ (x_1,x_2)\neq(0,0)
\right\}.
\]
It satisfies the general bounds
\[
\frac{\min\{\lambda^p,\mu^p\}}{2^{p-3}(\lambda^p+\mu^p)}
\le
C_{L_-}(\lambda,\mu,X)
\le
2,
\]
and is stable under Banach–Mazur distance:
\[
\frac{1}{d(X,Y)^p}\,C_{L_-}(\lambda,\mu,Y)
\le
C_{L_-}(\lambda,\mu,X)
\le
d(X,Y)^p\,C_{L_-}(\lambda,\mu,Y).
\]
The same source shows that \(X\) is uniformly non-square if and only if
\[
C_{L_-}(\lambda,\mu,X)
<
\frac{(1+\mu)^p}{2^{p-2}(\lambda^p+\mu^p)},
\]
and derives a sufficient condition for normal structure from an inequality involving \(C_{L_-}(\lambda,\mu,X)\) and the weak orthogonality coefficient \(\omega(X)\) [2505.17628].

Another equivalent reformulation uses **isosceles orthogonality**. For \(0\le \alpha\le \frac12\), \(1\le p<\infty\), and \(x_1\perp_I x_2\), the constant
\[
C^{I}_{NJ}(\alpha,p,X)
=
\sup\left\{
\frac{\|\alpha x_1+(1-\alpha )x_2\|^{p}+\|(1-\alpha)x_1+\alpha x_2\|^{p}}
{\|x_1+x_2\|^{p}}
:
x_1,x_2\in X,\ (x_1,x_2)\neq(0,0),\ x_1\perp_{I}x_2
\right\}
\]
is shown to satisfy the exact identity
\[
C_{NJ}^I(\alpha,p,X)=\frac12\,\gamma_X^p(1-2\alpha),
\]
so it is equivalent to the generalized \(p\)-th von Neumann–Jordan constant through the one-parameter function \(\gamma_X^p\). The map \(\alpha\mapsto C_{NJ}^I(\alpha,p,X)\) is convex, non-decreasing, and continuous on \([0,\frac12]\), with
\[
(1-\alpha)^p+\alpha^p
\le
C^I_{NJ}(\alpha,p,X)
\le
2(1-\alpha)^p.
\]
The upper bound is attained in \(\ell_1^n\), \(\ell_\infty^n\), and \(C[a,b]\), while for \(2\le p<\infty\) the lower bound is attained in \(\ell_p\). The same framework yields a characterization of uniform smoothness via the limit of \(\sqrt[p]{2^{p-1}C^I_{NJ}(\alpha,p,X)}\) as \(\alpha\to\frac12\) [2506.13142].

Exact evaluation is also available in specific finite-dimensional model spaces. For the Banaś–Frączek space \(R_\lambda^2\), with \(\lambda>1\) and under the condition
\[
\lambda^2\left(1-\frac1{\lambda^2}\right)^{p/2}\ge 1,
\]
the skew generalized von Neumann–Jordan constant is
\[
C_{RJ}(\xi,\eta,R_\lambda^2)
=
\frac{
(\xi+\eta)^p
+
\big[(\xi+\eta)^2-\tfrac{4\xi\eta}{\lambda^2}\big]^{p/2}
}{
2^{p-1}(\xi^p+\eta^p)
}.
\]
In the symmetric specialization \(\xi=\eta=1\), this gives
\[
C_{NJ}^{(p)}(R_\lambda^2)
=
1+\left(1-\frac1{\lambda^2}\right)^{p/2},
\]
and for \(p=2\) it yields
\[
C_{NJ}(R_\lambda^2)=2-\frac1{\lambda^2}.
\]
These formulas place the skew constant within a long-standing program of exact geometric constant computation in nonclassical two-dimensional norms [2412.11665].

## 5. Quasi-Banach, weak Orlicz, and weak Lebesgue extensions

In quasi-Banach spaces, the relaxed triangle inequality forces a renormalization of the skew constant by quasi-triangle factors. Starting from the classical quasi-Banach adaptation of the von Neumann–Jordan constant, the weak-space paper introduces two skewed quasi-triangle constants
\[
C_{X_1}
=
\sup\left\{
\frac{\|\lambda f+\mu g\|_X}{\lambda\|f\|_X+\mu\|g\|_X}
:
f,g\in X,\ (f,g)\neq(0,0)
\right\},
\]
\[
C_{X_2}
=
\sup\left\{
\frac{\|\mu f-\lambda g\|_X}{\mu\|f\|_X+\lambda\|g\|_X}
:
f,g\in X,\ (f,g)\neq(0,0)
\right\},
\]
and then defines a **skew von Neumann–Jordan constant in a quasi-Banach space** by normalizing the numerator
\[
\|\lambda f+\mu g\|_X^{2}+\|\mu f-\lambda g\|_X^{2}
\]
with
\[
C_{X_1}C_{X_2}(\lambda^2+\mu^2)(\|f\|_X^{2}+\|g\|_X^{2}).
\]
The paper also introduces the generalized \(p\)-th version \(L_{YJ}^{C,p}(\lambda,\mu,X)\), obtained by replacing the exponent \(2\) by \(p>0\) [2508.06999].

The principal applications are to weak Orlicz and weak Lebesgue spaces. For an \(N\)-function \(\Phi\), the weak Orlicz quasi-norm is
\[
\|f\|_{w L^{\Phi}(\mathbb{R}^n)}
:=
\inf \left\{
b>0:
\sup _{t>0} \Phi(t)\left|\left\{x \in \mathbb{R}^n: \frac{|f(x)|}{b}>t\right\}\right| \leq 1
\right\},
\]
and a key lemma gives, for measurable \(E\subset\mathbb R^n\) with \(0<|E|<\infty\),
\[
\|\chi_E\|_{wL^{\Phi}(\mathbb{R}^{n})}
=
\frac{1}{\Phi^{-1}\!\left(\frac{1}{|E|}\right)}.
\]
Using this formula and disjointly supported characteristic functions, the paper derives lower bounds for \(L_{YJ}^{C}(wL^\Phi)\) and \(L_{YJ}^{C,p}(wL^\Phi)\) in terms of the doubling parameters
\[
\bar{\alpha}_{\Phi}
:=
\inf_{t>0}\frac{\Phi^{-1}(t)}{\Phi^{-1}(2t)},
\qquad
\bar{\beta}_{\Phi}
:=
\sup_{t>0}\frac{\Phi^{-1}(t)}{\Phi^{-1}(2t)},
\]
together with the quasi-triangle constants \(C_{\Phi_1}\) and \(C_{\Phi_2}\) [2508.06999].

When \(\Phi(u)=u^p\), one has \(wL^\Phi(\mathbb R^n)=wL^p(\mathbb R^n)\) and
\[
\bar{\alpha}_{\Phi}=\bar{\beta}_{\Phi}=2^{-1/p}.
\]
This specialization yields explicit lower bounds for both the quadratic skew constant and its generalized \(p\)-th version in weak Lebesgue spaces. The paper thereby extends the skew von Neumann–Jordan formalism from Banach spaces to quasi-Banach function spaces, and the correction by \(C_{X_1}C_{X_2}\) isolates the geometric distortion not already forced by quasi-norm subadditivity [2508.06999].

## 6. Entropy-skewness interpretation and the vanishing asymptotic constant

A completely different use of the phrase arises in random quantum states. For a bipartite system with reduced density matrix \(\rho_A\) and eigenvalues \(\lambda_1,\dots,\lambda_m\), the von Neumann entanglement entropy is
\[
S=-\mathrm{Tr}(\rho_A\ln\rho_A)=-\sum_{i=1}^m\lambda_i\ln\lambda_i.
\]
The paper computes the exact third cumulant \(\kappa_3\) of \(S\), complementing the previously known mean \(\kappa_1\) and variance \(\kappa_2\), and defines the skewness in the standard Pearson sense by
\[
\gamma_1
=
\mathbb{E}\!\left[\left(\frac{S-\kappa_1}{\sqrt{\kappa_2}}\right)^3\right]
=
\frac{\kappa_3}{\kappa_2^{3/2}}.
\]
This is an exact finite-size skewness of the von Neumann entropy distribution, not a Banach-space geometric constant [1910.01199].

The asymptotic result is decisive. In the regime
\[
m\to\infty,\qquad n\to\infty,\qquad \frac{m}{n}=c\in(0,1],
\]
the paper shows
\[
\kappa_2=\Theta\!\left(\frac1{n^2}\right),
\qquad
\kappa_3=\Theta\!\left(\frac1{n^4}\right),
\qquad
\gamma_1(m,n)\sim \frac{\mathrm{const}(c)}{n}\to 0.
\]
It therefore concludes that if one defines a “Skew von Neumann Constant” as the limiting skewness
\[
\gamma_1^\ast(c)
=
\lim_{m,n\to\infty,\ m/n\to c}\gamma_1(m,n),
\]
then the only possible constant is
\[
\gamma_1^\ast=0.
\]
In this interpretation, the phrase refers not to a nontrivial universal modulus but to the asymptotic disappearance of skewness as the standardized entropy becomes Gaussian [1910.01199].

The available literature therefore supports no single canonical definition of the skew von Neumann constant. The dominant usage is geometric: a family of weighted von Neumann–Jordan-type constants on Banach and quasi-Banach spaces, with exact formulas, comparison theorems, and consequences for uniform nonsquareness, smoothness, normal structure, and isomorphic classification. A second, operator-algebraic usage treats the central symmetry \(Z\) classifying strong skew commutativity preserving maps as the relevant “constant.” A third, entropy-theoretic usage identifies the asymptotic limiting skewness of the von Neumann entropy distribution, yielding the trivial limit \(0\). Across these settings, the common theme is the replacement of a symmetric von Neumann-type quantity by a skewed, weighted, or asymmetry-sensitive analogue.

Source: https://www.emergentmind.com/topics/skew-von-neumann-constant