Orthogonally additive polynomials on convolution algebras associated with a compact group (Q2414751)

Let $X$ and $Y$ be complex linear spaces and $A$ be an algebra. A map $P : X \to Y$ is said to be an $n$-homogeneous polynomial if there exists an $n$-linear map $\varphi: X^n \to Y$ such that $P(x) = \varphi (x, \ldots , x)$ for all $x \in X$. A map $P: A \to Y$ is said to be orthogonally additive if $ab = ba = 0$ implies that $P(a + b) = P(a) + P(b)$ for all $a, b \in A$. Given a linear map $\Phi: \mathcal{P}_n(A)\to Y$ , the map $P : A \to Y$ defined by \[ P(a) = \Phi(a^n)\quad (a \in A)\qquad{\diamond} \] yields a significant example of $n$-homogeneous polynomial. \par The authors ask the following three interesting questions: \par (Q1) Let $A$ be a specified algebra. Is it true that every orthogonally additive $n$-homogeneous polynomial $P$ from $A$ into each linear space $Y$ can be expressed in the standard form ($\diamond$) for some linear map $\Phi: \mathcal{P}_n(A) \to Y$? \par (Q2) Let $A$ be a specified Banach algebra. Is it true that every orthogonally additive continuous $n$-homogeneous polynomial $P$ from $A$ into each Banach space $Y$ can be expressed in the standard form ($\diamond$) for some continuous linear map $\Phi: \mathcal{P}_n(A) \to Y$? \par (Q3) Let $A$ be a specified Banach algebra. Is there any norm $|||\cdot|||$ on $\mathcal{P}_n(A)$ with the property that the orthogonally additive continuous $n$-homogeneous polynomials from $A$ into each Banach space $Y$ are exactly the polynomials of the form ($\diamond$) for some $|||\cdot|||$-continuous linear map $\Phi: \mathcal{P}_n(A) \to Y$? \par These questions were investigated by several mathematicians; see [\textit{Y. Benyamini} et al., Bull. Lond. Math. Soc. 38, No. 3, 459--469 (2006; Zbl 1110.46033)], [\textit{C. Palazuelos} et al., Q. J. Math. 59, No. 3, 363--374 (2008; Zbl 1159.46035)], and [\textit{Y.-S. Wang}, J. Math. Anal. Appl. 459, No. 2, 1222--1232 (2018; Zbl 1388.43002)] and references therein. In the paper under review, the authors show that the answer to Question (Q1) is positive for the two-sided ideal $\mathcal{T}(G)$ of $L^1(G)$ consisting of trigonometric polynomials on a compact group $G$ with a normalized Haar measure. They prove that the answer to Question (Q2) is positive for the group algebra $L^1(G)$. They, however, give a negative answer to Question (Q2) for any of the convolution algebras $L^p(\mathbb{T})$ for $1< p \leq \infty$, and $C(\mathbb{T})$, where $\mathbb{T}$ denotes the circle group. In addition, they show that for each Banach algebra A, there exists a largest norm topology on the linear space $\mathcal{P}_n(A)$ for which the answer to Question (Q3) can be positive. They show that the answer to Question (Q3) is positive for most of the significant convolution algebras associated to $G$.