The relationship between two involutive semigroups \(\mathcal S\) and \(\mathcal S_T\) is defined by a left multiplier \(T\) (Q403263)

Let \(\mathcal{S}\) denote a (multiplicative) semigroup, \(\mathrm{Mul}_t(\mathcal{S})\) the unital semigroup of left multipliers on \(\mathcal{S}\). Henceforth \(T\) will denote a left multiplier. A mapping \(T:\mathcal{S} \to \mathcal{S}\) is called a multiplier if it is both a left and a right multiplier. The unital semigroup of multipliers is denoted \(\mathrm{Mul}(\mathcal{S})\). If \(\mathcal{S}\) is itself a (complex) Banach algebra, the authors denote it by \(\mathfrak{A}\); the mapping in this case is assumed to be linear. A semigroup \(\mathcal{S}_{T}\) is defined with composition of \(a,b \in \mathcal{S}_{T}\) by \(a \circ b = aT(b)\). The authors select their semigroups to be involutive and consider complex-valued positive definite functions on \(\mathcal{S}\). The mapping \(T\) is presumed to preserve involution and is called a *-mapping. The authors' Theorem 6 is an extension to \((s_1,\dots ,s_n) \in \mathcal{S}\) of the well-known property of positive linear forms \(\phi\), viz. \(|\phi(s*t)|^{2} \leq |\phi(s*s)| |\phi(t*t)|\) for \(s,t \in \mathcal{S}\). For a continuous left multiplier \(T\), which is a *-mapping, and a positive definite function \(f\) on \(\mathcal{S}\), \(fT\) is a positive definite functional on elements with variables \((s_1,\dots,s_n)\). For a *-mapping \(T\), which is a left multiplier on \(\mathcal{S}\), then for each positive definite functional \(f\) on \(\mathcal{S}\) \(fT\) is a positive definite functional on \(\mathcal{S}\), i.e., \(f\) is a positive functional on \(\mathcal{S}_{T}\). The authors are vague on specifying \(\mathfrak{A}\). The reviewer holds that the authors are concerned only with \(\mathfrak{A}\) which are commutative Banach *-algebras and for the main theorem they are semisimple, or at least have a bounded approximate identity. The involution is adapted to fit the algebraic structure. They make an assumption that the use of \(T\) preserves the spectrum of \(\mathfrak{A}\), here made up of the characters, viz. multiplicative linear functionals on \(\mathfrak{A}\). A remark of the authors seems to mean that the spectra of corresponding elements of \(\mathfrak{A}_{T}\) and \(\mathfrak{A}\) may differ by multiplicative constants. Corresponding results hold also on the dual of \(\mathfrak{A}\) with its Arens product. The authors are concerned with conditions under which both Banach *-algebras \(\mathfrak{A}_{T}\) and \(\mathfrak{A}\) are also \(C^*\)-algebras. The right multiplier \(T^{\Delta}(a)\) is defined as \(T(a^*)^*\). The pair \((T,T^{\Delta})\) is a double centralizer, i.e., \(aT(b) = T^{\Delta}(a)b\) for all \(a,b \in \mathfrak{A}\); cf. [\textit{J. Dixmier}, Duke Math. J. 35, 375--382 (1968; Zbl 0179.18004)]. The authors make no attempt to explain the significance of the double centralizer; their reference to a paper by J. Laali does not exist as cited in the bibliography (the paper is, probably, [\textit{J. Laali}, ``The multipliers related products in Banach algebras'', Quaest. Math. 37, No. 4, 507--523 (2014; \url{doi:10.2989/16073606.2013.779988})]). However, the Banach *-algebras are algebras of functions and the existence of the double centralizer implies that the sup norm is equivalent to the operator norm. The authors' main theorem, where \(T\) is a *-mapping, states: {\parindent=8mm \begin{itemize}\item[(i)] \(\mathfrak{A}\) is a \(C^*\)-algebra if and only if \(\mathfrak{A}_{T}\) is. \item[(ii)] The bidual (with the second Arens product) \(\mathfrak{A}^{**}_{T^{**}}\) admits an involution. \item[(iii)] If \(\mathfrak{A}_{T}\) is a \(C^*\)-algebra, then so is \(\mathfrak{A}\); the converse holds if the mapping between \(\mathfrak{A}\) and \(\mathfrak{A}_{T}\) is an isometry. This converse is true even if \((T,T^{\Delta})\) is not a double centralizer. \end{itemize}} The article would have been clearer if the authors had illustrated their work with an appropriate example of \(\mathfrak {A}\).