Jordan axioms for C*-algebras (Q1116056)

Let A be an associative complex algebra. If, in addition, A is a \(C^*\)- algebra then it is well-known that the algebra \(A^+\), equipped with the symmetric product \(a\circ b:=(ab+ba)/2\), (a,b\(\in A)\), is a JB*-algebra with respect to the norm and the involution of the original C*-algebra A. The author's main theorem is the converse of the afore mentioned fact, i.e. if A is an associative complex algebra, and if \(A^+\) is a JB*- algebra for some norm and involution then A, equipped with the same norm and involution, is a C*-algebra. Hence the author obtains a purely JB*- algebraic characterization of C*-algebras. This result relies on an associative improvement of a theorem for unital alternative algebras given by \textit{R. B. Braun} [Math. Z. 185, 225-242 (1984; Zbl 0527.46054)]. In the rest of his paper the author investigates some consequences of his main theorem for quasiassociative JB*-algebras and for H*-algebras. Using notations and results of \textit{K. McCrimmon} [Trans. Am. Math. Soc. 158, 1-33 (1971; Zbl 0229.17002)] the author proves the following fact for quasiassociative JB*-algebras: Let A be a noncommutaive JB*-algebras which is split quasiassociative over its centroid C(A), furthermore assume that the JB*-algebra A admits no one-dimensional factor representation, then there exist an associative C*-algebra B and an element F in the centroid C(B) with \(0\leq F\leq 1\) such that \(A=B^{(F)}\), where \(B^{(F)}\) is the F-mutation of B. Finally the H*-algebra analogue to the authors' main theorem can be stated as follows: for an associative complex algebra H assume that \(H^+\) is an H*-algebra with zero annihilator for some inner product and some involution, then the original algebra H equipped with the same inner product and involution, is a H*-algebra.