Associatively amalgamatable varieties of rings (Q1907401)

A variety \(\mathcal V\) of rings is associatively amalgamatable if any amalgam of rings of \(\mathcal V\) can be embedded in an associative ring. The main result of the paper is the following: Theorem. A variety \(\mathcal V\) of rings is associatively amalgamatable if and only if \(\mathcal V\) consists of associative rings and satisfies either the identities \(xy=xy^n=x^ny\) for some natural \(n > 1\), or the identities \(klmx=0\), \(kx(y-y^n)=0\), \(l(x-x^n) y=0\) for natural numbers \(k\), \(l\), \(m\), and \(n\) such that \(n > 1\), while \(kl\) is square free and relatively prime to \(m\). Corollary. A variety of algebras over a finite field is associatively amalgamatable if and only if it consists of associative algebras and for some natural \(n > 1\), one of the identities \(xy=xy^n\) or \(xy=x^ny\) is satisfied. Among the nontrivial varieties of algebras over an infinite field, only the variety of algebras with zero multiplication is associatively amalgamatable. We note that earlier \textit{M. V. Volkov} [Mat. Zametki 33, No. 1, 3-13 (1983; Zbl 0539.16013)] obtained a description of varieties \(\mathcal V\) of associative rings with the property that any amalgam of rings of \(\mathcal V\) can be embedded in a ring of \(\mathcal V\).