On bowtie rings, universal survival rings and universal lying-over rings (Q693216)

In this review all rings are commutative and with identity. A ring \(R\) is called a \textit{universal lying-over ring} (shortly a \textit{ULO ring}) if every ring extension \(R\subseteq S\) satisfies the lying-over property. A ring extension \(R\subseteq S\) is called a \textit{survival extension} if \(PS\neq S\) for every \(P\in \text{Spec}(R)\). A ring \(R\) is called a \textit{universal survival ring} (shortly a \textit{US ring}) if every ring extension \(R\subseteq S\) is a survival extension. The following facts about the relation between the US and the ULO rings are known: (a) every ULO ring is a US ring; (b) in dimension \(0\), every ring is ULO, hence US; (c) in dimension 1, ULO \(\Leftrightarrow\) US; (d) in any dimension \(n\geq 2\), there exists a reduced ring of dimension \(n\) which is a US ring, but not a ULO ring. In the statement (d) two types of examples are known: (i) the so-called A+B construction; (ii) rings of continuous real functions on an infinite compact Hausdorff space. They were constructed by \textit{D.~E.~Dobbs, R.~Levy} and \textit{J.~Shapiro} in [``A universal survival ring of continuous functions which is not a universal lying-over ring'', to appear in Rocky Mt. J. Math]. In the paper under review, the authors introduce a new (third) type of reduced \(n\)-dimensional US rings which are not ULO rings (for any \(2\leq n\leq\infty\)). They start with an old construction by \textit{J.~L.~Dorroh} from the paper [Bull. Am. Math. Soc. 38, 85--88 (1932; Zbl 0003.38701; JFM 58.0137.02)]. Let \(R\) be a ring and \(E\) an \(R\)-algebra (possibly without identity). Let \(T=R\times E\) with the componentwise addition and with the multiplication \((r_1, e_1)\cdot (r_2, e_2)= (r_1r_2,\, r_1e_2+r_2e_1+e_1e_2)\). Then \(T\) is called a \textit{bow-tie ring} and denoted by \(T=R\bowtie E\). Following an idea of \textit{W.~V.~Vasconcelos} from the paper [Trans. Am. Math. Soc. 138, 505--512 (1969; Zbl 0175.03603)], the authors now for any \(2\leq n\leq \infty\) take \(R\) to be a quasilocal \(n\)-dimensional integral domain with the maximal ideal \(M\) and \(E\) to be a direct sum of countably infinitely many copies of the field \(R/M\). They prove that \(T=R\bowtie E\) is a reduced \(n\)-dimensional US ring which is not a ULO ring. They also prove that this ring could not be obtained by an A+B construction, nor as a ring of continuous functions on a Tychonoff space.