No label defined (Q4948442)

A radical \(\alpha\) is a radical hereditary if for every radical \(\gamma\), when \(R\) is in \(\alpha\), then so is \(\gamma(R)\). An example is given to show that this is definitely a weaker condition than being hereditary (i.e., if \(R\) is in \(\alpha\) then so is any ideal of \(R\)). For radicals \(\alpha\) and \(\gamma\), \(\alpha\cap\gamma\) is again a radical, but \((\alpha\cap\gamma)(R)\) may be strictly contained in \(\alpha(R)\cap\gamma(R)\). Two interesting descending chains are: \[ R\supseteq\alpha(R)\supseteq\gamma\alpha(R)\supseteq\alpha\gamma\alpha(R)\supseteq\cdots\quad\text{and}\quad R\supseteq\gamma(R)\supseteq\alpha\gamma(R)\supseteq\gamma\alpha\gamma(R)\supseteq\cdots\;. \] The chains terminate when a term is equal to \((\alpha\cap\gamma)(R)\). Examples are given to show that all possible terminations of these chains are possible, i.e., at any member, or never. For any two elements \(a\) and \(b\) of a lattice \(L\), a pseudocomplement of \(a\) with respect to \(b\) exists if there is a unique largest element \(c\) in \(L\), among the set of elements \(x\) in \(L\) such that \(a\wedge x\leq b\). The difficulty is that such pseudocomplements need not exist. It is known that if \(\alpha_1\) and \(\alpha_2\) are hereditary radicals, and \(L\) is the lattice of all radical properties, then the pseudocomplement of \(\alpha_1\) with respect to \(\alpha_2\) does exist. The author proves that if \(\alpha_1\) is radically hereditary, with no restriction on \(\alpha_2\), then the pseudocomplement of \(\alpha_1\) with respect to \(\alpha_2\) does exist (Theorem 2). An example is given where \(\alpha_2\) is hereditary, with no condition on \(\alpha_1\), where the pseudocomplement of \(\alpha_1\) with respect to \(\alpha_2\) does not exist.