On a poset algebra which is hereditarily but not canonically well generated (Q1424045)

A Boolean algebra \(B\) is well-generated if it contains a well-founded sublattice \(L\) which generates \(B\). If \(B\) is well-generated then \(B\) is superatomic. \(B\) is canonically well-generated if it has a system \(H\) representing all generalized atoms such that the sublattice of \(B\) which is generated by \(H\) is well-founded. If \(C\) is a chain then \(B(C)\) denotes the interval algebra over \(C\). Let \(B\) and \(C\) be Boolean algebras; then \(B* C\) denotes their free product. The authors show: (1) Every subalgebra of \(B(\aleph_1)*B(\aleph_0)\) is well-generated. (2) \(B(\aleph_1)*B(\aleph_1)\) contains a non-well-generated Boolean algebra. (3) \(B(\aleph_1)* B(\aleph_0)\) contains a subalgebra which is not canonically well-generated. The authors give an example of a superatomic Boolean algebra which is a subalgebra of \({\mathcal P}(\omega)\) and cannot be embedded into a well-generated algebra.