No label defined (Q4952914)

In 1963, Paravičenko proved that every Boolean algebra of size \(\aleph_1\) can be embedded into \({\mathcal P}({\mathbf N})/fin\), a result that subject to the Continuum Hypothesis (CH) implies the universality of \({\mathcal P}({\mathbf N})/fin\) for Boolean algebras of size \(c\), the continuum, or less. Since that time it was an interesting open problem whether this universality remains true without assumptions beyond ZFC. Though it is known since a long time that \(\sigma\)-centered Boolean algebras embed into \({\mathcal P}({\mathbf N})/fin\), for more general c.c.c. Boolean algebras this problem seems to be much more difficult. The main result of the present paper states that subject to the Open Colouring Axiom (OCA) the measure algebra, i.e., the Borel \(\sigma\)-algebra over \([0, 1]\) modulo Lebesgue null sets, cannot embed into the Boolean algebra \({\mathcal P}({\mathbf N})/fin\). The method of proof for this result falls under a class of similar procedures for maps between Boolean algebras or topological spaces which were known to hold true under CH but failed under OCA since in that case OCA implies that the corresponding maps should be of simple structure, whereas in ZFC this can't be true. In this paper the crucial maps are liftings of embeddings of the measure algebra, while simplicity means completeness and simplicity. The informed reader will notice immediately that the above main result contradicts to an earlier one of R. Frankiewicz and A. Gutek, but the authors of the present paper could point out a serious gap in the paper of R. Frankiewicz and A. Gutek. Unfortunately the same is with another paper of R. Frankiewicz where he suggested in a sketch of proof the non-imbeddability of the measure algebra by using Shelah's oracle c.c. method. So in fact the latter has to be left as an interesting open problem for future research.