Definability of Boolean algebras in \({\mathbb {HF}}\)-superstructures (Q2714038)

Informally, a model \(\mathfrak M\) of a finite language is said to be \(\Sigma\)-definable over an admissible set \(A\) if there exist a \(\Sigma\)-subset \(B\) of \(A\) and an equivalence relation \(\eta\) on it which is simultaneously \(\Sigma\)- and \(\Pi\)-definable over \(A\) and there exists a 1-1 mapping \(f\) from \(M\) onto the quotient \(B/\eta\) such that the sets \(\{\langle x_1,\dots,x_n\rangle\in B^n\mid {\mathfrak M}\models P(f^{-1}(x_1/\eta),\dots,f^{-1}(x_n/\eta))\}\) are \(\Sigma\)- and \(\Pi\)-definable over \(A\) for all basic predicates \(P\) of \(\mathfrak M\); a similar property holds for the operations of \(\mathfrak M\).NEWLINENEWLINENEWLINEThe main results of the article under review are as follows: 1) There exist an admissible set and a \(\Sigma\)-definable model over it such that it is not definable with trivial equivalence \(\eta\). 2) Assume \(L\) is a linear order which is definable over \({\mathbb {HF}}(B)\), where \(B\) is a superatomic Boolean algebra. Then \(L\) has a computable presentation. 3) There exists a superatomic Boolean algebra whose Frechét rank is not \(\Sigma\)-definable over its hereditarily finite superstructure.