Additivity of the two-dimensional Miller ideal (Q992034)

The study of finite-dimensional Miller forcings \(\mathbb{M}^n\) for \(n>0\), which consist of all \(n\)-tuples of super-perfect trees with the coordinatewise ordering, was started by the first author fifteen years ago. The properties of \(\mathbb{M}^n\) seem to diverge at \(n=2\). For instance, both \(\mathbb{M}\), \( \mathbb{M}^2\) add no Cohen reals, while \(\mathbb{M}^3\) does. Also, letting \(\mathcal{J}(\mathbb{M}^n)\) denote the ideal associated with \(\mathbb{M}^n\), the additivity of \(\mathcal{J}(\mathbb{M}^n)\) is \(\omega\) for \(n\geq 3\), while both \(\mathcal{J}(\mathbb{M})\) and \(\mathcal{J}(\mathbb{M}^2)\) are \(\sigma\)-ideals. In this paper the authors further investigate the additivity of \(\mathcal{J}(\mathbb{M}^2)\). In an earlier work, \textit{S. Jossen} and \textit{O. Spinas} [``A two-dimensional tree ideal'', Lect. Notes Log. 19, 294--322 (2005; Zbl 1105.03047)] showed that it is consistent that the additivity of \(\mathcal{J}(\mathbb{M}^2)\) is strictly less than the covering number of \(\mathcal{J}(\mathbb{M}^2)\), i.e. \(\mathsf{add}(\mathcal{J}(\mathbb{M}^2)) < \mathsf{cov}(\mathcal{J}(\mathbb{M}^2))\). But in their model \(\mathsf{add}(\mathcal{J}(\mathbb{M}^2))\) equals \(\mathsf{cov}(\mathcal{M})\), the covering number of the meager ideal \(\mathcal{M}\). In Section 1, the authors construct a version of amoeba forcing with Laver property, via which they show that it is consistent that \(\mathsf{cov}(\mathcal{M}) < \mathsf{add}(\mathcal{J}(\mathbb{M}^2))\). This forcing increases \(\mathsf{add}(\mathcal{J}(\mathbb{M}^2))\) without adding Cohen reals. In Section 2, the authors continue to study the size of \(\mathsf{add}(\mathcal{J}(\mathbb{M}^2))\) in the context of Martin's Axiom. Under Martin's Axiom, the additivity of \(\mathcal{J}(\mathbb{M})\) is the continuum. Here they show that, assuming Martin's Axiom for \(\sigma\)-centered forcings, the additivity of \(\mathcal{J}(\mathbb{M}^2)\) is the continuum. Moreover, the same assumption implies that \(\mathbb{M}^2\) is \(\sigma\)-centered and forcing with \(\mathbb{M}^2\) preserves cardinals. In the last section, the authors extend a result of Mathias (the case \(n=1\)) on maximal antichains in the Boolean algebra \(\mathcal{P}(\omega)/\mathrm{fin}\) to all finite dimensions. They use an \(n\)-dimensional version of Mathias forcing with the Laver property, introduced by Shelah and Spinas, which is similar to the amoeba forcing in Section 1. Their theorem is that for any \(n>0\) there are no analytic infinite maximal antichains in \((\mathcal{P}(\omega)/\mathrm{fin})^n.\) However, this conclusion is false in dimension \(\omega\): there is a perfect (thus closed) maximal antichain in \((\mathcal{P}(\omega)/\mathrm{fin})^\omega.\)