Hindman's theorem is only a countable phenomenon (Q2413413): Difference between revisions

The author contemplates higher cardinal generalizations of the following Ramsey-type theorem of \textit{N. Hindman} [J. Comb. Theory, Ser. A 17, 1--11 (1974; Zbl 0285.05012)]: For every partition \([\mathbb N]^{<\aleph_0}=A_1\cup A_2\) of the collection \([\mathbb N]^{<\aleph_0}\) of finite subsets of \(\mathbb N\) into two cells, there exists an infinite family \(X\subseteq [\mathbb N]^{<\aleph_0}\) of pairwise disjoint finite subsets of \(\mathbb N\) and an \(i\in 2\) such that the set \(\mathrm{FU}(X)\) of finite unions of elements of \(X\) is a subset of \(A_i\). For \(\kappa\), \(\lambda\) cardinal numbers with \(\lambda\leq\kappa\), the author considers generalizations of Hindman's theorem [loc. cit.], which he denotes by \(\mathrm{HIND}(\kappa, \lambda)\), where for every partition of \([\kappa]^{<\aleph_0}=A_0\cup A_1\) of the finite subsets of \(\kappa\) into two cells, there exists a family \(X\subseteq [\kappa]^{<\aleph_0}\) of cardinality \(\lambda\) consisting of pairwise disjoint finite subsets of \(\kappa\), such that for some \(i\in 2\) we have \(\mathrm{FU}(X)\subseteq A_i\). In this notation, Hindman's theorem is the assertion that \(\mathrm{HIND}(\aleph_0, \aleph_0)\) holds. The author's main theorem is that for any uncountable \(\lambda\) and any \(\kappa\geq\lambda\), \(\mathrm{HIND}(\kappa, \lambda)\) fails. The ``counterexample'' partition of the finite subsets of \(\kappa\) is clever and surprisingly easy to describe. In the final two sections of the paper, the author takes up possible generalizations of Hindman's theorem [loc. cit.] to semigroups and groups. These also fail for uncountable \(\lambda\).