A new and simpler noncommutative central sets theorem (Q2344323): Difference between revisions

\noindent The Hales-Jewett theorem and the Szemeredi theorem (which generalizes the 1927 van der Waerden theorem asserting that in any finite partitioning of the integers, there is a part containing arbitrarily long arithmetic progressions) are examples of Ramsey theorems. Neil Hindman proved in 1974 that in any finite partitioning of the positive integers, at least one part contains all finite subsums of some infinite sequence. Later, it was realized that the best way to look at these problems (in any commutative semigroup \(S\) like the natural numbers) is to look at the Stone-Čech compactification \(\beta S\) of \(S\). The important notion of ``\textit{central subsets of natural numbers}'' was introduced in 1981 by Furstenberg via ergodic theory and led to applications in Ramsey-type problems. The new perspective enabled him and others to re-prove results like van der Waerden's and Hales-Jewett's theorems and to generalize them. A key property the central subsets possess is ``\textit{partition regularity}''; that is, in any finite partitioning of the natural numbers, one part must contain a central subset. Furstenberg also proved a `Central sets theorem' from which some strong combinatorial results can be deduced such as the fact that a central subset contains solutions to all partition regular systems of homogeneous equations. \noindent The notion of central sets has been generalized to any discrete semigroup \(S\) in place of the natural numbers. The Stone-Čech compactification \(\beta S\) of \(S\) plays a key role. Recall that \(\beta S\) is the collection of all ultrafilters on \(S\) and \(S\) is thought of as the sub-collection of the principal ultrafilters. A subset \(C\) of \(S\) is central if and only if there is an idempotent \(p\) in the smallest ideal of \(\beta S\) such that \(C \in p\). Furstenberg proved the so-called central sets theorem which has been generalized by Hindman and others. The strongest generalization was obtained by Dibyendu De, Neil Hindman and Dona Strauss who also obtained a version for arbitrary (noncommutative) semigroups. \noindent A very special case of their commutative version is the following assertion: Let \((S,+)\) be a commutative semigroup and \(C \subseteq S\) a central subset. Then, for every sequence \(F\) in \(S\), there exists \(c \in C\) and a finite subset \(H\) of the natural numbers such that for each \(f \in F\), we have \[ c + \sum_{t \in H} f(t) \in C. \] In order to pull out an element \(c\) as above (or perform an analogous splitting in the general case), De, Hindman and Strauss introduced, for each positive integer \(m\), the sets \(I_m\) of \(m\)-tuples \((H(1),H(2), \cdots, H(m))\) of nonempty, finite sets of positive integers such that max \(H(i) <\) min \(H(i+1)\) for \(i=1,2, \cdots, m-1\). Using the noncommutative version of the central sets theorem of De, Hindman and Strauss, it is possible to deduce: Let \((S,.)\) be any semigroup and \(C \subseteq S\), a central subset. Then, for every sequence \(F\) in \(S\), there is a positive integer \(m\) such that, for some \((c(1),c(2), \cdots, c(m+1)) \in S^{m+1}\) and some \((H(1),H(2), \cdots, H(m)) \in I_m\), and for every \(f \in F\), we have \[ \prod_{i=1}^m \bigg( c(i) \prod_{t \in H(i)} f(t) \bigg) c(m+1) \in C. \] \noindent In the proof of this, the sets \(H(i)\)'s may have different sizes. The main result of this paper is to show that one can take each \(H(i)\) to be a singleton, without any loss of generality. The author formulates and proves a simpler version of the noncommutative central sets theorem, which he shows to be equivalent to the more complicated formulation of De-Hindman-Strauss.