On modular homology in the Boolean algebra. III (Q5949387)

This paper is a continuation of a program [for part II see \textit{S. Bell} and the authors, ibid. 199, No. 2, 556-580 (1998; Zbl 0892.16007)] of considerable maturity at the time of the publication of this paper wherein the entire `two-step' homology process so applicable in many non-algebraic but algebraicized theories has become an `\(r\)-step' more strongly algebra flavored theory of which the `two-step' process is a special through very important situation. Thus, if \(F\) is a field, \(\Omega\) is an \(n\)-set and for \(0\leq k\leq n\) an integer, \(M^n_k\) is the \(F\)-space with basis the \(k\)-subsets of \(\Omega\), then \(M^n_k\) is naturally an \(FS_n\) permutation module of \(S_n\) acting on its basis as an \(F\)-space. For \(r>0\), \(\partial_r\colon M^n_k\to M^n_{k-r}\) acts on this basis by setting \(\partial_r(\Delta):=\sum\Gamma\), where the summation runs over all \(\Gamma\subseteq\Delta\), of size \(|\Delta|-r\) (\(=k-r\) in this case). Then \(\partial_r\) is the \(r\)-step inclusion map, which for \(\text{char }F=p>0\) yields \(\partial^p_r=0\), with many nicely computable consequences noted in earlier publications as well as this one, many of great combinatorial interest and part of a vaster landscape not fully explored even yet.