Signed posets (Q1803877)

One sure sign of a successful generalization of a particular theory is an extension of language and technique to a larger domain permitting one to reconstruct considerable parts of previous theory on such a new basis. Another such sign is to provide new interpretations and insights captured by such an extended theory, especially involving results previously unobtainable or mysterious surprises hitherto possibly not even guessed at. The author demonstrates that the second condition has been met by providing some examples of applications in this paper in a somewhat abbreviated form for the sake of conservation of space including an analogue of MacMahon's calculation of the distribution of the major index for permutations. On the other hand, the first condition stated is met abundantly in that he is able to extend the notion of poset and, having done so, he is able to reconstruct a theory of \(P\)-partitions à la Stanley, to take the posets of order-ideal analogues through its paces and to consider \(EL\)-shellability properties of analogues of upper- semimodular lattices. The key to the generalization is to view partial orders for finite posets as subsets \(P\) of root systems: \(\Phi=\{e_ i-e_ j\mid 1\leq i\neq j\leq n\}\subset R^ n\), with the \(e_ i\) standard basis vectors and \(e_ i- e_ j\in P\) iff \(i<j\) in the original poset. By changing this root system to \(\Phi=\{e_ i \mid 1\leq i\leq n\}\cup\{\pm e_ i\pm e_ j\mid 1\leq i< j\leq n\}\) and maintaining the conditions defining posets, i.e., \(\alpha\in P\Rightarrow-\alpha\not\in P\), \(\overline P^{PLC}=\bigl\{\alpha\in\Phi\mid \alpha=\sum_{\beta\in P} C_ \beta\beta,\;C_ \beta\geq 0\bigr\}= P\), the trick is accomplished. Analogues of diagrams associated with posets are constructed but do not necessarily overwhelm with naturality, indicating the importance of looking at this particular generalization just the right way. At a more primitive level one then learns that posets as special subsets of a root system do carry information of a representation kind which may not be as readily available otherwise, which it may be useful to remember in the light of what was done here. A standard embedding of posets into signed posets generated by them is provided. This presentation being a selection from a larger set of results contained in the author's dissertation one does well to go through them as presented here while referring back to original proofs of results being generalized to appreciate delicacies of fit when obstacles presented are overcome. At the same time one might note how far these results are removed from the constructive combinatorial types represented by those obtained in currently practized dimension theory for example. It seems therefore entirely reasonable to suppose that much more can yet be done to bring other known notions and results into `hyperoctahedral' being to the benefit of even old problems in need of newer insights possibly.