Spaces of lattice diagram polynomials in one set of variables (Q696808): Difference between revisions

Let \(M_\mu\) be the vector space spanned by all partial derivatives of the determinant \(\Delta_\mu(X;Y)=\det(x_i^{r_j}y_i^{c_j})_{i,j=1}^n\), where \(\mu\) is a partition of \(n\) and the \((r_j,c_j)\) run on the set of all cells of the diagram of \(\mu\). Then \(M_\mu\) is a bigraded \(S_n\)-module, with a bigraded character encoded by the bigraded Frobenius characteristic of \(M_{\mu}\) which is defined as the symmetric function \[ H_\mu=H_\mu(X;q,t)=\sum t^rq^s{\mathcal F}(\text{ch}{\mathcal H}_{rs}[M_{\mu}]), \] where \(\text{ ch}{\mathcal H}_{rs}[M_{\mu}]\) is the \(S_n\)-character of the homogeneous component of bidegree \((r,s)\) and \(\mathcal F\) is the Frobenius correspondence which maps the irreducible character \(\chi^{\lambda}\) to the Schur function \(s_{\lambda}(X)\). The Garsia-Haiman \(n!\)-conjecture states that \(H_\mu(X;q,t)\) is given by a renormalization of the Macdonald polynomial. By [\textit{M. Haiman}, J. Am. Math. Soc. 14, 941-1006 (2001; Zbl 1009.14001)] this is already the \(n!\)-theorem. As a tool for an elementary proof of the \(n!\)-theorem, \textit{F. Bergeron, N. Bergeron, A. M. Garsia, M. Haiman} and \textit{G. Tesler} [Adv. Math. 142, 244-334 (1999; Zbl 0934.05122)] introduced slightly more general spaces \(M_{\mu/ij}\) related with diagrams obtained by removing a single cell from a partition diagram. Various new conjectures were formulated there including a four-term recurrence for \(M_{\mu/ij}\) expressed in terms of \(M_{\mu/i,j+1},M_{\mu/i+1,j}\) and \(M_{\mu/i+1,j+1}\). In the paper under review the authors give an explicit basis of the \(Y\)-free component of \(M_{\mu/ij}\) and prove the \(Y\)-free analogue of the four-term recurrence. This basis has the nice feature that it is a natural generalization of the Artin basis for the space of harmonic polynomials for the symmetric group.