The dual form of Beck type identities (Q2127843)

A partition of a positive integer \(n\) is a weakly decreasing sequence \(\lambda = (\lambda_i)_i\) of positive integers whose sum is \(n\). The positive integers \(\lambda_i\) are called the parts of \(\lambda\), and the multiplicity of \(\lambda_i\), \(f_\lambda (\lambda_i)\), is the number of times \(\lambda_i\) appears in \(\lambda\). In this paper, the authors give analytic and combinatorial proofs for four identities similar to two others that George Beck conjectured to be true. In what follows we assume that \(t\) and \(k\) are integers such that \(0 \leq t \leq k-1\). Let \(\mathcal{M}_k (n)\) be the set of partitions of \(n\) with exactly one part divisible by \(k\), \(\mathcal{M}_{k,t} (n)\) be the set of partitions in \(\mathcal{M}_k (n)\) such that \(f_\lambda (\lambda_i) \equiv t \bmod k\) for the unique part \(\lambda_i\) divisible by \(k\), and \(\mathcal{M}_{k,\neq t} (n)\) be the set of partitions in \(\mathcal{M}_k (n)\) such that \(f_\lambda (\lambda_i) \not\equiv t \bmod k\) for the unique part \(\lambda_i\) divisible by \(k\). Let \(\mathcal{R}_k(n)\) be the set of partitions of \(n\) in which exactly one part occurs at least \(k\) times, \(\mathcal{R}_{k,t} (n)\) be the set of partitions in \(\mathcal{R}_k (n)\) in which \(\lambda_i \equiv t \bmod k\) for the unique part \(\lambda_i\) appearing at least \(k\) times, and \(\mathcal{R}_{k,\neq t} (n)\) be the set of partitions in \(\mathcal{R}_k (n)\) in which \(\lambda_i \not\equiv t \bmod k\) for the unique part \(\lambda_i\) appearing at least \(k\) times. Finally, let \(\mathcal{D}_k (n)\) be the set of partitions of \(n\) with no part occurring at least \(k\) times. The first and second identities, both in Theorem 1.11, state that for \(n \geq 1\), \(k \geq 2\), and \(0 \leq t \leq k-1\), \begin{align*} | \mathcal{M}_{k,t} (n) | & = | \mathcal{R}_{k,t} (n) |, \\ | \mathcal{M}_{k,\neq t} (n) | & = | \mathcal{R}_{k,\neq t} (n) |. \end{align*} The third and fourth identities, in Theorems 1.13 and 1.15, respectively, state that for \(n \geq 1\) and \(k \geq 2\), \begin{align*} | \mathcal{M}_{k,\neq 0} (n) | - (k-1) | \mathcal{M}_{k,0} (n) | & = \sum_{\lambda \in \mathcal{D}_k (n)} n_k (\lambda) = | \mathcal{R}_{k,\neq 0} (n) | - (k-1) | \mathcal{R}_{k,0} (n) |, \\ | \mathcal{M}_{k,1} (n) | - | \mathcal{M}_{k,0} (n) | & = \sum_{\lambda \in \mathcal{D}_k (n)} \overline{n}_k (\lambda) = | \mathcal{R}_{k,1} (n) | - | \mathcal{R}_{k,0} (n) |, \end{align*} where \(n_k (\lambda)\) is the number of parts divisible by \(k\) in \(\lambda\) and \(\overline{n}_k (\lambda)\) is the number of parts not divisible by \(k\) in \(\lambda\).