Embedding the flag representation in divided powers (Q847600)

Let \(\mathbb F\) be a finite field with \(q\) elements. An element in the category of functors \(\mathcal F\) takes a finite dimensional \(\mathbb F\)-vector space to an \(\mathbb F\)-vector space. The functors playing a role here are the divided powers \(\Gamma^k\), the flag representations \(\mathbb F[\mathfrak{Flag}_r]\) sending \(V\) to the vector space having basis all complete flags in \(V\) of length \(r\), and the projective functors \(P_{\mathbb F^r} = \mathbb F[\text{Hom}(\mathbb F^r, -)]\). For any multi-index \(\underline s\) given by \(s_1 > s_2 > \dots > s_r > s_{r+1} = 0\) there is a morphism \(\phi_{\underline s}: \mathbb F[\mathfrak{Flag}_r] \rightarrow \Gamma^{[\underline s]_q}\) where \([\underline s]_q = \sum (q^{s_i} -1)\). This morphism is defined globally in \(\mathcal F\) by factoring a morphism from \(P_{\mathbb F^r}\) through the obvious projective cover \(P_{\mathbb F^r} \rightarrow \mathbb F[\mathfrak{Flag}_r]\). The main theorem is then that \(\phi_{\underline s}(V)\) is a monomorphism if \(q^{s_i - s_{i+1}} -1 \geq (q-1)\text{dim} V-i+1\) for \(1\leq i \leq r\). The result is actually a little bit stronger and involves the Verschiebung morphism. When \(q=2\) the theorem has been proved by \textit{M. C. Crabb} and \textit{J. R. Hubbuck} in [Prog. Math. 136, 143--154 (1996; Zbl 0858.55014)].