For any monotone $g : \{0, \dots, n\} \to \mathbb{N}$ with step increments $\le 1$ and total increase $g(n) = g(0) + 1$, there is a unique index $j$ where $g$ jumps by $1$.
For each row $i > 0$ of the Schreier refinement of $\sigma$ by $\tau$, there is a unique column $j$ where the dimension jumps from $\dim(\sigma_{i-1} + \sigma_i \cap \tau_j)$ to $\dim(\sigma_{i-1} + \sigma_i \cap \tau_{j+1})$ by exactly $1$.
The Schreier refinement permutation: for each row $i$, jumpCol returns the unique
column $j$ where the refinement increases in dimension, giving the bijection between the
composition factors of $\sigma$ and those of $\tau$.
Instances For
Defining property: at column $j = $ jumpCol, the Schreier cell dimension jumps by exactly $1$.
Uniqueness: any column where the Schreier cell dimension jumps must equal jumpCol.
At the jump column, the Schreier cell strictly increases.
Key non-containment: $V_i \cap W_{j+1}$ is not contained in $V_{i-1}$ at the jump column $j$, ensuring existence of the Jordan-Hölder gap vector.
Existence of a Jordan-Hölder gap vector: a vector $v \in V_i \cap W_{j+1}$ with $v \notin V_{i-1}$.
Choice of the Jordan-Hölder gap vector at row $i$: an explicit witness in $V_i \cap W_{j+1} \setminus V_{i-1}$ where $j$ is the jump column.
Instances For
The Jordan-Hölder gap vector lies in $V_i \cap W_{j+1}$.
The Jordan-Hölder gap vector does not lie in $V_{i-1}$.
The Jordan-Hölder gap vector lies in $V_i$.
The Jordan-Hölder gap vector lies in $W_{j+1}$, where $j$ is the jump column.
The Jordan-Hölder gap vector is nonzero.