Algebraic geometers care about a basic question: given a geometric object (a "variety") sitting in some high-dimensional space, what does the set of all lines, planes, and higher-dimensional flats that are spanned by points on that object look like? The collection of all points lying on lines connecting pairs of points on the variety is called its "secant variety," and understanding its dimension tells you, among other things, how efficiently you can decompose tensors or polynomials into simple pieces. When the expected dimension (a naive count of free parameters) matches the actual dimension, life is straightforward. When they differ, the variety is called "defective," and those cases are both rare and mathematically interesting.
Flag varieties are geometric objects that encode chains of nested subspaces of a vector space. For example, the simplest non-trivial case keeps track of a line sitting inside a plane sitting inside a larger space. These varieties can be embedded into projective space in many different ways, controlled by a choice of integers (called a polarization). The paper develops a new algebraic toolkit, called "Schur apolarity," to study the secant varieties of flag varieties under these general embeddings. Classical apolarity is a well-known technique for Veronese varieties (which encode symmetric tensors), but flag varieties involve more complex, mixed-symmetry objects. A key technical discovery is that a naive algebraic operation (squaring an ideal) does not automatically capture the correct geometric information in this richer setting, and a more careful geometric substitute is needed.
As a concrete payoff, the authors completely classify the dimensions of all secant varieties for the flag variety of lines-in-planes in an n-dimensional space, embedded by the simplest mixed polarization. Out of infinitely many cases, only two turn out to be defective: the second secant variety when n equals 3, and the third secant variety when n equals 4, each falling exactly one dimension short of expectations. This clean classification demonstrates the power of the new framework and opens the door to tackling more complicated flag varieties systematically.