 Solution

The two-vector Gram matrix isThe upper bound follows because the subtracted square is nonnegative. If either vector is zero, the determinant is zero. Otherwise apply the non-strict version of the calculation in i: $\Vert v_1-t{v}_2{\Vert }^2\ge 0$ with $t=⟨v_1,v_2⟩/\Vert v_2{\Vert }^2$ gives $⟨v_1,v_2{⟩}^2\le \Vert v_1{\Vert }^2\Vert v_2{\Vert }^2$. ThereforeEquality on the left holds exactly when the two vectors are linearly dependent; equality on the right holds exactly when they are orthogonal. The Gram determinant measures the squared area of their parallelogram: in the nonzero case, $\det A=\Vert v_2{\Vert }^2\Vert v_1-t{v}_2{\Vert }^2$, the square of base times perpendicular height.