Linear Algebra 48 | Laplace Expansion [dark version]

Laplace (cofactor) expansion turns determinant calculation from the brute-force Leibniz method into a recursive “shrink the matrix” procedure. Instead of summing 24 products for a 4×4 determinant, the method picks one row or one column, multiplies each chosen entry by the determinant of the remaining (n−1)×(n−1) submatrix, and then applies the correct plus/minus sign (the cofactor sign). This matters because it reduces the work dramatically—especially when the chosen row or column contains many zeros.

The core mechanism is easiest to see on a 4×4 matrix. Every term in the Leibniz expansion uses exactly one entry from each row and each column. If the expansion starts with the entry a11, then row 1 and column 1 are “used up,” leaving only the lower-right 3×3 submatrix to account for the remaining permutations. That accounts for six of the 24 permutation products. Repeating the same idea with a21 forces the use of row 2 and column 1, which means column 1 and row 2 are removed, again leaving a 3×3 determinant—covering the next six products. Continuing through a31 and a41 completes all 24 products.

The missing piece is the sign. The cofactor expansion doesn’t just add the sub-determinants; it alternates signs according to whether the corresponding permutation is even or odd. For the 4×4 case, the signs come out as a checkerboard pattern: start with a plus in the top-left corner, then flip the sign when moving one step right or one step down. This sign rule generalizes to any size matrix and can be encoded as (−1)^(i+j), where i and j are the row and column indices of the chosen entry.

In the general n×n case, Laplace expansion is written as a sum of n terms. Fix a column j: det(A) = Σ_{i=1..n} a_{ij} (−1)^(i+j) det(A_{ij}), where A_{ij} is the (n−1)×(n−1) matrix formed by deleting row i and column j. The same structure holds if expanding along a row i instead of a column. Conceptually, the method reduces the determinant size step by step: each expansion produces smaller subproblems until reaching 3×3, 2×2, or even 1×1 determinants.

A worked example demonstrates the practical advantage. For a specific 4×4 matrix with many zeros, choosing the second row—containing only one non-vanishing entry—collapses the expansion to a single term. That 3×3 determinant is then expanded again using a row with only one nonzero entry, and the final 2×2 determinant is computed directly (yielding 6−4=2). Putting the factors and checkerboard signs together gives the original 4×4 determinant as 4. The takeaway is straightforward: Laplace expansion is less about memorizing complicated formulas and more about choosing the row or column with the most zeros and applying the (−1)^(i+j) sign pattern carefully.