The cofactor expansion (Laplace expansion) generalizes determinant computation to n×n matrices recursively: det(A) = Σⱼ aᵢⱼ Cᵢⱼ where Cᵢⱼ = (−1)^(i+j) Mᵢⱼ is the signed (i,j) minor and Mᵢⱼ is the determinant of the (n−1)×(n−1) submatrix formed by deleting row i and column j. This expansion can be performed along any row or column, and choosing a row or column with many zeros minimizes computation. The checkerboard sign pattern of cofactors (alternating + and −) is a key feature. For large n, cofactor expansion is computationally expensive but conceptually fundamental.
Practice expanding along different rows and columns of the same matrix to verify you get the same determinant. Choose expansions along rows with zeros to reduce work. Build up from 3×3 to 4×4 manually before trusting computational tools.
From your work with 2×2 and 3×3 determinants, you already know the pattern: det[a b; c d] = ad − bc, and the 3×3 determinant can be computed by expanding along the top row, multiplying each entry by the 2×2 determinant of what remains when that entry's row and column are deleted. Cofactor expansion generalizes this recipe to matrices of any size. The key insight is that determinants are defined recursively: a 4×4 determinant is defined in terms of four 3×3 determinants, which are each defined in terms of three 2×2 determinants, and so on down to the base case.
The formal setup introduces two related objects. The (i, j) minor Mᵢⱼ is the determinant of the (n−1)×(n−1) submatrix formed by deleting row i and column j. The (i, j) cofactor Cᵢⱼ is the signed minor: Cᵢⱼ = (−1)^(i+j) × Mᵢⱼ. The sign factor (−1)^(i+j) produces the checkerboard pattern — the (1,1) cofactor has a + sign, (1,2) has −, (2,1) has −, (2,2) has +, and so on. This sign pattern ensures that no matter which row or column you choose to expand along, you get the same determinant. det(A) = Σⱼ aᵢⱼ Cᵢⱼ for any fixed row i (or the analogous column version).
The freedom to choose any row or column is not just a mathematical curiosity — it is the key to efficiency. Expanding along a row with many zeros requires fewer actual multiplications because any term aᵢⱼ Cᵢⱼ with aᵢⱼ = 0 contributes nothing. For a 4×4 matrix like [[0, 0, 3, 0], ...], expanding along the first row reduces to just one 3×3 determinant instead of four. In practice, you should always scan for the row or column with the most zeros before starting. If the matrix has no zeros, you can sometimes create them via row operations first (which scale the determinant in a known way).
The conceptual payoff of cofactor expansion goes beyond computation. The same cofactors define the adjugate matrix adj(A), where adj(A)ᵢⱼ = Cⱼᵢ (note the transposed indices). The adjugate satisfies A × adj(A) = det(A) × I, which immediately gives the formula for the matrix inverse: A⁻¹ = adj(A) / det(A) when det(A) ≠ 0. Cramer's Rule, which expresses the solution to Ax = b in terms of determinants, also follows directly from cofactor theory. So while cofactor expansion is computationally expensive for large matrices (O(n!) in the naive recursive form), it is conceptually indispensable: it is the algebraic foundation on which determinant properties, the inverse formula, and Cramer's Rule all rest.
No topics depend on this one yet.