Skip to content
TheChatArchiveThe Chat Archive
Grok 4.1 Fast· xaiHugging Face Frontier Reasoning Traces

Show that the determinant of a product of matrices equals the product of determinants.

  1. User

    Show that the determinant of a product of matrices equals the product of determinants.

  2. Assistant

    Proof that det⁡(AB)=det⁡(A)det⁡(B)\det(AB) = \det(A) \det(B) for n×nn \times n matrices A,BA, B

    We use the Leibniz formula for the determinant of an n×nn \times n matrix M=(mij)M = (m_{ij}):

    det⁡(M)=∑σ∈Snsgn⁡(σ)∏i=1nmi,σ(i),\det(M) = \sum_{\sigma \in S_n} \operatorname{sgn}(\sigma) \prod_{i=1}^n m_{i,\sigma(i)},

    where SnS_n is the set of all permutations of {1,…,n}\{1, \dots, n\} and sgn⁡(σ)∈{+1,−1}\operatorname{sgn}(\sigma) \in \{+1, -1\} is the sign of σ\sigma.

    Let C=ABC = AB, so cij=∑k=1naikbkjc_{ij} = \sum_{k=1}^n a_{ik} b_{kj}. Then,

    det⁡(C)=∑τ∈Snsgn⁡(τ)∏i=1nci,τ(i)=∑τ∈Snsgn⁡(τ)∏i=1n(∑ki=1naikibkiτ(i)).\det(C) = \sum_{\tau \in S_n} \operatorname{sgn}(\tau) \prod_{i=1}^n c_{i,\tau(i)} = \sum_{\tau \in S_n} \operatorname{sgn}(\tau) \prod_{i=1}^n \left( \sum_{k_i=1}^n a_{i k_i} b_{k_i \tau(i)} \right).

    Expanding the product of sums gives

    det⁡(C)=∑τ∈Sn∑k1=1n⋯∑kn=1nsgn⁡(τ)∏i=1naiki⋅∏i=1nbkiτ(i).\det(C) = \sum_{\tau \in S_n} \sum_{k_1=1}^n \cdots \sum_{k_n=1}^n \operatorname{sgn}(\tau) \prod_{i=1}^n a_{i k_i} \cdot \prod_{i=1}^n b_{k_i \tau(i)}.

    Rearranging the sums,

    det⁡(C)=∑k1,…,kn=1n(∏i=1naiki)[∑τ∈Snsgn⁡(τ)∏i=1nbkiτ(i)].\det(C) = \sum_{k_1,\dots,k_n=1}^n \left( \prod_{i=1}^n a_{i k_i} \right) \left[ \sum_{\tau \in S_n} \operatorname{sgn}(\tau) \prod_{i=1}^n b_{k_i \tau(i)} \right].

    For fixed (k1,…,kn)(k_1, \dots, k_n), define the matrix R=(rij)R = (r_{ij}) by rij=bkijr_{i j} = b_{k_i j} (i.e., row ii of RR is row kik_i of BB). The inner sum is then exactly det⁡(R)\det(R):

    ∑τ∈Snsgn⁡(τ)∏i=1nriτ(i)=det⁡(R).\sum_{\tau \in S_n} \operatorname{sgn}(\tau) \prod_{i=1}^n r_{i \tau(i)} = \det(R).

    Thus,

    det⁡(C)=∑k1,…,kn=1n(∏i=1naiki)det⁡(R),\det(C) = \sum_{k_1,\dots,k_n=1}^n \left( \prod_{i=1}^n a_{i k_i} \right) \det(R),

    where RR depends on (k1,…,kn)(k_1, \dots, k_n).

    Key observation: det⁡(R)=0\det(R) = 0 unless (k1,…,kn)(k_1, \dots, k_n) is a permutation of {1,…,n}\{1, \dots, n\}. Indeed:

    • If any kp=kqk_p = k_q for p≠qp \neq q, then rows pp and qq of RR are identical, so det⁡(R)=0\det(R) = 0 (determinants vanish on matrices with repeated rows).
    • Thus, nonzero terms occur only when i↦kii \mapsto k_i is a bijection, i.e., there exists ρ∈Sn\rho \in S_n with ki=ρ(i)k_i = \rho(i) for all ii.

    For such a ρ\rho, RR is BB with rows permuted by ρ\rho (row ii of RR is row ρ(i)\rho(i) of BB). Permuting the rows of a matrix by ρ\rho multiplies the determinant by sgn⁡(ρ)\operatorname{sgn}(\rho):

    det⁡(R)=sgn⁡(ρ)det⁡(B).\det(R) = \operatorname{sgn}(\rho) \det(B).

    Moreover,

    ∏i=1naiki=∏i=1naiρ(i).\prod_{i=1}^n a_{i k_i} = \prod_{i=1}^n a_{i \rho(i)}.

    Summing only over these permutations ρ∈Sn\rho \in S_n,

    det⁡(C)=∑ρ∈Sn(∏i=1naiρ(i))⋅sgn⁡(ρ)det⁡(B)=det⁡(B)∑ρ∈Snsgn⁡(ρ)∏i=1naiρ(i)=det⁡(B)⋅det⁡(A).\det(C) = \sum_{\rho \in S_n} \left( \prod_{i=1}^n a_{i \rho(i)} \right) \cdot \operatorname{sgn}(\rho) \det(B) = \det(B) \sum_{\rho \in S_n} \operatorname{sgn}(\rho) \prod_{i=1}^n a_{i \rho(i)} = \det(B) \cdot \det(A).

    Thus, det⁡(AB)=det⁡(A)det⁡(B)\det(AB) = \det(A) \det(B).

    Notes

    • This holds over any commutative ring (e.g., R\mathbb{R}, C\mathbb{C}).
    • By induction on the number of factors, it extends to det⁡(A1⋯Ak)=det⁡(A1)⋯det⁡(Ak)\det(A_1 \cdots A_k) = \det(A_1) \cdots \det(A_k).
    • Base case n=1n=1: Trivial, as det⁡([ab])=ab=det⁡([a])det⁡([b])\det([a b]) = ab = \det([a]) \det([b]).

    This proof relies only on the Leibniz formula and basic properties of permutations/row operations.

Responses were produced in different, uncontrolled contexts (system prompts, settings, prior turns). This is not a benchmark.