- Practice
- About 16 min
- 5 practice questions
Transpose, symmetric and skew-symmetric matrices
Find the transpose of a matrix, use its properties including the reversal law (AB)ᵀ = BᵀAᵀ, identify symmetric and skew-symmetric matrices, and split any square matrix into the two.
If something looks wrong, please tell us.
By the end of this lesson you can
- Write Aᵀ for any matrix and state its order.
- Use (Aᵀ)ᵀ = A, (kA)ᵀ = kAᵀ, (A + B)ᵀ = Aᵀ + Bᵀ and (AB)ᵀ = BᵀAᵀ.
- Test whether a square matrix is symmetric or skew-symmetric.
- Express a square matrix as the sum of a symmetric and a skew-symmetric matrix.
Turning rows into columns sounds like a small change, but it gives one of the most used operations in matrix algebra. The transpose leads straight to two important families of square matrices, the symmetric and the skew-symmetric ones, and to a neat theorem: every square matrix is built from one of each.
The transpose
The transpose of A, written Aᵀ, is obtained by interchanging the rows and columns of A. If A = [aᵢⱼ]ₘₓₙ then Aᵀ = [bᵢⱼ]ₙₓₘ with bᵢⱼ = aⱼᵢ, so the (i, j)th entry of Aᵀ is the (j, i)th entry of A. For example, if A = [1 4 −2; 0 3 5] (2 × 3), then Aᵀ = [1 0; 4 3; −2 5] (3 × 2).
| Property | In words |
|---|---|
| (Aᵀ)ᵀ = A | transposing twice gives back A |
| (kA)ᵀ = kAᵀ | a scalar comes out unchanged |
| (A + B)ᵀ = Aᵀ + Bᵀ | transpose of a sum is the sum of transposes; similarly for A − B |
| (AB)ᵀ = BᵀAᵀ | reversal law: the order of the factors is reversed |
Worked example 1: checking the reversal law
Let A = [1 2; 3 0] and B = [2 −1; 1 4]. Working: AB = [1·2 + 2·1, 1·(−1) + 2·4; 3·2 + 0·1, 3·(−1) + 0·4] = [4 7; 6 −3], so (AB)ᵀ = [4 6; 7 −3]. Now Bᵀ = [2 1; −1 4] and Aᵀ = [1 3; 2 0]. BᵀAᵀ = [2·1 + 1·2, 2·3 + 1·0; −1·1 + 4·2, −1·3 + 4·0] = [4 6; 7 −3]. The two agree, so (AB)ᵀ = BᵀAᵀ. For contrast, AᵀBᵀ = [1·2 + 3·(−1), 1·1 + 3·4; 2·2 + 0, 2·1 + 0] = [−1 13; 4 2], which is different.
Symmetric and skew-symmetric matrices
A square matrix A is symmetric if Aᵀ = A, that is aᵢⱼ = aⱼᵢ for all i, j: the matrix is a mirror image across its main diagonal. For example [2 5 −1; 5 0 3; −1 3 7] is symmetric. A square matrix A is skew-symmetric if Aᵀ = −A, that is aᵢⱼ = −aⱼᵢ for all i, j. Putting i = j gives aᵢᵢ = −aᵢᵢ, so 2aᵢᵢ = 0: every diagonal entry of a skew-symmetric matrix is 0. For example [0 4 −2; −4 0 6; 2 −6 0] is skew-symmetric.
Symmetric or skew-symmetric?
-
Symmetric: Aᵀ = A
- aᵢⱼ = aⱼᵢ
- Diagonal entries can be anything
- Example [1 3; 3 2]
- A + Aᵀ is always symmetric
-
Skew-symmetric: Aᵀ = −A
- aᵢⱼ = −aⱼᵢ
- Diagonal entries all 0
- Example [0 3; −3 0]
- A − Aᵀ is always skew-symmetric
A matrix that is both symmetric and skew-symmetric satisfies A = Aᵀ = −A, so 2A = O and A is the zero matrix.
Every square matrix splits into the two
Theorem 7.1. For any square matrix A with real entries, A + Aᵀ is symmetric and A − Aᵀ is skew-symmetric. Proof: (A + Aᵀ)ᵀ = Aᵀ + (Aᵀ)ᵀ = Aᵀ + A = A + Aᵀ, so it is symmetric. (A − Aᵀ)ᵀ = Aᵀ − A = −(A − Aᵀ), so it is skew-symmetric.
Theorem 7.2. Any square matrix can be written as the sum of a symmetric and a skew-symmetric matrix: A = ½(A + Aᵀ) + ½(A − Aᵀ). The first part P = ½(A + Aᵀ) is symmetric and the second Q = ½(A − Aᵀ) is skew-symmetric, because multiplying by the scalar ½ does not change either property, and P + Q = ½A + ½A = A.
Worked example 2: splitting a 3 × 3 matrix
Let A = [3 5 −1; 1 2 4; 7 −2 0], so Aᵀ = [3 1 7; 5 2 −2; −1 4 0]. Working: A + Aᵀ = [6 6 6; 6 4 2; 6 2 0], so P = [3 3 3; 3 2 1; 3 1 0], which is symmetric. A − Aᵀ = [0 4 −8; −4 0 6; 8 −6 0], so Q = [0 2 −4; −2 0 3; 4 −3 0], which is skew-symmetric with zero diagonal. Check P + Q: [3 + 0, 3 + 2, 3 − 4; 3 − 2, 2 + 0, 1 + 3; 3 + 4, 1 − 3, 0 + 0] = [3 5 −1; 1 2 4; 7 −2 0] = A.
Worked example 3: unknown entries
Find p, q, r if [0 p 3; −5 0 q; r 2 0] is skew-symmetric. Working: aᵢⱼ = −aⱼᵢ. a₁₂ = −a₂₁ gives p = 5; a₂₃ = −a₃₂ gives q = −2; a₃₁ = −a₁₃ gives r = −3. The diagonal is already all zeros, as it must be.
(AB)ᵀ = BᵀAᵀ. Symmetric: Aᵀ = A. Skew-symmetric: Aᵀ = −A, diagonal all zero. A = ½(A + Aᵀ) + ½(A − Aᵀ).
Key takeaways: the transpose swaps rows and columns and reverses the order of a product; symmetric means equal to its own transpose, skew-symmetric means equal to the negative of it, with zeros on the diagonal; and every square matrix is ½(A + Aᵀ) plus ½(A − Aᵀ).
Check your understanding
-
1. If A is a 2 × 5 matrix, what is the order of AᵀA?
- 2 × 2
- 5 × 2
- 2 × 5
- 5 × 5
Show answer
Answer: D. Aᵀ is 5 × 2 and A is 2 × 5. The product (5 × 2)(2 × 5) is defined and has order 5 × 5.
-
2. For matrices of suitable orders, (AB)ᵀ equals
- AᵀBᵀ
- BᵀAᵀ
- BA
- (BA)ᵀ
Show answer
Answer: B. The reversal law says the transpose of a product is the product of the transposes in reverse order: (AB)ᵀ = BᵀAᵀ.
-
3. For which value of x is [0 x 4; −3 0 −1; −4 1 0] skew-symmetric?
- 3
- −3
- 0
- 4
Show answer
Answer: A. Skew-symmetry needs a₁₂ = −a₂₁, so x = −(−3) = 3. The other pairs already agree: a₁₃ = 4 and a₃₁ = −4; a₂₃ = −1 and a₃₂ = 1; and the diagonal is zero.
-
4. The 3 × 3 matrix with aᵢⱼ = i − j is
- symmetric
- the zero matrix
- skew-symmetric
- a unit matrix
Show answer
Answer: C. aⱼᵢ = j − i = −(i − j) = −aᵢⱼ for every i, j, and the diagonal is i − i = 0. The matrix is [0 −1 −2; 1 0 −1; 2 1 0], which is skew-symmetric.
-
5. When A = [4 −2; 6 1] is written as P + Q with P symmetric and Q skew-symmetric, what is P?
- [4 4; 4 1]
- [4 −2; −2 1]
- [0 −4; 4 0]
- [4 2; 2 1]
Show answer
Answer: D. Aᵀ = [4 6; −2 1], so A + Aᵀ = [8 4; 4 2] and P = ½(A + Aᵀ) = [4 2; 2 1]. Then Q = A − P = [0 −4; 4 0], which is skew-symmetric.
Sign in to keep track of the lessons you finish.