Class 12 Maths, Worked in Full Past papersGet updates
Chapter 3 · Matrices

Example 21: Verify (AB)′=B′A′(AB)'=B'A'

ExamplesTranspose · Section 3.5
The question

If

A=[−245],B=[ 13−6 ],A = \begin{bmatrix} -2 \\ 4 \\ 5 \end{bmatrix}, \qquad B = \left[\, 1 \quad 3 \quad -6 \,\right],

verify that (AB)′=B′A′(AB)' = B'A'⁠.

Exam answer AB=[(−2)⋅1(−2)⋅3(−2)⋅(−6)4⋅14⋅34⋅(−6)5⋅15⋅35⋅(−6)]=[−2−612412−24515−30]\begin{aligned} &AB \\ &= \begin{bmatrix} (-2){\cdot}1 & (-2){\cdot}3 & (-2){\cdot}(-6) \\ 4{\cdot}1 & 4{\cdot}3 & 4{\cdot}(-6) \\ 5{\cdot}1 & 5{\cdot}3 & 5{\cdot}(-6) \end{bmatrix} \\ &= \begin{bmatrix} -2 & -6 & 12 \\ 4 & 12 & -24 \\ 5 & 15 & -30 \end{bmatrix} \end{aligned} (AB)′=[−245−6121512−24−30](AB)' = \begin{bmatrix} -2 & 4 & 5 \\ -6 & 12 & 15 \\ 12 & -24 & -30 \end{bmatrix} A′=[−245], B′=[13−6]A' = \begin{bmatrix} -2 & 4 & 5 \end{bmatrix},\ B' = \begin{bmatrix} 1 \\ 3 \\ -6 \end{bmatrix} B′A′=[1⋅(−2)1⋅41⋅53⋅(−2)3⋅43⋅5(−6)⋅(−2)(−6)⋅4(−6)⋅5]=[−245−6121512−24−30]\begin{aligned} &B'A' \\ &= \begin{bmatrix} 1{\cdot}(-2) & 1{\cdot}4 & 1{\cdot}5 \\ 3{\cdot}(-2) & 3{\cdot}4 & 3{\cdot}5 \\ (-6){\cdot}(-2) & (-6){\cdot}4 & (-6){\cdot}5 \end{bmatrix} \\ &= \begin{bmatrix} -2 & 4 & 5 \\ -6 & 12 & 15 \\ 12 & -24 & -30 \end{bmatrix} \end{aligned}

∴ (AB)′=B′A′\therefore\ (AB)' = B'A'⁠. Hence verified.

Want the reason behind every step, the basics and the checks?

The one idea this problem is built on

The transpose of a product is the product of the transposes in the reverse order: (AB)′=B′A′(AB)' = B'A'⁠, not A′B′A'B'⁠. The orders tell the story. AA is 3×13\times1 and BB is 1×31\times3⁠, so ABAB is 3×33\times3⁠. Their transposes are 1×31\times3 and 3×13\times1⁠, so B′A′B'A' is again 3×33\times3⁠, while A′B′A'B' would be only 1×11\times1⁠. Only the reversed product can match.

Given, to show, and the plan
Given
A column matrix AA of order 3×13\times1 and a row matrix BB of order 1×31\times3⁠.
To show
(AB)′=B′A′(AB)' = B'A'⁠.
Plan
(1) Check ABAB is defined and compute all nine entries. (2) Transpose it. (3) Separately find A′A'⁠, B′B' and compute all nine entries of B′A′B'A'⁠. (4) Compare entry by entry.

Basics you need

1When a product is defined, and its order

ABAB is defined when the number of columns of AA equals the number of rows of BB⁠. If AA is m×nm\times n and BB is n×pn\times p⁠, then ABAB is m×pm\times p⁠. Basics: pair up, multiply, add

2Each entry of a product

The (i,k)(i,k) entry of ABAB is row ii of AA times column kk of BB⁠: multiply matching entries and add. When the inner size is 1, each row and column has a single entry, so each entry of ABAB is just one product ai1b1ka_{i1}b_{1k}⁠.

3Transpose

A′A' is AA with rows turned into columns; its (i,j)(i,j) entry is ajia_{ji}⁠. The transpose of a column is a row and vice versa. Basics: reading a table sideways

Solution

  1. Check that ABAB is defined and find its order

    AA is 3×13\times1 and BB is 1×31\times3⁠. The number of columns of AA (which is 1) equals the number of rows of BB (also 1), so ABAB is defined, and its order is (rows of AA⁠) ×\times (columns of BB⁠) =3×3= 3\times3⁠.

    A product can only be formed when each row of the first matrix has exactly as many entries as each column of the second, so that they can be paired off. Checking the orders first also tells us the shape of the answer before we compute anything.
    Section 3.4.5: if AA is m×nm\times n and BB is n×pn\times p⁠, then ABAB is m×pm\times p⁠. Lesson: matrix multiplication
  2. Compute all nine entries of ABAB

    Row ii of AA is the single number ai1a_{i1} and column kk of BB is the single number b1kb_{1k}⁠, so cik=ai1b1kc_{ik} = a_{i1}b_{1k}⁠. With a11=−2a_{11} = -2⁠, a21=4a_{21} = 4⁠, a31=5a_{31} = 5 and b11=1b_{11} = 1⁠, b12=3b_{12} = 3⁠, b13=−6b_{13} = -6⁠:

    c11=(−2)(1)=−2,c12=(−2)(3)=−6,c13=(−2)(−6)=12,c21=(4)(1)=4,c22=(4)(3)=12,c23=(4)(−6)=−24,c31=(5)(1)=5,c32=(5)(3)=15,c33=(5)(−6)=−30.\begin{aligned} c_{11} &= (-2)(1) = -2,\\ c_{12} &= (-2)(3) = -6,\\ c_{13} &= (-2)(-6) = 12,\\ c_{21} &= (4)(1) = 4,\\ c_{22} &= (4)(3) = 12,\\ c_{23} &= (4)(-6) = -24,\\ c_{31} &= (5)(1) = 5,\\ c_{32} &= (5)(3) = 15,\\ c_{33} &= (5)(-6) = -30. \end{aligned} AB=[−2−612412−24515−30].AB = \begin{bmatrix} -2 & -6 & 12 \\ 4 & 12 & -24 \\ 5 & 15 & -30 \end{bmatrix}.
    With an inner size of 1, "row times column" has only one pair to multiply, so there is nothing to add. Row ii of ABAB is therefore ai1a_{i1} times the row BB⁠: row 1 is −2-2 times 1,3,−61, 3, -6⁠, and so on. Watch the signs: (−2)(−6)=+12(-2)(-6) = +12⁠.
    The (i,k)(i,k) entry of ABAB is ∑jaijbjk\sum_j a_{ij}b_{jk}⁠; here jj takes only the value 1. Basics: pair up, multiply, add
  3. Transpose ABAB⁠: the left side

    ABAB is 3×33\times3⁠, so (AB)′(AB)' is 3×33\times3⁠. Row 1 of ABAB (−2,−6,12-2, -6, 12⁠) becomes column 1, row 2 (4,12,−244, 12, -24⁠) becomes column 2, and row 3 (5,15,−305, 15, -30⁠) becomes column 3:

    (AB)′=[−245−6121512−24−30].(AB)' = \begin{bmatrix} -2 & 4 & 5 \\ -6 & 12 & 15 \\ 12 & -24 & -30 \end{bmatrix}.
    For a square matrix the order does not change under transpose, but the entries off the main diagonal swap across it: the (1,2)(1,2) entry −6-6 of ABAB becomes the (2,1)(2,1) entry of (AB)′(AB)'⁠. The diagonal entries −2,12,−30-2, 12, -30 stay in place.
    Definition 3, Section 3.5: (i,j)(i,j) entry of X′X' is the (j,i)(j,i) entry of XX⁠. Lesson: transpose
  4. Find A′A' and B′B'⁠, and check B′A′B'A' is defined

    The column AA becomes a row, and the row BB becomes a column:

    A′=[ −245 ] (1×3),B′=[13−6] (3×1).A' = \left[\, -2 \quad 4 \quad 5 \,\right] \ (1\times3), \qquad B' = \begin{bmatrix} 1 \\ 3 \\ -6 \end{bmatrix} \ (3\times1).

    B′B' is 3×13\times1 and A′A' is 1×31\times3⁠: the columns of B′B' (1) match the rows of A′A' (1), so B′A′B'A' is defined and has order 3×33\times3⁠, the same as (AB)′(AB)'⁠.

    Checking the order now is a first test of the rule: if B′A′B'A' were not 3×33\times3 it could not possibly equal (AB)′(AB)'⁠. It also shows why the reverse order is needed (see the Common mistakes box).
  5. Compute all nine entries of B′A′B'A'⁠: the right side

    Write dikd_{ik} for the (i,k)(i,k) entry. Each is (entry ii of the column B′B'⁠) times (entry kk of the row A′A'⁠):

    d11=(1)(−2)=−2,d12=(1)(4)=4,d13=(1)(5)=5,d21=(3)(−2)=−6,d22=(3)(4)=12,d23=(3)(5)=15,d31=(−6)(−2)=12,d32=(−6)(4)=−24,d33=(−6)(5)=−30.\begin{aligned} d_{11} &= (1)(-2) = -2,\\ d_{12} &= (1)(4) = 4,\\ d_{13} &= (1)(5) = 5,\\ d_{21} &= (3)(-2) = -6,\\ d_{22} &= (3)(4) = 12,\\ d_{23} &= (3)(5) = 15,\\ d_{31} &= (-6)(-2) = 12,\\ d_{32} &= (-6)(4) = -24,\\ d_{33} &= (-6)(5) = -30. \end{aligned} B′A′=[−245−6121512−24−30].B'A' = \begin{bmatrix} -2 & 4 & 5 \\ -6 & 12 & 15 \\ 12 & -24 & -30 \end{bmatrix}.
    This is computed from scratch, without looking at (AB)′(AB)'⁠, so the comparison in the next step is a real check rather than a copy.
  6. Compare

    Both matrices are 3×33\times3⁠, and all nine entries agree: row 1 is −2,4,5-2, 4, 5⁠, row 2 is −6,12,15-6, 12, 15⁠, row 3 is 12,−24,−3012, -24, -30 in each. Hence (AB)′=B′A′(AB)' = B'A'.

    Why it works in general: the (i,k)(i,k) entry of (AB)′(AB)' is the (k,i)(k,i) entry of ABAB⁠, that is ∑jakjbji\sum_j a_{kj}b_{ji}⁠. The (i,k)(i,k) entry of B′A′B'A' is ∑jbij′ajk′=∑jbjiakj\sum_j b'_{ij}a'_{jk} = \sum_j b_{ji}a_{kj}⁠. These are the same products of numbers, only written in the other order, and multiplication of numbers is commutative.
    Property (iv) of Section 3.5.1: (AB)′=B′A′(AB)' = B'A'⁠.
Result (AB)′=B′A′=[−245−6121512−24−30].(AB)' = B'A' = \begin{bmatrix} -2 & 4 & 5 \\ -6 & 12 & 15 \\ 12 & -24 & -30 \end{bmatrix}.

The property (AB)′=B′A′(AB)' = B'A' is verified for the given AA and BB⁠.

Check it with numbers

A quick sum test: the nine entries of ABAB add up to −2−6+12+4+12−24+5+15−30=−14-2-6+12+4+12-24+5+15-30 = -14⁠, and so do the nine entries of B′A′B'A'⁠, as they must, since transposing only moves entries. More directly, ABAB multiplies "each entry of AA" by "each entry of BB": the sum of all entries is (−2+4+5)(1+3−6)=7×(−2)=−14(-2+4+5)(1+3-6) = 7\times(-2) = -14⁠. Both checks agree.

Common mistakes
  1. Writing (AB)′=A′B′(AB)' = A'B'⁠. Here A′A' is 1×31\times3 and B′B' is 3×13\times1⁠, so A′B′A'B' is 1×11\times1⁠: it equals [(−2)(1)+(4)(3)+(5)(−6)]=[−20][(-2)(1)+(4)(3)+(5)(-6)] = [-20]⁠. A 1×11\times1 matrix cannot equal a 3×33\times3 one. The order of the factors must reverse.
  2. Thinking a column times a row is a single number. A 3×13\times1 matrix times a 1×31\times3 matrix is 3×33\times3⁠. It is BABA (1×31\times3 times 3×13\times1⁠) that is 1×11\times1⁠.
  3. Sign slips. (−2)(−6)=12(-2)(-6) = 12 and (−6)(−2)=12(-6)(-2) = 12 are positive.
What to take away
  1. Transpose reverses products: (AB)′=B′A′(AB)' = B'A'⁠. Like taking off socks and shoes, the last thing done is the first undone.
  2. Always check orders. They tell you at once that A′B′A'B' cannot be the answer here.
  3. A column times a row (an "outer product") gives a full matrix whose (i,k)(i,k) entry is simply ai1b1ka_{i1}b_{1k}⁠.

Every fact used, and where it comes from

FactUsed inSource
Column and row matrices; order m×nm\times nSteps 1, 4Sections 3.2.1 and 3.3 (Lesson: types of matrices)
ABAB defined when columns of AA = rows of BB⁠; order m×pm\times p⁠; entry ruleSteps 1, 2, 4, 5Section 3.4.5 (Lesson: matrix multiplication)
Transpose A′=[aji]A' = [a_{ji}]Steps 3, 4Definition 3, Section 3.5 (Lesson: transpose)
Equal matrices: same order, equal entriesStep 6Definition 2, Section 3.3.1
(AB)′=B′A′(AB)' = B'A'Step 6Section 3.5.1, property (iv)
Found an error, or want more?

Tell us if something on this page is wrong or unclear, or if you would like a fuller explanation, another method or a new feature. Send feedback on this problem →