Matrix operations — addition, subtraction (conditions): Question bank

This bank hunts the misconceptions, not the arithmetic. Every answer gives a reason, never a bare "yes"/"no". If you can defend each answer out loud, you have truly understood matrix addition and subtraction.

Before you start, recall the tools you should already own from Matrix Notation and Terminology (order $m \times n$ ), Zero Matrix and Identity Matrix (the all-zero matrix $O$ ), Scalar Multiplication of Matrices (multiplying every entry by a number $k$ ) and Transpose of a Matrix (flipping rows into columns, written $B^{T}$ ).

True or false — justify

True or false: If

A

and

B

have the same number of total entries, then

A + B

is defined.

False. A

2 \times 3

and a

3 \times 2

matrix both have 6 entries, but position

(1, 3)

exists in the first and not the second — count is not structure.

True or false: Matrix addition is commutative, i.e.

A + B = B + A

whenever both are defined.

True. Each entry obeys

a_{ij} + b_{ij} = b_{ij} + a_{ij}

because real-number addition is commutative, and this holds slot-by-slot.

True or false: Matrix subtraction is commutative, i.e.

A - B = B - A

False. Subtraction is anti-commutative:

A - B = - (B - A)

, so unless

A = B

the two results differ by a sign in every entry.

True or false:

A + O = A

for any matrix

A

, using a single fixed zero matrix

O

False as stated —

O

must have the same order as

A

. The zero matrix is not one object; there is a

2 \times 3

zero, a

3 \times 3

zero, etc.

True or false: For any square matrix

A

, both

A + A^{T}

and

A - A^{T}

are defined.

True.

A^{T}

of a square

n \times n

matrix is also

n \times n

, so orders match; note

A - A^{T}

has all-zero diagonal since

a_{ii} - a_{ii} = 0

True or false: If

A

2 \times 3

and

B

3 \times 2

, then

A + B^{T}

is a valid matrix addition of

A

and

B

False.

A + B^{T}

is a legal operation (both

2 \times 3

), but it adds

A

to the transpose of

B

, not to

B

— the entries of

B

have been relocated, changing their meaning.

True or false: Addition is associative:

(A + B) + C = A + (B + C)

when all are the same order.

True. Slot-by-slot,

(a_{ij} + b_{ij}) + c_{ij} = a_{ij} + (b_{ij} + c_{ij})

by real-number associativity, so grouping never matters.

True or false:

k (A + B) = k A + k B

for a scalar

k

(see Scalar Multiplication of Matrices).

True. Each entry gives

k (a_{ij} + b_{ij}) = k a_{ij} + k b_{ij}

by the distributive law of real numbers.

True or false: The additive inverse of

A

A^{T}

False. The additive inverse is

- A = [- a_{ij}]

, obtained by negating every entry;

A + (- A) = O

. Transposing rearranges entries, it does not negate them.

Spot the error

Claim: "

A

3 \times 4

B

4 \times 3

; since a

3 \times 4

times a

4 \times 3

is allowed,

A + B

is allowed too." — What's wrong?

The person borrowed the rule for Matrix Multiplication (inner dimensions match). Addition needs identical orders, so

3 \times 4

plus

4 \times 3

is undefined.

Claim: "

[12] + [12] = [24]

." — What's wrong?

1 \times 2

row and a

2 \times 1

column have different orders, so the sum is undefined; the writer silently reshaped one to force an answer.

Claim: "

A + B

where

A

2 \times 2

and

B

2 \times 2

: I added the diagonals only and left off-diagonals unchanged." — Error?

Addition is element-wise for every slot, not just the diagonal — off-diagonal entries must also be summed as

a_{ij} + b_{ij}

Claim: "Since

A - B

needs same order, I can still do it by padding the shorter matrix with zeros." — Error?

Padding invents entries that were never in the original data, breaking the deterministic definition; the operation stays undefined, not "fixable".

Claim: "

- A

means flip the sign of the first entry and leave the rest." — Error?

- A = [- a_{ij}]

negates every entry, not just one; only then does

A + (- A) = O

hold in every slot.

Claim: "Because addition is commutative,

A - B = B - A

." — Error?

Commutativity belongs to

+

, not

-

. Subtraction is defined as

A + (- B)

, and

A + (- B) \neq = B + (- A)

in general.

Why questions

Why must the two matrices share both dimensions, not just one?

Every entry needs a partner at the same

(i, j)

; matching only rows leaves unpaired columns (or vice-versa), so some slot has no term to add.

Why is "same total element count" a tempting but false condition?

Total count ignores where each number lives; a

2 \times 3

and

3 \times 2

share 6 numbers but the slot

(1, 3)

exists in one grid and not the other.

Why does subtraction inherit associativity and identity from addition but lose commutativity?

Because

A - B

is defined as

A + (- B)

, so all

+

properties carry over — except that swapping operands flips the sign, killing commutativity.

Why is the zero matrix the additive identity rather than the identity matrix

I

For addition we need

a_{ij} + x_{ij} = a_{ij}

, forcing

x_{ij} = 0

everywhere;

I

has ones on its diagonal and belongs to multiplication, not addition (see Zero Matrix and Identity Matrix).

Why does transposing

B

before adding change the meaning even when it fixes the dimensions?

Transposing swaps each entry's row and column role — if

B

meant "products × stores",

B^{T}

means "stores × products", so you are combining different relationships.

Why can matrix addition model "combining two linear systems" (link to System of Linear Equations)?

Because coefficient matrices of the same shape add slot-by-slot, mirroring how equation-by-equation you add matching coefficients of the same variables.

Edge cases

A + A

defined, and what is it?

Yes — same matrix, same order — and

A + A = 2 A

, the scalar-2 multiple, since each slot gives

a_{ij} + a_{ij} = 2 a_{ij}

O + O

defined, and what is the result?

Yes, provided both zero matrices share the order; the result is again that same zero matrix, since

0 + 0 = 0

in every slot.

Can two

1 \times 1

matrices be added, and how does it relate to ordinary numbers?

Yes; a

1 \times 1

matrix

[a]

plus

[b]

[a + b]

, so scalar addition is just the smallest case of matrix addition.

A - A

defined, and what is it?

Yes; every slot gives

a_{ij} - a_{ij} = 0

, so the result is the zero matrix

O

of the same order — the additive-inverse law in action.

Can a

3 \times 3

matrix be added to its own transpose, and what special structure appears?

Yes (

3 \times 3

plus

3 \times 3

);

A + A^{T}

is always symmetric because slot

(i, j)

and

(j, i)

both equal

a_{ij} + a_{j i}

A

is any

m \times n

matrix, is

A - O

ever different from

A

No — as long as

O

is the

m \times n

zero,

A - O = A

in every slot; a shape mismatch would make it undefined, not "different".

Two matrices have orders

m \times n

and

n \times m

with

m \neq = n

; can any of

A + B

A - B

A + B^{T}

A - B^{T}

be computed?

Only

A + B^{T}

and

A - B^{T}

: transposing

B

makes it

m \times n

to match

A

. The raw

A \pm B

stay undefined because orders differ.

Recall One-line summary

Same order in, same order out — every slot pairs with its twin, $+$ keeps all the nice laws, and $-$ keeps them all except the freedom to swap sides.

Connections

Matrix operations — addition, subtraction (conditions) — parent topic these traps drill.
Matrix Notation and Terminology — the order $m \times n$ that decides "defined vs undefined".
Zero Matrix and Identity Matrix — why $O$ (not $I$ ) is the additive identity.
Scalar Multiplication of Matrices — powers the $A + A = 2 A$ and distributive traps.
Transpose of a Matrix — the $A + B^{T}$ misconception.
Matrix Multiplication — the rival dimension rule people wrongly borrow.
System of Linear Equations — where combining same-shape matrices has meaning.