Introduces the matrix as an object with its own arithmetic. Covers dimensions and notation, addition and scalar multiplication entry by entry, and the row-times-column rule for products — including why the dimension condition exists and why matrix multiplication is not commutative.
Subject: Precalculus · 65 slides · symbolic lesson
Open the interactive version of this deck
Title
Precalculus · Chapter 9 — Systems of Equations and Inequalities
§9.5 Matrices and Matrix Operations, pp. 1116-1129
Objectives
Five things, each one you can check yourself on paper.
OpenStax, Precalculus, §9.5 Matrices and Matrix Operations §9.5, pp. 1116-1129 — the pages these objectives are drawn from
Warm-up
The letters carry no information once the positions are fixed.
Discussion prompt
In the equations 2x plus 3y equals 8 and x minus y equals 1, what is actually needed to solve them?
Hint: Which symbols could be dropped?
Answer:
The coefficients and constants — 2, 3, 8 and 1, negative 1, 1. The letters x and y are placeholders whose positions already say which variable is which.
So the whole system is captured by a rectangular array of numbers, with columns for each variable and a column for the constants.
That array is a matrix. Once systems are written this way, solving them becomes a set of operations on rows — which is what §9.6 does.
Concept
A matrix is a grid of numbers, and the operations defined on matrices are chosen so that they represent operations on the systems and transformations the matrices stand for.
matrix — a rectangular array of numbers arranged in rows and columns, with dimensions given as rows by columns
\[ A=\begin{bmatrix}2&3\\1&-1\end{bmatrix} \quad(2\times 2) \]
Dimensions are always quoted rows first, then columns. That order is a convention but a rigid one, and getting it backwards makes every dimension check give the wrong answer.
Figure (svg): Three cards giving the rules for adding matrices, scaling them, and multiplying them
OpenStax, Precalculus, §9.5 Matrices and Matrix Operations §9.5, pp. 1116-1119
Section
Section 1
Concept
A matrix's size is given as its number of rows by its number of columns, and an individual entry is named by its row and column position in that order.
The rows-first convention runs through everything: the dimension rule for products, the notation for entries, and the row operations of the next section. Reversing it silently makes every subsequent check wrong.
Figure (svg): Three cards giving the rules for adding matrices, scaling them, and multiplying them
OpenStax, Precalculus, §9.5 Matrices and Matrix Operations §9.5, pp. 1116-1120
Picture it
Two are routine and the third is not.
Figure (svg): Three cards giving the rules for adding matrices, scaling them, and multiplying them
The caption divides the section. Addition and scaling behave like ordinary arithmetic; multiplication is a new operation with new rules.
Worked example
Rows first.
\[ \text{For } A=\begin{bmatrix}2&1&3\\0&-4&5\end{bmatrix}, \text{ give its size and the entry in row 2, column 3.} \]
Count the rows
Why: Two horizontal lines of numbers.
\[ 2 \]
Count the columns
Why: Three vertical ones.
\[ 3 \]
State the dimensions
Why: Rows by columns.
\[ 2\text{ by } 3 \]
Locate the entry
Why: Second row, third column.
\[ 5 \]
Figure (svg): The solution to Worked example state dimensions and locate an entry shown as a ladder of expressions, one row per legal move
\[ 2\times 3, \; a_{23}=5 \]
Verify: count the entries
Why: A 2 by 3 matrix has six entries, and the array shown has six — consistent. Multiplying the dimensions always gives the entry count, which is a quick check that the dimensions were read correctly.
OpenStax, Precalculus, §9.5 Matrices and Matrix Operations §9.5, pp. 1117-1118
Prediction
A matrix is 4 by 5.
Predict first
How many entries does it have?
Correct: Twenty.
Why: The entry count is the product of the dimensions, since each of four rows contains five entries. Checking that count against the visible array confirms the dimensions were read correctly.
Worked example
Every entry and both dimensions.
\[ \text{Can a } 2\times 3 \text{ matrix equal a } 3\times 2 \text{ one?} \]
Compare the row counts
Why: Two versus three.
Conclude on shape
Why: Different dimensions.
Note the entry count
Why: Both have six entries.
State the rule
Why: Shape must match too.
Figure (svg): The solution to Worked example when are two matrices equal shown as a ladder of expressions, one row per legal move
\[ \text{different shapes are never equal} \]
Verify: consider why the entry count is not enough
Why: Both matrices hold six numbers, but the arrangement carries meaning — in a system, rows are equations and columns are variables. Two arrangements of the same numbers describe different things, so shape is part of identity.
OpenStax, Precalculus, §9.5 Matrices and Matrix Operations §9.5, pp. 1118-1120
Trap
\[ \text{three columns and two rows, so it is }3\times 2 \]
State the larger or the more visible count first
Why: The convention is not applied.
Every subsequent dimension check for a product is then wrong.
Rows first, always. Two rows and three columns is a two by three matrix.
The convention is rigid because the product rule depends on it — the first matrix's columns must match the second's rows.
Say 'rows by columns' aloud when reading a size. It is a habit that costs nothing and prevents a whole class of errors.
Faded example
Three rows and two columns.
Fill in the blanks
\text3=2\times___
Why: Rows come first and columns second, always. Reversing the order gives a different matrix shape and makes every product dimension check fail.
Sorting
Both dimensions and every entry must match.
Sort into buckets
Sort each comparison.
Step zero
You are asked to perform a matrix operation.
Discussion prompt
What do you check before computing anything?
Hint: Not every operation is defined.
Answer:
The dimensions, and whether the operation is defined for them. Addition needs identical shapes and multiplication needs matching inner dimensions.
Unlike ordinary arithmetic, a matrix operation can simply fail to exist — and computing anyway produces nonsense rather than an error.
So the check is the first step, not a formality. Writing the two sizes down side by side takes a second and settles both whether the operation exists and what shape its result will be.
Section
Section 2
Concept
Two matrices of the same size are added by adding corresponding entries. Multiplying by a number scales every entry, and is always defined.
These operations hold no surprises: addition is commutative and associative, scaling distributes over addition, and everything works as expected. It is worth noticing that explicitly, because multiplication then breaks the pattern so sharply.
Figure (svg): Three cards giving the rules for adding matrices, scaling them, and multiplying them
OpenStax, Precalculus, §9.5 Matrices and Matrix Operations §9.5, pp. 1120-1123
Picture it
The first two cards are this idea.
Figure (svg): Three cards giving the rules for adding matrices, scaling them, and multiplying them
Both are position-by-position, which is why both require identical shapes and why neither holds any surprise.
Worked example
Corresponding entries.
\[ \text{Add } \begin{bmatrix}2&1\\0&3\end{bmatrix} \text{ and } \begin{bmatrix}4&-1\\5&2\end{bmatrix}. \]
Check the dimensions
Why: Both two by two.
Add the top row
Why: Position by position.
\[ 6\text{ and } 0 \]
Add the bottom row
Why: Position by position.
\[ 5\text{ and } 5 \]
Assemble
Why: Same shape as the inputs.
\[ 2\text{ by } 2 \]
Figure (svg): The solution to Worked example add two matrices shown as a ladder of expressions, one row per legal move
\[ \begin{bmatrix}6&0\\5&5\end{bmatrix} \]
Verify: check the shape is preserved
Why: The sum has the same dimensions as both inputs, which addition always guarantees. Any operation that changed the shape would signal that something other than addition had been performed.
OpenStax, Precalculus, §9.5 Matrices and Matrix Operations §9.5, pp. 1121-1122
Prediction
A 3 by 2 matrix and a 2 by 3 matrix.
Predict first
Can they be added?
Correct: No, the dimensions differ.
Why: Addition pairs entries by position, so the shapes must be identical. Equal entry counts are irrelevant, since there is no entry in the third row of a two-row matrix to pair with.
Worked example
Scale first, then add.
\[ \text{Compute } 3A-2B \text{ for } A=\begin{bmatrix}1&2\\0&1\end{bmatrix}, \; B=\begin{bmatrix}0&1\\2&1\end{bmatrix}. \]
Scale the first
Why: Every entry tripled.
\[ 3, 6, 0, 3 \]
Scale the second
Why: Every entry doubled.
\[ 0, 2, 4, 2 \]
Subtract entry by entry
Why: Position by position.
\[ 3, 4, -4, 1 \]
Assemble
Why: Same shape throughout.
\[ 2\text{ by } 2 \]
Figure (svg): Three cards giving the rules for adding matrices, scaling them, and multiplying them
\[ \begin{bmatrix}3&4\\-4&1\end{bmatrix} \]
Verify: check one entry directly
Why: The bottom-left entry is three times zero minus two times two, which is negative four — matching. Every entry can be checked independently, since the operation is entirely positional.
OpenStax, Precalculus, §9.5 Matrices and Matrix Operations §9.5, pp. 1122-1123
Error analysis
A student adds a 2 by 3 matrix to a 3 by 2 one.
Annotate
On: \( \begin{bmatrix}1&2&3\\4&5&6\end{bmatrix}+\begin{bmatrix}1&2\\3&4\\5&6\end{bmatrix} \)
An undefined operation produces no error message when done by hand — it produces a plausible-looking array. Checking the dimensions first is what prevents an answer that means nothing.
Faded example
Two entries in the same position.
Fill in the blanks
2+4=6, \quad 1+(-1)=0
Why: Each entry of the sum comes from the two entries in that position, added as ordinary numbers. The operation is entirely positional, which is why it holds no surprises.
Sorting
Addition and scaling have different requirements.
Sort into buckets
Sort each operation.
Explain it to yourself
Addition requires identical dimensions.
Discussion prompt
Explain why that requirement exists.
Hint: How is each entry of the sum produced?
Answer:
Each entry of the sum comes from the two entries in the same position. That pairing needs a partner for every entry.
With different shapes some positions have no partner — a three-row matrix has entries in row 3 that a two-row matrix simply lacks.
So the operation is not merely awkward but undefined. A good explanation notes the contrast with scalar multiplication, which needs no partner and is therefore defined for every matrix regardless of shape.
Section
Section 3
Concept
A product exists only when the first matrix's column count equals the second's row count. When it does, the product has the first's row count and the second's column count.
The last point is worth dwelling on. A 2 by 3 times a 3 by 4 is defined and gives a 2 by 4, but the reverse order pairs 4 with 2 and does not exist at all. Order matters before any arithmetic happens.
Figure (svg): A diagram showing the dimension rule for matrix multiplication: the inner dimensions must match and the outer ones give the product's size
OpenStax, Precalculus, §9.5 Matrices and Matrix Operations §9.5, pp. 1123-1126
Picture it
Inner match, outer give the size.
Figure (svg): A diagram showing the dimension rule for matrix multiplication: the inner dimensions must match and the outer ones give the product's size
Writing the two sizes side by side makes the check visual rather than a matter of recall, and it produces the product's shape at the same time.
Worked example
Write the sizes side by side.
\[ \text{Does } AB \text{ exist for } A \text{ of size } 2\times 3 \text{ and } B \text{ of size } 3\times 4? \]
Write them in order
Why: First then second.
\[ 2 x 3\text{ and } 3 x 4 \]
Check the inner pair
Why: Three and three.
Conclude it exists
Why: The condition holds.
Read the outer pair
Why: Two and four.
\[ 2\text{ by } 4 \]
Figure (svg): A diagram showing the dimension rule for matrix multiplication: the inner dimensions must match and the outer ones give the product's size
\[ AB \text{ is } 2\times 4 \]
Verify: check the reverse order
Why: For BA the sizes read 3 by 4 and 2 by 3, whose inner numbers are 4 and 2 — they do not match, so BA does not exist. One order works and the other does not, which is the starkest form of non-commutativity.
OpenStax, Precalculus, §9.5 Matrices and Matrix Operations §9.5, pp. 1124-1125
Prediction
A 4 by 2 matrix times a 2 by 5 matrix.
Predict first
What is the product's size?
Correct: Four by five.
Why: The inner dimensions are both two, so the product exists, and the outer dimensions four and five give its size. Writing the two sizes side by side makes both facts readable at once.
Worked example
Square matrices always allow both.
\[ \text{For two } 2\times 2 \text{ matrices, are } AB \text{ and } BA \text{ the same?} \]
Check both products exist
Why: Inner dimensions match either way.
Check the shapes
Why: Both two by two.
Compute a small example
Why: Entries generally differ.
Conclude
Why: Same shape, different entries.
Figure (svg): The solution to Worked example both orders defined but different shown as a ladder of expressions, one row per legal move
\[ AB\ne BA \text{ in general} \]
Verify: consider why the shapes agree but the entries do not
Why: Each entry of a product comes from one row of the first matrix and one column of the second. Swapping the order swaps which matrix supplies rows and which supplies columns, so the sums being formed are entirely different even though the grid has the same shape.
OpenStax, Precalculus, §9.5 Matrices and Matrix Operations §9.5, pp. 1125-1126
Trap
\[ A\text{ is }3\times 4, \; B\text{ is }2\times 3, \text{ so compute }AB \]
Multiply because both matrices are available
Why: The dimension check is skipped.
The inner dimensions are 4 and 2, so the product does not exist at all.
Check the inner dimensions before computing. Four and two do not match, so this product is undefined.
The reverse order would work: B times A pairs 3 with 3 and gives a 2 by 4 result.
Which order is defined is part of the problem, not a detail. In matrix arithmetic the question 'can these be multiplied' has to be asked before the question 'what is the product'.
Faded example
A 3 by 2 times a 2 by 6.
Fill in the blanks
3\times2 \text6 2\times 6 \;\Longrightarrow\; \text___3\times___
Why: The inner twos match so the product exists, and the outer three and six give its dimensions. The rule delivers both the existence check and the result's shape in one reading.
Sorting
Check the inner dimensions.
Sort into buckets
Sort each pair, in the order written.
Explain it
Not every pair of matrices can be multiplied.
Discussion prompt
Explain to a classmate where the rule comes from.
Hint: How is one entry computed?
Answer:
Each entry of the product is formed by pairing off a row of the first with a column of the second, term by term, and adding.
Pairing requires them to be the same length — the row's length is the first matrix's column count and the column's length is the second's row count.
So the rule is not a convention but a requirement of the mechanism. A good explanation adds that this is also why order matters: swapping the matrices swaps which supplies rows and which supplies columns.
Section
Section 4
Concept
The entry in a given row and column of the product comes from that row of the first matrix and that column of the second, multiplied term by term and summed.
Working systematically — across the first row of the product, then the second, and so on — keeps track of which row and column are in play. Jumping between positions is where errors enter.
Figure (svg): A diagram showing one entry of a matrix product being formed from a row of the first matrix and a column of the second
OpenStax, Precalculus, §9.5 Matrices and Matrix Operations §9.5, pp. 1126-1129
Picture it
One row and one column produce one number.
Figure (svg): A diagram showing one entry of a matrix product being formed from a row of the first matrix and a column of the second
The two highlighted strips have the same length, which is what makes the pairing possible. The result of all that multiplying and adding is a single entry of the product.
Worked example
One row, one column.
\[ \text{Find the entry in row 1, column 1 of } \begin{bmatrix}2&1&3\end{bmatrix} \text{ times a column } \begin{bmatrix}4\\0\\5\end{bmatrix}. \]
Pair the first terms
Why: Two and four.
\[ 8 \]
Pair the second terms
Why: One and zero.
\[ 0 \]
Pair the third terms
Why: Three and five.
\[ 15 \]
Add
Why: The three products.
\[ 23 \]
Figure (svg): A diagram showing one entry of a matrix product being formed from a row of the first matrix and a column of the second
\[ 23 \]
Verify: count the products
Why: Three products were formed, matching the inner dimension of three. A shorter or longer sum would mean the row and column had different lengths, which the dimension rule forbids.
OpenStax, Precalculus, §9.5 Matrices and Matrix Operations §9.5, pp. 1126-1127
Prediction
A 2 by 3 matrix times a 3 by 4 matrix.
Predict first
How many products are added for each entry?
Correct: Three, the inner dimension.
Why: Each row of the first matrix has three entries and each column of the second has three, so three products are formed and added. The inner dimension sets the length of every such sum.
Worked example
Four entries, four row-column pairings.
\[ \text{Compute } \begin{bmatrix}1&2\\3&0\end{bmatrix}\begin{bmatrix}4&1\\2&5\end{bmatrix}. \]
Row 1 with column 1
Why: Four plus four.
\[ 8 \]
Row 1 with column 2
Why: One plus ten.
\[ 11 \]
Row 2 with column 1
Why: Twelve plus zero.
\[ 12 \]
Row 2 with column 2
Why: Three plus zero.
\[ 3 \]
Figure (svg): The solution to Worked example a full product shown as a ladder of expressions, one row per legal move
\[ \begin{bmatrix}8&11\\12&3\end{bmatrix} \]
Verify: compute the reverse order and compare
Why: Multiplying the other way gives a different matrix entirely — its top-left entry is 4 plus 3, which is 7, not 8. The two products are both defined and both two by two, and they are not equal, which is the concrete form of non-commutativity.
OpenStax, Precalculus, §9.5 Matrices and Matrix Operations §9.5, pp. 1127-1129
Error analysis
A student computes an entry of a product.
Annotate
On: \( \text{row 1 of }A\text{ times row 1 of }B \)
The asymmetry between the two matrices is what makes multiplication non-commutative. Rows come from the first and columns from the second, and swapping the matrices swaps those roles.
Faded example
Pairing a row with a column.
Fill in the blanks
2(4)+1(0)+3(5)=8+0+15=23
Why: The three products are formed term by term down the row and column and then added. That single number becomes one entry of the product matrix.
Matching
Position determines which row and column.
Match the pairs
Why: The entry's row index picks the row from the first matrix and its column index picks the column from the second. That correspondence is the whole rule, and working through the positions systematically keeps it straight.
Explain it to yourself
Rows come from one matrix and columns from the other.
Discussion prompt
Explain what that asymmetry causes.
Hint: What happens when the order is swapped?
Answer:
The first matrix supplies rows and the second supplies columns. Those are different roles, and the two matrices are not interchangeable in them.
Swapping the order swaps the roles, so entirely different sums are formed — even when both products are defined and have the same shape.
That is why matrix multiplication is not commutative. A good explanation notes that this is not an oddity to be memorised but a direct consequence of how each entry is built, which makes it predictable rather than surprising.
Section
Section 5
Concept
Addition behaves as expected, but multiplication breaks several rules that hold for numbers. Knowing which ones fail prevents importing them by habit.
The last row matters as a contrast: everything that goes wrong goes wrong on the multiplication side. Addition and scalar multiplication can be used with ordinary intuition, and only products need care.
| rule for numbers | for matrices |
|---|---|
| ab equals ba | generally false |
| any two can be multiplied | only if dimensions allow |
| a product is zero only if a factor is | false; nonzero matrices can multiply to zero |
| every nonzero element has a reciprocal | false; many have no inverse |
| a plus b equals b plus a | true, addition is well behaved |
Figure (svg): A contrast between the arithmetic of numbers and the arithmetic of matrices, highlighting where the familiar rules fail
OpenStax, Precalculus, §9.5 Matrices and Matrix Operations §9.5, pp. 1124-1129
Picture it
The right column lists the surprises.
Figure (svg): A contrast between the arithmetic of numbers and the arithmetic of matrices, highlighting where the familiar rules fail
The caption is the summary worth carrying: addition is safe and multiplication is not, and every unexpected behaviour in this chapter comes from the second.
Worked example
The zero-product property fails.
\[ \text{Multiply } \begin{bmatrix}1&0\\0&0\end{bmatrix} \text{ and } \begin{bmatrix}0&0\\0&1\end{bmatrix}. \]
Row 1 with column 1
Why: One times zero plus zero times zero.
\[ 0 \]
Row 1 with column 2
Why: One times zero plus zero times one.
\[ 0 \]
Row 2 with both columns
Why: Every product is zero.
\[ 0\text{ and } 0 \]
Assemble
Why: All entries zero.
Figure (svg): A contrast between the arithmetic of numbers and the arithmetic of matrices, highlighting where the familiar rules fail
\[ \begin{bmatrix}0&0\\0&0\end{bmatrix} \]
Verify: check both factors are nonzero
Why: The first has a 1 in its top-left and the second has a 1 in its bottom-right, so neither is the zero matrix. Yet their product is. For numbers this is impossible, and it is why the factoring techniques of chapter 3 do not transfer to matrices.
OpenStax, Precalculus, §9.5 Matrices and Matrix Operations §9.5, pp. 1127-1129
Sorting
Addition behaves; multiplication does not.
Sort into buckets
Sort each rule.
Worked example
Both products defined, both different.
\[ \text{Compare } AB \text{ and } BA \text{ for } A=\begin{bmatrix}1&1\\0&1\end{bmatrix}, \; B=\begin{bmatrix}1&0\\1&1\end{bmatrix}. \]
Compute the first product
Why: Row by column.
\[ [[2, 1], [1, 1]] \]
Compute the reverse
Why: Row by column again.
\[ [[1, 1], [1, 2]] \]
Compare
Why: Different entries.
Note both are defined
Why: Both two by two.
Figure (svg): The solution to Worked example order changes the answer shown as a ladder of expressions, one row per legal move
\[ AB\ne BA \]
Verify: locate the difference
Why: The top-left entries are 2 and 1 respectively. Swapping the order changed which matrix supplied rows and which supplied columns, so different sums were formed at every position — the shapes agree and nothing else does.
OpenStax, Precalculus, §9.5 Matrices and Matrix Operations §9.5, pp. 1128-1129
Trap
\[ AB=AC \;\Longrightarrow\; B=C \]
Cancel the common factor
Why: The matrix on the left is divided out of both sides.
This is invalid unless the cancelled matrix has an inverse.
Cancellation requires an inverse, and many nonzero matrices do not have one.
The same failure produces zero products from nonzero factors: if a matrix has no inverse, multiplying by it can destroy information.
Section 9.7 identifies which matrices can be cancelled. Until then, treat cancellation as unavailable rather than as a rule with exceptions.
Prediction
A times B equals A times C, with A not the zero matrix.
Predict first
Does it follow that B equals C?
Correct: Only if A has an inverse.
Why: Cancelling means multiplying both sides by an inverse, which many nonzero matrices lack. Without one, a matrix can destroy information and two different matrices can give the same product.
Faded example
Multiplying two matrices with disjoint nonzero entries.
Fill in the blanks
1\cdot 0+0\cdot 0=0, \text0___\text___
Why: Every row-column pairing puts a zero against each nonzero entry, so every product vanishes. Neither factor is the zero matrix, which shows the zero-product property failing.
Explain it
Some familiar arithmetic transfers and some does not.
Discussion prompt
Explain to a classmate what can be trusted.
Hint: Which operations are positional?
Answer:
Addition and scalar multiplication are entry by entry, so at each position it is ordinary arithmetic. Every rule you know for numbers holds.
Multiplication is a different operation entirely — rows against columns — so nothing transfers automatically. Commutativity, cancellation and the zero-product property all fail.
The practical rule is: be relaxed about sums and careful about products. A good explanation adds that this is why the section spends most of its time on multiplication and almost none on addition.
Comparison
Fill the blanks from memory. Two are positional and one is not.
Comparison matrix
| addition | scalar multiple | matrix product | |
|---|---|---|---|
| requirement | identical dimensions | none | inner dimensions match |
| result's shape | the same | the same | the outer dimensions |
| commutative | yes | yes | no |
| computed | entry by entry | entry by entry | row against column |
The last row explains the rest. Positional operations inherit ordinary arithmetic; the row-against-column operation does not.
Pattern
Five steps, and the first decides whether the rest happens.
Drawing the empty result grid first is worth the ten seconds: it makes the number of entries explicit and keeps the row-column bookkeeping visible.
OpenStax Algebra and Trigonometry 2e, §11.5 Matrices and Matrix Operations §11.5
Check
The dimension rule.
Check your understanding
A 3 by 2 matrix times a 2 by 5 matrix gives a matrix of what size?
Answer: A
Why: The inner dimensions are both two, so the product exists, and the outer dimensions three and five give its size. Writing the sizes side by side makes both readable at once.
Check
Computing an entry.
Check your understanding
Which ingredients produce the entry in row 2, column 3 of a product?
Answer: A
Why: The entry's row index picks a row from the first matrix and its column index picks a column from the second. Rows always come from the first and columns from the second, which is the asymmetry that makes the operation non-commutative.
Check
What fails.
Check your understanding
Which of these is generally false for matrices?
Answer: A
Why: Multiplication is not commutative, because swapping the order swaps which matrix supplies rows and which supplies columns. The other three concern addition and scaling, which are entry-by-entry and behave like ordinary arithmetic.
Real world
Every rotation and scaling in computer graphics is a matrix product.
Discussion prompt
Why does the order of two graphics transformations matter?
Hint: What operation combines them?
Answer:
Each transformation is a matrix, and applying two in sequence is their product. Since matrix multiplication is not commutative, the order changes the result.
Rotating a shape and then moving it right gives something different from moving it right and then rotating — the rotation carries the displaced shape around a different arc.
So the non-commutativity is not a mathematical curiosity but a visible geometric fact. Every graphics library documents its transformation order carefully for exactly this reason, and mismatched conventions between libraries are a standard source of bugs.
Commit first
State your confidence along with your answer.
Predict first
Why must the inner dimensions match for a product to exist?
Correct: Each entry pairs a row with a column term by term, so they must be the same length.
Why: A row of the first matrix has as many entries as that matrix has columns, and a column of the second has as many as it has rows. Pairing them term by term requires those counts to agree, which is exactly the dimension rule.
Explain it
The test of understanding is being able to say why, not just what.
Discussion prompt
A classmate is surprised that AB and BA differ. Explain why they should not be.
Hint: Where does each entry come from?
Answer:
Each entry of a product comes from a row of the first matrix and a column of the second. Those are different roles.
Swapping the order swaps which matrix plays which role, so entirely different numbers get paired and summed at every position.
So equality would be the surprise, not the difference. A good explanation adds the starker case: with non-square matrices, one order can be defined and the other not exist at all, which makes the asymmetry impossible to miss.
Exit ticket
One honest answer, so the next lesson can start in the right place.
Predict first
Which idea from this lesson would you most want to see again?
Correct: Any of these is a legitimate answer; the useful one is the honest one.
Why: There is no correct choice here. The fourth is where the whole section's difficulty concentrates, and the third is the check that has to become automatic before the next three sections make sense.
Connect it up
One page, drawn from memory, is worth more than rereading the section.
Draw it
Write the dimension rule with two sizes side by side, marking the inner and outer pairs. Beside it, compute one full two by two product showing which row and column produce each entry. Underneath, list the arithmetic rules that fail for matrices and give a one-line reason for each.
If your list of failures traces each one back to the row-against-column mechanism, the section's surprises are explained rather than memorised.
Recap
Five things, and the third is the one to check before every product.
| if you remember one thing | it should be this |
|---|---|
| about dimensions | rows first, always — the product rule depends on it |
| about products | inner dimensions match; outer ones give the size |
| about entries | row of the first against column of the second |
| about the rules | addition is safe; every surprise is in multiplication |
Section 9.6 puts matrices to work, turning the elimination of §9.2 into a mechanical sequence of row operations that reaches triangular form without any choices to make.
OpenStax, Precalculus, §9.5 Matrices and Matrix Operations §9.5, pp. 1116-1129 — everything on these slides traces back here
Want this taught 1-on-1? Alexander tutors Precalculus — $55/session, free consultation.