Exercise 3.2 · Q14: Show that for two pairs of matrices
Show that
(i)
(ii)
(i) Let
.
(ii) Let
Multiplying row by column, the entries of are
Similarly, the entries of are
. Hence proved.
Want the reason behind every step, the basics and the checks?
In , the rows of meet the columns of ; in the roles swap, so completely different sums are formed. Usually the results differ. To show that two matrices are unequal it is enough to find one position where they differ, but we compute both products in full so that you see every entry.
- Given
- Part (i): two matrices. Part (ii): two matrices. In each part call the first matrix and the second .
- To show
- in each part.
- Plan
- For each part: (1) check both products are defined and find their orders; (2) compute every entry of ; (3) compute every entry of ; (4) point to a position where they differ.
Basics you need
If is and is , then is and = row of times column of : multiply matching entries, then add. Basics: pair up, multiply, add
Matrices of the same order are equal only if every pair of corresponding entries is equal. So one position with different entries proves they are unequal. Lesson: equality of matrices
and are different calculations. Even when both are defined and of the same order, they need not be equal. Lesson: matrix multiplication
Part (i): the pair
Here
Both products are defined
and are both . For : columns of (2) = rows of (2), so is defined and is . For : columns of (2) = rows of (2), so is defined and is .
If one product were undefined, or the two had different orders, they could not be equal and there would be nothing to compute. Here both are , so we must compare entries.Compute , entry by entry
Rows of : and . Columns of : and , read top to bottom.
Entry uses row of the first factor and column of the second. The figure below shows .Compute , entry by entry
Now the rows come from : and ; the columns come from : and .
Swapping the factors swaps which matrix supplies the rows, so every entry is a different sum.Compare
Both products are , but in position , has 7 and has 16. Since , . (In fact all four positions differ.)
Equality would need all four entries to agree; one disagreement is enough to rule it out.
Part (ii): the pair
Here
Both products are defined
Both matrices are . Columns of (3) = rows of (3), so is defined and ; columns of (3) = rows of (3), so is defined and .
As in Part (i), Step 1: same order, so the comparison must be made entry by entry.Compute , row by row
Columns of , top to bottom: , , .
Row 1 of is :
Row 2 of is :
Row 3 of is :
For example and . Row 2 of is , so it simply picks out row 2 of : .Compute , row by row
Now the columns come from , top to bottom: , , .
Row 1 of is :
Row 2 of is :
Row 3 of is :
For example . Every entry is a different sum from the one in the same position of , because rows now come from .Compare
Both are , but in position , has 5 and has . Since , .
A single differing position settles inequality. (Comparing position by position, in fact all nine entries differ: for instance has in and in .)
(i) has rows , , while has rows , , so .
(ii) has rows , , , while has rows , , , so .
In both parts matrix multiplication is not commutative. (This is a "show that" question, so the key gives no separate answer.)
A quick independent check: the total of all entries of equals (column sums of ) times (row sums of ), paired up and added, because every product appears exactly once.
Part (i): column sums of are ; row sums of are ; so the total for is , and indeed ✓. For : column sums of are , row sums of are , giving , and ✓. The totals differ (75 and 85), which on its own shows .
Part (ii): column sums of are ; row sums of are ; total for is , and the entries of do add up to 27 ✓. Column sums of are ; row sums of are ; total for is , and the entries of add up to 19 ✓.
- Assuming as with numbers. For numbers always; for matrices this usually fails, as here.
- Taking rows from the second factor. In the rows always come from , the factor on the left.
- Sign slips with . For example , not 3.
- uses rows of with columns of ; uses rows of with columns of . These are different calculations.
- To show two matrices are unequal, one differing position is enough; to show they are equal, every position must be checked.
- can happen for special pairs (Question 13's rotation matrices, diagonal matrices), but it is not a rule.
Every fact used, and where it comes from
| Fact | Used in | Source |
|---|---|---|
| defined when columns of = rows of ; order | Part (i), Step 1; Part (ii), Step 1 | Section 3.4.5 (Lesson: multiplication) |
| Entry of : row of times column of | Part (i), Steps 2, 3; Part (ii), Steps 2, 3 | Section 3.4.5 |
| Equal matrices: equal corresponding entries | Part (i), Step 4; Part (ii), Step 4 | Definition 2, Section 3.3.1 |
| Matrix multiplication is not commutative | Result | Section 3.4.5, Examples 13 and 14 |
| Diagonal matrices of the same order commute | Take-away 3 | Section 3.4.5, note after Example 14 |
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 →