Section II.
Geometry of Determinants
347
1.15 If t : R3 R3 changes volumes by a factor of 7 and s : R3 R3 changes volumes
by a factor of 3/2 then by what factor will their composition changes volumes?
1.16 In what way does the definition of a box differ from the definition of a span?
X 1.17 Why doesnt this picture contradict Theorem 1.5?
2 1
0 1
area is 2
determinant is 2
area is 5
X 1.18 Does |T S| = |ST |? |T (SP)| = |(T S)P|?
1.19 Show that there are no 22 matrices A and B satisfying these.
1 1
2 1
AB =
BA =
2 0
1 1
X
X
X
1.20 (a) Suppose that |A| = 3 and that |B| = 2. Find |A2 BT B2 AT |.
(b) Assume that |A| = 0. Prove that |6A3 + 5A2 + 2A| = 0.
1.21 Let T be the matrix representing (with respect to the standard bases) the map
that rotates plane vectors counterclockwise thru radians. By what factor does T
change sizes?
1.22 Must a transformation t : R2 R2 that preserves areas also preserve lengths?
1.23 What is the volume of a parallelepiped in R3 bounded by a linearly dependent
set?
1.24 Find the area of the triangle in R3 with endpoints (1, 2, 1), (3, 1, 4), and
(2, 2, 2). (Area, not volume. The triangle defines a plane what is the area of the
triangle in that plane?)
1.25 An alternate proof of Theorem 1.5 uses the definition of determinant functions.
(a) Note that the vectors forming S make a linearly dependent set if and only if
|S| = 0, and check that the result holds in this case.
(b) For the |S| 6= 0 case, to show that |T S|/|S| = |T | for all transformations, consider
the function d : Mnn R given by T 7 |T S|/|S|. Show that d has the first
property of a determinant.
(c) Show that d has the remaining three properties of a determinant function.
(d) Conclude that |T S| = |T | |S|.
1.26 Give a non-identity matrix with the property that AT = A1 . Show that if
AT = A1 then |A| = 1. Does the converse hold?
1.27 The algebraic property of determinants that factoring a scalar out of a single
row will multiply the determinant by that scalar shows that where H is 33, the
determinant of cH is c3 times the determinant of H. Explain this geometrically,
that is, using Theorem 1.5. (The observation that increasing the linear size of a
three-dimensional object by a factor of c will increase its volume by a factor of c3
while only increasing its surface area by an amount proportional to a factor of c2
is the Square-cube law [Wikipedia, Square-cube Law].)
1.28 We say that matrices H and G are similar if there is a nonsingular matrix P
such that H = P1 GP (we will study this relation in Chapter Five). Show that
similar matrices have the same determinant.
348
Chapter Four. Determinants
1.29 We usually represent vectors in R2 with respect to the standard basis so vectors
in the first quadrant have both coordinates positive.
~
v
+3
RepE2 (~v) =
+2
Moving counterclockwise around the origin, we cycle thru four regions:
+
+
.
+
+
Using this basis
B=h
0
1
,
i
1
0
gives the same counterclockwise cycle. We say these two bases have the same
orientation.
(a) Why do they give the same cycle?
(b) What other configurations of unit vectors on the axes give the same cycle?
(c) Find the determinants of the matrices formed from those (ordered) bases.
(d) What other counterclockwise cycles are possible, and what are the associated
determinants?
(e) What happens in R1 ?
(f) What happens in R3 ?
A fascinating general-audience discussion of orientations is in [Gardner].
1.30 This question uses material from the optional Determinant Functions Exist
subsection. Prove Theorem 1.5 by using the permutation expansion formula for
the determinant.
X 1.31
(a) Show that this gives the equation
x x2
y y2
1 1
of a line in R2 thru (x2 , y2 ) and (x3 , y3 ).
x3
y3 = 0
1
(b) [Petersen] Prove that the area of a triangle with vertices (x1 , y1 ), (x2 , y2 ), and
(x3 , y3 ) is
x
x2 x3
1 1
y1 y2 y3 .
2
1
1
1
(c) [Math. Mag., Jan. 1973] Prove that the area of a triangle with vertices at
(x1 , y1 ), (x2 , y2 ), and (x3 , y3 ) whose coordinates are integers has an area of N or
N/2 for some positive integer N.
Section III. Laplaces Formula
III
349
Laplaces Formula
Determinants are a font of interesting and amusing formulas. Here is one that is
often used to compute determinants by hand.
III.1
Laplaces Expansion
The example shows a 33 case but the approach works for any size n > 1.
1.1 Example Consider the permutation expansion.
t
1,1
t2,1
t3,1
t1,2
t2,2
t3,2
1
t1,3
t2,3 = t1,1 t2,2 t3,3 0
0
t3,3
0
1
0
0
+ t1,2 t2,1 t3,3 1
0
0
+ t1,3 t2,1 t3,2 1
0
1 0
0
0 + t1,1 t2,3 t3,2 0 0
0 1
1
0
1 0
0 0 + t1,2 t2,3 t3,1 0
1
0 1
0
0 1
0 0 + t1,3 t2,2 t3,1 0
1
1 0
0
1
0
1
0
0
0
1
0
0
1
0
1
0
0
Pick a row or column and factor out its entries; here we do the entries in the
first row.
1 0
= t1,1 t2,2 t3,3 0 1
0 0
+ t1,2 t2,1 t3,3 1
0
+ t1,3 t2,1 t3,2 1
0
1 0
0
0 + t2,3 t3,2 0 0
0 1
1
0
1 0
0 0 + t2,3 t3,1 0
1
0 1
0
0 1
0 0 + t2,2 t3,1 0
1
1 0
0
1
0
1 0
0 1
0 0
0 1
1 0
0 0
In those permutation matrices, swap to get the first rows into place. This
requires one swap to each of the permutation matrices on the second line, and
two swaps to each on the third line. (Recall that row swaps change the sign of
350
Chapter Four. Determinants
the determinant.)
1 0
= t1,1 t2,2 t3,3 0 1
0 0
t1,2 t2,1 t3,3 0
0
+ t1,3 t2,1 t3,2 0
0
1 0
0
0 + t2,3 t3,2 0 0
0 1
1
1
0 0
1 0 + t2,3 t3,1 0
0
0 1
1
0 0
1 0 + t2,2 t3,1 0
0
0 1
0
1
0
0 0
0 1
1 0
0 0
0 1
1 0
On each line the terms in square brackets involve only the second and third row
and column, and simplify to a 22 determinant.
t
t
t
2,2 t2,3
2,1 t2,3
2,1 t2,2
= t1,1
t1,2
+ t1,3
t3,2 t3,3
t3,1 t3,3
t3,1 t3,2
The formula given in Theorem 1.5, which generalizes this example, is a recurrence the determinant is expressed as a combination of determinants. This
formula isnt circular because it gives the nn case in terms of smaller ones.
1.2 Definition For any nn matrix T , the (n 1)(n 1) matrix formed by
deleting row i and column j of T is the i, j minor of T . The i, j cofactor Ti,j of
T is (1)i+j times the determinant of the i, j minor of T .
1.3 Example The 1, 2 cofactor of the matrix from Example 1.1 is the negative of
the second 22 determinant.
t
2,1 t2,3
T1,2 = 1
t3,1 t3,3
1.4 Example Where
T = 4
7
these are the 1, 2 and 2, 2 cofactors.
4 6
T1,2 = (1)1+2
=6
7 9
2
5
8
T2,2
6
9
1
= (1)2+2
7
3
= 12
9
Section III. Laplaces Formula
351
1.5 Theorem (Laplace Expansion of Determinants) Where T is an nn matrix, we
can find the determinant by expanding by cofactors on any row i or column j.
|T | = ti,1 Ti,1 + ti,2 Ti,2 + + ti,n Ti,n
= t1,j T1,j + t2,j T2,j + + tn,j Tn,j
Proof Exercise 25.
QED
1.6 Example We can compute the determinant
1 2 3
|T | = 4 5 6
7 8 9
by expanding along the first row, as in Example 1.1.
5 6
4 6
4 5
|T | = 1 (+1)
+ 2 (1)
+ 3 (+1)
= 3 + 12 9 = 0
8 9
7 9
7 8
Or, we could expand down the second column.
4 6
1 3
1
|T | = 2 (1)
+ 5 (+1)
+ 8 (1)
7 9
7 9
4
3
= 12 60 + 48 = 0
6
1.7 Example A row or column with many zeroes suggests a Laplace expansion.
1 5 0
1 5
1 5
2 1
= 16
+ 0 (+1)
+ 1 (1)
2 1 1 = 0 (+1)
2 1
3 1
3 1
3 1 0
We finish by applying Laplaces expansion to derive a new formula for the
inverse of a matrix. With Theorem 1.5, we can calculate the determinant of a
matrix by taking linear combinations of entries from a row with their associated
cofactors.
ti,1 Ti,1 + ti,2 Ti,2 + + ti,n Ti,n = |T |
()
Recall that a matrix with two identical rows has a zero determinant. Thus,
weighing the cofactors by entries from row k with k 6= i gives zero
ti,1 Tk,1 + ti,2 Tk,2 + + ti,n Tk,n = 0
because it represents the expansion along the row k of a matrix with
to row k. This summarizes () and ().
t1,1 t1,2 . . . t1,n
T1,1 T2,1 . . . Tn,1
|T | 0
t2,1 t2,2 . . . t2,n T1,2 T2,2 . . . Tn,2 0 |T |
=
..
..
..
.
.
.
tn,1 tn,2 . . . tn,n
T1,n T2,n . . . Tn,n
0
0
()
row i equal
...
...
...
|T |
352
Chapter Four. Determinants
Note that the order of the subscripts in the matrix of cofactors is opposite to
the order of subscripts in the other matrix; e.g., along the first row of the matrix
of cofactors the subscripts are 1, 1 then 2, 1, etc.
1.8 Definition The matrix adjoint to the square matrix T is
T1,1 T2,1 . . . Tn,1
T1,2 T2,2 . . . Tn,2
adj(T ) =
..
T1,n T2,n . . . Tn,n
where Tj,i is the j, i cofactor.
1.9 Theorem Where T is a square matrix, T adj(T ) = adj(T ) T = |T | I. Thus if
T has an inverse, if |T | 6= 0, then T 1 = (1/|T |) adj(T ).
Proof Equations () and ().
QED
1.10 Example If
T = 2
1
0
1
0
1
1
then adj(T ) is
T1,1
T1,2
T1,3
T2,1
T2,2
T2,3
1
0
T3,1
2
T3,2 =
1
T3,3
2
1
1
1
1
1
1
0
0
0
1
1
1
1
4 0
1 1
1
4
2
1
0
1
2
0
4
1
1
4
= 3
1
1
0
1
0
3
0
and taking the product with T gives the diagonal matrix |T | I.
1 0 4
1
0 4
3 0
0
2 1 1 3 3 9 = 0 3 0
1 0 1
1 0
1
0
0 3
The inverse of T is (1/ 3) adj(T ).
1/3
0/3 4/3
1/3 0
T 1 = 3/3 3/3 9/3 = 1
1
1/3 0/3
1/3
1/3 0
4/3
3
1/3
9
1
Section III. Laplaces Formula
353
The formulas from this subsection are often used for by-hand calculation
and are sometimes useful with special types of matrices. However, for generic
matrices they are not the best choice because they require more arithmetic than,
for instance, the Gauss-Jordan method.
Exercises
X 1.11 Find the cofactor.
1
T = 1
0
(a) T2,3
(b) T3,2
0
1
2
(c) T1,3
X 1.12 Find the determinant by expanding
3 0
1 2
1 3
(a) on the first row
2
3
1
1
2
0
(b) on the second row
(c) on the third column.
1.13 Find the adjoint of the matrix in Example 1.6.
X 1.14 Find the
2
(a) 1
1
X 1.15 Find the
matrix adjoint to each.
1 4
1 4 3
3
1
1
1
(b)
(c)
(d) 1 0 3
0 2
2 4
5 0
1 8 9
0 1
inverse of each matrix in the prior question with Theorem 1.9.
1.16 Find the matrix adjoint to this one.
2 1
1 2
0 1
0 0
0
1
2
1
0
0
1
2
X 1.17 Expand across the first row to derive the formula for the determinant of a 22
matrix.
X 1.18 Expand across the first row to derive the formula for the determinant of a 33
matrix.
X 1.19 (a) Give a formula for the adjoint of a 22 matrix.
(b) Use it to derive the formula for the inverse.
X 1.20 Can we compute a determinant by expanding down the diagonal?
1.21 Give a formula for the adjoint of a diagonal matrix.
X 1.22 Prove that the transpose of the adjoint is the adjoint of the transpose.
1.23 Prove or disprove: adj(adj(T )) = T .
1.24 A square matrix is upper triangular if each i, j entry is zero in the part above
the diagonal, that is, when i > j.
(a) Must the adjoint of an upper triangular matrix be upper triangular? Lower
triangular?
354
Chapter Four. Determinants
(b) Prove that the inverse of a upper triangular matrix is upper triangular, if an
inverse exists.
1.25 This question requires material from the optional Determinants Exist subsection. Prove Theorem 1.5 by using the permutation expansion.
1.26 Prove that the determinant of a matrix equals the determinant of its transpose
using Laplaces expansion and induction on the size of the matrix.
? 1.27 Show that
1
1
Fn = 0
0
.
1
1
1
0
.
1
0
1
1
.
1
1
0
1
.
1
0
1
0
.
1
1
0
1
.
. . .
. . .
. . .
. . .
. . .
where Fn is the n-th term of 1, 1, 2, 3, 5, . . . , x, y, x + y, . . . , the Fibonacci sequence,
and the determinant is of order n 1. [Am. Math. Mon., Jun. 1949]
Topic
Cramers Rule
A linear system is equivalent to a linear relationship among vectors.
!
!
!
x1 + 2x2 = 6
1
2
6
x1
+ x2
=
3x1 + x2 = 8
3
1
8
In the picture below the small parallelogram is formed from sides that are
the vectors 13 and 21 . It is nested inside a parallelogram with sides x1 13
and x2 21 . By the vector equation, the far corner of the larger parallelogram is
6
8 .
6
8
1
x1
3
1
3
2
1
2
x2
1
This drawing restates the algebraic question of finding the solution of a linear
system into geometric terms: by what factors x1 and x2 must we dilate the sides
of the starting parallelogram so that it will fill the other one?
We can use this picture, and our geometric understanding of determinants,
to get a new formula for solving linear systems. Compare the sizes of these
shaded boxes.
6
8
1
x1
3
1
3
2
1
2
1
2
1
356
Chapter Four. Determinants
The second is defined by the vectors x1 13 and 21 and one of the properties of
the size function the determinant is that therefore the size of the second
box is x1 times the size of the first. The third box is derived from the second by
shearing, adding x2 21 to x1 13 to get x1 13 + x2 21 = 68 , along with 21 . The
determinant is not affected by shearing so the size of the third box equals that
of the second.
Taken together we have this.
1 2 x 1 2 x 1 + x 2 2 6 2
1
1
2
x1
=
=
=
3 1 x1 3 1 x1 3 + x2 1 1 8 1
Solving gives the value of one of the
6
8
x1 =
1
3
variables.
2
1
10
=
=2
5
2
1
The generalization of this example is Cramers Rule: if |A| 6= 0 then the
system A~x = ~b has the unique solution xi = |Bi |/|A| where the matrix Bi is
formed from A by replacing column i with the vector ~b. The proof is Exercise 3.
For instance, to solve this system for x2
1 0 4
x1
2
2 1 1 x2 = 1
1 0 1
x3
1
we do this computation.
1
2
1
x2 =
1
2
1
2
1
1
0
1
0
4
1
1
18
=
3
4
1
1
Cramers Rule lets us by-eye solve systems that are small and simple. For
example, we can solve systems with two equations and two unknowns, or three
equations and three unknowns, where the numbers are small integers. Such
cases appear often enough that many people find this formula handy.
But using it to solving large or complex systems is not practical, either by
hand or by a computer. A Gausss Method-based approach is faster.
Exercises
1 Use Cramers Rule to solve each for each of the variables.
