MATH431: Complex Numbers

Justin Wyss-Gallifent

September 27, 2021

1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.1 Definition . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2 Graphical Representation . . . . . . . . . . . . . . . . . . 2
1.3 Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.4 Translations . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.5 Scaling . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.6 Rotations . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.7 Reflections . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.8 Combinations . . . . . . . . . . . . . . . . . . . . . . . . . 9
2 Angle and Closest Point Representations . . . . . . . . . . . . . . 10
2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.2 Lines Through the Origin . . . . . . . . . . . . . . . . . . 10
2.3 Closest Point Representation . . . . . . . . . . . . . . . . 10
2.4 Rotations of these Lines . . . . . . . . . . . . . . . . . . . 11
2.5 Reflections of these Lines . . . . . . . . . . . . . . . . . . 11
2.6 Translations of these Lines . . . . . . . . . . . . . . . . . 12
2.7 More on Transformation of these Lines . . . . . . . . . . . 17
2.8 Locating Points on these Lines . . . . . . . . . . . . . . . 17
2.9 Parametrization . . . . . . . . . . . . . . . . . . . . . . . 19
3 Complex Affine Transformations . . . . . . . . . . . . . . . . . . 20
4 Matrix Representation of Complex Numbers . . . . . . . . . . . . 22
5 Taking this to 3D . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

1
1 Introduction
1.1 Definition
√
Definition 1.1.1. The complex numbers C are generated by defining ı̂ = −1
and then creating the set of all numbers of the form a + bı̂ where a, b ∈ R. The
value a is the real part and the value b is the imaginary part. For a complex
number z these are denoted Re(z) and Im(z) respectively.

Note 1.1.1. Note that we’ve used a Calculus 3-type ı̂ here, which looks like a
vector. This is because it will be treated as such later so it’s best to bring in
the notation now, as peculiar as that may seem.


1.2 Graphical Representation

A complex number z = a + bı̂ can be graphed by plotting the number in the
plane using the x-axis as the real axis and the y-axis as the imaginary axis and
plotting z at the location (a, b). For example here is the complex number 3 + 2ı̂:
Im

4i

2i 3 + 2ı̂

Re
−4 −2 2 4

-2i

-4i

It’s this graphical representation that allows us to write manipulations of the

plane as operations on complex numbers. These operations turn out to be quite
simple and convenient.

1.3 Properties
Standard operations on complex numbers arise obviously from those of real
numbers and keeping in mind that ı̂2 = −1.
Example 1.1. For example (2 + 3ı̂)(4 − 5ı̂) = 8 − 10ı̂ + 12ı̂ − 15ı̂2 = 23 + 2ı̂.

2
 
Exercise 1.1. Calculate (4 + 2ı̂)(8 − 4ı̂).

Definition 1.3.1. The magnitude (or absolute value or√norm) of a complex
number z = a + bı̂ is denoted |z| and is defined by |z| = a2 + b2 .

Definition 1.3.2. A unit complex number is a complex number with norm
equal to 1.

Definition 1.3.3. For a complex number z = a + bı̂ we define the complex
conjugate of z, denoted z, by z = a − bı̂.

Note that for z = a + bı̂ we have zz = a2 + b2 = |z|2 .
Theorem 1.3.1. For z, w ∈ C we have zw = zw.

Proof. Omitted.

Exercise 1.2. Complete the above proof.


The complex conjugate allows us to do division and resolve the result into an
obvious complex number. In general for z1 , z2 ∈ C we calculate and simplify
z1 z1 z2
= = ...
z2 z2 z2

Example 1.2. To divide 2 + 3ı̂ by 4 − 5ı̂ we do:

 
2 + 3ı̂ 2 + 3ı̂ 4 − 5ı̂ (2 + 3ı̂)(4 − 5ı̂) 23 + 2ı̂ 23 2
= = = = + ı̂
4 − 5ı̂ 4 + 5ı̂ 4 − 5ı̂ 41 41 41 41


3−4ı̂
Exercise 1.3. Calculate 1+ı̂ .


As a special case of the above we can always take the reciprocal (or multiplicative
inverse) 1/z of a nonzero complex number.
Exercise 1.4. Find the reciprocal of 5 − 2ı̂.

3
 
Both Re(z) and Im(z) can be defined in terms of a complex number and its
conjugate.
Theorem 1.3.2. For z ∈ C we have:
Re(z) = 12 (z + z) and Im(z) = 1
2ı̂ (z − z)

Proof. Obvious.

Theorem 1.3.3. (Euler’s Formula)

If the complex number a + bı̂ makes an angle of θ with the positive real axis
and has norm r then we may rewrite it:

a + bı̂ = reθı̂

Proof. There are many ways to define ez for z ∈ C. One classic way is via the
Taylor expansion:
∞ k
X z
ez =
k!
k=0

which converges for all z ∈ C. Under this definition as well as the Taylor
expansions for sine and cosine we get the result.

Visually we have:
Im

a + bii = reθı̂
r
θ
Re

Corollary 1.3.1. As a result of the previous theorem:

(a) For all θ,r we have reı̂θ = r cos θ + r sin θı̂
(b) For all θ we have eiθ = cos θ + sin θı̂
(c) If z = a + bı̂ = reθı̂ then z = a − bı̂ = re−θı̂ .
(d) We can also manage expressions such as ea+bı̂ by noting that ea+bı̂ = ea ebı̂ =
ea (cos b + sin bı̂) but in reality this need rarely arises.

4
 √
Example 1.3. The complex number z = 3 + 1ı̂ makes an angle of θ = π/6
with the positive real axis and has magnitude 2. Therefore:
√
z = 3 + 1ı̂ = 2eı̂(π/6)


Exercise 1.5. Rewrite v = 5 + 5ı̂ in exponential form.

√
Exercise 1.6. Rewrite v = 4 + 4 3ı̂ in exponential form.

Exercise 1.7. Rewrite 5e(5π/6)ı̂ in standard form.


1.4 Translations
So now given a point represented by a complex number it’s clear that we can
translate the point by adding another complex number. In other words to
translate the point represented by z ∈ C by a units in the x-direction (real
diretion) and b units in the y-direction we simply add:

z 7→ z + (a + bı̂)

1.5 Scaling
By “scaling” we mean scaling away from the origin. If we take a complex number
and multiply it by a real number then obviously we scale that complex number,
moving it away from or towards the origin.
Example 1.4. Multiplying 2 + 3ı̂ by 4 results in 4(2 + 3ı̂) = 8 + 12ı̂ which is
four times as far from the origin.


1.6 Rotations
But what happens graphically if we multiply by a complex number?
Consider the most basic example, multiplication by ı̂. Look at the point (2, 3)
represented by 2 + 3ı̂. We calculate:

ı̂(2 + 3ı̂) = 2ı̂ + 3ı̂2 = 2ı̂ + 3(−1) = −3 + 2ı̂

π
Does this remind you of anything? It appears to rotate the point by 2 counter-
clockwise and in fact it does exactly this for any point:

5
 ı̂(a + bı̂) = −b + aı̂

If we stop and think for a second we might see a connection here. The point
ı̂ itself makes an angle of π/2 with the positive real axis and the result is a
rotation by π/2 radians about the origin.
So how about multiplying by an arbitrary complex number?
Well for starters let’s look at an arbitrary unit complex number that makes an
angle of θ with the real axis. What happens if we multiply it by the complex
number a + bı̂?
That unit complex number will have the form q = cos θ + sin θı̂ so let’s check
what happens if we multiply such a unit complex number by a + bı̂.

(cos θ + sin θı̂)(a + bı̂) = (a cos θ − b sin θ) + (a sin θ + b cos θ)ı̂

This is really familiar. If we phrase this with points it states:

The point (a, b) goes to the point (a cos θ − b sin θ, a sin θ + b cos θ).
And if we phrase this with vectors it states:
      
a a cos θ − b sin θ cos θ − sin θ a
The vector goes to the vector = .
b a sin θ + b cos θ sin θ cos θ b
So complex numbers can play the role of vectors in R2 and complex multiplica-
tion by unit complex numbers can substitute in for matrix multiplication when
it comes to rotation.
Thus we have:

Rotθ (a + bı̂) = (cos θ + sin θı̂)(a + bı̂)

Or alternately:

Rotθ (z) = eθı̂ z

This latter form is really handy and makes it completely clear why the product
of two rotations is a rotation.
We summarize all this in a theorem:
Theorem 1.6.1. Multiplication by the unit complex number eθı̂ = cos θ +
sin θı̂ rotates the complex plane counterclockwise about the origin by θ radians.
Alternately put, multiplication by a unit complex number α rotates the complex
plane counterclockwise about the origin by the angle that α makes with the
positive real axis.

6
Proof. In the notes above.

Exercise 1.8. Find the result when z = 10+7ı̂ is rotated clockwise by an angle
of π/6 about the origin.

(π/5)ı̂
Exercise 1.9. Find the result when z = 3e is rotated counterclockwise by
5π/3 about the origin.

Exercise 1.10. What happens if we multiply by a complex number which not
a unit complex number? Give an example to clarify your answer.

Corollary 1.6.1. For any z ∈ C, multiplication by z/|z| rotates the complex
plane so that z lands on the positive real axis.

Proof. If z makes an angle of θ with the positive real axis then z makes an
angle of −θ with the positive real axis. Then since z/|z| is a unit complex
number making that same angle of −θ, multiplication by z/|z| rotates the plane
counterclockwise about the origin by −θ which is the same as clockwise about
the origin by θ, thus placing z on the positive real axis.
√
Example 1.5. Multiplication by (2 − 3ı̂)/ 13 rotates the plane clockwise so
that 2 + 3ı̂ moves to the positive real axis.


1.7 Reflections
How about reflections, specifically those in lines through the origin. Well com-
plex conjugation reflects in the real axis since it negates the imaginary compo-
nent. Consequently we can reflect in a line through the origin as we did with
matrices and vectors, by first rotating the line to the real axis, then reflecting
in the real axis, then rotating back.
A line through the origin in the complex plane can be represented by a single
nonzero complex number just like a vector. So to reflect in the line represented
by a + bı̂ 6= 0 we first find the corresponding angle θ and then for any z we
can reflect it by first rotating the complex plane by −θ, then conjugating, then
rotating back.

7
 Reflθ (z) = (cos θ + sin θı̂)(cos(−θ) + sin(−θ)ı̂)z
= (cos θ + sin θı̂)(cos(θ) − sin(θ)ı̂)z
= (cos θ + sin θı̂)(cos θ + sin θı̂)z
= ...multiply and trig...
= (cos(2θ) + sin(2θ)ı̂)z

Which can also be written as:

Reflθ (x + yı̂) = (cos(2θ) + sin(2θ)ı̂)(x − yı̂)

or far more easily as:

Reflθ (z) = eθı̂ e−θı̂ z = eθı̂ eθı̂ z = e2θı̂ z

This is quite interesting because it asserts that reflecting in the line with angle
θ is equivalent to taking the conjugate and rotating it by an angle of 2θ. This
should not be at all surprising as we saw the same behavior with matrices earlier.
Example 1.6. To find the result when z = 2 + 5ı̂ is reflected in the line
represented by√ 3 + 1ı̂ notice that although
√ θ is not obvious or nice we do know
that sin θ = 1/ 10 and cos θ = 3/ 10. Consequently:
 √  √ 
sin(2θ) = 2 sin θ cos θ = 2 1/ 10 3/ 10 = 3/5
 √ 2  √ 2
cos(2θ) = cos2 θ − sin2 θ = 3/ 10 − 1/ 10 = 4/5

Thus reflection is given by the product:

 
4 3
Reflθ (x + yı̂) = + ı̂ (x − yı̂)
5 5

and so  
4 3 23 14
Reflθ (2 + 5ı̂) = + ı̂ (2 − 5ı̂) = − ı̂
5 5 5 5


Exercise 1.11. Find the result when z = 10 + 7ı̂ is reflected in the line which
makes an angle of θ = π/3 with the positive real axis.

Exercise 1.12. Find the result when z = 3 − 1ı̂ is reflected in the line through
the origin and through 5 + 8ı̂.

8
 
Exercise 1.13. As with matrices and vectors composing two reflections results
in a rotation. Show how this works with complex numbers.

Exercise 1.14. Show algebraically that if z ∈ C is on the line represented by
z0 ∈ C that reflecting z in that line just returns z.

Exercise 1.15. Assuming a + bı̂ makes an angle of θ with the positive real-axis
rewrite the formula:

Reflθ (x + yı̂) = (cos(2θ) + sin(2θ)ı̂)(x − yı̂)

without any sines or cosines.


Exercise 1.16. Consider the assertion:
 
2 −θ )ı̂
−( π ( π
−θ )ı̂
Reflθ (z) = e −e 2 z

(a) Show algebraically that this assertion is true.

(b) Provide a geometric explanation for the assertion.
Hint: The original development used complex conjugation as a key point
to reflect over the real axis. How would we go about reflecting over the
complex axis and using this instead?


1.8 Combinations
We may of course combine these operations, for example to rotate about a point
other than the origin we translate, rotate, and translate back.
Example 1.7. To rotate by π/4 about 1 + 2ı̂ we first subtract 1 + 2ı̂, multiply
by e(π/4)ı̂ , and then add 1 + 2ı̂. That is:

z 7→ e(π/4)ı̂ (z − (1 + 2ı̂)) + (1 + 2ı̂)

Alternately we can rewrite a + bı̂ in place of z, and cos(π/4) + sin(π/4)ı̂ in place

of e(π/4)ı̂ and then distribute and simplify.

Exercise 1.17. For the previous example do the rewrites and distribute and
simplify.


9
Exercise 1.18. Find the resulting point when z = 6 + 2ı̂ is rotated by 11π/6
clockwise around 5 + 3ı̂. Write this in the form a + bı̂.

Exercise 1.19. Find the resulting point when z = −2 + 1ı̂ is reflected in the
line given in Euclidean coordinates by y = 3x + 2. Write this in the form a + bı̂.


2 Angle and Closest Point Representations

2.1 Introduction
There are several ways to represent linex in C and we’ll investigate two of them
first. For lines through the origin we’ll use an angle and for lines not through
the origin we’ll use the closest point.

2.2 Lines Through the Origin

A line through the origin may be represented by the angle θ0 that it makes with
the positive real axis. This is not unique unless we put some restrictions on θ0
which we won’t do for now.

2.3 Closest Point Representation

Lines not through the origin may (as in R2 ) be represented by the unique point
on the line z0 closest to the origin.
For example the point z0 = −1 + 2ı̂ represents the line shown:
Im

z0 = −1 + 2ı̂

Re

10
2.4 Rotations of these Lines
Clearly rotating a line through the origin represented by θ0 about the origin by
θ radians counterclockwise results in another line through the origin represented
by θ0 + θ.
The mapping is then:

θ0 7→ θ0 + θ

Example 2.1. Rotating the line represented by θ0 = π/6 by an angle of θ = π/3

results in a line represented by θ0 + θ = π/6 + π/3 = π/2.

Similarly to rotate a line using the closest point representation we simply rotate
the point that represents the line and we get the point that represents the
rotated line, hence the mapping:

z0 7→ eθı̂ z0 = (cos θ + sin θı̂)z0

Example 2.2. Consider the line represented by z0 = 5 + 2ı̂. If we rotate this

by θ = π/3 about the origin the resulting line will be represented by:

e(π/3)ı̂ (5 + 2ı̂) = (cos(π/3) + sin(π/3)ı̂)(5 + 2ı̂)

√ !
1 3
= + ı̂ (5 + 2ı̂)
2 2
√ √
5−2 3 2+5 3
= + ı̂
2 2

2.5 Reflections of these Lines

Suppose now we have a line through the origin represented by θ0 and we wish
to reflect it in another line through the origin represented by some θ.
One way we can do this is think of our θ0 line as being a line through the origin
0 and through the point eθ0 ı̂ . We then reflect the point eθ0 ı̂ and look at the
angle associated to the resulting point.
We’ve seen that in general reflecting a point z is done via:

z 7→ (cos(2θ) + sin(2θ)ı̂)z = e2θı̂ z̄

so in our case reflecting eθ0 ı̂ result in:

11
 eθ0 ı̂ 7→ e2θı̂ eθ0 ı̂ = e2θı̂ e−θ0 ı̂ = e(2θ−θ0 )ı̂

The resulting line then has angle 2θ − θ0 and so overall the mapping is:

θ0 7→ 2θ − θ0

Example 2.3. Reflecting the line represented by θ0 = π/6 in a line through the
origin represented by θ = π/3 results in a line represented by 2(π/3)−π/6 = π/2.

Similarly to reflect a line represented by a closest point z0 we simply reflect z0 .
That is:

z0 7→ (cos(2θ) + sin(2θ)ı̂)z0 = e2θı̂ z0

Example 2.4. Reflecting the line represented by z0 = 2 − 5ı̂ in a line through

the origin represented by θ = π/6 results in a line represented by:

e2(π/6)ı̂ (2 − 5ı̂) = (cos(2π/6) + sin(2π/6)ı̂)(2 − 5ı̂) = ...

2.6 Translations of these Lines

There are four cases to deal with when it comes to translations:
(a) A line through the origin could translate to another line through the origin.
(b) A line not through the origin could translate to a line through the origin.
(c) A line not through the origin could translate to a line not through the origin.
(d) A line through the origin could translate to a line not through the origin.
Exercise 2.1. Under which conditions could each of the cases arise? Give a
specific example of each.

We’ll tackle these cases one-by-one.
(a) Suppose L passes through the origin and we translate by α ∈ C such that
T (L) also passes through the origin.
In this case L is represented by θ and α must translate along the same line
in order for the line to remain on the origin. Therefore T (L) can also be
represented by θ because it’s the same line. Of course we could add any
integer multiple of π to this, too.

12
 Exercise 2.2. Suppose a line represented by θ = π/4 is translated by
α = −3 − 3ı̂. Which angle represents the translated line?

(b) Suppose L does not pass through the origin and we translate by α such that
T (L) passes through the origin.
Suppose L is represented by closest point z0 .
Consider the following picture:
L

z0 θ T (L)

In this case the angle θ that T (L) makes with the positive real axis is the
same angle that L makes with the horizontal. This angle equals the angle
that z0 makes minus π/2. Of course we can add any integer multiple of π
to this:
π π
θ = arg(z0 ) − = atan2 (Im(z0 ), Re(z0 )) + + kπ
2 2

Note 2.6.1. The atan2(y, x) function returns the angle from the positive
x-axis to the ray from the origin to (x, y). It’s more succinct to use arg but
often atan2 is found in programming languages.
Note 2.6.2. You may wish to draw some pictures where z0 is in other
quadrants in order to convince yourself that the formula still works.
Exercise 2.3. Suppose the line represented by 2−2ı̂ is translated by α = 3.
Which angle represents the translated line?

(c) Suppose L does not pass through the origin and we translate by α such that
T (L) does not pass through the origin.
Theorem 2.6.1. Suppose L does not pass through the origin and has
closest point z0 . Let T denote translation by α ∈ C. Assuming T (L) does

13
not contain the origin show that the closest point representing T (L) is:
z0
z0 + Re(z0 α)
|z0 |2

Proof. Consider the following picture where the line L has been translated
by α to the line T (L).
The goal is to find the point z.

Im
z0 + α
α
z =???
T (L)
z0
Re

We proceed by first rotating the picture so that the line joining 0 to z0 to z

lies along the positive real axis. We can do this by applying the clockwise
rotation resulting from multiplication by z0 /|z0 |.
Observe that our desired z has moved to the rightmost point on the positive
real axis.

Im

z0
|z0 | (z0 + α)

Re

14
 We then get z by taking the real part of our |zz00 | (z0 + α) and rotating it
back. The result is as follows and can be simplified as shown:
 
z0 z0
z= Re (z0 + α)
|z0 | |z0 |
z0
= Re (z0 (z0 + α))
|z0 |2
z0
Re |z0 |2 + z0 α


= 2
|z0 |
z0  2 
= 2
|z0 | + Re (z0 α)
|z0 |
z0
= z0 + Re(z0 α)
|z0 |2

Exercise 2.4. Suppose the line represented by closest point z0 = −1 − 3ı̂

is translated by α = 2 + ı̂. Which point represents the translated line?

Exercise 2.5. Suppose the line represented by closest point z0 = 2e(π/5)ı̂
is translated by α = 3e(π/3)ı̂ . Which point represents the translated line?

(d) Supose L passes through the origin and we translate by α such that T (L)
does not pass through the origin.
Here’s a warm up:
Exercise 2.6. Suppose L is represented by 1 + 1ı̂. If T is translation by
1 + 0ı̂, which closest point represents T (L)? This can be done by inspection
so draw a picture.

Exercise 2.7. Suppose L is represented by 1 + 2ı̂. If T is translation by
1 − 7ı̂, which closest point represents T (L)? This probably cannot be done
by inspection but a picture can help you figure it out.

Theorem 2.6.2. Suppose L lies through the origin and makes an angle of
θ with the positive real axis. Let T denote translation by α ∈ C. Assuming
T (L) does not contain the origin show that the closest point representing
T (L) is:

z = eθı̂ Im αe−θı̂ ı̂

Proof. Omit.

15
 Exercise 2.8. Complete the above proof.
Hint: This is not hard and is based on the following picture. The goal in the
picture is to find z given α and θ0 . Think about rotating L to the positive
real axis first. If you’re totally confused you might look at (d) first since
the proof is given. Note that (c) was in fact a harder proof and not quite
the same but it might give you some ideas.
Im

z α L

θ0
Re

Exercise 2.9. Suppose a line L represented by θ0 = 2π/3 is translated by

α = 4 − 3ı̂. Which point represents the translated line?

Exercise 2.10. Suppose a line L passes through the origin and through
2 − 1ı̂. If this line is translated by α = 5 + 2ı̂ which point represents the
translated line?

Summary:
In summary we have:

Translation Type Corresponding Mapping

Through Origin to Through Origin θ0 7→ θ = θ0 + kπ

θ0 7→ z = ı̂eθ0 ı̂ Im αe−θ0 ı̂


Through Origin to Not Through Origin
π
Not Through Origin to Through Origin z0 7→ θ = arg(z0 ) + 2 + kπ
z0
Not Through Origin to Not Through Origin z0 7→ z = z0 + |z0 |2 Re(z0 α)

Note 2.6.3. We have not addressed here the issue of how we know which case
we have!

16
 
Exercise 2.11. Suppose you have a line not through the origin and you trans-
late it. Not knowing whether the result is through the origin you assume it isn’t
and use that approach. What should happen intuitively? Cross-check this with
the calculation to see that it does, in fact, happen. Is this useful?

Exercise 2.12. Suppose you have a line through the origin and you translate
it. Not knowing whether the result is through the origin you assume it isn’t and
use that approach. What should happen intuitively? Cross-check this with the
calculation to see that it does, in fact, happen. Is this useful?


2.7 More on Transformation of these Lines

Of course this can get even more contorted if, for example, We want to rotate a
line about a point which is not the origin. Naturally we know how to do it, we
translate so the point is at the origin, then rotate, then translate back. However
we need to understand that either of these two translations could result in any of
the four complications which arose in the previous section. Of course it’s worth
noting that most lines (in some statistical sense) don’t intersect the origin.
Exercise 2.13. Find the resulting line (either represented by an angle or point,
whichever is appropriate) when the line represented by 3 + 2ı̂ is rotated by π/6
about the point 1 + 0ı̂.

Exercise 2.14. Find the resulting line (either represented by an angle or point,
whichever is appropriate) when the line represented by 3 + 2ı̂ is reflected in the
line with Euclidean equation y = 2x + 1.

Exercise 2.15. Let L be the line represented by 1 + 2ı̂. If we rotate L about
1 + 1ı̂ by θ radians, what must θ be to ensure that the rotated line meets the
origin?


2.8 Locating Points on these Lines

Using the angle representation of a line through the origin makes it fairly easy
to see if a point is on that line, we simply check if the point makes the same
angle as the line.
However for the closest point representation of a line not through the origin it’s
not so obvious. It’s not hard, it’s just not obvious.

17
Example 2.5. As an example if −1 + 2ı̂ represents a line, how can we tell if
10 + 18ı̂ is on the line? One way would be to subtract them and see if the result
is perpendicular to −1 + 2ı̂ treating complex numbers as vectors. To frame this
differently, we could find the slope between the points and see if it equals the
negative reciprocal of the slope joining −1 + 2ı̂ to the origin.

The downside to this approach is that we move out of the complex numbers
to do the work. Another approach would be to rotate the point of interest by
π/2 about the representing point and see if the result is a scalar multiple of the
representing point.
Example 2.6. To check if 10 + 18ı̂ is on the line represented by −1 + 2ı̂ we
calculate:
i(10 + 18ı̂ − (−1 + 2ı̂)) + (−1 + 2ı̂) = −17 + 13ı̂
and observe that this is not a scalar multiple of −1 + 2ı̂ and hence 10 + 18ı̂ is
not on this line.

Example 2.7. To check if 5 + 5ı̂ is on the line represented by −1 + 2ı̂ we
calculate:
ı̂(5 + 5ı̂ − (−1 + 2ı̂)) + (−1 + 2ı̂) = −4 + 8ı̂
and observe that this is a scalar multiple of −1 + 2ı̂ and hence 5 + 5ı̂ is on this
line.

Exercise 2.16. Use this approach to check if the following points are on the
line represented by 3 − 1ı̂:
(a) 6 + 18ı̂
(b) −4 − 8ı̂
(c) −2 − 6ı̂

Exercise 2.17. Use this approach to find a generic criteria under which x + yı̂
is on the line represented by closest point a + bı̂ 6= 0. Your criteria should read
something like:
x + yı̂ is on the line represented by a + bı̂
iff
???? is a scalar multiple of ???

Note that checking whether v and w are scalar multiples is an existence check
but it can be rephrased as an equality check since v and w are scalar multiples
if when normalized they are either the same or opposites.

18
2.9 Parametrization
As a final comment we can represent a line by a point z0 ∈ C and an additional
α ∈ C which indicates a direction. This pair (z0 , α) then represents a line in
that all points z(t) = z0 + αt with t ∈ R.
The disadvantage to this is the introduction of another variable in the way we’re
describing the line in the sense that calculations involving the line almost always
require this other variable. For example checking if a point is on the line involves
solving for t. We’re not implying this is good or bad, just that it’s an added
factor.
Example 2.8. The line represented by the pair:

(z0 , α) = (−3 − 2ı̂, 4 + 1ı̂)

is shown here:
Im

Re
−3 − 2ı̂


Interestingly this representation is strongly tied into the closest point parametriza-
tion in that if z0 is the closest point on L to the origin then the line can be
written as the set of points
z(t) = z0 + z0 tı̂
This is fairly clear. If z0 is the closest point then the line through the origin and
z0 is perpendicular to L. This perpendicular line consists of points z0 t and when
we do z0 tı̂ we rotate by π/2 and get points through the origin perpendicular to
the line. The z0 + simply translates those points to lie on the point z0 .
This representation is easily understood because it’s similar to the traditional
parametrization of lines used in courses like Calculus 3:

r(t) = (x0 + at)ı̂ + (y0 + bt)̂ + (z0 + ct)k̂

19
One major downside to this representation is the fact that it involves another
variable t. Consequently finding points on the line and checking if points are on
the line invariably results in solving for t, which can be tricky.
A second major downside is that the representation is not unique. One line is
represented by many different z0 and α.
On the other hand it’s fairly easy to apply transformatons to parametrized lines
because the parametrization z = z0 + αt explicitly returns complex numbers
which can then be operated on individually.
Example 2.9. Consider the line parametrized by z = (2 + 3ı̂) + (4 − 5ı̂)t. If we
rotate this about the origin by π/4 radians the result can be written a number
of was. For example here’s one way:

(cos(π/4) + sin(π/4)ı̂)z = (cos(π/4) + sin(π/4)ı̂) ((2 + 3ı̂) + (4 − 5ı̂)t)

√ √
= ( 2/2 + ( 2/2)ı̂) ((2 + 3ı̂) + (4 − 5ı̂)t)
√ √ √ √
= (− 2/2 + 5( 2/2)ı̂) + (9 2/2 − ( 2/2)ı̂)t

The rewrite is done so that it’s clear from the result what the anchor point and
the direction complex number are. In other words the result has been rewritten
in the standard form.

Exercise 2.18. Find the result when the line z = (5+2ı̂)+(4−3ı̂)t is translated
by 3 + 7ı̂.

Exercise 2.19. Find the result when the line z = (5 + 2ı̂) + (4 − 3ı̂)t is reflected
in the line represented by v = 4 + 1ı̂.
Note: If you choose your approach carefully this isn’t particularly bad and the
result simplifies.


3 Complex Affine Transformations

As we’ve seen, rotations are performed by multiplication by unit complex num-
bers, scaling by multiplication by real numbers, and translation by addition of
complex numbers.
It follows that this comprehensively covers all transformations which satisy the
following definition:
Definition 3.0.1. A complex affine transformation is a transformation φ : C →
C which may be written in the from φ(z) = az + b where a, b ∈ C.


20
Theorem 3.0.1. The composition of two affine transformations is an affine
transformation.

Proof. Omitted.

Exercise 3.1. Show that the above theorem is true.


From this we get the following:
Theorem 3.0.2. Every combination of rotations, scalings and translations is
affine and every affine transformation is a combination of those.

Proof. The first part follows from the previous theorem. To see the second part
we simply take an affine transformation and rewrite it:
a
z 7→ az + b = |a| z+b
|a|

From here we can see that z undergoes first a rotation, then a scaling, then a
translation.

Not all transformations C → C are affine, however. In fact reflections are not,
and even complex conjugation is not.
Exercise 3.2. Show that complex conjugation is not affine. One way to do this
is to assume it is and find a problem. If it were, this would mean that there are
a, b ∈ C such that φ(z) = az + b satisfies φ(0) = 0, φ(1) = 1, and φ(ı̂) = −ı̂.
Show why this would be a problem.

Exercise 3.3. Show that reflections in general are not affine.

Exercise 3.4. Show that the following transformations are not affine:
(a) The mapping x + yı̂ 7→ x − yı̂.
(b) The mapping x + yı̂ 7→ x + 2yı̂.
(c) The mapping x + yı̂ 7→ (x + y) + yı̂.

2 2
It’s worth noting that in R all of the above are affine (we’ll say R affine) in
the sense that they may be written as:
φ(v) = Av + b with A a 2 × 2 matrix and b ∈ R2

21
For example the second from the exercises above is easy to see:
    
 1 0  x   x 
  = 
0 2 y 2y

It follows that the notion of being affine depends upon how the space is being
managed.
Exercise 3.5. Show that the three problems from the last exercise are R2 affine.

Exercise 3.6. Show that the mapping x + yı̂ 7→ x2 + yı̂ is not complex affine
nor R2 affine.

Which really goes to show that the complex numbers are in some ways at a bit
of a disadvantage. There are parallels between R2 and C but they’re certainly
not equivalent in terms of the ease of computations that we need to do.

4 Matrix Representation of Complex Numbers

We’ve seen in the chapter some parallels of how matrices and complex numbers
may achieve the same goals although phrased somewhat differently.
Interestingly complex numbers themselves may be represented as matrices. If
we set up the following representation:
 
a −b
a + bı̂ ←→ 
 

b a

Then we have:
Theorem 4.0.1. Under the correspondance given above we have:
(a) Addition of complex numbers is represented by addition of matrices.
(b) Multiplication of complex numbers is represented by multiplication of ma-
trices.
(c) The inverses of a complex number is represented by the inverse of a matrix.
(d) Division of complex numbers is taken care of by inverses.
(e) Complex conjugation is represented by the matrix transpose.

22
(f) The magnitude of a complex number is represented by the determinant of
a matrix.

Proof. Here’s a proof of (b): Observe that:

(a + bı̂)(c + dı̂) = (ac − bd) + (ad + bc)ı̂

and that:     
 a −b   c −d   ac − bd −ad − bc 
  = 
b a d c ad + bc ac − bd

We immediately see that the representation of the product matches the product.

For example rotation of a + bı̂ by θ degrees was given by:

Rotθ (a + bı̂) = (cos θ + sin θı̂)(a + bı̂)

In the language of matrices this becomes:

    
 a −b   cos θ − sin θ   a −b 
Rotθ   =   
b a sin θ cos θ b a

Not only is this useful because it reduces much of the handling of complex num-
bers to basic matrix calculations but this property will extend to quaternions,
as we will see later.
Exercise 4.1. Prove (a),(c),(d),(e),(f) of the theorem above. All of them are
fairly straighforward.


5 Taking this to 3D
Given that complex numbers can at most represent two dimensions via the real
and imaginary parts it’s not clear how this could extend to three dimensions.
For that we need quaternions, which extend complex numbers.

23