Imaginary Numbers Are Real

WORKBOOK
2
Hello!
"Imagine you had an art class in which they taught you how to paint a fence, but never
showed you the great masters. Of course, you would say; I hate art. You were bad at
painting the fence but you wouldnt know what else there is to art. Unfortunately, that is
exactly what happens with mathematics. What we study at school is a tiny little part of
mathematics. I want people to discover the magic world of mathematics, almost like a
parallel universe, that most of us arent aware even exists."

- Edward Frenkel

1. You Rock!
Thanks for checking out the Welch Labs Imaginary Numbers Are Real Workbook. Here you'll find everything you need to get the
most out of my Imaginary Numbers are Real YouTube series. For each video in the series, you'll find a workbook section complete
with the text and key figures from the video, more in-depth features covering interesting areas, and most importantly, exercises.

2. About Them Exercises.

As the quote above says much better than I can, the exercises we're given in math class are all too often terrible. At some point you've
likely been assigned the kinds of fence-painting exercises where you're asked to do the same type of problem over and over and over
with minor tweaks to the numbers. I hate this stuff. It's uninspiring and a bad use of your brain. That said, I promise I have done my
very best to avoid this type of problem here. Each exercise is here for a reason. No fence painting.

Of course, this is only half the equation. Like anything else, to get real value out, you must put real work in. And if you do, I promise
it's worth it. Imaginary numbers are rich and beautiful, and their history is fascinating. Really understanding this stuff will give you
tremendous perspective on the power and beauty of modern mathematics and science. The exercises for each section are divided into 4
parts:

Exercises Description

Designed to stimulate, you guessed it,

Discussion
discussion! No wrong or right answers.
Imaginary Numbers can be tough, and missing
key concepts can really make this stuff way less
Drill
fun - the Drills are designed to ensure your grasp
of the key concepts is sound.
The Critical Thinking exercises are where things
get good - these exercises are specifically
Critical Thinking
designed to question your mathematical
assumptions and grow your skills.

Challenge These questions are like, hard.

3. About Them Solutions.

I'm a firm believer in quick feedback when learning new concepts. This is why you'll find answers to the exercises in the back of the
book. I encourage you to check your work as you go. Of course, this doesn't mean you should flip to the back of the book as soon as the
going gets tough - a little suffering and uncertainty is good.

Enjoy!
@stephencwelch

Stephen Welch, 2016

 3
Table of Contents
"Hey, so what's like, in this book?"

Part Title Pages

1 Introduction 4-9

2 A Little History 10-17

3 Cardan's Problem 18-21

4 Bombelli's Solution 22-26

Numbers are Two

5 28-32
Dimensional

6 The Complex Plane 34-38

Complex
7 40-45
Multiplication

8 Math Wizardry 46-50

9 Closure 52-57

10 Complex Functions 58-64

Wandering In Four
11 66-71
Dimensions

12 Riemann's Solution 72-78

13 Riemann Surfaces 80-85

Solutions To Exercises 86-end

4
Imaginary Numbers Are Real
Part 1: Introduction

Lets take the function

y = f(x)

(1)
4
If we graph our function, we obtain the friendly
parabola of Figure 1. Now lets say we want to figure out 3

where the equation equals zero - we want to find the roots.

2
On our plot this should be where the function crosses the
x-axis. 1
As we can see, our parabola actually never crosses the
x-axis, so according to our plot, there are no solutions to
-3 3 x
the equation x 2 + 1 = 0. -2 -1 1 2

But theres a small problem. A little over 200 years ago -1

a smart guy named Gauss proved1 that every polynomial 2
Figure 1 | Graph of f (x) = x +1.
equation of degree n has exactly n roots. Our polynomial
has a highest power, or degree, of two, so we should have considered impossible for over two thousand years: the
two roots. And Gauss discovery is not just some random square root of negative one.2
rule, today we call it the Fundamental Theorem of Algebra. When we include this missing dimension in our
So our plot seems to disagree with something so analysis - our parabola ends up looking more like Figure 2,
important its called the Fundamental Theorem of Algebra, and needless to say, is a bit more interesting.
which might be a problem. What Gauss is telling us here is Now that our input numbers are in their full two
that there are two perfectly good values of x that we could dimensional form, we see how our function x 2 + 1 really
plug into our function, and get zero out. Where could these behaves. And we can now see that our function does cross
two missing roots be? the x-axis!3 We were just looking in the wrong dimension.
The short answer here is that we dont have enough So, why is this extra dimension that numbers possess
numbers. We typically think of numbers existing on a one not common knowledge? Part of this reason is that it has
dimensional continuum - the number line. All our friends been given a terrible, terrible name. A name that suggest
are here: zero, one, negative numbers, fractions, even that these numbers arent ever real! 4

irrational numbers like 2 show up. In fact, Gauss himself had something to say about this
But this system is incomplete. And our missing numbers naming convention:
are not just further left or right, they live in a whole
That this subject [imaginary numbers] has
new dimension. Algebraically, this new dimension has
hitherto been surrounded by mysterious obscurity,
everything to do with a problem that was mathematically is to be attributed largely to an ill adapted notation.
If, for example, +1, -1, and the square root of -1
a b had been called direct, inverse and lateral units,
instead of positive, negative and imaginary (or even
impossible), such an obscurity would have been out
of the question.
- Carl Friedrich Gauss (1777-1855)

So yes, this missing dimension is comprised of numbers

c d that have been given ridiculous name imaginary. Gauss
proposed these numbers should instead be given the name

2More on this later

3If youre paying attention, you should be thinking what the heck Stephen,
you said the graph would cross the axis twice (there would be two solutions),
Figure 2 | Graph of f (x) = x 2+1 where x includes imaginary numbers. Panels and the graph in figure 2 crosses like a million times! Great point Greg! The
a-d show "pulling" the function out of the page. reason for this is weve only plotted the real part of the graph to keep things
simple (ish) well cover the complete solution (which does have exactly 2
answers) in Part 13. Get excited.
1Mostly, the full proof took a little longer 4 J
2:45-3:20 Part One: Introduction 5
lateral1 - so from here on, lets let lateral mean imaginary.
To get a better handle on imaginary/lateral numbers
and really understand whats going on in Figure 2, lets
spend a little time thinking about numbers.
Early humans really only had use for the natural
numbers (1, 2, 3). This makes sense because of how
numbers were used as a tool for counting things. So to Figure 3 | From here on, lateral = imaginary. Gauss preferred the term lateral to
early humans, the number line would have just been a imaginary, we'll see exactly why in Part 6.
series of dots.
basically cutting edge technology for a couple thousand
As civilizations advanced, people needed answers to
years.
more sophisticated math questions like when to plant
The next big innovations to hit the number line were
seeds, how to divide land, and how to keep track of financial
the number zero and negative numbers, but it took some
transactions. The natural numbers just werent cutting it
time2 to get everyone on board. Since its not obvious3 what
anymore, so the Egyptians innovated and developed a new,
these numbers mean or how they fit into the real world,
high tech solution: fractions.
Fractions filled in the gaps in our number line, and were
2 Like a lot of time. Like thousands of years time.
1 Well explain why Gauss preferred the name imaginary in part six. 3 Remember 3rd grade?

Relevant Example Zero As Imaginary

Civilization Fractions Zero Negatives Number Line
Time Numerals Placeholder Numbers

Prehistory <3000 BC
1 2 3

Ancient Egypt 1740BC

1 2 3

Babylonia 300BC
1 2 3

Olmec 700-400BC
1 2 3
500BC-
Greek
100AD 1 2 3
200BC-
China
200AD -3 -2 -1 1 2 3

27 BC-
Roman
476AD 1 2 3

Cambodia 700AD
1 2 3
600-
India + Persia
1000AD -3 -2 -1 0 1 2 3

500-
Medieval Europe
1400AD 1 2 3
Renaissance 1300-
Europe 1700AD 1, 2, 3 -3 -2 -1 0 1 2 3

i
Modern Era >1700 AD 1, 2, 3 -1 1
-i

Table 1 | A brief overview of the history of numbers. It has taken quite some time for modern numbers to come to be. Only in the last couple hundred years do we see
imaginary/lateral numbers really accepted. The dates here are approximate, and keep in mind that most of these civilizations didn't actually have number lines! The
point here is visualize how numbers developed over time. Finally, notice the difference between zero and zero as a placeholder. In a positional number systems like
ours, the location of a digit carries meaning. The three in 23 means 3 units, while the three in 32 means 3 "tens", or thirty. We run into a problem if we need to tell the
difference between 30 and 300, and we don't have zero! This is the placeholder zero, it is not a concept alone, but a notational tool.
6 Part One: Introduction 3:20-end

zero and negative numbers were met with skepticism, Its not surprising that most of the people who have
and largely avoided or ignored. Some cultures were more lived on our planet would be suspicious of questions like
suspicious that others, depending largely on how people this. These problems dont make any sense. Even brilliant
viewed the connection between mathematics and reality. A mathematicians of the 18th century, such as Leonard Euler,
great example is here Greek civilization despite making didnt fully know what to do with negatives he at one
huge strides in geometry, the Greeks generally didnt accept point wrote that negatives were greater than infinity.1
negative numbers or zeros, after all how could nothing be So its fair to say that negative and imaginary numbers
something? raise a lot of very good, very valid questions, such as:
Whats even wilder is that this is not all ancient
Why do we require students to understand and
history - just a few centuries ago, mathematicians would
work with numbers that eluded the greatest
intentionally move terms around to avoid having negatives
mathematical minds for thousands of years?
show up in equations. Suspicion of zero and negative
numbers did eventually fade - partially because negatives Why did we even come accept negative and
are useful for expressing concepts like debt, but mostly imaginary numbers in the first place, when they
because negatives just kept sneaking into mathematics. dont really seem connected to anything in the real
It turns out theres just a whole lot of math you just cant world?
do if you dont allow negative numbers to play. Without
negatives, simple algebra problems like x + 3 = 2 have How do these extra numbers help explain the
no answer. Before negatives were accepted, this problem missing solutions to Equation 1?
would have no solution, just like we thought Equation 1 Next time, well begin to address these questions by
had no solution. going way back to the discovery of complex numbers.
The thing is, its not crazy or weird to think problems
like this have no solutions to solve x + 3 = 2, we subtract
3 from both sides, resulting in x = 2 - 3. In words, this
algebra problem basically says: if I have 2 things and I take
away 3, how many things do I have left?

- = -( ) ?
Figure 4 | Negative numbers don't always make sense. Two apples, take away three
apples equals...the anti-apple?

1 Sketchy.
 Exercises 1 7

Discussion Drill

1.1 Why do you think most people who have lived Plot each function.
on our planet would have been suspicious of negative 1.6 f(x) = x 2-4x+4
numbers?

1.2 Why do you think negative numbers have been

so widely accepted today, despite being somewhat
sketchy?

1.3 Do you think negative numbers should be taught

to elementary school students? Why or why not?

1.7 g(x) = -x 2+x+6

1.4 How would you explain negative numbers to a 5th

grader?

1.5 Which historical civilizations do you think em-

braced mathematics? Which didnt?
8 Exercises 1

Critical Thinking 1.9 So far weve seen a function with 2 roots, g, and a
function with no obvious roots, f (Equation 1). Find
1.8 Consider the better-behaved sibling of Equation 1: and plot a highest power of two (quadratic) function
with one obvious root. Call it h(x). Or else.
g(x) = x 2-1 y

a) Plot g(x). 4
y
3

4 2

3 1

2
-4 -3 -2 -1 1 2 3 4 x
1 -1

-2
-4 -3 -2 -1 1 2 3 4 x
-1 -3

-2 -4

-3

-4

1.10 You cut x feet from a 10 foot rope, leaving y feet

b) For what x-values does g(x)=0? g(x)=0 when x of remaining rope.
=_____ or x =______. These are called the 10
roots, zeros, or x-intercepts of g.
x y
c) The roots of g are much easier to find than the
roots of f (Equation 1), why is this the case?

a) Write an equation relating x and y.

b) Solve for y when x=7, x=10, and x=13.

c) One of your answers from part b should be neg-

ative. Is this result meaninfgul? What does a negative
answer tell you about your remaining rope? Does the
specific value of your answer matter, or is knowing
your answer is negative enough to reach a conclusion
about the remaining rope?
 Exercises 1 9

1.11 You have exactly 10 minutes to get to class for Challenge

your big test on lateral numbers! Your friend Gus
(whos weirdly in to this sort of thing) says that class 1.12 Consider the function:
is exactly 10,000 feet away. Lets say you travel to
class at a rate of r feet/minute. You would like to get
there a few minutes early. More specifically, lets say p(x) = x 4-10x 3+35x 2-50x +24
you will arrive t minutes early.
a) According to the fundamental theorem of algebra,
a) Write an equation that relates r and t. how many roots should p(x) have?

b) Find all the roots of p(x). You may use technology

if you would like, but its not necessary to complete
the problem.

b) Compute t for r=1250 feet/minute, r=1000 feet/

minute, and r=625 feet/minute.

c) One of your answers from part b should be nega-

tive. Is this result meaningful? What does a negative
answer tell you about when you will arrive to class?
10
Imaginary Numbers Are Real
Part 2: A Little History

Last time, we left off wondering how imaginary/later-

al numbers could help us find the roots of our equation
x 2 + 1, and further, how imaginary and negative numbers
even became a part of modern mathematics after be-
ing avoided and ignored for a couple of thousand years,
because, lets be honest here they dont really make that
much sense.1
But something happened in Europe around five cen-
turies ago that would no longer allow mathematicians to Figure 5 | Highly-paid professional artist rendering of Scipione del Ferro. 1465-
ignore these numbers. An Italian mathematician, Scipione 1562.
del Ferro, was trying to solve a problem not that different
than ours. other side. This is pretty easy in linear7 equations, we can
At some point youve likely seen the quadratic formu- just add, multiply, divide, or subtract until we get x alone.
la. This formula is super useful because gives us the roots2 Quadratics are a bit harder you may have learned to
of any equation with a highest power of two3- all you have do this in school it requires some cleverness, and factor-
to do is plug in a, b, and c,and out pops the answer: ing by completing the square.
Del Ferro was trying to do the same thing for his cubic
, (2) equation, and through some very clever substitution, he
eventually found a solution: 8
where:

(4)

Del Ferro was trying to find a formula like this for

equations with a highest power of three - cubics. The gen- Just like the quadratic formula, del Ferros new formu-
eral case is pretty tough: la allowed him to find the solution to cubic equations by
simply plugging in values. Table 2 shows a summary of
these types of polynomial equations and solutions.
Now, for some reason, the way mathematicians earned
so Del Ferro first considered the case where the x 2 term is money in the 16th century was through challenging other
missing4, and the last term is negative: mathematicians to what were basically math duels so
del Ferro kept his new formula a secret to use in his next
duel.
.

In the 16th century, negative terms were way to sketchy to

write, so del Ferro wrote his cubic as

(3)
and required c and d to be positive.5
Now that we have our equation set up, the game here is
Figure 6 | High-stakes math duel. Because thats obviously what math is for.
to get x by itself on one side, and all the constants6 on the Dueling.

1 Imaginary numbers: Hey, were having a little get together next week, were
hoping you can make it! Mathematicians: ehhh, were a little busy doing real What happens next is a bit of a long story heres the
math. quick version. Del Ferro kept the formula secret until he
2 Aka solutions, aka zeros lets just say it lets you find x!
3 Quadratic, hence the name was on his death bed, when he finally told his student
4 Making the equation easier to solve, this is called a depressed cubic
5 Notice we lost the a here as well. Were allowed to do this by dividing
through by a, and letting the new c be c/a and the new d be d/a. After all, 7 highest power 1, this shown in more detail in the first row of Table 2.
theyre just constants! 8 If you think del Ferro wasnt that clever, try solving equation for x yourself.
6 constants = a, b, c A full derivation is available at welchlabs.com/blog.
2:02-3:20 Part Two: A Little History 11

Type of Highest Solution

General Form Graphs Look Like General Solution
Equation Power Discovered

A long
Linear 1 time ago

Quadratic 2 ~2000 BC

Early
Cubic 3 * 1500s

Quartic 4 1540 AD

Quintic 5 Proven not to exist in 1824!

Table 2 | Polynomials and solutions. *For cubic functions, the general solution shown is to this simplified case. Note that as we increase our highest power,
polynomials become significantly harder to solve! The quartic case gets a little ridiculous - just one of the 4 solutions to the quartic equation is shown here - as you
can see - it doesnt quite fit.
surviving work of the original discoverer, del Ferro,4 he
figured that it wasnt such a big secret, and published the
formula in his book Ars Magna. 5
Cardan went on to improve on his borrowed formula,
even making it work for cubics that included an x 2 term.6
However, along the way Cardan came across a problem.
In a slightly different version of the equation written as x 3
= cx + d, under certain values of c and d7, the formula
Figure 7 | Niccol Fontana Tartaglia. 1500-1557. If you think your nickname would break.
sucks, it doesnt this guy got his jaw sliced by a soldier as a kid, leading to Lets take the innocent looking8
a stammer for the rest of his life, and being called stammerer (Tartaglia)
even in the equations he helped develop. (5)
Antionio Fior. Foir immediately thought he was invinci- when we plug into Cardans formula we get a result
ble, or at least invincible in a math duel,1 and challenged that involves the square root of negative numbers.9
a way more skilled mathematician, Fontana Tartaglia, to
a duel. Tartaglia had successfully solved similar cubics,
but had thus far been unable to solve cubics of del Ferro's
form. Suspecting Foir would be able to solve these tougher
problems, Tartaglia freaked out before the math-off and
figured out how to solve the equation at the last minute,
and proceeded to completely dominate Fior.2
Tartaglia then went on to share the formula with the
world! Not really, he kept it super secret so he could keep
kicking butt in math duels.3 That is, until a very talented
mathematician named Girolamo Cardano heard about Figure 8 | Plugging into Cardans formula. When we try to use Cardans formula
to evaluate the simple cubic of Equation 5, we run into a small problem.
the formula, and pressured Tartaglia to share he eventu-
ally went along, but only after Cardan swore to an oath of 4 Remember him? ...from like...the last page...
secrecy. Fortunately for us, after Cardan came across the 5 Ars Magna = The Great Art (referring to algebra, instead of the lesser
arithmetic). Fontana was not so happy about Cardan sharing his formula and
accused him of plagiarism and such. Drama ensued.
1 So not really invincible 6 Cardan did this through clever substitution. Given f (x) = x3+ bx2 +cx
2 This paragraph is a bit different in the accompanying video, where I got this +d, substitute x = x-b/3.
detail wrong. I said that Tartaglia had falsely claimed to be able to solve these 7 d 2/4-c3/27 < 0
problems, the corrected story above is what actually happened. 8 Cardan used this example in Ars Magna
3Mwhahahaha 9 Note that were plugging into Cardans modified version (shown in Table 2)
12 Part Two: A Little History 3:20-end

The square roots of a negative number created enough

of problem to stop Cardan in his tracks. Square roots ask
us to find a number, that when multiplied by itself, yield
the number inside the root sign. The square root of nine is
three because three times three is nine. Importantly, the
square root of nine is also negative three, because negative
three times negative three is also positive nine.

But what about roots of negative numbers? What is the

square root of negative nine? Positive three wont work,
and neither will negative three, so were stuck.

Cardan was stuck too - he didnt know of any num-

bers, that when multiplied by themselves resulted in a
negative. 1
Now this certainly wasnt the first time the square root
of a negative had shown up usually mathematicians
would interpret this as the problems way of saying there
are no solutions, and in many cases this is true.2 However,
in this case we know there is at least one solution, because
of the way cubics are shaped.3
Regardless of their coefficients, cubic functions will
always cross the x-axis at least once, meaning that our
equation x 3 = 15x+4 will have at least one real solution.
So what we have here is a problem that must have an
answer, a formula that has been proven to work. But when
we put these together and try to solve the problem with
our formula we quickly arrive at what appears to be
impossible - the square roots of negative numbers.
Sometimes, when things break in math and science
it means just that - theyre broken but there are other,
more interesting situations in which broken mathematics
give us the keys to unlock new insights. The way in which
Cardan's formula was broken turned out to be incredibly
important to mathematics and science, and thats what
well begin to discuss next time.

of del Ferros original equation. Cardans modification allows us to solve cases

involving negative values of del Ferros constant, c.
1 He actually did kinda know about these, but wasnt sure how to apply them
here. See exercise 3.17.
2 See Exercises 2.11 and 2.14
3 in fact, one solution to Equation 5 is just 4. Check out Figure 13 for an
example of how cubics are shaped.
 Exercises 2 13

Discussion

2.1 Why did del Ferro keep his discovery secret?

2.2 Why are problems like -16 strange?

2.3 Why do you think math duels are way less popular today than in the 16th century?

Drill

Plot Solve by Factoring (If Possible) Solve by Quadratic Formula (If Possible)

2.4 f(x) = x 2 - x - 2 0 = x2 - x - 2 0 = x2 - x - 2
y
x y
-2 4

3
-1
2
0 1

1 -4 -3 -2 -1 1 2 3 4 x
2 -1

3 -2

-3

-4

2.5 g(x) = x 2 + 3x + 2 0 = x 2 + 3x + 2 0 = x 2 + 3x + 2
y
x y
4
-3
3
-2 2

-1 1

0 -4 -3 -2 -1 1 2 3 4 x
1 -1

-2

-3

-4
14 Exercises 2

Plot Solve by Factoring (If Possible) Solve by Quadratic Formula (If Possible)

2.6 f(x) = x 2 - 2x + 1 0 = x 2 - 2x + 1 0 = x 2 - 2x + 1
y
x y
-1 4

3
0
2
1 1

2 -4 -3 -2 -1 1 2 3 4 x
3 -1

-2

-3

-4

2.7 g(x) = x 2 + x - 4 0 = x2 + x - 4 0 = x2 + x - 4
y
x y
4
-3
3
-2 2

-1 1

0 -4 -3 -2 -1 1 2 3 4 x
1 -1

2 -2

-3

-4

2.8 h(x) = x 2 + 1 0 = x2 + 1 0 = x2 + 1
y
x y
-2 4

3
-1
2
0 1

1 -4 -3 -2 -1 1 2 3 4 x
2 -1

-2

-3

-4
 Exercises 2 15

Critical Thinking 2.11 Sometimes problems like 2.10 are less straightfor-
ward.
2.9 Why is the Quadratic Formula useful? Does it
allow you to solve any types of problems that you a) Solve for x:
couldnt otherwise? (Hint: compare exercise 2.7 to 2.4-
2.6)

15
x

17

2.10 Solve for x: b) How did part a go? :) If you used the Pythagore-
an Theorem to solve for x, the result should have been
the square root of a negative number. Sketchy. Whats
going on here?
17
x

15
16 Exercises 2

2.12 Where do the parabola y = x 2-4 and the line y 2.14 a) Where do the parabola y = x 2-1 and the line
= 2x-1 intersect? y = 2x -3 intersect?

2.13 Where do the parabola y = x 2-4 and the line y b) Using the Quadratic Formula to solve part a should
= 2x-3 intersect? have resulted in the square root of negative number.
What does this mean about the problem? Plot the
parabola and line from part a below.

-4 -3 -2 -1 1 2 3 4 x
-1

-2

-3

-4
 Exercises 2 17

2.15 Solve the equation x 3 = 8 using del Ferros for- 2.17 Solve the equation x 3 = x + 2 using Cardans
mula (Equation 4). modified version of Del Ferros formula (shown in the
3rd row of Table 2).

2.16 Solve the equation x 3 + 6x = 20 using del Fer- Challenge

ros formula (Equation 4).
2.18 Derive the Quadratic Formula from
ax2+bx +c=0 by completing the square.

2.19 Derive del Ferros formula from x3+cx=d by

witchcraft.

2.20 What exact values of c and d result in the square

roots of negative numbers in Cardans formula (row
3 of, column 5 of Table 2)? Exactly how many real
solutions must this type of cubic have?
18
Imaginary Numbers Are Real
Part 3: Cardans Problem

We left off last time with Cardan and his broken

formula for finding the roots of cubic functions. Cardan
knew his problem had to have a solution but didnt
know what to do with the square roots of negative num-
bers that kept popping up in his equations.
Cardan came close to finding way to make his formula
work, but got stuck in an algebraic loop, where a bunch of
work would just lead him right back to where he started.
It took one more generation of mathematicians to get
to the bottom of this. Cardans student, Rafael Bombel-
li, made some incredible insights about what was really
going on here. Figure 9 | Rafael Bombelli. 1526-1572. Rafael was apparently pretty good at
draining swamps, which is cool...I guess.
Lets remember why Cardan was stuck the square
roots of negative numbers ask us to find a number, that
when multiplied by itself, will yield a negative. Neither like our old numbers. But remember - it does have the
positive, nor negative numbers will work. exact special property we need - when we square it, the
Bombellis first big insight1 was simply to accept that if result is negative. Further, since this number is neither
positive numbers wont work, and negative numbers wont negative nor positive, it must be something new.
work, then maybe theres some other kind of number out
there that will. Now, if there is some other type of number (6)
out there, a good follow up question is: what are we going
to call it? After all, we need to use it in our equations. Now if this all seems a bit fishy to you - like a slightly
Bombellis approach was a very practical one.2 Rather too convenient algebra trick, youre in good company.4
than dream up a new name and symbol3 - Bombelli sim- In fact, its hard to introduce imaginary5 numbers
ply let the square root of negatives be their own thing. In without them sounding like an arbitrary invention. How-
the past, mathematicians would have thrown in the towel ever, before we dismiss the square root of minus one as
here and declared the problem impossible, but Bombelli some abstraction invented to torture students, lets review
was able to press on simply by allowing the square root of what weve learned thus far.
negatives to exist. Cardan and Bombelli were genuinely stuck on a tough
Lets take a simple example of our new numbers the problem that they knew had a solution. What Bombelli
square root of negative one. Now, for being a new kind of was able to see, is that if he extended the existing number
number, it doesnt look very exciting and kind of seems system, as had been done so many times before,6 he could
solve the problem. 7 Just as people needed fractions, zero,
and negative numbers to solve new problems in past; to
solve this problem, Bombelli now needed the square root
of negative one to be its own, brand new, number.
Lets make sure were clear about what it means for the
square root of negative one to be its own number. If our
new number is truly a discovery and not an invention, it
should behave like the other numbers we already know
about it should follow the established rules of algebra
and arithmetic. And it turns out it the square root of mi-
nus one does, for the most part.
Figure 10 | The Square Root of Minus One. Heres the brand new number we need
Just as we can split apart the root of the product of
to fix our problems. It doesnt look like much, which is why our artist has added 2 positive numbers, 8 we can also split apart our square
these bold, dramatic lines.
4 Cardan and Bombelli felt the same way!
5 Lateral
6 From integers to fractions to zero to negatives.
1 which wasnt totally new - Cardan kinda knew this 7 Remember that extending the number system before allowed us to solve
2 Although, not that imaginative. problems we wouldnt have been able to otherwise, like x-3=1.
3 Behold the great mystery number kweuasdktst whose symbol is 8 Like sqrt 2* sqrt 3 = sqrt 6, but be careful, sqrt -1 * sqrt -1 does not equal 1!
and whos name must not be said. (not all rules apply to imaginary numbers).
2:42-end Part 3: Cardan's Problem 19
roots of negatives. The square root of minus 25 splits into
the square root of 25 times the square root of negative If youre really a
one.
number, act like one!
(7)

12
This process is important because it allows us to
express the root of any negative using the square root of
minus one. The square root of minus 25 becomes 5 times
the square root of minus one.
Figure 11 | Suspicion. Here we see twelve, an already established number,
interacting with our new number for the first time. As we can see, its not going
(8) well.

All this is important, but isnt enough to solve Car-

dans problem we still need to figure out how to deal
We can use this process to expand the root of any with the cube roots of these numbers.2 Bombelli was able
negative number, and write it as some number we already to solve our problem through one more powerful insight
know about, times the square root of minus one. here, and thats what well discuss next time.

Algebra with x Algebra with

(9)

Lets quickly make sure our new numbers follow the

same rules as other numbers. In algebra problems with
x, only like terms can be added and subtracted. 2x+3x is

5x, but 2+3x is just 2+3x.1 Likewise, 2 -1+3 -1=5 -1,

but 2 + 3 -1 is just 2 + 3 -1. Finally, unlike terms can
be multiplied just as in algebra with x: 5 times x is just 5

x, and 5 times -1 is just 5 -1.
Now, there are some cases where our new numbers
behave a little strangely, but these can often be avoided by
first separating out the square root of minus one. Table 3
shows some examples.
Table 3 | Algebra With the Square Root of Minus One. The square root of minus
Now that we have a grasp on how our new numbers one behaves like the numbers we already know about, for the most part. The
work, we can see how they fix one of our problems from one thing we are not allowed to do is shown in the bottom right square. Instead,
last time. We now have a strategy for dealing with the we should: rewrite the expression as -1
5 .
-1
2 = -1-1
52 = -1.10
=-10
roots of negatives. We can evaluate the square root of neg-

ative 9 we were stuck on, and obtain 3 times -1.

(10)

2 Cardans formula in Figure 8 involve the cube root of the square root of
1 Because 2 and 3x are not like terms negatives!
20
Exercises 3

Discussion 3.6 -5 + -5

3.1 Rafael Bombelli was born in 1526. For thousands

of years before Bombellis time, people used math-
ematics to solve all kinds of problems without any

need for -1. What do you think was so compelling

about Cardans problem that led Bombelli to consid- 3.7 -5 + 5
er using a completely new type of number to solve
it?

3.8 16 + -16

3.2 Imagine youre working on a hard math problem.

In fact, its so hard, that no one on the planet has

solved it yet. You work and work and work and get 3.9 -44
nowhere. Your friend Gus, who has been working on
the same problem, excitedly tells you that hes found
a solution! He shows you his work, and tells you that
if works perfectly with one minor catch. To solve the
problem he has allowed an entirely new kind of num-
ber to exist. A type of number that no one has used
3.10 -512
before. How would you respond to Gus?

3.11 7/
-7

Drill

Simplify the following. Helpful examples can be

found in Table 3 and Equation 9.
3.12
-4 -4
3.3 -16

3.4 -51
3.13 1/
-1

3.5 -4
 Exercises 3 21
Critical Thinking b) How did part a go? This problem dates back to
Cardan himself. In his book Ars Magna, Cardan
3.14 Why do all cubic functions have to have at least gives the task of dividing 10 into parts whose
one root? product is 40, and calls this manifestly impossible.
You may have reached a similar conclusion. However,
Cardan does push on in Ars Magna and gives

the solution 5+-15, and 5--15. Show that these
numbers add to 10 and multiply to 40.

3.15 Find two numbers that multiply to 15 and add

to 8.

Interestingly, even though Cardan gave this solution,

he didn't think very highly of it, saying: So
progressing arithmetic subtlety the end of which, as
3.16 Find two numbers that multiply to 11,187 and is said, is refined as it is useless.
add to 212.
Challenge

q(x) = 6x 4-5x 3+20x 2-20x -16

3.18 One solution to q(x)=0 is -4. Find all other
solutions.

3.17 a) Find two numbers that multiply to 40 and

add to 10.
22
Imaginary Numbers Are Real
Part 4: Bombellis Solution

Last time we decided to let -1 be its own new type of
(5)
number, hoping it would help us solve Cardans problem.
Doing this is helpful, but finding a solution requires one
more insight from our friend Rafael Bombelli.
Bombelli knew that because of the way cubics are (4)
shaped, as shown in Figure 13, Equation 5 had to have a

solution that didnt involve -1, it had to be a plain old
regular positive or negative number as he had seen before.

Figure 13 | Some Cubics. The end behavior of cubics mean they must have at
least one real zero. More specifically, as we move to the left or right on our graph Figure 12 | Reminder of Cardans problem. When we try to use Cardans
and follow our cubic curve, it must go up on one side and down on the other. otherwise functional formula (Equation 4) to evaluate the simple cubic of
Technically: as x, f(x) and as x-, f(x)-, or as x, equation 5, we run into a problem.
f(x)- and as x-, f(x).
We can cube these to show that these are in fact
equivalent to the left sides of Equation 11 and, more
His second big insight here was that for this to be the importantly, when we add the two parts, as Equation 4
case, the root of minus one parts of each half of the equa- tells us to do, we just get 4 which we know is a solution
tion must cancel out when added together! 1 to our original equation.4 We have found the solution to
Bombelli used this idea to equate the two parts of Cardans problem!5

the equation shown in Figure 12 to a+b -1 and ab -1, And whats really interesting is that our problem had
where a and b are constants we need to find.2 nothing to do with the square root of minus one and
neither did our answer however, along the way, we
found that by extending our number system to include the
(11) square root of minus one, we were able to find the solu-
tion.6 And it turns out that extending the number system
in this way is helpful in lots and lots of other problems as
We can first eliminate that pesky cubed root by cubing well.7
both sides of Equation 11. The result is a particularly So what did Bombelli do to celebrate after discovering
tough system of equations: a number crucial to the future of science and mathemat-
ics?8
He actually did nothing. He discounted his discovery
(12) and basically said it was a hack.9
As ridiculous as that seems now, Bombelli drew pretty
reasonable conclusion at the time. It just seemed a little
Bombelli was able to get around this through some
4The -1 parts cancel!
clever guessing and checking. If we look at our original
5 If this math seems a bit hacky, fear not, we'll learn a much more robust
equation (Equation 5), and test a few integers3 we eventu- approach soon.
ally see that 4 is a solution! If we substitute 4 into our new 6 The shortest path between two truths in the real domain passes through the
complex domain. ---Jacques Hadamard
equations, we can solve for a and b and obtain a = 2 and 7 Lots of problems. Like lots and lots of problems. And not just math prob-
b= 1. These values make the two parts of Equation 11 lems. Science problems. Engineering problems. Relationship problems. Ok,

equal to 2+ -1and 2- -1. not that last one, but the rest are legit.
8 MATH RAVE?!?!
9 The whole matter seemed to rest on sophistry rather than truth. - Rafael
1 They must be complex conjugates! Bombelli. In case you aren't quite up to speed on your random words from the
2 Were allowed to do this because a+b
-1+ a-b
-1 = 2a 17th century, Sophistry is the use of fallacious arguments, especially with the
3 1...nope, 2...nope, 3...nope intention of deceiving.
2:06-end Part 4: Bombelli's Solution 23
too convenient like this is a little trick devised just to
solve problems like this.1 Area Associated Sqaure Side Length
Squaring numbers, had, up until that point, largely
been associated with what the operation is named for 16 4
16=4
squares. A squares area is equal to the length of its side, 4
squared. So positive areas make sense but what could a

negative area be? What even is -1? 9 3
9= 3
Questions like these slowed down the development of 3

imaginary numbers. It turns out there is a much deep-

er and richer meaning lurking below the surface, but it 4 2 4=2
2
would take long after Bombellis death to be revealed.

1 1 1=1
1

?
-1 ?
-1=?

Table 4 | Sketchy Squares. Historically, the operation of squaring was associated

with what it was named for, squares. But what could a square with a minus area
possibly look like?

1 This is how a lot of student seem to feel when they first meet -1, and its a
completely legitimate reaction if you feel this way about it youre in good
company.
24
Exercises 4

Discussion 4.6 (5-2 -1)2

4.1 Why do you think Cardan and Bombelli mistrust-

ed -1?

4.7 (5 -1-2 -1)2

Solve for a and b:

4.8 2+-121 = a+b -1

4.2 The Greeks were arguably the most sophisticated

mathematicians of the ancient world, and today are
remembered for their many contributions to geometry.
If Bombelli were to travel back in time, do you think

he could convince the ancient Greeks to adopt -1? 4.9 -4 = a+b -1

4.10 4 = a+b -1

Drill Find a solution by guessing and checking:

4.11 a2 + b2 = 50

4.3 -1.(1+ -1) a - b3 = 6

4.4 (1- -1).(1+ -1)
4.12 a(a2 + b2) = 0
a+b=1

4.5 (2+3 -1).(2+3 -1)
 Exercises 4 25
4.13 a(a2 + b2) = 2 4.16 Show that:
a+b=2

(12)

by equating the "normal" and " -1" parts from
exercise 4.15.

Critical Thinking

In this section, we'll try to get a better feel for

exactly how Bombelli solved Cardan's problem,
x 3=15x+4.

4.14 How did Bombelli justify writing the equation

below? 4.17 Back in exercise 4.15, we only cubed the top line
of Equation 11. Cube both sides of the bottom line of
11, and turn the result into a system of equations as
the did in Exercise 4.16. What do you notice?
(11)

4.15 Show that cubing the top line of Equation 11

results in:

4.18 Using Equation 12, and the fact that x = 4 is a

solution to our Cardan's problem, x 3=15x +4, derive
that a = 2 and b = 1.
26 Exercises 4

4.19 The solution we found to Equation 12, a = 2 4.21 So, what is the expression below
and b = 1, mean that this must be true: (remarkably) equivalent to?

cube both sides to each equation to prove this.

If you feel that Bombelli's method is a bit

hacky, you're in good company. The sketchiness
here really surrounds taking the cube root of
our new types of numbers that involve -1.
Problems like this have confused lots of very smart
mathematicians. In the next few sections, we'll
find a nice way to think about these problems
that makes this process much simpler and
intuitive.1

Challenge

4.22 Solve for a and b:

4.20 Finally, show that using our solution above

does result in a solution to Cardan's problem,
x 3=15x+4.

4.23 Perhaps not surprisingly, Equation 12 has

more than one solution. Show that other solutions
to Equation 12 lead to more solutions to Cardan's
original problem, x 3=15x +4. You may use
technology. (Unless you're a complete bad-ass and
don't need it)

1 Once we cover this exciting new ground, we'll revisit this problem in the
Critical Thinking exercises in Section 8.
Notes . Doodles . Musings 27
28
Imaginary Numbers Are Real
Part 5: Numbers are Two Dimensional

We left off with Bombellis discovery that if he allowed

the square root of minus one to be its own number, he
could solve problems that had been stumping mathemati-
cians for decades.
Despite the usefulness of his discovery, Bombelli and
other mathematicians generally regarded it as a hack
after all, what could it possibly mean to take the square
root of a negative number? Just like our friends zero and
negative numbers before, the square root of negative one
was generally regarded with suspicion because it didnt
correspond to anything people could think of in the real Figure 14 | Leonard Euler. 1707-1783. Brilliant mathematician who got sick of
world. writing -1 all the time.

For this reason, the square root of minus one was giv- What is remarkable about this time period3 is that
en the terrible names imaginary or impossible. A century although imaginary and complex numbers were used in
or so later Euler began using the symbol i to indicate calculations and derivations4, the deeper meaning behind
the square root of negative one, making the algebra less these numbers was left undiscovered for over 200 years
clunky.1 after Bombellis death.
(13) Before we dive into this deeper meaning, lets think
about i algebraically for a moment.
Unfortunately, the name imaginary stuck around, and If we raise i to higher and higher powers, it doesnt get
thats still what we call these guys. In response, everything bigger as other numbers would. We know i squared is -1
on the original number line gets the name real.2 from the definition, and if we keep multiplying i by itself,
we see a pattern that repeats every four multiplications.5
Over and over and over and over. Hold on to that fact for
a few paragraphs.
-4 -3 -2 -1 0 1 2 3 4

The Numbers
Real Numbers Imaginary Numbers Work

^
Real 21=2 i 1=i
Figure 15 | The numbers get re-branded. Because everything thats not imaginary
must be real, obviously... 22=4 i 2=-1

23=8 i 3=-i i 3=i .i 2=i .-1

When we put together a real and imaginary part, we
24=16 i 4=1 i 4=i 2.i 2=-1.-1
get what we now call a complex number.
25=32 i 5=i i 5=i 4.i =1.i

Complex Number 26=64 i 6=-1 i 6=i 4.i 2=1.-1

3 + 2i 27=128

28=256
i 7=-i

i 8= 1
i 7=i 4.i 3=1.-i

i 8=i 4.i 4=1.1

Real Part Imaginary Part 29=512 i 9= i i 9=i 4.i 4.i =1.1.i

Figure 16 | Complex Numbers. When we put together a real and imaginary/lateral Table 5 | Patterns. When we raise real numbers to higher and higher powers, they
number, the result is what well call a complex number. get bigger (for real numbers greater than 1). Interestingly, this is not the case
with imaginary numbers. Instead, a pattern emerges.

1 Well start doing this too! Im really sick of saying and writing root minus 1 3 ~1600-1800 AD
2 Imagine how this must make lateral numbers feel. "Hey, we're going to call 4 Pensively
all the numbers, except for you, real." 5 It has a period of 4.
1:32-3:40 Part 5: Numbers are Two Dimensional 29
Lets return to our friend, the number line. Remember

3.3=9
that all the numbers we know about1 show up here, except
imaginary numbers. They are nowhere to be found.
If we think back to our original problem with roots of 3
negative numbers, we can visualize this using the number
line. Remember the issue we had was finding a number, -9 -6 -3 0 3 9 6 9
that when multiplied by itself, would yielded a negative.
To see this more clearly, well use arrows instead of
dots to indicate numbers (Fig 17). Multiplying a positive
by itself maintains direction on the number line it stays
positive. If we multiply by a negative, we flip directions, or
3.-2=-6
rotate by 180 o. Squaring a negative lands us in the positive
numbers because we start on the left side with our first -9 -6 -6 -3 0 3 3 6 9
negative and rotate 180 o when we multiply by the second
negative. So theres no way to land on a negative number
when squaring, a positive squared results in a positive and
negative squared requires starting in negative territory, -3.-3=9
and when we multiply by the other negative we arrive
back in the positive numbers. -9 -6 -3 -3 0 3 6 9 9
So what we need is something in the middle. A num-
ber that when we multiply by it, only rotates 90 o, not 180 o
as negatives do.
This is exactly what imaginary numbers do i squared
is negative one, meaning that the first i puts us 90 o from
the positive real numbers, and multiplying by i rotates Figure 17 | Multiplication using the number line. Multiplying by a positive
us 90 o further, exactly where we wanted to be, firmly in number maintains direction on the number line, while multiplying by a negative
switches direction. Squaring can never result in a negative number, because
negative number territory. 2 multiplying a positive time a positive maintains direction, while multiplying a
negative by a negative flips us back in the positive direction.

matches a 90 o rotation from i. As we keep raising i to

higher and higher powers, we keep rotating around, with
our values repeating every fourth power, just as they did
Imaginary
-1 -1 0 1 i
i
i.i = -1 i 2=i 6=-1 i=i5
Figure 18 | Multiplication by i. One way to understand what multiplying by i is as
a rotation by 90o. Multiplying i by itself moves us a total of 180o, exactly what we
need to land on -1.

Back to that fact youre hanging on to from Table 5.

Since multiplying by i corresponds to one 90 o rotation, if -1 -1 1 1 Real

we place our imaginary axis at a right angle to our num-

ber line, our algebra will perfectly fit with our geometry
(Fig 19).
If we start with the real number 1 and multiply by i,
algebraically we get i which geometrically corresponds i 3=i 7=-i i 4=i 8=1
to a 90 o rotation from 1 to i. Multiplying by i again results -i -i
in i squared, which, by definition, is minus 1 which again
Figure 19 | Pattern Matching. Understanding multiplication by i as a 90o rotation
1 Integers, zeros, fractions, rational numbers, irrational numbers, transcen- perfectly matches the behavior we see in Table 5. Crazy.
dental numbers
2 Minus i does the same thing!
30 Part 5: Numbers are Two Dimensional 3:40-end

algebraically.
So the insight here is that imaginary numbers do not
exist apart from the real numbers, but right on top of
them, hiding in a perpendicular dimension.
This is the deeper meaning beneath imaginary num-
bers. They arent just some random extra number or hack
- they are the natural extension of our number system
from 1 dimension to 2.
Numbers are 2 dimensional.1
And whats even more remarkable, is that if we accept
this that numbers have a hidden dimension we end up
not only with more complete mathematics, but incredibly
powerful tools for science and engineering.
Next time well show how and why thinking of num-
bers this way is useful.

1 If anyone tells you this is obvious or easy theyre lying. Despite using i in
calculations, generations of very bright mathematicians missed this for over
200 years.
 Exercises 5 31

Discussion 5.9 i 1000

5.1 Why might interpreting multiplication by a nega-

tive number as a 180 o rotation make sense? 5.10 i 1001

5.11 i 1002

5.12 i 1003

5.13 i -1
5.2 Why might it make sense for imaginary numbers
to be placed at a right angle to real numbers?
5.14 i -20

Critical Thinking

5.15 As shown in Figure 18, multiplying i by itself

(i 2 = -1) can be interpreted as a 180o rotation from
the positive real axis to the negative real axis. Since
5.3 Do you think numbers are two-dimensional? (-i )2 also equals -1, what could multiplying by -i
Why or why not? represent?

5.16 True or false?

Drill

5.4 i 2

5.5 i 20

5.6 i 21

5.7 i 22

5.8 i 23
32 Exercises 5

The central idea of this section, that imaginary numbers 5.17 Let's say a = 8 in Carnot's problem. Solve for x, the length
should be placed at a right angle to the real numbers, of one of our segments.
is not obvious. We know this because it took over 200
years after Bombelli's death to discovered. This idea was
finally unearthed by two non-mathematicians separately,
Casper Wessel and Jean-Robert Argand, around 1800.
However, like most good ideas in math and science, this
one took quite some time to be accepted. In 1831, the
mathematician Augustus deMorgan said of the topic:

"We have shown the symbol -1 to be void of
meaning, or rather self-contradictory and absurd."

It's interesting to think about how exactly How did Exercise 5.17 go? When Carnot solved this equation for
mathematicians struggled with and argued about these x, his result was a complex number (hopefully your result was
ideas during this period. A long standing practice in complex as well, specifically 4+4i and 4-4i). Carnot interpreted
mathematics, (dating most notably back to the Greeks), the fact that his result was complex as meaning that the cut
is the idea of geometric proof. Speaking very roughly: point we're looking for does not lie on AB, thus the problem
if we can represent a mathematical idea visually, it was physically impossible. However, other mathematicians (such
must be true. Whether or not this is the right way to as the Frenchman Abb Adrien-Quentin Bue) interpreted
approach mathematics is an open question, but this way this result differently. Roughly speaking, Bue agreed that the
of thinking is certainly valuable, and fueled mathematical problem was physically impossible, but argued that the results
was nonetheless meaningful. More specifically, if we allow
development for millennia. When -1 began showing up
in mathematics in the 17th century, many mathematicians imaginary numbers to exist in the dimension perpendicular to
attempted and failed to find a way to explain or the real axis, we obtain some type to picture like this:

understand -1 visually/geometrically. This difficulty C

contributed to the mistrust of -1.

Now that -1 has a strong visual interpretation
(existing at a right angle to the real numbers) and has
4i
4+ 4i
become firmly rooted in mathematics, it's interesting to
=
x

look back at historical attempts to show visually what

the -1 could possibly mean. One interesting example 4
comes from an 1803 publication from Lazare Carnot: A B
8
Given a line segment AB of length a, how can it be
divided into two shorter segments so that the product 5.18 What is the distance between A and C, shown above?
of their lengths is equal to one-half the square of the
original length?

Let's temporarily assume that drawing pictures to

represent mathematics is a good idea and draw a picture.
5.19 Does your answer to 5.18 make sense in terms of Carnot's
a original problem?
x a-x

A B
From Carnot's problems statement, we can write the formula: So, if we allow imaginary numbers to live perpendicularly to
the real numbers, the answer we obtain does make some type of
sense. Sure, there's still no way to cut the string the achieve what
we want physically, but the results also are not meaningless.
They tell us there is an answer, but we must move 4 units in the
perpendicular (imaginary) direction to find it. Problems like

this don't make a very strong case for -1, which helps me
understand why this matter was so contentions for so

long. Only later, when our geometric interpretation -1
was shown to be indispensable to math and science did
these ideas become widely accepted.
Notes . Doodles . Musings 33
34
Imaginary Numbers Are Real
Part 6: The Complex Plane

Last time we arrived at an incredibly powerful tool THE COMPLEX PLANE THE XY PLANE
y
in mathematics, science, and engineering: the complex Imaginary

plane. The complex plane is an extension of the number -3-3i

4

(-3,-3)
4

line, where we include the imaginary dimension vertical-

3 3

2
1+2i 2
(1, 2)
ly. Just as we can plot xy coordinates on the xy plane, we 1 1

can plot complex numbers on the complex plane. -4 -3 -2 -1 1 2 3 4 Real -4 -3 -2 -1 1 2 3 4 x

This arrangement is why Gauss preferred the term -1

1-i
-1
(1, -1)
lateral over imaginary and inverse instead of negative.1 -2 -2

Gauss suggested we should call the numbers to the right

-3 -3

-4 -4

of the origin are direct, numbers to the left inverse, and

numbers up or down lateral. Figure 20 | The complex plane and its xy cousin. These planes have a lot in
common. They both come outfitted for adventure with cartesian coordinate
systems. However, the complex plane has has a special twist that makes it way
GUASS NAMES OUR (DUMB) NAMES hotter than its boring cousin.
Lateral Imaginary
Imaginary

Inverse Direct Negative Positive 4

3
4+i
Lateral
Imaginary 5+4i + 1+3i
2 1+3i
5+4i
Figure 20 | Gauss names for numbers and ours. That this subject [imaginary
numbers] has hitherto been surrounded by mysterious obscurity, is to be

4+i
attributed largely to an ill adapted notation. If, for example, +1, -1, and the
square root of -1 had been called direct, inverse and lateral units, instead 1
of positive, negative and imaginary (or even impossible), such an obscurity
would have been out of the question. -Gauss
1 2 3 4 Real
Figure 21 | Adding complex numbers. We add complex numbers by adding the
Now that weve seen the complex plane, lets discuss real and imaginary parts seperately. This has a nice visual interpretation, adding
complex numbers is exactly like adding vectors, we place the second arrow at the
why its so powerful. Weve seen two dimensional planes end of our first, and wherever its tip lands is our result.
before2 where each axis represents a different quantity, in
fact we started our whole series with one. the same exact thing with vectors. Where complex num-
In a normal xy plane, theres no required connection bers really get interesting is through multiplication.
between the dimensions, no rules about how they should We can multiply complex numbers together by foiling,
relate to one another. On the complex plane however, just as we do with binomials in algebra, with the minor
we have the rules of algebra with complex numbers we complication that we know i 2 can be replaced with -1.
discussed earlier. These rules impose a very specific and This is a perfectly valid algebraic solution to our prob-
useful relationship between our two dimensions.
This first rule is how complex numbers add and sub-
tract. The real and imaginary parts add independently,
making complex numbers and the complex plane useful (14)
for problems involving movement in two dimensions. If
we travel in one direction for a certain distance, and then
in another direction, we can add the components of each
part of our trip together to find the total distance we have
traveled in each direction (Fig 21).3 lem but is only half the picture. There is another, equally
So thats cool, but as you may already know, we can do valid, way to think about multiplying complex numbers.
And it has everything to do with the complex plane.
1 We first saw this idea from Gauss back in Part 1. Instead of just telling you what this interpretation is,
2 Like the x, y (Cartesian coordinate system) plane
3 AKA Displacements
it will be more way more fun to try to figure it out with
2:04-end Part 6: The Complex Plane 35

some examples.1 To discover this deeper meaning for

yourself2, all you need know is the following: how to mul-
tiply complex numbers algebraically as we just did, how to c a2+b2=c2
plot numbers on the complex plane, the Pythagorean the- b
orem, and finally how to use arctangent to find angles. 3

a
Whats pretty cool here is that if youre able to figure
out what were after here the interpretation of complex
number multiplication using the complex plane youll
have figured out a super useful bit of math that was un-
known to the smartest mathematicians on the planet until
only 2 centuries ago.
y
Next time well uncover this interpretation of complex
multiplication using the complex plane and were going
to do it with only 4 examples:
x
Figure 22 | Tools youll need. These are the tools youll need for your mission,
should you choose to accept. You will also need to use complex multiplication
as shown in Equation 14 and to plot points on the complex plane as shown in
Figure 20.

Through considering each of these problems on the

complex plane and looking at the patterns that naturally
emerge, well arrive at the deeper meaning were looking
for.
Do try this at home, you'll find these problems nicely
laid out for you in the Critical Thinking section of the fol-
lowing exercises. Even if youre already a boss at complex
numbers or have no idea what Im talking about I prom-
ise its a valuable process - and well sort our all the details
next time. 4

1 ...I promise.
2 Just as Wessel, Argand, and Gauss did a couple centuries ago but Bom-
belli failed to (!).
3 or you could just google. But I promise youll learn less that way.
4 Do it. Seriously. Do it. Do the math problem. Sit down and do it. It will take
like half an hour tops. And if you figure it out, you can tell all your friends
your smarter than the 16th Century mathematician Gerolamo Cardano. And
theyll definitely think youre really cool. For sure. Do it.
36
Exercises 6
Discussion 6.6 -i(3+5i)

6.1 Why did Guass prefer the names direct, inverse,

and lateral over positive, negative, and imaginary?

6.7 (x+(1+i))((x +(1-i))

6.2 What do the xy plane and the complex plane

have in common? How are the different? Fill in your
answers in the super cool Venn diagram below. 6.8 (-2-i)2

Simplify (your answer should not have an i in the

denominator):

6.9

XY Plane Complex Plane

6.10
Drill
Simplify:
6.3 (1+i)(1+i)

6.4 (1+i)(1-i)
6.11

6.5 (2+2i)(1-2i)
 Exercises 6 37

Problem Solve Algabraically Solve Graphically

6.12 (2+i) + (1+2i)

Imaginary

1 2 3 4 Real

6.13 (-3+i) + 2i Imaginary

-4 -3 -2 -1 1 2 3 4 Real
-1

-2

-3

-4

6.14 (2+i) - (1+2i) Imaginary

-4 -3 -2 -1 1 2 3 4 Real
-1

-2

-3

-4
*Vaguely helpful hint: Try writ-
ing as an addition problem.

6.15 2.(2+i) Imaginary

1 2 3 4 Real
38 Exercises 6

Critical Thinking

6.16 What is the connection between complex multiplication and the complex plane? To figure this
out, we'll look at four examples. This is by far my favorite exercise in the entire workbook. It's cool because if
you figure it out, you'll have discovered something that mathematicians missed for over 200 years.

a) To get started, for each example compute the product, plot the two numbers we're multiplying together, and
plot the result. The first example has been done for you.1

1.(4+3i).i = Imaginary 2.(4+3i).2i Imaginary

8 8
4i+3i 2=
-3+4i 6 6

4 4
-3+4i 4+3i
i
2 2

-8 -6 -4 -2 2 4 6 8 Real -8 -6 -4 -2 2 4 6 8 Real

3.(4+3i).(4+3i) Imaginary 4.(2+i).(1+2i) Imaginary

24 8

18 6

12 4

6 2

-24 -18 -12 -6 6 12 18 24 Real -8 -6 -4 -2 2 4 6 8 Real

b) Use your results from a to complete the table below.

Problem Result Angle 1 Angle 2 Result Angle Distance 1 Distance 2 Result Distance

(4+3i).i -3+4i 36.9 90 126.9 5 1 5

(4+3i).
2i
(4+3i).
(4+3i)
(2+i).
(1+2i)

c) What patterns do you see? What is the connection between complex multiplication and the complex plane?

1 Wow, that Stephen Welch is such a nice guy, doing a whole problem for you!
Notes . Doodles . Musings 39
40
Imaginary Numbers Are Real
Part 7: Complex Multiplication
Imaginary Imaginary
8 8
-6+8i
6 6

4 4
-3+4i 4+3i 4+3i
i
2 2
2i
-8 -6 -4 -2 2 4 6 8 Real -8 -6 -4 -2 2 4 6 8 Real

Imaginary Imaginary
24 8
7+24i
18 6

5i
12 4

6 2
1+2i
4+3i 2+i
-24 -18 -12 -6 6 12 18 24 Real -8 -6 -4 -2 2 4 6 8 Real

Figure 23 | Our four examples from last time plotted on the complex plane. Well use these examples to figure out the connection between the complex plane and complex
multiplication.

Last time we left off with a real math problem: what is For each example, well add the angle of each com-
the connection between complex multiplication and the plex number to Table 5. Now lets look for a connection
complex plane? between our three angles.
To get to the bottom of this well use the four exam-
ples we mentioned last time. For each example, well plot Problem Angle 1 Angle 2 Result Result Angle

the two numbers were multiplying together in Figure 23.

Well also compute the result algebraically and add it to (4+3i).i 36.9 90 -3+4i 126.9
each plot.
Our job now is to look for patterns. Back in part five,
(4+3i).2i 36.9 90 -6+8i 126.9
we learned that i has something to do with rotation on
the complex plane. So a good thing to keep track of here
(4+3i).
will be the angle our complex numbers make with the real 36.9 36.9 7+24i 73.8
axis. (4+3i)
We can determine our angles using a little trigonome- (2+i).
try, specifically the arctangent function. 26.6 63.4 5i 90
(1+2i)
Imaginary
Table 6 | Patterns? What is the connection between our angles for each example?

4 3
3
4+3i 4 After a little pondering1, we see that the angle of our
2
36.9 result is exactly equal to the angles of the numbers were
3 multiplying, added together.
1
This is the first of half the connection were looking
Real
for: when multiplying on the complex plane, the angle of
1
4
2 3 4

Figure 24 | Arctangent. Well use the arctangent function to find the angle each of
our complex numbers makes with the real axis.
1 Hmmmm...
:55-3:00 Part 7: Complex Multiplication 41
our result is equal to the sum of the angles of the numbers multiplying, we obtain the distance from the origin of the
were multiplying. result!
Lets now have a closer look at our first two examples. We now have the complete picture. When we multi-
Notice that the angles are identical, but the resulting ply complex numbers on the complex plane, their angles
complex numbers are not. This means that just keeping from the real axis add, and their distances from the origin
track of angles alone is not enough to sufficiently describe multiply.
complex multiplication in the complex plane there is This is the connection we were looking for between
something else going on. complex multiplication and the complex plane.
So what is the difference between these examples? It
looks like multiplying by 2i has pushed our result further Imaginary
from the origin than multiplying by i.
A good follow up question is how much further? A + B= C
We can measure the distance between the origin and A.B=C
our complex numbers by forming right triangles and
using the Pythagorean theorem. C
Imaginary
B
C A
4
c2 = 32+42 A
4+3i B Real
3
c2 = 9+16
2 c
3
c2 = 25 Figure 26 | Our Result. When we multiply two complex numbers, their angles
add and distances to the origin multiply.
1 c=5
Real We now have completely separate, but completely
1
4
2 3 4
equivalent interpretations of complex multiplication. To
Figure 25 | Pythagorean Theorem. Well use the Pythagorean Theorem to find the
distance between each of our complex numbers and the origin.
multiply two complex numbers together, we can fol-
low the rules of algebra, OR, we can find each numbers
distance from the origin and angle to the real axis on the
Just as before, lets compute our measurement for each complex plane and multiply and add each.
example and look for patterns Well put the results in And whats really cool here is that although these
Table 7. approaches look and are totally different, but they do the
same exact thing. What were seeing here is the same un-
Result
Problem Distance 1 Distance 2 Result
Distance derlying process from two separate vantage points. I really
like this idea, because it reminds me that theres more
(4+3i).i 5 1 -3+4i 5 to math than what we see on the page. There are deeper
truths embedded in our universe, and math is one way of
expressing them.
(4+3i).
5 2 -6+8i 10
2i
(4+3i).
5 5 7+24i 25
(4+3i)
(2+i).
5 5 5i 5
(1+2i)
Table 7 | More Patterns? What is the connection between our distances for each
example?

After some more pondering1, we see that if we mul-

Figure 27 | Two perfectly good ways to multiply complex numbers. We can
tiply the distances from the origin of the numbers were multiply complex numbers algebraically as shown on the left, or we can use the
complex plane as shown on the right.

1 Also hmmmmm...
42 Part 7: Complex Multiplication 3:00-end

Now that weve made our discovery, lets formalize our Imaginary
results a bit. We found that the quantities we should keep i 2=-1
track of when multiplying complex numbers on the com- 4
plex plane are the distance from the origin and the angle
3
4+3i Rectangular
Form Polar Form
from the real axis.
These quantities turn out to be so important, that 2
5 4+3i = 5 36.9
we use them as another way to write complex numbers. Magnitude or Angle or
Instead of writing complex numbers as the sum of their 1
Modulus Argument
real and imaginary parts1, we instead write them as their 36.9
distance from the origin and the angle they make with the 1 2 3 4 Real
real axis. This is called polar form, and the distance from
Figure 28 | Two ways to write complex numbers. We can write complex numbers
the origin gets a special name, magnitude.2 in rectangular or polar form.
Multiplying complex numbers in polar form is super
easy we just multiply the magnitudes and add the an-
gles. Division is pretty simple too, especially compared to
dividing in rectangular form - to divide in polar form we
divide our magnitudes, and subtract our angles.
Next time well show that this discovery is not only
cool, but useful. Well use the complex plane to make hard
algebra problems easier, faster, and more intuitive.

1 Rectangular form!
2 Magnitude also goes by the name modulus, and the angle is also called the
argument.
 Exercises 7 43

Discussion Convert to rectangular form:

7.6
7.1 The main idea of this section is that there are
two very different, but completely equivalent methods
we can use to multiply complex numbers. How could
such seemingly different approaches can yield the
same exact results?

7.7

7.2 Can you think of an example where using the po-

lar form of complex numbers would make life easier?
7.8

7.9
7.3 Why do you think it took over 200 years for the
connection between complex multiplication and the
complex plan to be discovered?

Critical Thinking
Solve these suckers:
7.10 (1+i)(1+i)

Drill
Convert to polar form:
7.4 1+i 7.11

7.5 2-3i

7.12 How are 7.10 and 7.11 related?

7.5
44 Exercises 7

Convert to polar form, solve, and convert your 7.16 (1+i)(1+i)(1+i)

answer back to rectangular form. For extra
credit draw a pretty picture that shows what's
happening.1
7.13

7.17 (1+i)(1+i)(1+i)(1+i)

7.14

7.18

7.15

7.19

1 Each problem is worth 1 extra credit point. 5 extra credit points my be con-
verted into 1 cool point at the discretion of the point holder. Upon achieving 3
cool points, demand some type of prize from your teacher/professor/friends/
family members. Be persistent.
 Exercises 7 45

Challenge 7.21 Convert to rectangular form using sin and

cos:
In school you may have been required to
memorize some formulas like these: a)

(1)
(2) b)
(3)
(4)
c)
Since memorization is basically the worst thing
ever1, let's try to find a way to never have to
memorize these formulas again.
7.22 Substitute your answers from 7.21 into2:
7.20 Let's begin by considering two complex
numbers, each with a magnitude of 1:
Expand and simplify your result.

To make things a little more visual, let's say our

two complex numbers look something like this
on the complex plane:

Imaginary

7.23 Derive Equations 1 and 2 by equating the

real and imaginary parts of each side of your
result to 7.22.

Real

7.24 Derive Equations 3 and 4 by considering

the special case where = .
In terms of and , compute the product:

and add it to the plot above in the generally

correct location.

1 Memorization does not equal learning! 2 This is hopefully the equation you obtained in 7.20!
46
Imaginary Numbers Are Real
Part 8: Math Wizardry

Lets solve a simple equation:

Imaginary

(15)
1
What value of x would make this equation work?
If you said 1, great job, 13=1. Weve found one answer.
Now, are there any more answers? 1
-1 Real
Way back in part one we introduced the Fundamental
Theorem of Algebra, which says that a polynomial must
-1
have as many roots1 as its highest power. We can rear-
range our equation as x 3 -1=0 to make it a more obvious
polynomial, and since our highest power2 is 3, this equa- Figure 30 | The number one. The real number one shown on the complex plane.
tion must have 3 solutions.
There is a way to find all three answers without using 1 is...14, if we give each our xs a magnitude of 1,5 our
complex numbers along the way, but it involves perfect resulting magnitude will work out to 1 easy. But what
cube factoring and the quadratic formula and takes like about our angles?
7 steps. We know that when multiplying complex numbers,
our angles add, so we need an angle, that when added
together 3 times gives 0.6 Or 360. 360 seems a little more
reasonable, so whats the correct angle here?
Well, since were dividing 360 into three even parts,
the right answer here is 360 divided by 3, or 120. When
we put this together with the magnitude of 1, we have a
second solution! Its the complex number with a magni-
tude of 1 and an angle of 120.
This answer makes a lot of sense on the complex
plane:
Imaginary

1
Figure 29 | Solving Equation 15 by factoring. We can solve Equation 15 by
factoring, but doing it this way is hella long. And you have to remember how to
factor the difference of perfect cubes. And the Quadratic Formula. Lame. 120

Instead, lets try to solve the problem visually using the -1 120 1
Real

complex plane. Our question, in words, is: what numbers, 120

when multiplied by themselves 3 times, equal 1?
-1

Figure 31 | A solution! Our second solution is shown in green, the complex

Lets think about this problem using the polar form of number with and angle of 120 degrees and magnitude of 1. Here, we can see that
complex numbers we discussed last time. We can think multiplying this number by itself 3 times results in 1.

about 1, in the complex domain, as a number with a mag-

nitude of 1, and an angle of zero. Or 360.3
When we multiply numbers in the complex domain, Multiplying our answer by itself once results in a mag-
their magnitudes multiply and their angles add. Our nitude of 1 at an angle of 240, and multiplying by our
result should have a magnitude of 1, since 1 times 1 times answer again lands us exactly where we wanted to be at
a magnitude of one and an angle of 360 also known

1 AKA Solutions. Aka Zeros. 4 Or 1 to whatever power you want

2 AKA Degree 5 Because 1 = cubed root of 1. 1*1*1 =1!
3 or -360, or -720, 720, or 1080, or if youre up to it, 360*n, where n an integer. 6 This would actually yield our first answer, 1!
2:20-end Part 8: Math Wizardy 47
as...one.
So weve found an answer, the complex number with
a magnitude of one and an angle of 120. This may seem
like a strange answer for algebra class, so lets put it back
in Cartesian form.
Our good friend the unit circle can save us some time
here. We would like to know the rectangular coordinates
of a point on the unit circle at an angle of 120.1 According (-1,0)
to the unit circle, our answer is -1/2 for the real part and

3/2 for the imaginary part.2
Ok, so now that we have an answer in rectangular
form, lets try it! If we multiply out our result:

Figure 32 | Your BFF the unit circle. Yeah, its kind of a big deal. Today its going to
help us convert between polar and rectangular coordinates.

We find our missing answer going the other way around

our circle. If we start at negative 120 and multiply our
number by itself 3 times, we also land at the purely real

number one. So our missing answer is -1/2-( 3/2)i.3
So weve found all the answers to our problem x 3=1,
and shown how the complex plane allows us to find these
visually. In this case, using complex numbers saved us
some time over the algebraic approach and for more
complicated problems, the complex plane becomes even
more useful. For example, what if we change the power
in our original problem to say 8? That is x 8 -1=0. We
Pretty cool, right? We were able to solve a tough
could try to factor this4 or just have a quick look at our
algebra problem visually using the complex plane, and get
complex plane and realize that, just like our last problem
the same exact answer we obtained through factoring, as
where we needed to divide the unit circle into 3 equal
shown in Figure 29.
portions we need to, in this case, divide our circle into 8
Finally, weve only found two answers so far, and the
equal portions so our solutions need to be at 45, again
fundamental theorem of algebra demands that we have 3.
along the unit circle. All 8 answers. Done.5
Imaginary Next time well show how imaginary numbers are the
missing puzzle piece that make algebra complete.
1

-120
-120 Real
-1 1
-120

-1

Figure 33 | Our final solution to x3=1. We find our final solution, the complex
number with a magnitude of 1 and angle of -120, by moving clockwise around
our circle in three equal steps. 3 Notice that our 2 complex roots are complex conjugates, this will always be
the case when our polynomials have real coefficients.
1 We know were on the unit circle because our number has a magnitude of 1 4 (x^4-1)(x^4+1)= (x^2-1)(x^2+1)(x^4+1)= (x-1)(x+1)(x^2+1)(x^4+1)=0...
5 One solution would be 2/2+( 2/2)i . This is called an nth root of unity

2 You could also use a 30/60/90 special right triangle or sin and cosinex =
1*cos(120), y=1*sin(120). problem and has lots of cool applications, like the Fourier Transform!
48
Exercises 8

Discussion 8.5 (1 + i 3)6

8.1 Why must the equation x 3=1 have 3 solutions?

8.6

8.2 How does the complex plane make tough prob-

lems easier?

8.7 i

8.3 Why to you think the mathematician Jacques

Hadamard wrote:

"The shortest path between two truths in the real

domain passes through the complex domain."

-Jacques Hadamard, ~1945

8.8

Drill
Answer in rectangular form: 8.9
8.4 (1+i)6
 Exercises 8 49
Critical Thinking Your answer to 8.13 should be something like:

Let's say we're given a given complex number with a

magnitude1 of 1, and an angle2 of . We can convert
our number to rectangular form as follows: This is known as de Moivre's theorem.4 Note that the
right side of the equation is sometimes abbreviated
using cis: cis(x) = cos(x) + isin(x). De Moivre's
theorem can be expanded to numbers that have
8.10 Using a picture, show why the above equation is magnitudes other than 1 without too much headache.
legit. Let's say we're given some complex number with a
magnitude or r and an angle of . We can convert our
number to rectangular form as follows

8.14 Using a picture, show why the above equation is

legit.

Now let's consider what happens when we raise our

number to various powers. Write each answer in
rectangular form, using sin, cos, and .3

8.11 ( )2

Now let's consider what happens when we raise our

new number to various powers. Write each answer in
rectangular form, using r, sin, cos, and .

8.15 ( )2

8.12 ( )1/3

8.16 ( )1/3

8.13 ( )n

8.17 ( )n

1 or modulus
2 or argument
3 Feel free to explain you answer with a picture for one extra credit point per
problem. 4 Although it never shows up in his work...makes sense...
50 Exercises 8

Way the heck back in Part 4 (Exercises 4.14-4.21), we 8.20 Add together your answers to 8.19 a and b. If
discussed how Bombelli solved Cardan's problem by your result is 4, nice job! Think about how im-
allowing imaginary numbers to exist. A key part of pressed Cardan and Bombelli would by your math
Bombelli's argument was this rather shocking fact: wizardry. If you're answer isn't 4, check out the
solutions in the back of the book.

Now, with the help of the 200+ years of mathematical

advancement we've covered since Part 4, we'll show
that the above equation is must be true without break-
ing a sweat! This is remarkable when we consider how
much trouble this equation gave Cardan and Bombelli. Challenge

8.18 To begin showing that the above equation is true, The brilliant mathematician Gottfried Wilhelm
convert these two parts of the equation to polar form. Leibniz (1646-1716), co-father to calculus, is known
Round your answers to 4 decimal places. to have struggled deeply with the topics we're cov-
ering here.

a) 2 + -121
"I do not understand how...a quantity could be real,
when imaginary or impossible numbers are used to
express it."

-Leibniz

After his death, Leibniz's unpublished work re-

b) 2 - -121 vealed that he calculated some special cases of Car-
dan's formula again and again, presumably looking
for deeper insights like the ones that have been
handed to us. In solving the cubics x 3=13x+12 and
x 3=48x+72 using Cardan's formula, Leibniz discov-
ered that:

8.19 Using your results from the previous page,

compute the cube root of your answers from
8.18, and convert your results to rectangular and:
form.
a)

8.21 Show that the two equations above are true.

b)
Notes . Doodles . Musings 51
52
Imaginary Numbers Are Real
Part 9: Closure

Before we finish up the series and solve our problem So the set of natural numbers is not closed under
from part one, lets talk about how complex numbers are subtraction, but the set of all integers is. By expanding our
the missing puzzle piece that make algebra complete. number system, we can guarantee that any subtraction
Back in part two we saw how the definition of what question we can ask will have an answer.
a number is has evolved over time, beginning with the
natural numbers. The Egyptians figured out that these NUMBERS SYMBOL EXAMPLES CLOSED UNDER
numbers were missing something1, and its pretty obvious
to us today that the natural numbers are incomplete. Natural 1,2,3... +
However, as we saw with complex numbers, its not
always obvious when our numbers are missing something. ...-2, -1, 0,
Fortunately, there is a more sophisticated way to deter- Integers +, -
1, 2...
mine if we have all the types of numbers we need the
Table 8 | Closure under subtraction. To ensure we can handle all subtraction
mathematical idea of closure. problems, we must expand our number system to include integers.
Lets play a game. Ill give you a set numbers, and an
algebraic operation. I want you to tell me if any two num-
bers in the set, when combined with the operation, give a As we include more algebraic operations, we must
number not in the set. continue to expand our number system. Division requires
Our first set is the natural numbers and our operation us to expand our number system to include fractions
is addition. also known as rational numbers. Rational comes from
the word ratio3 rational numbers are numbers can be
SET OF NUMBERS OPERATION expressed as the ratio of two integers.4
We can show the relationship between the numbers
Natural: 1, 2, 3... + weve covered using another invention of Eulers.5 Using
Figure 34 | Are the natural numbers closed under addition?
an Euler diagram,6 we can visually express the idea that
one set includes another all integers are rational num-
bers, because we can always express them as a ratio of two
So the question is: are there any two natural numbers integers, but not all rational numbers are integers.
that when added together, produce something that is not a
natural number? NATURAL NUMBERS
After a little noodling, it should seem pretty reason- 1, 2, 3...
able that any two natural numbers, added together, result
in another natural number. Mathematically, we can say INTEGERS
that the set of natural numbers is closed under addition. ...-2, -1, 1, 2, 3...
Next, lets try the set of natural numbers and the oper- RATIONAL NUMBERS
ation of subtraction. 1.2, -5/2, 0.08...
SET OF NUMBERS OPERATION Figure 36 | Relationship between natural numbers, integers, and rational numbers.
All natural numbers are integers, but not all integers are natural numbers.

Natural: 1, 2, 3... - Lets recap. So far weve made it to rational numbers,

Figure 35 | Are the natural numbers closed under subtraction? which includes numbers like 1, 0, -5.1, and -2/3. What
operations are the rational numbers closed under?
For some pairs of natural numbers, like 6 and 4, things Well, any two rational numbers added together yield
work just fine 6 minus by 4 is 2, which is a natural num- another rational number, so we can say that rational num-
ber. But what about 2 minus 6? This results in an answer bers are closed under addition. We can say the same for
that is nowhere to be found in our set of natural numbers, subtraction, multiplication, and division.7
so our set is not closed under subtraction. We need to
expand our set to include zero2 and negative numbers for 3 Which comes from the Greek word logos, which means word! Word.
4 Like 2/3
this to be the case.
5 His other invention were making use of is the notation i = -1.
6 These are basically more flexible Venn diagrams.
1 Fractions! 7 With the notable exception of dividing by 0 this is a new can of worms.
2 Because, for example 3-3=0. called calculus.
2:48-4:25 Part 9: Closure 53
Now, what about powers and roots? Does a rational
number raised to a rational power always yield a rational SET OF NUMBERS OPERATION
result?
It turns out that for problems like (2/9)2 this is no Real:
problem our result is rational.1 Where we get into trou- Figure 38 | Are the real numbers closed under roots? Time for one last round of
everyones favorite game.
ble is things like (2)1/2 . Raising something to the power of
1/2 is the same thing as taking the square root, so this is real numbers, and our operation is taking roots. Do we

equivalent to 2. Well save the full argument2 for another have closure? Are there any real numbers, that when we

day, but it turns out that 2 is not rational there are no take some root, yield a result that is not a real number?

two integers, that when divided, equal exactly 2. Since The answer is, that despite all the types of numbers
numbers like this are not rational, we give them the name weve included along the way, were still missing some-
irrational. thing. We can write an expression only using real num-

Theres one more class of numbers that are even cooler bers and roots for example, -9, that has no solution in
than irrationals - the transcendental numbers like and the real numbers.
e, well also save these for another day. For this problem to have an answer, we must expand
our number system once more to include imaginary
NATURAL NUMBERS IRRATIONAL NUMBERS numbers. And taking all our real numbers from before,
1, 2, 3... together with imaginary numbers, we arrive at our broad-
INTEGERS est class of numbers the complex numbers.
...-2, -1, 1, 2, 3... TRANSCENDENTAL NUMBERS
RATIONAL NUMBERS NUMBERS SYMBOL EXAMPLES CLOSED UNDER
1.2, -5/2, 0.08... REAL NUMBERS
Natural 1,2,3... + x ()x
Figure 37 | The real numbers. The real numbers are an inclusive group! Real
numbers include the natural numbers, integers, rational numbers, irrational
...-2, -1, 0,
numbers, and even transcendental numbers. But it still feels like were missing Integers +-x
something... 1, 2...
So, weve expanded our number system again to in- Rational .4, 1/2, -16 +-x
clude irrationals, and all these numbers, taken together,
form what we call the real numbers.
.4, 2, 1/2,
Lets play our game one more time. Our set is now the Real +-x
, 51

+ - x
1 (2/9)2 =(4/81), which is hella-rational! 'Hella' is a super fun word Complex 2, .5, , 2i
()x
I learned while at graduate school in northern California, I think it
means something like "really". Use hella in a conversation for 1 extra Table 9 | Closure. Expanding our number systems to include closure under more
credit point. algebraic operations.
2 Its pretty cool, it means there are holes in the rational number line! See
exercise 9.19.

NATURAL NUMBERS IRRATIONAL NUMBERS

1, 2, 3...
INTEGERS
...-2, -1, 1, 2, 3... TRANSCENDENAL NUMBERS
RATIONAL NUMBERS
1.2, -5/2, 0.08... REAL NUMBERS
IMAGINARY NUMBERS i, 2i, -i/2... COMPLEX NUMBERS
Figure 39 | Finally. All the numbers we need to answer any algebra question we can think of using addition, subtraction, multiplication, division, powers and
roots. Complex numbers are closed under all algebraic operations.
54 Part 9: Closure 4:25-end

Initially, mathematicians were concerned that even

complex numbers were not sufficient that problems like

-i would result in an even more complex number
perhaps even a three dimensional number, instead of our
two-dimensional complex numbers.
Fortunately, this turned out not to be the case. In fact,

we can evaluate -i, again using our good friend, the
complex plane. Since -i has a magnitude of 1 and an angle
of -90, we just need a number with a magnitude of 1 and

an angle of -45, according to the unit circle, 2/2-( 2/2)
i.
So the square root of negative i is just another complex
number - theres no need for some wild new three-dimen-
sional number. In fact, theres no operation in the world
using addition, subtraction, multiplication, division, pow-
ers, and roots that the complex numbers cant handle.1 Figure 40 | The Complex numbers are closed under the operations we care about.

Imaginary numbers are the exact missing piece that Even crazy sounding problem like -i result in complex numbers. As long as we
stick to multiplication, division, addition, subtraction, powers, and roots, theres
make algebra complete. no need for some crazy, more complex number.

1 Except dividing by zero this leads us to calculus.

 Exercises 9 55

Discussion
9.6 Real Numbers, division.
9.1 What is mathematical closure?

9.7 Irrational Numbers, addition.

9.2 Why might closure be important for scientists,

engineers, and other people who use mathematics?

9.8 Irrational Numbers, multiplication.

9.3 The real numbers seem pretty complete. What

evidence do we have that indicates this is not true?

9.9 Transcendental Numbers, subtraction.

Drill

Time for a few more rounds of the closure game!

Determine if the set is closed under the given
operation. If the set is not closed, provide a counter-
example. 9.10 Imaginary Numbers, multiplication.

9.4 Integers, multiplication.

9.5 Integers, division.

9.11 Integers, powers.
56 Exercises 9

Critical Thinking Consider the completely made-up operation:

The idea of mathematical closure can be applied (a,b) = (a + b) modulo 4

to sets beyond those we've discussed thus far. For
example, let's consider the set of all even numbers Where modulo 4 means the remainder after dividing
(just in case you forgot, 0 is even). by four. For example, 9 modulo 4 = 1, because 4 goes
into 9 twice, leaving a remainder of 1. Numbers less
9.12 Are the even numbers closed under addition? than 4 remain unchanged, for example, 3 modulo 4 =
What about subtraction? For 1 extra credit point, 3.
prove that your answer it true for any two even
numbers. 9.15 Show that (5, 3) = 0, and (2, 1) = 3.

9.13 Are the even numbers closed under 9.16 Is the set {1, 2, 3, 4} closed under ?

multiplication? What about division? For 2 extra
credit points, prove that your answer is true for any
two even numbers.

9.17 Name a set that would be closed under .

9.14 Under what algebraic operations are the odd

numbers closed?
 Exercises 9 57
Challenge b. Since m is even, let m = 2k, for some integer k.
By plugging in 2k for m into your result from part
Using the kick-ass polar form of complex numbers we a, show that n must also be even.
learned back in Part 7, solve the problem below and
confirm that results belong to the set of Complex
Numbers.

9.18 (-i)1/6

c. Why can m and n not both be even?

9.19 2 is irrational. This means that 2 cannot be
expressed as a ratio of two integers. This fact didn't
fit so well with the orderly world view of the ancient
Greeks, and legend has it even lead to the murder of
Hippasus of Metapontum. Let's prove this dangerous Since m and n cannot both be even, we've reached a
fact by contradiction. To do this, let's assume, as contraction, indicating that our original assumption,

the Greeks did, that 2 is rational. If 2 is rational, that 2 is rational must be false, leaving us with the

then we can write 2 = m/n, for integers m and inevitable fact that 2 is irrational.
n, and let's say that m and n make our fraction in
lowest terms - so they have no common factors.

a. Square both sides of our equation 2 = m/n, and
show how we can conclude from our result that m
must be even.
58
Imaginary Numbers Are Real
Part 10: Complex Functions

Sowhat the heck is going on in Figure 2? a b

We saw this shape in part one and then proceeded to
not talk about it for like ten chapters. And at this point,
I wouldnt really blame you if you thought this was some
science fiction designed to get you excited about doing
boring algebra problems.
But before we chalk this up to unnecessary special
effects,1 lets remember where this shape came from. We c d
began our conversation in part one with an equation that
appeared to have no solution: x 2+1=0.
After all the work weve done so far, you can probably
to see how to find the answer algebraically: subtract one
from both sides and take the square root, resulting in x=i
and x=-i, done. Figure 2 | Graph of f (x) = x 2+1 where x includes imaginary numbers. Panels
a-d show "pulling" the function out of the page.

our x number line facing one way on a piece of paper, and

put our y number line one the same piece of paper, just
facing the other way. This forms a two-dimensional grid
known as the Cartesian coordinate system.
Apparently invented by Rene Descartes in the 16th cen-
Figure 41 | Solving Our Equation From Part One. After all the work we've done so tury after watching flies crawl around, the Cartesian coor-
far, this isn't so bad.
dinate system is a super powerful tool for understanding
the relationship between two variables.
So thats cool, but to really make sense of our shape The Cartesian coordinate system is powerful because
from part one we need to dig a little deeper, and talk it allows us to take abstract ideas, like functions, and turn
about functions of complex variables. then into something our brains can grasp much more
The kinds of functions most of us are used to, func- intuitively shapes. By giving each point on the plane
tions of real numbers2, have inputs and outputs that can its very own coordinates, Descartes was able to bring
each be visualized using a single dimension. This means together the two largest areas of mathematics at the time:
that all our x values fit on a single number line, and so do algebra and geometry. This greatly aided early efforts at
our y values. It seems pretty reasonable then3, that if we classifying functions by Isaac Newton and others, and
want to figure out how x and y are related, we should put today the Cartesian coordinate systems shows up every-
where; helping us do all kinds of things, like spot trends
y = f(x)
in data.
So thats all fun and wonderful but the Cartesian
4 coordinate system does come with a disclaimer.
It only works in 2 dimensions.4
3 y

DISCLAIMER
! x

-3 -2 -1 1 2 3 x Only works for 2 dimensions

-1

Figure 42 | It's easy to take the cartesian coordinate system for granted. Placing our
x and y number lines at a right angle to each other like this is not obvoius! Figure 43 | Limits of the Cartesian Coordinate System. It only works in two
dimensions :(

1 Transformers, anyone?
2 AKA Real-valued functions century.
3 Although no one figured this our until Descartes and Fermat in the 17th 4 Well, also 3, well get to that.
2:19-5:20 Part 10: Complex Functions 59
This limitation becomes a real problem when we start Imaginary Imaginary

to think about functions of complex variables. 4 4

These functions take in complex numbers for inputs, 3 3

2 2
and for the most part, also output complex numbers. This 1 1
means the numbers we put in and get out of our function -4 -3 -2 -1 1 2 3 4 -4 -3 -2 -1 1 2 3 4
Real Real
no longer fit on number lines. We now need two complex -1 -1

planes to keep track of our numbers: one for our input, -2 -2

-3 -3
one for our output. -4 -4
This raises an important question if, when visualiz-
ing these functions, what were really interested in is the
connection between the input and the output, how do we
Inputs Outputs
visualize whats happening on both planes simultaneous- Figure 44 | A Tale of Two Planes. To keep track of the inputs and outputs of a
function of complex variables, we need two complex planes.
ly?
We could try to fit our input and output planes to- and obtain our result:
gether somehow, as we did with our number lines for
real-valued functions, but we quickly run into a pretty
serious issue. As you likely know, the universe we live in
has1 three spatial dimensions so theres no way to fit the
four spatial dimensions we need into a single structure
that our brains can comprehend we simply run out of
dimensions.
Fortunately, there are some very clever ways to see the Plotting our inputs and outputs, we see that the point
relationship between two complex variables but before 1+i on our input z-plane was pushed, or mapped, by our
we can get to these, we need to think about the mathemat- function to 1+2i in our output w-plane.
ics of complex functions.
y v
Even though using separate planes for our input
and output is not a perfect solution, this approach can 4
f(z) 4

1+2i
3 3
still help us get started. Lets try it out with our original
1+i
2 2
function, f(x)=x 2+1. Before we begin, lets make a quick 1 1

variable name change make things easier down the road. -4 -3 -2 -1 1 2 3 4 x -4 -3 -2 -1 1 2 3 4

u
Well change the name of our input variable from x to z,2 -1 -1

and call our output variable w. Since z and w each have a

-2 -2

-3 -3
real and imaginary part, lets go one step further and give -4 -4

these parts names well let z = x + iy, so x and y repre- z w

sent the real and imaginary parts of z,3 and well let w = u Inputs Outputs
+ iv.
Figure 45 | Mapping. One way to visualize our complex function, f(z)=z 2+1,
w = f(z) = z +1 2 is by following how individual points are mapped from the input to
output plane.
z = x + iy (16)
w = u + iv Lets plug in a few more points to see if we can find a
pattern. If we test points along a straight line in our input
Just as we can use tables to keep track of our inputs space, we see that in our output space, our straight line is
and outputs for real-valued functions, we can also make transformed into a curved line, as shown in Figure 46.
tables to keep track of the inputs and outputs of complex Now, as you can imagine, plugging in points like this
functions. However, we now need four columns to keep can get pretty tedious. To speed things up, lets get a com-
track of our four variables: x, y, u, and v. puter do it for us. And instead of having our computer just
We can now experiment with our function: map certain points lets have it mapall the points.4
f(z)=z 2+1. If we plug a complex number into our func- Well take advantage of the fact that the images are
tion, for example z equals 1+i, we can do a little algebra just collections of pixels that happen to be arranged on a
grid. Well use some code written in the programming
1 Probably. language python to move every single pixel in an input
2 Many resources use x to denote a real number, and z for a complex number.
I have no idea why.
3 Respectively. 4 ALL THE POINTS!
60 Part 10: Complex Functions 5:20-end

y v z w
4 4
x y u v
3

2
3

2
1 1 1 2
1 1 -2 1 4 -4
-4 -3 -2 -1
-1
1 2 3 4 x -4 -3 -2 -1
-1
1 2 3 4 u -1 1 1 -2
-2 -2 0 1 0 0
2 1 4 4
-3 -3

-4 -4
z w

Figure 46 | Straight Lines Get All Curvy. If we map points along a straight line in our input space, the result is a curved line in our ouput space!
image to its proper location in an output image.
To make this work, well assign each pixel in our input
video a complex number that corresponds to its location
on the complex plane. We can then let our code take care
of the tedious work of moving each pixel to its new loca-
tion, as dictated by our function, z 2+1.
Our code will move points exactly as we did by hand
before - if we have a blue pixel at the 1+i location on our
input graph, this blue pixel will be moved to the 1+2i loca-
tion on our output graph, because f(1+i) is equal to 1+2i. 1
We saw before that our function warped a straight
horizontal line into a curvy one, so it should have some Figure 47 | Every point gets a number. To map all the points in the left image,
interesting effects on our video. Well include some ref- we'll assign each pixel a complex number that corresponds to its location on the
complex plane. Our code will then take care of the tedious work of mapping each
erence markers on top of our input and output planes to pixel to its proper location on the output plane.
keep track of our numbers, but we wont transform these
pixels.
Alright, ready?
Lets start by drawing some simple lines in a grid. In
our output plane we can see that our family of straight
lines is turned into a family of curved
lines. Cool, right?2
So weve found a pattern, but how is
this pattern explained by our function
z 2+1? And more importantly, how does
this fit with everything else weve learned
about complex numbers?
What else would be interesting to
draw in our input space to test our
mapping? What shape would you draw
to learn more about what our function is
doing?
Next time, more shapes.

Figure 48 | Mapping an Entire Image. Here we've mapped all the points on our z-plane to the w-plane.
1 Python code available on github.
2 Notice that the angles between our lines have been preserved! This means
our mapping is conformal.
 Exercises 10 61

Discussion 10.5 g(i)

10.1 Why is the Cartesian Coordinate System so
freaking cool?

10.6 g(1+i)

10.2 Why are functions of complex variables so

freaking hard to visualize?

10.7 g(1-i)

10.3 What is the maximum number of dimensions

the Cartesian Coordinate System will work in?
Why?

Let h(z)=z 3+iz. Calculate the following:

10.8 h(1)

10.4 What shape would you draw on the left plane

of Figure 48 to better understand what our func-
tion, f(z)=z 2+1, is doing? 10.9 h(-i)

10.10 h(1-i)

Drill
Let g(z)=z 2+2z. Calculate the following:
10.4 g(1)

10.11 h(2+2i)
62 Exercises 10

Critical Thinking Time (2 days Time (day

Water Depth (m) Water Depth (m)
before murder) before murder)
10.12 The Cartesian Coordinate System is so 12:00 AM 2.16 2:00 AM Gus Fell Asleep
common today, it's easy to forget just how useful 2:00 AM 1.95 4:00 AM 0.29
and non-obvious it is! Let's consider an example.
4:00 AM 1.35 6:00 AM 0.77

You and your friend Gus are planning a murder 6:00 AM 0.59 8:00 AM 1.33
in a quiet beach town.You and Gus agree to 8:00 AM 0.30 10:00 AM 1.81
hide the body in a cave located off Sunset Bay. 10:00 AM 0.20 12:00 PM 2.26
However, there's a small problem. The cave can 12:00 PM 0.77 2:00 PM 1.85
only be accessed at low tide. Being thorough pre- 2:00 PM 1.34 4:00 PM 1.27
meditators, you and Gus devise a plan to predict
4:00 PM 1.89 6:00 PM 0.51
exactly when low tides will occur, allowing you to
schedule the perfect murder. Over the course of 6:00 PM 2.07 8:00 PM 0.07
the days leading up to the Murder, Gus records 8:00 PM 1.96 10:00 PM 0.17
the water depth at Sunset Bay every two hours. 10:00 PM 1.43 12:00 PM 0.60
Check out Gus's data over there -----------------> 12:00 AM 0.66

a) The morning of the Murder is here! You and

c) Based on your plot, when should you hide
Gus pour over his notebook, trying to find the
the body? Did you answer to part a and b
perfect time to commit the act and hide the body.
match? Which method is more convincing? What
Based on the data alone, (without drawing any
advantages might using the Cartesian Coordinate
pictures) when do you think You and Gus should
hide the body? System have?

b) Gus is NOT happy with your answer to part

a. Let's use the Cartesian Coordinate System
to convince Gus. Using the grid below, plot the
water depth and time of day.
 Exercises 10 63
Figure 46 shows how one line was transformed from straight to curved by our function f(z)=z 2+1. In the exercises
below, we'll experiment with mapping other shapes. For each exercise complete the table using f(z)=z 2+1, plot the
input and output points, and speculate wildly about what exactly happened to the input shape.

Exercise Table Input Plane Output Plane

10.13 y v
z w
x y u v 4 4

3 3
-2 -1 2 2

1 1
-1 -1
-4 -3 -2 -1 1 2 3 4 x -4 -3 -2 -1 1 2 3 4
u
0 -1 -1 -1

-2 -2
1 -1 -3 -3

-4 -4
2 -1 z w
10.14 y v
z w
x y u v 4 4

3 3
1 -2 2 2

1 1
1 -1
-4 -3 -2 -1 1 2 3 4 x -4 -3 -2 -1 1 2 3 4
u
1 0 -1 -1

-2 -2
1 1 -3 -3

-4 -4
1 2 z w
10.15 y v
z w
x y u v 4 4

3 3
-2 -2 2 2

1 1
-1 -1
-4 -3 -2 -1 1 2 3 4 x -4 -3 -2 -1 1 2 3 4
u
0 0 -1 -1

-2 -2
1 1 -3 -3

-4 -4
2 2 z w
10.16 y v
z w
x y u v 4 4
Make Your Own Shape!

3 3

2 2

1 1

-4 -3 -2 -1 1 2 3 4 x -4 -3 -2 -1
-1
1 2 3 4
u
-1

-2 -2

-3 -3

-4 -4
z w
64 Exercises 10

Challenge
y
w = f(z) = z 2+1
(16) 4
z = x + iy
3
w = u + iv
2
10.17 a) Using Equation 16, draw curves/lines
1
on the z-plane to the right where u = 0.

b)Using Equation 16, draw curves/lines on

-4 -3 -2 -1 1 2 3 4 x
-1
the z-plane to the right where v = 0.
-2
c)What do your curves from parts a and b
tell you about the zeros of f? -3

-4
z
Notes . Doodles . Musings 65
66
Imaginary Numbers Are Real
Part 11: Wandering in Four Dimensions

Last time, we left off trying to think of shapes to draw Imaginary

on our input plane that would help us better understand
our function, f(z)=z 2+1. A + B= C
Since z 2 means multiply z by itself, and z is a com- A.B=C
plex number our functions behavior should have some
connection to complex multiplication. Back in part seven C
we saw that one way to interpret complex multiplication B
is as a rotation and scaling of our input values when we
multiply two complex numbers together, their magnitudes C A
multiply and their angles add. A
So the z 2 part of our function should take our complex
B Real
number, z, square its distance to the origin and double its
angle. The plus one portion of our equation is a little less Figure 49 | Complex Multiplication. When we multiply two complex numbers,
thier angles add and distances to the origin multiply.
exciting - adding a positive real number will move all our
points in the positive real direction, so to the right, in this
This turns out to be a straight line through the origin.
case by one. Since this shift to the right doesnt affect the
If we draw a few green lines through the origin as shown
behavior that were interested in, well leave it out of our
in the left pane of Figure 50, we see that the outputs are
equation for now.
also straight lines, at what looks like double the angle!
Lets test the idea that our function will double the
So weve shown that our transformation doubles the
angle of its input values. What kind of shape should we
angle of our input values now what about the magnitude
draw to test this idea?
our inputs? We said earlier that when squaring a complex
Ideally, we want to draw a shape that is made up points
number, the magnitude of the complex number should
that are all at the same angle, to see if our function chang-
also be squared. So the distance to the origin from our
es all points of the same angle in the same way.
input points should be squared in our mapping.1
So what kind of shape is made of points all at the same
angle? 1 Just don't forget that numbers<1 get smaller when squared!

Figure 50 | Mapping Pixels According to the Function f(z)=z 2+1. By carefully choosing the shapes we draw on z, we can learn more about our complex function. Here
we see that quarter circles are mapped to half circles with different radii, and lines through the origin are mapped to other lines through the origin with larger angles.
1:55-4:07 Part 11: Wandering in Four Dimensions 67
What kind of shape should we draw to test this idea? As we continue our circles, our new points have no-
We want to test magnitude alone so it would be nice to where to go except directly on top of our old points. This
have a shape with a constant magnitude. What kind of makes sense algebraically - because of the way squaring
shape has the same magnitude, or distance to the origin, works, points like 1+i and -1-i will map to the same exact
everywhere? output value:
This shape we need now is exactly a circle! If we add f(1+i) = (1+i)2
sections of a few circles to our picture as shown in Figure
= (1+i)(1+i)
50, we see that these result in new circle sections but
= 1 + i + i + i2
now at different distances to the origin. So as we expected,
our circles are preserved, but their radii are changed. = 1 + 2i - 1
So now were really getting somewhere by carefully = 2i
choosing our input shape, we were able to better under-
stand exactly what our function does to complex num- f(-1-i) = (-1-i)2
bers. = (-1-i)(-1-i)
Wonderful. = 1 + i + i + i2
But before we celebrate, lets keep a couple things in = 1 + 2i - 1
mind. For one, were dealing with a really simple function. = 2i
And secondly, even for this simple function we can still
run into trouble with our two complex plane setup. So its not that our function is broken or anything
For example, we know that our mapping doubles the its just that the technique were using to visualize it cant
angle of the input points. This is fine, until we use up too really handle multiple values being mapped to the same
much of our input space. If we continue the circles we location on the output plane after all which pixel
started earlier once we arrive at 180 degrees1, we begin should we display, the one at the 1+i location, or the one
to see a problem. at the -1-i location?
Our shapes have been expanded to fill up the entire
output space, but weve only used half the input space!

2i
1+i
-1-i

Figure 52 | Double Trouble. Two input points map to a single output point,
meaning the bottom left part of our circle ends up directly on top of the upper
left part. Not cool.

Mappings like this create problems in mathematics,

although typically in the reverse direction. The reverse of
a function, where our inputs become outputs and outputs
become inputs, is called the inverse. The inverse of our
function is pretty straight forward to find we simply
need to solve our equation for z: 2

Figure 51 | Out of Room! If we continue our circles on the z-plane, our mapped (17)
circles have no where to go on w, except directly on top of our other points.

2 Note, were adding the plus or minus here for clarity, some resources omit it
and take root(z) = +/- root(z), and some use the 1/2 power as an alternative to
1 AKA pi. represent specific parts of the function.
68 Part 11: Wandering in Four Dimensions 4:07-6:01

We now we run into the real math problem.

Our inverse function represents the same exact con- Function
nections as our forward function, just in the opposite
y v
direction. So our two inputs that mapped to the same
output are now a single input that maps to two outputs. 4 4

The w value of 2i maps to both z=1+i and z=-1-i. This 3 3

2i
2 2
mapping from a single input to multiple outputs is a big
enough problem that our functions inverse to not even
1
1+i 1

technically considered a function the definition of a -4 -3 -2 -1 1 2 3 4 x -4 -3 -2 -1 1 2 3 4 u

-1 -1
function requires that each input be mapped to one and -1-i -2 -2
only one output. -3 -3
To make things nice and confusing, non-functions -4 -4
z w
like this are called multi-functions. So our mapping is the
same in either direction, but taken from z to w is consid-
Inverse Multi-Function
ered a function, but from w to z is considered a multifunc-
tion.
Lets experiment with our multifunction. When we Figure 53 | Our Function and Its Inverse. We can think about our mapping in two ways,
left to right or right to left.
draw shapes on our w-plane, our shapes are duplicated
and shrunk own onto our z-plane:
Our shapes are copied because each point in w is
mapped to two points in z, and shrunk because the square
root function takes the square root of the magnitude of
our w values and divides each angle by 2.
Lets experiment with one more type of shape a path.
Well pick a starting point on our w plane, wander around
for while, and return to where we started. By following the
paths on both our w and z-planes, we can get some idea
of what happens as we wander around the 4-dimensional
space of our complex multifunction.
Our blue path returns us right back to where we start-
ed on both the z and w-plane. No surprise there. But if
change our path a little, something weird happens.
Figure 54 | Double Trouble for Realsies. When we draw on our w-plane, and map
our pixels to our z-plane, we end up with 2 copies of our drawing! Along our green path, our w path returns to where we
started, but our z pathsdont.

Figure 55 | What the heck? Our blue path ends up right where it started...but our green path jumps! What's going on here?
6:01-end Part 11: Wandering in Four Dimensions 69
Somehow our z values somehow jumped to a whole
new part of the plane! Somehow weve wandered our way
into a completely new part of our multifunction.
So it seems that some paths on w lead us back to where KIND OF
we started, but others dont. What could be going on here?
How is the complex landscape of our multifunction tak- A BIG
ing such similar paths in such different directions?
One reason I like math is that, for many problems,1
someone much smarter than me has already given them
DEAL
some serious thought, and quite often found an elegant Bernhard Riemann
solution. (1826-1866)

In this case, that person was one of Gauss students - Figure 56 | Bernhard Riemann. The work of Berhnard Riemann continues to have
a huge impact on modern mathematics. One of his ideas on prime numbers,
Bernhard Riemann. Well discuss his solution next time. known today as the Riemann Hypothesis, remains unsolved and is so important
to mathematics that a correct solution will earn you a cool $1M from the Clay
Mathematics Institute.

1 But certainly not ALL problems!

70
Exercises 11
Discussion Drill

11.1 Why did we choose to draw circles and lines Using the polar form of complex numbers we learned
through the origin in Figure 50? back in Parts 7 and 8, find all solutions to the the
multifunction:

-1
11.5 f (-2i)

11.2 In Figure 51 we run into a problem. Mutliple

oints from our z-plane are mapped to the same
location on our w-plane. Why does this happen?
-1
11.6 f (-1)

11.3 In Figure 54, why do we end up with 2 copies

of the shapes we draw on the w-plane?
11.7 i

-1
11.8 f (0)
11.4 Why do you think our paths behave so
strangely in Figure 55?
 Exercises 11 71
Critical Thinking

It's your turn to experiment with some paths! In each exercise below, compute the x and y values for each
point according to the function , plot the input and output points, and connect consecutive points
with a line. You may use technology.

Exercise Table z-plane w-plane

11.9 z1 z2 y v
w
x1 y1 x2 y2 u v 4 4

3 3
4 0 2 2

1 1
0 2
-4 -3 -2 -1 1 2 3 4 x -4 -3 -2 -1 1 2 3 4
u
0 4 -1 -1

-2 -2
2 23 -3 -3

-4 -4
4 0 z w
11.10 z1 z2 y v
w
x1 y1 x2 y2 u v 4 4

3 3
4 0 2 2

1 1
0 2
-4 -3 -2 -1 1 2 3 4 x -4 -3 -2 -1 1 2 3 4
u
-4 0 -1 -1

-2 -2
0 -2 -3 -3

-4 -4
4 0 z w
11.11 z1 z2 y v
w
x1 y1 x2 y2 u v 4 4

3 3
Make Your Own Path!!

2 2

1 1

-4 -3 -2 -1 1 2 3 4 x -4 -3 -2 -1
-1
1 2 3 4
u
-1

-2 -2

-3 -3

-4 -4
z w

Challenge
11.12 Explain why our paths jump in Figure 55.
72
Imaginary Numbers Are Real
Part 12: Riemann's Solution
Last time we left off wondering why some paths on our v
w-plane led us to completely new values on our z-plane, 4

while others didnt. 2i 3

Gauss student, Bernhard Riemann, made some 2

1
powerful insights into problems like this in the mid-nine- y
1

teenth century. -4 -3 -2 -1
-1
1 2 3 4
u
4
The first part of Riemanns contribution is the idea 3
-2

that, for functions like this, we need more than two 2

-3

complex planes to visualize our function. Since each point -1+i -4

w1
1+i
1

on our w-plane maps to two points on our z-plane, we -4 -3 -2 -1 1 2 3 4 x v

-1
can begin to resolve our ambiguity by adding a second -1-i -2
4

w-plane, and letting each of our two points on z map to its -3 2

2i
3

very own copy of the w-plane. -4

z
2

1
So thats fine, but it immediately raises an important
question: how do we pick which z values to map to each
-4 -3 -2 -1
-1
1 2 3 4
u
plane? A simple and effective approach here is to simply -2

divide the z plane into two halves well let the right half
-3

-4
map to our first w-plane, and our left half map to our w2
second w-plane.1 These restricted versions of our multi- Figure 57 | Riemann's Idea, Part I. The first part of Riemann's idea is to allow
function are called branches. each of our two solutions on z to map to its very own copy of the w-plane. This
Lets draw a path again, but this time just on our first does raise an issue though - which plane do we map other points, such as -1+i
to?
w-plane. Things look just fine until we cross the negative
real axis, and our path on the z-plane suddenly jumps! whats perhaps more disappointing here is that we havent
gained any insight into the interesting loop behavior we
saw last time, in fact we cant even recreate it with these
setup no matter what kind of loop we draw, as shown in
Figure 59, we always end up exactly where we started on
both the z and either w-plane, it seems that weve legal-
ized this behavior out of existence.

Figure 58 | Crap! Since we've decided to map our first w-plane to the right side of
our z-plane, when we cross the negative real axis on w, our path on z suddenly
jumps!

This, of course, is what must happen weve required

points from our first w-plane to only map to the right side
of our z-plane. Almost every2 point on w has two possible
solutions on z, and with our first branch, weve decided to Figure 59 | What Happened? By splitting up our function like this, we no longer
see the cool loop behavior in Figure 55. Lame.
always pick the one on the right.
So our path now jumps around on the z-plane, but
Further, the fact that our function jumps across the
1 These are called branches! Weve made a branch cut
z-plane means that our branches are discontinuous - a
2 Which point(s) on w dont map to two points on z?
1:59-end Part 12: Riemann's Solution 73
huge problem mathematically. Functions of complex vari- Lets pay careful attention to where our discontinuities
ables are a big part of modern mathematics and science, and show up. Well follow the points along a single path, and
if our functions are jumping around like this, we cant do to make sure we can tell the points of our path apart, well
important things like take derivatives and integrals.1 continuously change its color.
So weve fixed the multi-valued problem by splitting our As we move from quadrant one to quadrant two on z in
multifunction into branches, our function is now one-to- Figure 60, we switch branches. We switch back to our first
one2, but in the process, weve introduced some serious branch when moving from quadrant three to four. For our
issues - thus far Riemanns solution is not looking so great. function to be continuous, we need to somehow connect our
Fortunately, that was just part one, and part two is much two w-planes at the exact points where our path jumps.
cooler. What Riemann saw here was a way to bring together our
Lets consider our discontinuity problem in a bit more two complex planes in such a way that our multifunction
detail. Well switch back to our forward function momentar- would be perfectly continuous3, while maintaining the nice
ily, and draw again on our z-plane. one-to-one properties of our two w-plane solution.
To see Riemann's Solution, grab some scissors and tape,
1 In the language of calculus or real analysis, differentiable. Or in complex and check out the instructions below.
analysis, holomorphic or analytic.
2 Every input has exactly one unique output 3 And differentiable/analytic/holomorphic!

Figure 60 | Make your own Reimann Surface!

The great mathematician Bernhard Riemann saw a way to fit to- Imaginary
gether our two w-planes in such a way that our colored path would
be made continuous. Your job is to recreate Riemanns idea by cut- 2
ting out the two w-planes to the right, and positioning/cutting/tap-
ing them into a form that will make our curve continuous. This can
be accomplished by making a single cut in each plane. If you suc-
ceed, you will have created a truly beautiful piece of mathematics, 1
and the topic for our next section: A Riemann Surface.

-2 -1 1 2 Real
Imaginary

2 -1

1 -2
w1

-2 -1 1 2 Real

Imaginary
-1

-2
z 1

-2 -1 1 2 Real

-1

-2
w2
74 Notes . Doodles . Musings
 Exercises 12 75

-1
Discussion 12.5 f (2i)

12.1 What's good about Riemann's two complex plane

solution?
y
4

12.2 What's not so good about Riemann's two com-

-4 -3 -2 -1 1 2 3 4 x
-1
plex plane solution? -2

-3

-4
z
-1
12.6 f (-4)

12.3 Why do our paths jump in Figures 58 and

59?
y
4

Drill -4 -3 -2 -1 1 2 3 4 x
-1
Find all plot solutions to the the multifunction: -2

-3

-4
-1 z
12.4 f (4)
-1
12.7 f (-2i)

y y
4 4

3 3

2 2

1 1

-4 -3 -2 -1 1 2 3 4 x -4 -3 -2 -1 1 2 3 4 x
-1 -1

-2 -2

-3 -3

-4 -4
z z
76 Part 12 Exercises

Critical Thinking

12.8 Let's try to get a better feel for what's happening in Figure 58. For each w value in the tables below, compute the
corresponding z values. Each w value will yeild two z values, so choose the appropriate one based on which side of the plane
(left of right) it's on. Plot all points, and connect consecutive points to form a path when possible. For points that end up on
on the imaginary z (y) axis (these could be interpreted as left or right side), keep both solutions. For one cool point, make
your left and right paths different colors.
zleft w2 zright w1
xleft yleft u v xright yright u v
4 0 4 0
0 2 0 2
0 4 0 4
2 23 2 23
4 0 4 0

v
4

y -4 -3 -2 -1
-1
1 2 3 4
u
4
-2
3
-3
2
-4
1 w1
-4 -3 -2 -1 1 2 3 4 x v
-1
4
-2
3
-3
2
-4
z 1

-4 -3 -2 -1
-1
1 2 3 4
u
-2

-3

-4
w2
 Part 12 Exercises 77
12.9 For each w value in the tables below, compute the corresponding z values. Each w value will yeild two z
values, so choose the appropriate one based on which side of the plane (left of right) it's on. Plot all points, and
connect consecutive points to form a path when possible. For points that end up on on the imaginary z (y) axis
(these could be interpreted as left or right side), keep both solutions. For another cool point, make your left and
right paths different colors.

zleft w2 zright w1
xleft yleft u v xright yright u v
4 0 4 0
0 2 0 2
-4 0 -4 0
0 -2 0 -2
4 0 4 0

v
4

y -4 -3 -2 -1
-1
1 2 3 4
u
4
-2
3
-3
2
-4
1 w1
-4 -3 -2 -1 1 2 3 4 x v
-1
4
-2
3
-3
2
-4
z 1

-4 -3 -2 -1
-1
1 2 3 4
u
-2

-3

-4
w2
78 Part 12 Exercises

12.10 Now Forwards! Map each z value in the table below to its corresponding w value. Note that each z value
will map to either w1 or w2, with one exception: allow points along the imagary z (y) axis to map to both w1 and
w2 - this is a small hack (don't tell your mathematician friends) to make your graph easier to understand. Plot all
points, and connect consecutive points to form a path when possible.
z w1 w2
x y u v u v
2 0
1 2
0 2
-2 1
-2 0
-1 -2
0 -1
1 -1
2 0
v
4

y -4 -3 -2 -1
-1
1 2 3 4
u
4
-2
3
-3
2
-4
1 w1
-4 -3 -2 -1 1 2 3 4 x v
-1
4
-2
3
-3
2
-4
z 1

-4 -3 -2 -1
-1
1 2 3 4
u
-2

-3

-4
w2
 Part 12 Exercises 79

Challenge

12.11 In Figure 58, our first discontinuity on

z occurs when we first cross the negative real
axis on w. Why does this happen at this specific
location? Is this the result of how our function
behaves, or of a choice we made?
80
Imaginary Numbers Are Real
Part 13: Riemann Surfaces
This is a Riemann Surface.
Its going to help us think in four dimensions.
We made it by cutting our planes at the discontinuities
in our paths, and taping them together in a way that made
our paths from one plane to the other continuous.
Riemanns big insight here was that the domain, the
input values of our multifunction, should not be a flat, two
dimensional plane. Our domain should instead be this
a curved surface1 living in higher dimensional space. A
Riemann surface.
Whats incredible here is what the geometry2 of our
Riemann surface is going to allow us to do. Using our
Riemann surface as the input space of our multifunction,
we can fix literally all the problems weve encountered thus
far: our function will one to one, continuous, and our Rie-
mann surface will even help us elegantly explain the weird
loop behavior we saw in part eleven.
Lets see how.
Riemann envisioned these surfaces as sheets covering
the input plane, w.3 Our Riemann surface is constructed
from two copies of the complex plane, and the idea here
is that each input value on w lies directly below its corre-
sponding points on each layer of our Riemann surface.
If we follow a line straight up from the value 2i on our
w-plane, we find two points that correspond to w equals 2i.
Our Riemann surface fixes our one-to-one problem just as Figure 61 | The Riemann Surface for Our Function . We made
this surface by printing the paths from Figure 60 on transparencies, cutting the
our 2-complex plane solution did each of our two solu- planes at their discontinuities, and taping the paths together.
tions on z corresponds to its very own copy of the w-plane,
these are called braches.
So our Riemann surface makes our mapping one-
to-one, just as our two-complex plane approach did last
time, but what about continuity? As we saw last time, a
big problem with our two-complex plane solution is that it
introduced discontinuities.
We constructed our surface in such a way that our
colored path was continuous, but we encounter a weird
self-intersection along the negative real axis - our planes
have to pass right through eachother. Lets dig a little deep-
er. Remember the main idea here using our Riemann
surface as the input space, the domain, to our complex
multifunction should give us clarity. So instead of drawing
paths or shapes on our w-plane, we should really be draw-
ing on our Riemann surface.
Drawing on three-dimensional surfaces can be a little
challenging, so well make use of a tool that wasnt invent- Figure 62 | The Riemann Surface for Our Function . We made
ed until over a century after Riemanns death a computer! this surface using the programming language Python and the plotting library
Plotly.
Just as with our paper version, well start with our
w-plane lying flat on the ground, and place our Riemann Before we start drawing paths all over our surface, lets
surface directly above. make sure we know what were looking at. Were trying
to understand the complex function w = z 2, or taken in
1 Manifold!
2 Topology, really.
the other direction, z equals plus or minus the square
3 Or sphere! root of w. The mapping between w and z is the same both
3:13-5:55 Part 13: Riemann Surfaces 81
directions. The visualization challenge here is that our
mapping is four-dimensional, both z and w have real and
imaginary parts back in part 10 we called these x, y, u
and v.
What were seeing when we visualize our Riemann
surface is a two-dimensional surface in three-dimen-
sional space. In this case, since we positioned our surface
directly above the w-plane, two of our three dimensions
correspond exactly to the real and imaginary parts of w,
we called these u and v. Part of Riemanns idea is that our
third dimension should represent the z-value of our func-
tion. But as we know, z is complex number it has both a
real and imaginary part, so theres no way to show both of Figure 63 | Each point on our surface is a solution to our function. The point
these on a single axis. w=-2i, when plugged into our multifunction yields two
solutions, z=-1+i and z=1-i. These solutions show up as the points (u,v,x,y)
Whats often done when visualizing these surfac- = (0,-2,-1,1) and (0,-2,1,-1) on our Riemann Surface. 3 of our 4 coordinates
es, and what were doing here, is to simply pick the real show up as our points location in 3D space.
or imaginary part of z,1 and use this value as the third
dimension, the height, of our surface. In Figure 63, were run into a bit of a problem when we hit our self-inter-
using x, the real part of z. section, after all, should we stay on the same w-plane, or
Doing this has a nice visual result: each point on our hope to the other one?
Riemann surface lives in 3d space at a location corre- To answer this question lets try to figure out why our
sponding exactly to its u, v, and x values. So each point on surface self-intersects in the first place. This intersection
our surface represents a single solution to our equation, happens along the negative real axis on our w-plane. Lets
and 3 of the 4 values needed to describe the solution are consider a point on this axis: w =-1. Plugging in -1 for w
represented by the points location in 3D space. yields two solutions:
This is a nice result, but we must remember that this
is not the whole picture. Theres another variable the
imaginary part of z, we called this y that is not included
in our visualization at all.
This becomes important when trying to figure out if Our two solutions, z = +i, and -i are clearly different,
weve actually fixed our continuity problem. If we follow but have the same real part, zero. Since were only visu-
a path along a single branch of our Riemann surface, we alizing the real part of z we have no way of seeing that
these are in fact different points.
1 Or magnitude or angle!

Our w planes intersect here.

w1

w2

w
Figure 65 | Tricky Tricky. Along our self-intersection line, our two distinct
solutions are collapsed into one becuase we don't have enough dimensions to
show that out two points are, in fact, different.
Figure 64 | What to do about the intersection of our w planes? We would like to
think that we've fixed our continuity problem, but as we follow a path around
our surface, what should we do when we hit this point of intersection? Should This is the danger of visualizing high dimensional
we stay on the same plane or hop to the other one? How do we even figure this
out? mathematical concepts. What were really looking at here
82 Part 13: Riemann Surfaces 5:55-7:38

is a projection, a shadow, of our full four dimensional

surface.
So our line of self-intersection actuallyisnt. This is
exactly like the two-dimensional shadows of 3-dimen-
sional objects giving the appearance that the objects
intersect.
There are inherent limitations to the types of struc-
tures we can visualize in the 3D space we inhabit.
However, there are some clever ways to get a feel for
whats happening in our missing fourth dimension. One
approach is to expand our visualization to include anoth-
er dimensions of human perception,1 such as color. Figure 66 | Shadows are not the whole picture. The shadows of the pens
Well color each point on our surface with a color that intresect, but the pens don't! Our Riemann surface is just a shadow of our full
4-dimensional function, complete with false intersections.
corresponds to the value of the 4th dimension of our func-
tion in this case the imaginary part of z, y. To do this we
one missing piece of the puzzle what about the weird
need to decide which colors to match which numbers to
behavior we saw in Part 11 where some paths ended up in
this is called a colormap.
new locations and others, didnt?
v
2

y
-2 -1 1 2 u
-1

-2
w
Figure 67 | One way to visualize all 4 dimensions. If we color our suface according
to our missing value, y, we can get a better feel for what all 4 variables are
doing at once. Since our planes are different colors at our line of intersection,
this means that our surface actually doesn't self-intersect! The apparent self- y
intersection is a result of visualizing a 4D object in 3D space.

Our colors now give us a nice idea of whats happening 1

in our missing 4th dimension, y. If we look at our line of
self-intersection now, its much more clear that since our
colors2 are different, our four-dimensional function actu-
ally doesnt intersect itself. The apparent self-intersection -1 1 x
is just an artifact of our visualization technique.
So as we follow paths on our Riemann surface, the
right thing to do with these self-intersection lines is to ig- -1
nore them. If we now follow our path around our surface,
we see that its perfectly continuous, even at our weird
self-intersection line. z
Excellent. Figure 68 | Another Pair of Suspicious Paths. One more look at the suspicious
So thats great, we have continuity but theres still types of paths we first saw back in Part 11. Out blue paths return to where they
started, but our green paths don't! Hmmm...I wonder if there's a better way to
1 Sound?! think about this...
2 Any y values
7:38-10:37 Part 13: Riemann Surfaces 83

Figure 69 | Riemann Surfaces are Cool AF. Using our Riemann Surface, we can see exactly why our blue path returns to where it started, but our green path doesn't!

Lets recreate these suspicious paths, this time using behavior we saw! And this is just the beginning, Riemann
our Riemann Surface to help us out.1 To keep our surface surfaces are huge part of modern mathematics, and theres
from getting too crowded, well choose just one of our two way more to say than we have time for here.
paths to visualize first. Alright, were finally ready to answer the question
Well draw the same exact paths on w and see how what the heck was happing way back in Figure 2?
they show up on our Riemann surface as they are mapped
to the z-plane.
a b
So why does our green path start out in one location
on our z-plane and end up in another?

Simply because our green path leads us to the other

layer of our surface.
c d
From the perspective of our w-plane, it appears thats
weve returned exactly to our starting point, but we actu-
ally, havent. The w-plane is just a projection, a shadow.
In reality, our path has led us to a completely different
branch of our function, with different z-values. 2
Our Riemann surface allows us to clearly see that Figure 2 | Graph of f (x) = x 2+1 where x includes imaginary numbers. Panels
some paths on w lead the other branch, and some dont. a-d show "pulling" the function out of the page.

More specifically, paths that go around the central point

on our Riemann surface end up on new branches. This Our entire discussion has centered around a single
point is called a branch point branch points occur wher- function, f(z)=z 2+1. And so far, weve looked at one way
ever the two branches of our function have the same exact to visualize the Riemann surface for our function by plot-
value and can tell us a great deal about how our complex ting 3 of our 4 variables in 3D space.
function behaves. For us, this is the point w = 0. The two-dimensional plot we started our discussion
So our Riemann surface not only fixes the troubles we with way back in Figure 1, the one most of us see in math
ran into earlier, but beautifully explains the strange path class, only shows 2 of our 4 variables: the real parts of z
and w. The surface in Figure 2 is the result of including
1 But this time using a thicker marker to make things more clear. one more variable: the imaginary part of z as the vertical
2 The solution has undergone a monodromy along its path.
84 Part 13: Riemann Surfaces 10:37-end

dimension of our visualization. Now that we have some idea of what all four variables
When we first saw this surface in part one, people are doing, lets look for the two roots predicted by Gauss.
asked a very good question - if, according to the funda- Remember, roots are where the output of our function
mental theorem of algebra, our function is supposed to equals zero. For this to be the case, both the real and lat-
have exactly 2 roots, why does our surface appear to equal eral parts of our output variable w, u and v, must be zero.
zero at way more than two locations? Seeing where u equals zero isnt too bad this is where
This apparent contradiction has everything to do our surface intersects the z-plane.
with the shortcomings of living in 3 dimensional space Now, where does v, the imaginary part of our output
weve been discussing. When we visualize 4-dimensional variable w, equal zero? If we look at our colormap, this
functions in 3-dimensional space, we must remember that should be where our surface is green. Its difficult to see
what were really seeing is a projection, a shadow of the exactly which shade of green corresponds to zero,1 so lets
functions full 4-dimensional form. add an orange line to our surface where v equals exactly
Lets have a closer look at our surface from Figure 2. In zero.
our opening shot, half of our surface was hidden behind Now if we look closely, we see that both the real and
our paper and the colors we used were chosen somewhat imaginary parts of w equal zero at exactly two points:
arbitrarily to roughly correspond to the surface height. i and -i on our complex z plane, exactly as our algebra
predicted in Figure 41!
z-plane
u=0

v=0
ZEROS!!!!
z=i, -i v

Figure 70 | A Closer Look at Our Surface in Figure 2. In Figure 2, half of our

surface is hidden beneath the paper. Figure 72 | Finally. The zeros we've been looking for since part 1! We've rotated
around our surface from Figure 71 to get a better view. Our zeros occur where
the real and imaginary parts of w, u and v, equal zero. This happens where our
Now that we know a bit more about functions of orange a blue lines intersect!
complex variables, lets change the color of our surface to
correspond to the fourth variable we were forced to leave We have finally found our missing roots. So our friend
out of our 3-dimensional visualization v. Gauss was right all along. Our function does have exactly
2 roots.
z-plane Of course, finding these roots took some effort! We
had to journey deep into mathematics, and ask ourselves
what a number really is.2 This led us to the strange, but
necessary conclusion that numbers we should really be us-
ing in algebra are the two-dimensional complex numbers.
v
This result dragged us down deeper into the four-dimen-
sional world of complex functions and Riemann Sur-
faces. When we finally emerged, we saw that the algebra
many of us learn in school is only a shadow of an elegant,
powerful, and higher dimensional mathematics that has
everything to do with the numbers that have been given
the terrible name, imaginary. Thanks for reading.3

Figure 71 | Colors! To get a better feel for what's happening in all 4 dimensions, 1 A limitation of using color instead of a spatial dimension!.
here we've colored our surface to correspond to hour missing variable, v. 2 Man this feels like a movie montage.
3 You Rock!
 Exercises 13 85

Discussion 13.5 What would the Riemann Surface for

look like?
13.1 How did our Riemann Surface help us under-
stand our wierd loop behavior?

Challenge
13.2 Why was our line of self-intesection actually not?
13.6 For 1,000,000 cool points, construct a
Riemann Surface (in code or paper) for a func-
tion other than f(z)=z 2 , and tweet a picture to
@welchlabs.

Critical Thinking

13.3 We've shown 2 different surfaces in 3D

space that represent our function f(z)=z 2
(Figure 62 and Figure 70). Are there other
possible surfaces? If so, how many?

13.4 What kind of paths could we draw on our Rie-

mann Surface that would return us exactly to where
we started?
86 Solutions To Exercises
Solutions To Exercises 87
88 Solutions To Exercises
Solutions To Exercises 89
90 Solutions To Exercises
Solutions To Exercises 91
92 Solutions To Exercises
Solutions To Exercises 93
94 Solutions To Exercises
Solutions To Exercises 95
96

About the Author

Stephen Welch is a huge jerk. Just look at that smug
face. He thinks that, just like him, you should not
only be good at math, but enjoy doing it. He has the
audacity to think that math and science are perhaps
the most beautiful discoveries we as humans have
made, and they should be savored accordingly.
Needless to say, we must stop this monster. Help show
Stephen what a big jerk he is by watching and liking
his You Tube videos and buying his books. That will
show him.1

Figure -1 | The jerk who wrote this workbook. When he's not making videos,
Stephen Welch spends his enormous You Tube profits on his professional
Elephant Polo team.

1 Stupid Jerk.