An Ill-Advised

Linear Algebra
Tutorial

By Chad M. Topaz
 Welcome To My Little Flight Of Fancy
Hello. I am a little booklet. I was written at approximately 3:30 a.m. one day
by somebody who was heavily under the influence of coffee. Despite my dubious
origin story, I hope I am helpful.

I am a modest booklet and so I really have just one goal that I am secretly
aiming at: for you to be able to conceptualize the dimensions of a matrix's
image and kernel. If you are saying to yourself "wait, that's TWO goals," then
congratulations. That is just the sort of careful thinking that will benefit you
in mathematics. Just don't be a jerk about it. If instead you are saying to
yourself, "I don't know what image and kernel
mean," then I am the booklet for you. If you
are saying to yourself "I already know how to
conceptualize the dimensions of a matrix's image
and kernel," then please put me down and go
spend your time dismantling racism, sexism,
transphobia, homophobia, ableism, and other
forms of oppression.

If you are still reading, I like to make things clear, so let me say the secret
goal again. I will whisper to the best of my ability.

Secret Goal:
Conceptualize dimensions of
matrix image and kernel
Why is this my secret goal? Because some other day, I want to teach you a
little bit of algebraic topology. Algebraic topology makes much more sense if
you know some linear algebra. So I am going to teach you that teensy little bit
of linear algebra that you need. But even if you do not give a mouse’s backside

❤ 🧡 💛 💚 💙 💜 Page 2 of 14 💜 💙 💚 💛 🧡 ❤
about algebraic topology, I hope I still might be useful to you by helping you
understand some core ideas from linear algebra.

To accomplish our goal, we are going to have to build up some vocabulary.

Vocabulary
• Vector
• Matrix
• Linear combination
• Matrix multiplication
• Span
• Linear in/dependence
• Basis
• Dimension (of a span)
• Transformation
• Image (range, column space)
• Kernel (roots, null space)

As I talk about these words, I might be casual with my language because I

want to privilege conceptual understanding over strict use of terminology and
definitions. This is why mathematicians often hate me. Haha, just kidding.
It’s merely ONE of the reasons! If you are a capital-M Mathematician and
you are reading me, please put me down right now and retreat to the safety
of your perfectly objective utopia. 👀 Thank you!

❤ 🧡 💛 💚 💙 💜 Page 3 of 14 💜 💙 💚 💛 🧡 ❤
 Vector
A vector is a list of n numbers representing a point or an
arrow in n-dimensional space. I’ll underline a variable to
indicate it's a vector. When I write out specific vectors,
I’ll do it in a column, even though it makes me tired.

Also did you notice that you can make the word “vector”
out of “velociraptor?”

𝜋
Example. 𝑥" = $ ( ∈ ℝ! is a vector! But not a dinosaur.
√2

Question 1. What space does each vector “live” (exist) in?

1
1 1
1 1
$ ( , .0/ , .0/ , 0 1
0 0
1 0
0

Matrix

A matrix is a group of vectors listed side-by-side. I’ll double-underline a

variable to indicate it’s a matrix. Also, we’ll talk about the size of matrices as
“rows-by-columns.”

1 4
Example. 𝐴
3 = . 2 5/ is a 3 x 2 matrix. We can conceive of it as two vectors,
3 6
each of which lives in ℝ" .

Question 2. Describe the sizes of these matrices. Be creative! ❤

1 10 3 1 0 0 0
1 0 2 −2 −10 16 0 1 0 0
$ (,0 1,0 1
0 1 −7 −1 0 −11 0 0 1 0
4 0 4 0 0 0 1

❤ 🧡 💛 💚 💙 💜 Page 4 of 14 💜 💙 💚 💛 🧡 ❤
 Linear Combination
If you were baking with butter, flour, sugar, eggs, and
chocolate, You could make different things by taking
combinations of ingredients in different amounts, for
instance, 1 part butter, 3 parts flour, 2 parts sugar,
1 part eggs, 1 part chocolate.

Same with vectors. A linear combination of vectors is

just a recipe made up of the vectors.

Example. Our ingredients are the three vectors

𝑥"# = $ 1( , 𝑥"! = $ 2 ( , 𝑥"" = $3(.

2 −1 1
Then one linear combination of these is

2−2+9 9
2𝑥"# − 𝑥"! + 3𝑥"" = 2 $1( − 1 $ 2 ( + 3 $3( = $ ( = $ (.
2 −1 1 4+1+3 8
By the way, we could represent every linear combination of these vectors as

𝛼# 𝑥"# + 𝛼! 𝑥
" ! + 𝛼" 𝑥"" , 𝛼#,!," ∈ ℝ.

Question 3. I’m going to give you two vectors

and ask you to calculate a specific linear
combination of them. Also, I want you to write
an expression to represent every linear
combination of the two vectors. Are you ready for
your vectors? Ok, here they are! Come get ‘em!!!
1 2
𝑥"# = .0/ , 𝑥"! = . 2 /
6 −2
The specific linear combination you should calculate is −2𝑥"# + 4𝑥
" ! but don’t
forget to do the other part of the question (every linear combination).

❤ 🧡 💛 💚 💙 💜 Page 5 of 14 💜 💙 💚 💛 🧡 ❤
 Matrix Multiplication
When we multiply a matrix 3 𝐴 and a vector 𝑥" together, the answer 𝐴 3 𝑥" is the
linear combination of the columns of 𝐴
3 with coefficients specified by 𝑥" . If you
want to do this multiplication, there had better be as many numbers in your
vector as there are columns in your matrix.

1
1 3 5 22
Example. $ ( .2/ = 1 $1( + 2 $3( + 3 $5( = $ (.
2 4 6 2 4 6 28
3

Example. $1 3 5( $1( literally does not make sense because

2 4 6 2

1 $1 ( + 2 $3( +? $5(
2 4 6
and looking for what the “?” is will be is as fruitless as looking for fairness
from that Kyle Rittenhouse judge.

So far we did a matrix-vector product but we can also do a matrix-matrix

product. Just imagine doing the same thing as before on each column of the
second matrix, and then smooshing your answers together in a final matrix.

❤ 🧡 💛 💚 💙 💜 Page 6 of 14 💜 💙 💚 💛 🧡 ❤
 1 4
Example. Let’s calculate .2 3/ $1 3 5(. Well, let’s first calculate the
2 4 6
5 6
product for each column vector in the second matrix:
1 4 9
1
.2 3/ $ ( = . 8 /,
2
5 6 17
1 4 19
3
.2 3/ $ ( = .18/,
4
5 6 39
1 4 29
5
.2 3/ $ ( = .28/.
6
5 6 61
Now it’s smoosh time!

1 4 9 19 29
1 3 5
.2 3/ $ ( = . 8 18 28/.
2 4 6
5 6 17 39 61
Question 4. Calculate the following products. And yes, I know that previous
sentence was not a question.

2
1 1 2
$ ( .−1 /
0 1 −1
3
2 1
1 1 2
$ ( .−1 1/
0 1 −1
3 1
1 2 5
$ ($ (
3 4 6

Span
The span of a set of vectors is all possible linear combinations of them. Really.

❤ 🧡 💛 💚 💙 💜 Page 7 of 14 💜 💙 💚 💛 🧡 ❤
Example. The span of $1( and $ 1 ( is 𝛼# $1( + 𝛼! $ 1 (. By the way, if you
1 −1 1 −1
tried all the values you could think of for
the coefficients, you’d find out that you
could make any/every point the plane ℝ! .
I would be just as happy for you to tell
me that the span is the linear
combination above or the plane. Go with
what feels good to you. I am not picky.

Linear In/dependence
Suppose it was your birthday and I gave you some
vectors. I know — YOU’RE WELCOME. If you could
make any of the vectors I gave you by taking a linear
combination of any of the other vectors I gave you, you
might be like, stop re-gifting vectors. And in this case,
we’d say the vectors are linearly dependent. In your
head, you can also think “redundant.” Your vectors would
have the same span even if you through one away. If there’s no redundancy,
we say the vectors are linearly independent.

Example. The vectors 𝑥"# = $1( , 𝑥"! = $−2( are linearly dependent because
1 −2
#
𝑥"! = −2𝑥"# (or alternatively, 𝑥"# = − ! 𝑥"! ). That is, the span of the vector 𝑥"#
alone is the same as the span the vector 𝑥"! alone and is the same as the span
of both vectors together.

Question 5. Decide if each group of vectors is linearly independent. But don’t

tell me what you decide. Keep it secret, locked away in the darkest recesses of
your mind.

1 1
1 1
A$ ( , $0 (B, A$ ( , $0 ( , $ 2 (B, C.1/ , . 1 / D,
0 1 0 1 −4
1 −1

❤ 🧡 💛 💚 💙 💜 Page 8 of 14 💜 💙 💚 💛 🧡 ❤
 1 3 1 0 0 1 0 2
C.1/ , .3/D, C.0/ , .1/ , .0 /D, C.0/ , .1/ , .2 /D
1 3 0 0 1 0 0 0

Basis
A basis is a minimal set of vectors that have a given span.

1
Example. The vectors A$ ( , $ 0(B form a basis. They span the entire plane ℝ!
0 1
and there’s no redundancy.

1
Example. The vectors A$ ( , $ 0( , $ 2 (B are not a basis. They still span the
0 1 −4
!
entire plane ℝ but there is redundancy. That is, the vectors are linearly
dependent. You could throw away any one of the vectors and the remaining two
vectors would form a basis.

Question 6. Find a basis for the span of each set of

vectors given in Question 5. Because I love you a lot, you
might notice that I already did the first one two
examples up from here.

Dimension (of a span)

The dimension of a span is the number of vectors in the basis.

1
Example. The vectors A$ ( , $ 0(B are linearly independent so they form a basis.
0 1
The dimension of the span is 2 because the basis has 2 vectors.

A very useful fact to know is that to make a basis for ℝ% , you need exactly n
linearly independent vectors. Any fewer than n and you can’t span ℝ% . Any more
than n and there's redundancy.

❤ 🧡 💛 💚 💙 💜 Page 9 of 14 💜 💙 💚 💛 🧡 ❤
 1
Example. The span of A$ ( , $0( , $ 2 (B is 2-dimensional. Why? Well, the 3
0 1 −4
vectors don’t form a basis because you only need two linearly independent
vectors to span ℝ! . So you can throw away any one of the vectors above and
then the remaining two form a basis.

This might be a good time for a

useless and extended skin care analogy.
Even though I am just a modest
booklet, I do have a rigorous skin care
routine. But I believe in sensible skin
care. Like, I think my skin care routine
is five dimensional. The basis vectors
in my skin care routine are cleanser,
moisturizing sunscreen, an alpha hydroxy acid leave-on exfoliant, a gentle retinol
serum, and a peptide-containing moisturizing night cream. Some of you are
wondering why I did not mention an eye cream, and the reason is that I am
vehemently opposed because of linear dependence. That eye cream is simply
NOT doing anything that the other products aren’t already doing.

#
1 −! 3
Example. The span of the vectors E.1/ , 0 − #1 , .3/F is 1-dimensional. The
2 ! 6
−1
vectors are all multiples of each other so we really only need to keep one of
them. Keep whichever one sparks the most joy for you!

It’s important to get clear on the difference between the dimension of a span
and the overall dimension that vectors live in. In the last example, we started
with three vectors in three-dimensional space (since each vector has three
coordinates). But it turned out that the span of those vectors was only one-
dimensional (and in fact, it is a line cutting through three dimensional space).

Question 7. For each set of vectors, state the dimension of the space that
the vectors live in, and then find the dimension of the span of the vectors.

❤ 🧡 💛 💚 💙 💜 Page 10 of 14 💜 💙 💚 💛 🧡 ❤
State these quantities aloud as you gaze into a mirror and cultivate a sense
of self-compassion.

1 0 1 0 0 1
1 1 0
A$ ( , $2 (B, A$ ( , $ (B, C.0 / , .1/ D, C.0/ , .1/ , .0 / , .1/D
0 2 0 0
0 0 0 0 1 1

Transformation
From now on, let’s think of a matrix 𝐴 3 as doing something to a vector via
multiplication. To be honest, you can think of it as a process where you put a
vector 𝑥" into a box and out comes 3 𝐴𝑥" . We could even say that 𝐴
3 transforms 𝑥" .
We shove in a vector of coefficients stored in 𝑥 " and out comes the
corresponding linear combination of the columns of 𝐴 3. If we REALLY want to
get crazy we could even define this whole situation with a function, like
𝑓&( (𝑥") ≡ 𝐴
3𝑥" . This is maybe not the usual way Linear Algebra People think
about it, but I like functions, and I am not usual.

Some big questions in linear algebra involve asking

about the range and the roots of 𝑓&( (𝑥").

In linear algebra speak, the range is “all the

stuff that 𝐴 3 is capable of transforming to.” I
used the word range so far since it might be
familiar to you from calculus or discrete math, but
in linear algebra, we usually call it the “image.”

In linear algebra speak, the roots are “all the

stuff that 𝐴3 transforms into the zero vector, 0",”
that is, the solutions to 𝐴
3 𝑥" = 0
" . The linear
algebra word for the set of roots is the “kernel” or the “null space.”

Let’s keep it simple. We don’t really need to be able to fully describe the image
and kernel of a matrix. I just need for you to be able to appreciate their
dimensions.

❤ 🧡 💛 💚 💙 💜 Page 11 of 14 💜 💙 💚 💛 🧡 ❤
 Image
If we are asking about the dimension of the image of 𝐴3 then we are really
asking “what is the dimension of the span of the columns of 𝐴
3?”

1 0
Example. 𝐴
3=$ (. Well, the columns are linearly independent. So they form
0 1
a basis. And there are two of them. So the dimension of the image is 2. We
can even write dim Im 𝐴3 = 2.

1 0 −3
Example. 𝐴
3=$ (. Each column vector lives in ℝ!
0 1 3
and we know that we only need two linearly independent
vectors to make a basis. But our matrix has three
columns, and they are linearly dependent. So let’s throw
one of them away. Then dim Im 𝐴 3 = 2. Another way of
saying this is that we added a column onto the matrix
from the previous example but it was redundant. It’s like
we were packing for a trip and you put eye cream in your
suitcase and I was like NO, and took it out again.

1 2 3
Example. 𝐴
3 = . 2 4 6/. The columns have a lot of linear dependence! In
3 6 9
fact, we need to throw away two columns in order to have a basis. So
dim Im 𝐴
3 = 1.

Question 8. Find dim Im 𝐴

3 for each matrix 3
𝐴.

1 0 1 1 2 −8 1 0 0
1 −2 1 0
$ ( $ ( .0 1 1/ .0 0 0 / .0 2 0 /
0 0 0 −2
1 0 1 1 2 −8 0 0 3

❤ 🧡 💛 💚 💙 💜 Page 12 of 14 💜 💙 💚 💛 🧡 ❤
 Kernel
If we are asking about the dimension of the kernel
of 𝐴
3 , then we are asking "what is the dimension of
the span of all the solutions of 𝐴
3𝑥" = 0
" ?”

1 2
Example. Let’s find the kernel of 𝐴 3 = .2 4/. We’re
3 6
going to need to find the roots, that is, we are going
to need to solve 𝐴
3 𝑥" = 0". Written out, this equation is

1 2 𝑥 0
#
.2 4/ $𝑥 ( = . 0/, or
!
3 6 0
1 2 0
𝑥# .2/ + 𝑥! .4/ = .0/
3 6 0
Reading across each row gives us the system of linear equations

𝑥# + 2𝑥! = 0

2𝑥# + 4𝑥! = 0

3𝑥# + 6𝑥! = 0.

The second and third equations say nothing new! All we really have is the first
one, which tells us that 𝑥# = −2𝑥! . Then our vector 𝑥" of unknown coefficients
is

𝑥 −2𝑥
" = $𝑥#( = O 𝑥 !P = 𝑥! $−2(.
𝑥
! ! 1

I was sneaky and pulled 𝑥! out in front of the vector. Nothing has yet told us
the value of 𝑥! and NOTHING EVER WILL because any value works. If you
don’t believe me, just choose different values and go back and plug them in.

❤ 🧡 💛 💚 💙 💜 Page 13 of 14 💜 💙 💚 💛 🧡 ❤
What we’ve discovered is that the solution is ANY multiple of the vector
$−2(. But that is just the span of that vector. So the dimension of the
1
kernel is 1, which we can write as dim Ker 3
𝐴 = 1.

Goodbye, Friends
Can you do it? Can you conceptualize the
dimension of a matrix’s image, and its kernel?
GOOD JOB! I don’t even care if you know
how to calculate them (you can always use a
computer if you need to). I just want you to
understand what they mean.

p.s. Marko
Even though I am just a booklet, I have some friends.
One of my friends is called Marko. Marko is tall and
smart. Marko thinks I should tell you that the
dimensions of a matrix’s kernel and
its image add up to something
special. I have been up since 3:30
a.m., though, so I am tired.
Therefore, I am not going to tell you this. If you want to
know this secret fact, please go learn it in a linear algebra class.

❤ 🧡 💛 💚 💙 💜 Page 14 of 14 💜 💙 💚 💛 🧡 ❤