You are on page 1of 3

A Better Way of Looking at Vectors and Matrices

Philip Tromans
http://pjtlog.blogspot.com/
December 2009

This document was produced after a session teaching some year 13 further maths students in a school.
It is appropriate to any student familiar with vectors and matrices however, and certainly would have helped
me in my first year of university maths. NB: Obviously the word “better” is very subjective, but I hope that
people can benefit from knowing the contents of this if they haven’t seen it before.

1 Introduction
In school, matrices are often taught to be grids of numbers equipped with a standard way of multiplying
them with other matrices, and with vectors. The real meaning behind what is going on geometrically is
often missed out. Here we attempt to explain the geometrical meaning of what is happening, and show some
of the slightly deeper properties of matrices. This document assumes that the reader is familiar with basic
matrix algebra, and how to multiply matrices and vectors.

2 Linearity
Before progressing to matrices, it’s useful to consider “What does linear really mean?”. In school, it is taught
that linear essentially means “like a straight line”, which although (almost) true when considering functions
like y = mx + c, it’s not especially helpful in 3D, and in other areas of more advanced mathematics. The
true definition of linearity is: A function f (x) is linear if:

f (x + y) = f (x) + f (y)

f (λx) = λf (x)
which can be combined into the single statement:

f (λx + µy) = λf (x) + µf (y)

Using this definition, we can see that the function f (x) = mx + c is linear if c = 0:

f (x + y) = m(x + y) + c = m(x + y) = mx + my = f (x) + f (y)

f (λx) = m(λx) = λmx = λf (x)

2.1 Exercises
1. Show that f (x) = mx + c is linear only if c = 0 (i.e. c must equal 0, otherwise f (x) isn’t linear).
2. Run through the algebra explicitly to prove that if f (x) has the properties established previously, then
f (λx + µy) = λf (x) + µf (y).

1
3 Vectors
A vector is a column of numbers, but more importantly represents a direction in space. Let i,j,k be the
usual unit basis vectors of standard 3D space. So,
     
1 0 0
i = 0 , j = 1 , k = 0
0 0 1

and these represent moving along the x, y and z axes by one unit respectively. We can express a general
vector b as:
 
λ
b = µ
φ
= λi + µj + φk

So the vector tells you how much to move along each basis vector.

4 Matrices Action on Vectors


We shall consider the matrix:  
1 2 3
A = 4 5 6
7 8 9
Note that:      
1 2 3
Ai = 4 , Aj = 5 , Ak = 6
7 8 9
Now if we multiply A by our arbitrary vector b:

Ab = A(λi + µj + φk)
= λAi + µAj + φAk
     
1 2 3
= λ 4 + µ 5 + φ 6
7 8 9

which is almost the same as the vector b, but instead of using the basis vectors i, j and k, we are now using
the columns of A. What is happening, is that by applying the matrix A to the vector b, we are changing the
worldview in which b is expressed: we are distorting the coordinate frame in which b resides. All a matrix
is, is a way of changing the basis vectors in which a vector is expressed as a combination of. This is made
clearer in the next section, on geometrical transformations.

4.1 Exercises
1. Check the values of Ai, Aj, Ak.

5 Geometrical Transformations
Consider rotation:

2
From the diagram (and some trigonometry), we can see that:
       
1 cos θ 0 − sin θ
→ , → ,
0 sin θ 1 cos θ
So the rotation matrix Rθ is (just by writing the two vectors next to each other):
 
cos θ − sin θ
Rθ =
sin θ cos θ
This is the same as in the previous section, the columns of the matrix are what the basis vectors get mapped
to by the associated transformation.
Similarly, we can calculate the transformation matrix for reflection in the y-axis. For this transformation,
we require:        
1 −1 0 0
→ , → ,
0 0 1 1
So the reflection matrix, Ry , is:  
−1 0
Ry =
0 1
Again, all that the matrix says is that our old basis vectors get mapped to the columns of the transformation
matrix.
Bearing these things in mind, we can see why transformation by the identity matrix doesn’t change space.
The identity matrix (in 3D) is:  
1 0 0
I = 0 1 0
0 0 1
What we’ve been saying so far is that the columns of the matrix are the new basis vectors, but in the case of
the identity matrix, the columns give exactly the same set of basis vectors as we had before! Hence nothing
changes under transformation by the identity matrix.

5.1 Exercises
1. Write down the transformation for reflection in the x-axis.
2. Write down a matrix which given the vector b (defined earlier) would swap the values of λ and µ.
3. Show explicitly that the identity matrix doesn’t change an arbitrary vector (show Ib = b).
This work is licensed under a Creative Commons Attribution-Share Alike 3.0 Unported License, see
http://creativecommons.org/licenses/by-sa/3.0/ for details.

You might also like