Chapter 2

The Mirror Problem

2.1 Introduction
Mirror, mirror on the wall,
Who is the fairest one of all?
asks the jealous queen in the fairy tale Snow White and gets an
answer she doesnt like.
Since time immemorial, looking into a mirror has been very popular. An amusing question which keeps popping up at regular intervals
in magazines is the following:
How large does a mirror have to be so that a person can see themselves
completely in it?

Here are some typical answers:

1. It depends on how far I stand from the mirror.
2. The mirror has to be as tall as I am.
These seem to make sense; but, unfortunately, both answers are
wrong.

2.2 The Mirror Problem for Individuals

Lets look rst at a person who wants to buy a mirror.
N. Herrmann, The Beauty of Everyday Mathematics,
DOI 10.1007/978-3-642-22104-0_2, Springer-Verlag Berlin Heidelberg 2012

13

14

2 The Mirror Problem

Wall
/

M
i
r j
r
o
r

Hair

Eyes
Person

Feet

Fig. 2.1 A person standing in front of a mirror has only to look down
halfway to their feet and look up halfway to their hair in order to see themselves completely; thats why the mirror if attached correctly has to be
only half as tall as the person.

The image in Figure 2.1 will help explain the situation.

According to the law of reection, the angle of incidence equals
the angle of reection and, likewise, the angle of incidence equals
the angle of reection . Now, if I look into a mirror hanging vertically
on the wall directly opposite me; and I want to see my feet, then I have
to lower my eyes only halfway between my eyes and the oor. My line
of sight hits the mirror at the angle . The angle of reection has
the same size and automatically directs my line of sight towards my
feet.
The same thing happens if I want to look at my hair. My line of
sight needs to only be lifted halfway between my eyes and my hair,
allowing me to see that my gray hair doesnt get any darker in the
mirror either.
Thus, we can summarize this as follows:
1. Seeing yourself completely in a mirror doesnt depend on the distance you are from the mirror.

2.3 The Mirror Problem for Groups

15

2. The mirror needs to be exactly half as tall as you are.

Some care needs to be exercised when hanging up the mirror. As
mathematicians, we can summarize this in an algorithm.
Algorithm for Attaching a mirror for an Individual
1. The mirror has to hang on a vertical wall.
2. The mirror has to be half as tall as I am.
3. The mirror has to be hung in such a way that I can just see my hair
at the upper edge of the mirror. The upper edge of the mirror thus
has to be mounted at a height equal to
my eye level + half the distance to my eyes half the distance to
my hair

2.3 The Mirror Problem for Groups

If several persons within a group, for example, a family or a bowling
club, all want to look at themselves in the same mirror from top to toe,
then we have to plan a little bit more carefully.
If you take the tallest person in the group, then you might be able
to align the mirror at its upper edge with this person. The bottom edge
is determined by half the height of this person. But if the mirror is
actually attached this way, then a shorter person wont be able to see
their feet. So if we dont want to attach the mirror to a movable device
so that it can be pulled up and down like a hanging lamp, then we have
to give the sales clerk some more money to buy a larger mirror.
The tallest person determines where the mirror is mounted at the
top.
The length downwards from this point is dened by the shortest
person who is actually the person with the lowest eye level. At the
bottom, the mirror has to reach the halfway point of this persons eye
level. This results in the total length of the mirror being as follows
(gure 2.2:

16

2 The Mirror Problem

Wall
/

66
M

i
r j
r
o
r

?
?

Hair

Eyes
tallest
person
Hair
Eyes 6
shortest
person

Feet

Fig. 2.2 Here, weve placed the shortest person in the group next to the
tallest person. We can see that this person has to look down beyond the bottom edge of the mirror so that they can see themselves completely. Thus, the
mirror has to be taller. The arrows on the left indicate the length of the mirror
for an individual and, further left, the length of the mirror for a group.
Total Length of Groups Mirror
eye level + half the distance to the hair to eye level of the tallest person
halfway to the eye level of the smallest person.

Algorithm for Attaching a Mirror for a Group

1. The mirror has to hang on a vertical wall.
2. The mirror has to be attached in such a way that the tallest member
of the group is able to just see their own hair in the mirror.
3. Thus, the upper edge of the mirror has to be mounted at a height
equal to
eye level + half the distance to the eyes half the distance to the
hair of the tallest person
Well, this was just a rehearsal, which has hopefully triggered our
interest in the following, more difcult task.

2.5 The Mirror Problem Expressed Mathematically

17

2.4 The Problem

The question were now going to ask is a more puzzling one:
The Mirror Problem
If I shake my right hand, my mirror image shakes its
left hand. But if I shake my head, my mirror image
doesnt shake its feet. So why does a mirror exchange
right and left, but not top and bottom?

We havent found anyone yet who didnt nd this question interesting. But what is the answer? To nd it, we have to examine the mirror
a bit more precisely; actually, mathematically.
Those of my dear readers who are not familiar with vectors in a
plane might want to skip the following section. Well summarize the
result in Section 2.6. Thats where those readers can log on again,
because the result is understandable even to people who are mathematical amateurs.

2.5 The Mirror Problem Expressed Mathematically

We now enter the plane to look at vectors.
So that we dont have to carry too much of a burden with us, which,
in turn, could cause us to lose sight of the essentials, well choose a
situation where the mirrors axis is the y axis. The diagram in Figure
2.3 shows that a random point (x, y), which we identify with its position vector a = (x, y), is depicted in the mirror at a point a = (x, y).
Such a representation has very simple properties. If we take a multiple of the vector a, we can form the image of the vector a rst, followed by multiplication of the image vector, which we can express
mathematically as follows if we abbreviate reection with the symbol RE:

18

2 The Mirror Problem

6

Mirror Axis
a = (x, y)

a = (x, y)

QQ
k

Q
Q
Q
Q
Q

Q

Q




3




-

x
Fig. 2.3 Reection at the y Axis

RE(k a) = k RE(a).
We can play the same little game when we add two vectors together.
We can obtain the image of the sum if we rst form the images of
the two vectors separately and then add the images together, which is
expressed mathematically as follows:
RE(a + b) = RE(a) + RE(b).
Such representations are called linear.
Now, math teaches us that such a representation can best be described by a matrix. A matrix is a square eld into which we enter
numbers. Since were looking at vectors in the plane, we take a 2 2
eld. This eld describes the representation. In order to characterize
it as a mathematical rule, we enclose it in parentheses. This is done as
follows:



12
.
34

In order to extract the mapping rule, we look at the two unit vectors
e1 = (1, 0) and e2 = (0, 1) and image them:
e1 = (1, 0) (1, 0),

e2 = (0, 1) (0, 1).

2.5 The Mirror Problem Expressed Mathematically

19

Weve written this down in a relaxed manner. But take a look again
at the drawing in Figure 2.3. The rst unit vector e1 = (1, 0) points
exactly to the right from the origin. If its mirrored in the y axis, then
its image vector points exactly to the left from the origin. Thats the
vector e1 = (1, 0) = (1, 0). The second unit vector, e2 = (0, 1)
points vertically upward. It is located directly on the y axis. Thats
why its reection stays where it is.
Now comes the rule for the calculation scheme; that is, the matrix
of the representation.
We enter the image vectors into a (2 2) matrix as columns:


A=


1 0
.
01

(2.1)

Thus, the representation is written in the form of a matrix as



xx =

1 0
01


x.

Now, we have to introduce another term, which is, in principle,

quite difcult to explain. Since we want to linger only in this plane,
though, weve only got to consider 2 2 matrices. Our new term is in
fact quite simple.
Denition 2.1 By the determinant of a 2 2 matrix



ab
,
cd

we mean the expression

det (A) = a d b c.
Thats just a number thats a little bit hard to calculate. Lets use an
example:

20


A=

24
3 1

2 The Mirror Problem

det (A) = 2 1 4 (3) = 14.

The next example shows that the determinant of a matrix can become negative:

B=

24
31

det (A) = 2 1 4 3 = 10.

The sign of the determinant is actually the distinctive feature that

we are interested in, because mathematicians have found the following:
Theorem 2.1 Linear representations whose corresponding matrices
have a positive determinant retain orientations whereas linear representations whose matrices have a negative determinant reverse orientations.
Well, its slowly becoming clearer. Its all about orientation.
A brief comment for purists:
All of this shouldnt anger serious mathematicians. We have just
given some hints here and there; those who wish to get more information are advised to look in specialist textbooks on linear algebra.

2.6 Results of Analysis of the Mirror Problem

First, we note that there are signicant differences between the pairs
right and left and top and bottom.
The terms right and left have a signicance which depends on
the respective person. What is right for me is left for my counterpart, and vice versa. It is my personal orientation that determines
where right and left are.
In contrast, the terms top and bottom are the same for everyone,
at least for those who are in our immediate vicinity, for example all
Americans. We all look up towards the clouds when someone talks to
us from above.

2.6 Results of Analysis of the Mirror Problem

21

And heres an important insight:

For our reection matrix A in equation (2.1), we get
det (A) = (1) 1 0 0 = 1.
So the determinant is negative! And this means:
The orientation is reversed when mirrored.
The best way to grasp this concept is by looking at a circle and its
mirror image. When we move a pen along a circular line in a clockwise direction from the top, our mirror image moves the pen in a counterclockwise direction. If I were to show my mirror image a clock face,
then the clock would always run in the wrong direction, regardless of
where I was standing or where the mirror was located. Thats what
happens with orientation.
The mirror doesnt actually swap right and left; instead, it reverses
the orientation. Orientation is thus associated directly with persons.
My orientation tells me where to nd the right side. The mirror reverses this image, which explains why my mirror image sees this quite
differently and calls it left.
Top and bottom,however, are terms of an entirely different nature. Top and bottom are the same for everyone in our vicinity.
They are objective terms, which are not associated with our personal
orientation. These terms are, thus, not swapped by a mirror. In fact,
this applies to the terms east and west as well. If I point towards
the east, my mirror image also points towards the east in the mirror
realm. And if I ask my mirror image, when I am standing in Berlin, to
point towards the Eiffel Tower, then my mirror image will point in the
mirror in exactly the same direction as I am pointing: towards Paris
and not towards Moscow.

http://www.springer.com/978-3-642-22103-3