1.

3 AXIOMS FOR EUCLIDEAN GEOMETRY  21

Figure 1.8 Does triangle FEA have two right angles?

Our axiom set is a descendant of the five axioms provided by Euclid, but there
are some differences. There are two main reasons why we do not use Euclid’s axioms
as he originally gave them:

1. Euclid phrases his axioms in a way that is hard for the modern reader to
appreciate.
2. It has been necessary to add axioms to Euclid’s set in order to be able to give
rigorous proofs of many Euclidean theorems.

A number of individuals, and at least one committee, have taken their turn at
improving Euclid’s axiom set, notably David Hilbert in 1899, G. D. Birkhoff in 1932,
and the School Mathematics Study Group (SMSG) working during the 1960s. Even
though these axiom sets differ from one another and from Euclid’s, they all lead to
the well-known theorems in Euclid’s Elements. Consequently, we say that they are all
axiom sets for Euclidean geometry. The axioms we list in this section are a minor
rewording of the SMSG axiom set.
As we embark on our study of axiomatic Euclidean geometry, you will be asked
to consider proofs of some statements that seem obvious and many that you have
learned before. To enter into the spirit of our study, you must put aside what you have
learned before or find obvious. In earlier sections we have relied on some geometry
you have studied before. But in this section, our proofs are constructed only from the
22  CHAPTER 1 THE AXIOMATIC METHOD IN GEOMETRY

axioms we are about to list and any theorems we have previously proved from those
axioms. Keep in mind that our objective in our axiomatic discussion of Euclidean
geometry is not to learn the most facts of geometry, but to learn about the logical
structure of geometry and to practice the art of proof.

The Objects of Geometry: Points, Lines, and All That

However, whether we emphasize facts or logical structure, geometry is about geometic
figures. One would suppose that a deductive study of geometry would begin by
carefully defining the objects geometry deals with. Indeed, Euclid attempts to do just
that by telling us, for example, that a point is “that which has no part.” But to
understand this, we need to know the meaning of part. We could look part up in a
dictionary and find more words, which we could also look up, and so on. Logically
speaking, there is no end to this. We’ll get around this by leaving point and line and
plane as undefined terms in our axiomatic system.
This appears to ignore that, as a practical matter, we all have a pretty good idea
of what these words mean. But there is no contradiction here. Geometry exists as a
logical system (in which point is undefined) but also as a practical social enterprise —
something we all do together and talk about and show each other pictures of. In
geometry as a social enterprise, we have no trouble reaching a common understanding
of point. For example, a teacher says the word point and draws a very tiny dot on the
blackboard. Many experiences like this convey the customary interpretation of what
a point is. So it turns out there is no harm in leaving it undefined in our axiomatic
system.
Not every term in geometry needs to be left undefined. For example, we will
shortly define line segment in terms of points and lines. In fact, every other term in
geometry, from altitude to zonotope, can be defined in terms of our three undefined
terms. The most important geometric figures are shown in Figure 1.9. But here
is something you might not expect: We will need axioms to define some of these
figures.

Figure 1.9 The cast of characters.

 1.3 AXIOMS FOR EUCLIDEAN GEOMETRY  23

Axioms for Points on Lines

A X I O M 1 : The Point–Line Incidence Axiom
A line is a set of points. Given any two different points, there is exactly one line that
contains them.
←−→
We denote the line connecting A and B by AB. Our first theorem about lines uses a
style of proof called proof by contradiction or indirect proof. It is based on the idea that the truths of
Euclidean geometry do not contradict one another. In particular, if you reason correctly
on statements that are true, then you can never deduce a statement that contradicts
another statement known to be true. If you do find a contradiction, then one of the
statements you have been reasoning from must be false. In our proof we will make a
supposition and show it leads to a contradiction. This proves the supposition false.

* T H E O R E M 1.1
Two lines intersect in at most one point.

PROOF
Suppose lines L and M contain the two points A and B. Then A and B would have
two lines containing them, violating Axiom 1. This contradiction shows that our
supposition that L and M contain two points in common must be false. 
Our next axiom is just a mathematical way of saying what everyone who has
ever used a ruler will find familiar: A line “comes with” a set of numerical markings
that we can use for calculations and proofs.

A X I O M 2 : The Ruler Axiom

1. For any two points A and B, there is a positive number called the distance
from A to B and denoted AB. (This distance has properties we derive shortly.)
2. For each line L, there is a function f that assigns a real number to each point
of L in such a way that:
(a) Different points have different numbers associated with them.
(b) Every number — positive, negative, or zero — has some point with
which it is associated.
(c) If A and B are points of L, then fB − fA = AB, the distance from
A to B. Note that if A and B designate the same point, then AB = 0.
The function f is called the ruler function for line L and the number fA is called
the coordinate of A on line L. Properties a and b are often summarized by saying the
function is a 1:1 correspondence between the points of the line and the real numbers.
24  CHAPTER 1 THE AXIOMATIC METHOD IN GEOMETRY

This axiom allows us to use the properties of the real numbers to find out things
about lines. For example, there are infinitely many real numbers, so we must have
infinitely many corresponding points on a line. The ruler axiom also allows us to
define the key geometric idea of betweenness.
A line will have many possible ruler functions: For example, replacing fA by
fA + 7 for every A on a line gives another ruler function. But the theorems we
are going to prove come out the same even if we use a new set of ruler functions
conforming to Axiom 2.

DEFINITIONS
Let A, B, and C be three points on a line and f be the ruler function for that line.

1. We say B is between A and C if either fA < fB < fC or fC < fB < fA.
We write A-B-C to indicate that B is between A and C C-B-A has the same
meaning as A-B-C.
2. The segment from A to B, denoted AB, is defined to be the set consisting of
A, B, and all points X where A-X-B.
3. If AB = CD, then the segments AB and CD are called congruent.

* T H E O R E M 1.2
1. AB = BA.
2. If A-B-C, then AB + BC = AC.
3. If A B C are three different points on a line, then exactly one of them is
between the other two.
PROOF
1. AB =  fB − fA =  −  fA − fB  =  fA − fB = BA.
2. There are two cases. First, suppose fA < fB < fC. Then

AB + BC =  fB − fA +  fC − fB

= fB − fA + fC − fB
= fC − fA
=  fC − fA
= AC

We leave the second case, where fC < fB < fA, as an exercise.
 1.3 AXIOMS FOR EUCLIDEAN GEOMETRY  25

3. Let a b c be the coordinates of A B C according to the ruler function for

the line on which they lie. It is a well-known fact about numbers that,
out of three different numbers, exactly one can lie between the other two.
Consequently, by the definition of what it means for one point to be between
two others, the result follows. 

The following theorem was assumed as an axiom by Euclid. If Euclid had included
our ruler axiom among his axioms, then, of course, he would not have needed to
assume what we are about to prove.

* T H E O R E M 1.3: Extendibility
If A and B are any two points, then the segment AB can be extended by any positive
distance on either side of segment AB (Figure 1.10).

PROOF
Let e > 0 be the amount of extension wanted, and let’s say we want to extend past
B to a point C so that B is between A and C and BC = e. Let a and b be the real
←−→
numbers fA and fB under the ruler function for the line AB.

Case 1
a < b (Figure 1.10). Then define c = b + e. By part 3 of the ruler axiom, there is a
point C that corresponds to the number c. C is the point we want since:

(a) B is between A and C (since a < b < b + e).

(b) BC =  fC − fB = b + e − b  =  e  = e, since e > 0. To extend on the
other side of the segment, find the point corresponding to a − e.

Figure 1.10 The ruler function helps extend segment AB.

26  CHAPTER 1 THE AXIOMATIC METHOD IN GEOMETRY

Case 2
a > b. We leave this as an exercise. 

* T H E O R E M 1.4: The Midpoint Theorem

Every segment has a midpoint. That is, for any points A and B, there is a point M on
segment AB so that AM = MB.

PROOF
We leave this as an exercise. 

The midpoint theorem suggests that we think of points as infinitely small. If they
had any positive size, a line segment would be a bit like a necklace (Figure 1.11). If
the number of points on a segment were even, there would not be one exactly in the
middle as required by Theorem 1.4.

DEFINITION
−
→
If A and B are distinct points, the ray from A through B, denoted AB, is the set of all points
←−→
C on line AB such that A is not between B and C. We call A the endpoint of the ray.
The negative phrasing of this definition is sometimes awkward, so the following
theorem is sometimes handy. Its proof is based on one of our previous results. Can
you find the proof?

* T H E O R E M 1.5
−
→
AB consists of segment AB together with all points X where A-B-X. 

DEFINITION
→−
− →
Let A B, and C be any three distinct points. AB AC is called the angle BAC and
denoted ∠BAC (Figure 1.9). We may also denote this angle ∠CAB. A is called the

Figure 1.11 A segment AB of six “fat” points would not have a midpoint.
 1.3 AXIOMS FOR EUCLIDEAN GEOMETRY  27

vertex of the angle. If A B, and C lie on a line (are collinear), with A between B and
←−→
C, we call ∠CAB a straight angle. It is no different from the line BC.
When we think of an angle that is not a straight angle and not 0 , we often think
of it as the space between the rays that border it. Our definition of angle does not try
to capture that idea. We need a separate definition of the interior of an angle, and we
base it on the next axiom.

Separation
A X I O M 3 : Pasch’s Separation Axiom for a Line
Given a line L in the plane, the points in the plane that are not on L form two
non-Empty sets, H1 and H2 , called half-planes, so:

(a) If A and B are points in the same half-plane, then AB lies wholly in that
half-plane.
(b) If A and B are points not in the same half-plane, then AB intersects L.

H1 and H2 are also called sides of L L is called the boundary line of H1 and H2 .

Notice that the half-planes mentioned in Pasch’s axiom do not contain their
boundary line. For this reason they are sometimes referred to as open half-planes.

DEFINITION
Let A B, and C be three noncollinear points, as in Figure 1.12. Let HB be the half-plane
←−→ ←−→
determined by AC that contains B. Let HC be the half-plane determined by AB that
contains C. The inside, or interior, of ∠BAC is defined to be HB ∩ HC .

Figure 1.12 An angle and its interior.

28  CHAPTER 1 THE AXIOMATIC METHOD IN GEOMETRY

Pasch’s axiom was only added to the axiom set for Euclidean geometry in the late
nineteenth century, when geometers became aware that, for many geometric figures,
there was no way to rigorously define the inside or outside of the figure, much less
prove theorems about the insides and outsides. For example, if you had asked Euclid
to prove that if a line contains a point on the inside of an angle, then it must cross
at least one of the rays making the angle, he would have been unable to do so. He
would undoubtedly have been unconcerned about this, thinking this theorem to be
too obvious to bother with.

* T H E O R E M 1.6
−
→
If a ray AB has endpoint A on line L, but B does not lie on L, then all points of the
ray, except for A, lie on the same side of L as B.

PROOF
The proof is indirect. Assume there is a point C on the ray so that C and B are on
opposite sides of L. By Pasch’s axiom, BC crosses L at some point. This must be A,
←−→ ←−→
since BC ⊂ AB and, by Theorem 1.1, AB crosses L in just one point, namely, A. Since
A is not B or C, the fact that A is in BC means B-A-C. But this means C is not in
−
→
AB, by the definition of a ray. 

Triangles play a starring role in geometry, and now it is time to define them. Let
A B C be three points that are not collinear. In that case, we define the triangle ABC
to be AB ∪ BC ∪ AC. Can you see how to define the interior of a triangle?
Suppose we have a triangle ABC and we extend side AC to D, thereby creating
an exterior angle ∠BCD as in Figure 1.13. Pick any point M on BC that is not B or
C, and then extend AM past M to a point E. Will E be on the inside of the exterior

Figure 1.13
 1.3 AXIOMS FOR EUCLIDEAN GEOMETRY  29

angle ∠BCD? Our visual intuition says yes. But if we want the highest degree of rigor,
we need a proof. Here it is, but with reasons for some steps left out for you to supply.

* T H E O R E M 1.7
If A B, and C are not collinear and

1. A-C-D
2. B-M-C
3. A-M-E

then E is in the interior of ∠BCD.

PROOF
According to our definition of the interior of an angle, we need to show two things:
←−→
(a) that E is on the same side as D of line BC, and
←−→
(b) that E is on the same side as B of line CD.
←−→
(a) A and D are on opposite sides of BC . (Why?) A and E are on opposite
←−→
sides of BC . (Why?) Thus, we have shown both E and D to be on the
←−→
opposite side from A of line BC . Therefore, E and D must be on the
←−→
same side of BC , as we wished to prove.
←−→ ←−→ ←−→ ←−→
(b) By hypothesis, B is not on AC, so B is not on CD since AC and CD are the
←−→
same line. Thus, by the previous theorem, CB lies wholly on the B side of
←−→ ←−→
CD (except for C). But B-M-C means M is on CB , and so M and B lie
←−→ ←−→
on the same side of CD. Likewise, E and M lie on the same side of AC.
←−→
Thus, E and B lie on the same side of CD, as we wished to prove. 

Axioms for Measuring Angles

We have spoken of angles but not about measuring them. To fill this gap, we come
now to a group of axioms that does for angles what the ruler axiom does for lines.
We might refer to them, as a group, as the protractor axioms.

A X I O M 4 : The Angle Measurement Axiom

To every angle, there corresponds a real number between 0 and 180 called its measure
or size. We denote the measure of ∠BAC by m∠BAC.
30  CHAPTER 1 THE AXIOMATIC METHOD IN GEOMETRY

DEFINITION
If m∠BAC = m∠PQR, then we say ∠BAC and ∠PQR are congruent angles.

A X I O M 5 : The Angle Construction Axiom

−→
Let AB lie entirely on the boundary line L of some half-plane H. For every number r
−
→
where 0 ≤ r ≤ 180 , there is exactly one ray AC where C is in H ∪ L and m∠CAB = r.

A X I O M 6 : The Angle Addition Axiom

If D is a point in the interior of ∠BAC, then m∠BAC = m∠BAD + m∠DAC.

DEFINITION
←−→
If A-B-C, and D is any point not on line AC, then ∠ABD and ∠DBC are called
supplementary (Figure 1.14a).

A X I O M 7 : The Supplementary Angles Axiom

If two angles are supplementary, then their measures add to 180 .

Now consider two lines crossing at X, making four angles as in Figure 1.14b. Each
angle has two neighboring supplementary angles and one that is “across” from it.

Figure 1.14 (a) Supplementary angles:  +  = 180 . (b) ∠AXB and ∠A XB  are vertical
angles; ∠BXA and ∠B  XA are vertical angles.