Notes on geometry

December 25, 2020

Contents

1 The Algebra of the Real numbers 5

2 Incidence Geometry in Planes and Space 7

2.1 Incidence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

3 Distance and congruence 9

3.1 The distance function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
3.2 Betweenness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
3.3 Segments, Rays, Angles, and Triangles . . . . . . . . . . . . . . . . . . . . 15
3.4 Congruence of segments . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

4 Separation in Planes and Space 21

4.1 Convexity and separation . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
4.2 Incidence Theorems Based on the Plane-Separation postulate . . . . . . . 25
4.3 Incidence Theorems Continued . . . . . . . . . . . . . . . . . . . . . . . . 27
4.4 Convex Quadrilaterals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
4.5 Separation of Space by Planes . . . . . . . . . . . . . . . . . . . . . . . . . 29
4.6 The Seven Bridges of Königsberg . . . . . . . . . . . . . . . . . . . . . . . 29

5 Angular measure 31
5.1 Angular measure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31

6 Congruence Between Triangles 35

6.1 The Idea of a Congruence . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
6.2 The Basic Congruence Theorems . . . . . . . . . . . . . . . . . . . . . . . 35
6.3 Some Remarks on Terminology . . . . . . . . . . . . . . . . . . . . . . . . 38
6.4 The Independence of the SAS postulate . . . . . . . . . . . . . . . . . . . 38
6.5 Existence of Perpendiculars . . . . . . . . . . . . . . . . . . . . . . . . . . 38

7 Geometric Inequalities 41
7.1 Geometric Inequalities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41

8 The Euclidean Program: Congruence Without Distance 43

8.1 The Synthetic Postulates . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
8.2 The Laws of Inequality for Segments . . . . . . . . . . . . . . . . . . . . . 44
8.3 Right Angles, Synthetically Considered . . . . . . . . . . . . . . . . . . . . 44

3
Contents

8.4 The Synthetic Form of the Triangular Inequality. Addition of congruence

classes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44

9 Three Geometries 45
9.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
9.2 The Poincaré Model for Lobachevskian Geometry . . . . . . . . . . . . . . 45

4
1 The Algebra of the Real numbers

This chapter talks about the axioms of real numbers.

5
2 Incidence Geometry in Planes and Space

2.1 Incidence

The discussion of Euclidean geometry starts in the settings of a simpler structure.

The structure given by a triplet [Points, Lines, Planes]. Every line and every plane is a
set of points.
As is the case in any axiomatic foundations we take these three kinds of objects as
undefined to escape the circularity problem.

Definition 2.1.1 (collinear and coplanar)

We say that a set of points is collinear/coplanar if there is line/plane contains all
the points in the set.

Axioms 2.1.2 (Axioms of incidence)

I-1 Given any two different points there is exactly one line containing them.
I-2 Given any three different noncollinear points there is exactly one plane con-
taining them.
I-3 If two points lie in a plane, then the line containing them lies in the plane
I-4 If two planes intersect, then their intersection is a line.
I-5 Every line contains at least 2 points. Every plane contains at least 3 non-
collinear points. There is at least two planes.

Notation 2.1.3 (Line and Plane)

←→
We denote a unique line through two points P, Q by P Q. We denote a unique plane
←−−→
through three points P, Q, R by P QR.

7
2 Incidence Geometry in Planes and Space

Theorem 2.1.1
Two different lines intersect in at most one point.

Proof. By contradiction. If they intersect in two points they are the same line I-1.

Theorem 2.1.2
If the line intersects a plane not containing it, then the intersection is a single point.

Proof. If plane contains two points of the line then it contains the line I-2.

Theorem 2.1.3
Given a line and a point not on a line, there is exactly one plane containing them.

Proof. By I-5 we have three non-collinear points. By I-2 they uniquely determine the
plane.

Theorem 2.1.4
If two lines intersect, their union lies in exactly one plane.

Proof. Let intersection be Q (thm). Let P1 , P2 6= Q be on different lines (given by I-5).

Then P1 , P2 , Q uniquely determine the plane I-2, where both lines lie I-3.

8
3 Distance and congruence

The idea of this chapter is to make our geometric structure richer. We accomplish this
by introducing the notion of distance, which can be later used to define many common
notions, such as betweenness and congruence.

3.1 The distance function

In this section we will introduce a notion of distance and connect it to The ruler postulate,
which will tie our geometry closer to R.
Our new structure is [Points, Lines, Planes, d] where

d : Points × Points → R≥0

Axioms 3.1.1 (Axioms of distance)

The axioms of distance are:
D-1 If d(P, Q) = 0 then P = Q
D-2 d(P, Q) = d(Q, P ) for all P, Q
D-3 Every line has a coordinate system.

We could quite naturally want to add the axioms stating the equivalence of: the shortest
distance between two points is a straight line (triangle inequality), but we will deduce it
based on other axioms instead.

Notation 3.1.2 (distance)

We will use distance very often, thus we set P Q := d(P, Q)

Lets now address the coordinate system. The idea is that every line can be mapped to
real line to R. This, of course, has to play nicely with the defined distance function.

9
3 Distance and congruence

Definition 3.1.3 (Coordinate system)

The coordinate system of a line is a bijective correspondence f between the points
on the line and R, such that P Q = |f (P ) − f (Q)|.
We call f (P ) the coordinate of P .

Theorem 3.1.1
If f is a coordinate system, then −f also is.

Proof. Clearly −f is a bijection since x 7→ −x is a bijection and composition of bijections

is bijective. Also:

P Q = |f (P ) − f (Q)|
= |−1(f (Q) − f (P ))|
= |−f (Q) − (−f (P ))|

Theorem 3.1.2
If f is a coordinate system, then x 7→ f (x) + a for a ∈ R also is.

Proof. As before, since x 7→ x + a is a bijection, thus x 7→ f (x) + a is also a bijection.

P Q = |f (P ) − f (Q)|
= |(f (P ) + a) − (f (Q) + a)|

Theorem 3.1.3 (The ruler postulate)

Let L be a line and P, Q two point on L then there exists a coordinate system for
L s.t. f (P ) = 0 and f (Q) > 0.

Proof. Suppose f (P ) = a then x 7→ f (x)−a moves it to zero. Now reverse the coordinate
system x 7→ −(f (x) − a) if necessary and we are done.

We will also show that we can take the ruler placement theorem as an axiom instead of
the distance axioms.

10
 3.2 Betweenness

Theorem 3.1.4 (Ruler placement as an axiom)

All the Axioms of distance follow from the The ruler postulate.

Proof.
D-1 Take the coordinate system. We have |f (P ) − f (Q)| = 0 thus f (P ) = f (Q) since
f is bijection we have P = Q.
D-2 Follows immediately from |f (P ) − f (Q)| = |f (Q) − f (P )|
D-3 This is for free.

3.2 Betweenness

This section immediately utilities the notion of distance. We use it to define the notion
of betweenness. Naturally we use lines (collinearity) as our base.

Definition 3.2.1 (Betweenness)

We say that B is between A and C and write A − B − C iff AB + BC = AC

Naturally we expect:

Theorem 3.2.1
If A − B − C then C − B − A.

Proof. If AB + BC = AC then CB + BA = CA by D-2.

Now we want to build a bridge between the notion of betweenness and The ruler postu-
late.
To this end we define an auxiliary notion of betweenness for R.

Definition 3.2.2 (betweenness on R)

We write x − y − z iff x < y < z or z < y < x

11
3 Distance and congruence

Theorem 3.2.2 (betweenness and coordinates)

Given points P, Q, R on line L and coordinate system f .
f (P ) − f (Q) − f (R) iff P − Q − R.
Proven by 3.2.1 and 3.2.3

Lemma 3.2.1
f (P ) − f (Q) − f (R) =⇒ P − Q − R

Proof. Suppose f (P ) < f (Q) < f (R). Then (by definition of Coordinate system)

RQ = f (R) − f (Q)

and
QP = f (Q) − f (P )
thus
RQ + QP = f (R) − f (P ) = RP
This means R − Q − P , by 3.2.1 we are done.

Before proving the other direction we will try introducing some machinery to make our
work easier. In particular we will work a little bit more on coordinate systems.

Theorem 3.2.3
Given line L, if f, g are coordinate systems for L then for some a:

f = x 7→ g(x) + a or f = x 7→ −g(x) + a

Proof. For all P, Q ∈ L we have

|f (P ) − f (Q)| = P Q = |g(P ) − g(Q)|

This means

f (P ) − f (Q) = g(P ) − g(Q) or f (P ) − f (Q) = −(g(P ) − g(Q))

In the second case we could take −g as a starting point to get a first case, thus suppose

f (P ) − f (Q) = g(P ) − g(Q)

Then
(f − g)(P ) = (f − g)(Q)

12
 3.2 Betweenness

but this means that f − g is a constant function1 x 7→ a for some a ∈ R. Thus

f (x) = g(x) + a

Combining with potential swap of g for −g we get f (x) = ±g(x) + a

Lemma 3.2.2 (Change of coordinates preserves betweenness)

Given L, P, Q, R ∈ L and two coordinates systems f, g.
If f (P ) − f (Q) − f (R) then g(P ) − g(Q) − g(R).

Proof. Suppose f (P ) < f (Q) < f (R)

Using 3.2.3 we need to consider two cases.
1. If g(x) = f (x) + a
If f (P ) < f (Q) < f (R) then f (P ) + a < f (Q) + a < f (R) + a
2. If g(x) = −f (x)
Then we have g(R) < g(Q) < g(P ) which means g(P ) − g(Q) − g(R) by 3.2.2.

With the established machinery, we can show the second direction for the coordinate
system of our own choosing.

Lemma 3.2.3
P − Q − R =⇒ f (P ) − f (Q) − f (R)

Proof. We have
Suppose f (P ) = 0 and f (Q) > 0, we can choose this because of 3.2.2 and 3.1.3.
Then by P − Q − R we have
P Q + QR = P R
This means
|f (P ) − f (Q)| + |f (Q) − f (R)| = |f (P ) − f (R)|
Which simplifies to:
f (Q) + |f (Q) − f (R)| = |f (R)|
If f (R) < 0 then |f (Q) − f (R)| > |f (R)| (contradiction).
1
because it agrees on all arguments

13
3 Distance and congruence

Thus f (R) > 0. We can rewrite our equation as

f (R) = f (Q) + |f (Q) − f (R)|

It follows that
f (R) ≥ f (Q) + |f (Q) − f (R)|
and since Q 6= R, so |f (Q) − f (R)| =
6 0 and

f (R) > f (Q)

We have
f (R) > f (Q) > f (P )
therefore
f (R) − f (Q) − f (P )

Theorem 3.2.4
Given three points on a line, exactly one is between other two.

Proof. Let the points be named P, Q, R.

Take a coordinate system f . By ordering on R the f (P ), f (Q), f (R) are uniquely ordered.
Without loss of generality we can assume f (P ) − f (Q) − f (R), by betweenness and
coordinates we have P − Q − R as required.

Notation 3.2.3 (betweenness generalised)

For A1 , . . . , An by writing A1 − · · · − An we mean Aj − Ak − Al for

1≤j<k<l≤n

Lemma 3.2.4
Given n points they can be named in such an order that A1 − · · · − An

Proof. Take f as a coordinate system and then A1 as the smallest point in the coordinate
system, A2 as the second smallest and so on. Using betweenness and coordinates.

14
 3.3 Segments, Rays, Angles, and Triangles

Theorem 3.2.5
If A, B are points then there are points C, D st A − D − B and A − B − C.

Proof. Take coordinate system f

• D := f −1 ((f (A) + f (B))/2)
• Suppose f (A) < f (B) and take C := f −1 (f (B) + 1)

Lemma 3.2.5
If A − B − C and B − D − C then A − B − D.

Proof. Take coordinate system f (A) = 0 and f (B) > 0 (we can freely choose this
coordinates, because of 3.2.2). Then f (C) > 0 and f (B) < f (D) < f (C). Thus
f (A) < f (B) < f (D). Which means f (A) − f (B) − f (D), by 3.2.2 we are done.

3.3 Segments, Rays, Angles, and Triangles

Definition 3.3.1 (Segment)

We define segment AB to be a set consisting of A, B and of all points between A
and B.

Definition 3.3.2 (Ray)

−−→ ←→
We define ray AB to be a set of all points C ∈ AB such that C − A − B doesn’t
hold.
−−→
We call A an endpoint of AB

Definition 3.3.3 (Angle)

We define an angle to be a union of two rays sharing the same endpoint.
−−→ −→
Given rays AB, AC we denote the angel ∠BAC = ∠CAB.
We call A a vertex of the angle.

15
3 Distance and congruence

Few notes on the definition of angles. This definition is appropriate for the euclidean
geometry, but later on, when moving on to analytic geometry we will need to introduce
a notion of directed angle.

Definition 3.3.4 (Triangle)

Given three noncollinear points A, B, C we call AB ∪ BC ∪ AC a triangle. We use
words sides and vertices in the natural manner.
The triangle is denoted by 4ABC

Due to increase in the complexity of our objects we will slowly but surely become less
formal in our derivations. In particular we could show things like: given points A, B
−−→
there is unique ray AB, but since these things are very obvious and proving them all
would be extremely painful, we will generally treat them implicitly.

Lemma 3.3.1
Given two points A, B we have AB = BA

←→
Proof. Clearly A, B belong to both sets 3.3.1, and since for all C ∈ AB we have

A − C − B ⇐⇒ B − C − A

we have C ∈ AB ⇐⇒ C ∈ BA, thus AB = BA.

Lemma 3.3.2
For two points A, B, if
−−→ −−→ −→
If C ∈ AB and C 6= A then AB = AC.

Proof. We know that either A − B − C or A − C − B 3.2.4.

Because of symmetry we can assume A − B − C.
−−→
If P ∈ AB then not P − A − B.
−→
So, not P − A − C by 3.2.5 and therefore P ∈ AC by definition of Ray.
−→ −−→
If P ∈ AC then not P − A − C, again by 3.2.5 this implies not P − A − B. Thus P ∈ AB.
−−→ −→
We have established AB = AC.

16
 3.4 Congruence of segments

Lemma 3.3.3
−−→ −→
If B1 , C1 are points of AB, AC different than A, then

∠BAC = ∠B1 AC1

−−→ −−→ −→ −−→

Proof. Since AB = AB1 and AC = AC1 by 3.3.2. We have
−−→ −→ −−→ −−→
∠BAC = AB ∪ AC = AB1 ∪ AC1 = ∠B1 AC1

Lemma 3.3.4 (Vertices of triangle are uniquely determined)

If 4ABC = 4DEF then {A, B, C} = {D, E, F }

Proof. We want to start by showing that a vertex is not between any two other points
in a triangle. Lets take a vertex A. This is obviously true for segments AB and AC. If
A ∈ BC then A, B, C are collinear, which is a contradiction 3.3.4.
If P is not a vertex then P ∈ AB ∪ BC ∪ AC, which means

A − P − B or B − P − C or A − P − C

Thus vertices are uniquely determined by the fact that they are only three points not
between any other points.

3.4 Congruence of segments

Definition 3.4.1 (Congruence of segments)

We call two segments AB, CD congruent and write AB ∼
= CD if

AB = CD

Lemma 3.4.1
Congruence of segments is the equivalence relation

Proof. This is obvious from the axioms of equality and definition.

17
3 Distance and congruence

Theorem 3.4.1 (The segment construction)

−−→ −−→
Given a segment AB and a ray CD there is exactly one point E ∈ CD s.t.

AB ∼
= CE.

←→
Proof. Take a coordinate system f for CD s.t. f (C) = 0, f (D) > 0, then

E := f −1 (AB)

Theorem 3.4.2 (segments addition)

If
• A − B − C and A0 − B 0 − C 0
• AB ∼
= A0 B 0 and BC ∼
= B0C 0
then AC ∼
= A0 C 0

Proof. First we use 3.2.1 and then 3.4.1

AC = AB + BC = A0 B 0 + B 0 C 0 = A0 C 0

Theorem 3.4.3 (Segments subtraction)

If
• A − B − C and A0 − B 0 − C 0
• AB ∼
= A0 B 0 and AC ∼
= A0 C 0
then BC ∼
= B0C 0

Proof. As before:
BC = AC − AB = A0 C 0 − A0 B 0 = B 0 C 0

Definition 3.4.2 (Midpoint)

If A − B − C and AB ∼ = BC then B is the midpoint.

18
 3.4 Congruence of segments

Lemma 3.4.2 (Midpoint is unique)

Every segment AB has exactly one midpoint

Proof. Take coordinates f (A) = 0, f (B) > 0. Then f −1 (AB/2) is unique point equidis-
tant to A and B.

19
4 Separation in Planes and Space

The notes for this chapter are extremely sloppy. The fatigue of proving obvious things :<

4.1 Convexity and separation

Definition 4.1.1 (Convex set)

We call a set A convex if for any two points P, Q ∈ A we have P Q ⊆ A

Axioms 4.1.2 (The Plane-Separation postulate)

Given a line and a plane containing it, the set of all points of the plane that don’t
lie on the line is the union of two disjoint sets such that
• Each set is convex
• If P belongs to one set and Q to another then P Q intersect a line.
We call the sets half-planes and the line edge.

Lets prove one obvious lemma as a sanity check.

Lemma 4.1.1
In a plane, given a line L and points P, Q, if P Q ∩ L 6= ∅, then P, Q are in different
half-planes.

Proof. If P, Q are in the same half-plane H then the set can’t be convex since L ∩ H = ∅
but L ∩ P Q 6= ∅. This contradicts The Plane-Separation postulate.

21
4 Separation in Planes and Space

Theorem 4.1.1 (Pasch’s postulate)

Given a triangle 4ABC and a line L in the same plane.
If E ∈ L and A − E − C and then

either L ∩ AB 6= ∅ or L ∩ BC 6= ∅

or L = B.

Proof. By The Plane-Separation postulate L splits a plane into two half planes. By 4.1.1
we know that A, C are in different half-planes.
Lets denote the half-plane containing A: H.
We will now consider two cases AB is contained in H or it is not. If it is not then we
are done since AB ∩ L 6= ∅.
Otherwise B ∈ H, but C 6∈ H. Thus BC ∩ L 6= ∅

Now some obvious fun lemmas. In all of them it is understood that H1 , H2 are half-planes
and L is a line given by The Plane-Separation postulate.

Lemma 4.1.2
H1 , H2 are not both empty.

Proof. If they are then plane is a line and this is disallowed by I-5.

Lemma 4.1.3
H1 , H2 are both non-empty.

Proof. We already established that at least one is non-empty. Thus there is a line in the
plane that intersects L, by 3.2.5 we get a point in the other plane.

Lemma 4.1.4
H1 contains at least three non-collinear points.

Proof. Let A ∈ H1 and X ∈ H2 (given by 4.1.3).

←→
Using 3.2.5 on AX we get another point in H1 , lets call it B.
Are A, X, B collinear?

22
 4.1 Convexity and separation

If they are, then there must be a point Y such that A, B, Y are not collinear I-51 .
If Y ∈ H1 we are done, so suppose Y ∈ H2 .
←→
Pick a point C such that A − C − (AY ∩ L), then C ∈ H1 and A, B, C aren’t collinear
or A, Y, B, C would be, contradicting our assumption.

Lemma 4.1.5
H1 lies in exactly one plane. Thus a half-plane uniquely determines an plane.

Proof. If it lies in two planes then they share three non-collinear points 4.1.4 thus they
are the same plane I-2.

Lemma 4.1.6
A half-plane uniquely determines an edge.

←−−→
Proof. Suppose A1 ∈ H1 and A2 ∈ H2 . Then A1 A2 intersects both edges. If it intersects
them in a different points X, X 0 then there is Y st X − Y − X 0 and Y 6∈ H1 or it wouldn’t
be convex and for the same reason Y 6∈ H2 , so X = X 0 .
←−−→ ←−→
Then take B ∈ H1 such that B 6∈ A1 , A2 then BA2 intersects both lines in the same
point, but this intersection is different from the previous one.
Thus the two lines/edges share two points: they are the same line.

Theorem 4.1.2 (Intersection of convex sets is convex)

T
If A is non-empty family of convex sets then A is convex.

T
Proof. Let X, Y ∈ A, then for anyT set B ∈ A we have X, Y ∈ B and thus XY ⊆ B
since B is convex, therefore XY ⊆ A.

Definition 4.1.3 (Convex hull)

The convex hull of A is the intersection of all the convex sets containing A.

1
Actually we just know that there are three noncollinear points. But it must be that at least on of
←→
them is not on AB or they would be collinear.

23
4 Separation in Planes and Space

Lemma 4.1.7
Convex hull is convex.

Proof. By 4.1.2.

Lemma 4.1.8
H1 ∪ L is convex.

Proof. The only case to consider is P1 ∈ H1 and P2 ∈ L.

If P1 P2 6⊆ H1 ∪ L then it must intersect H2 , so there is a point Q ∈ H2 such that
Q ∈ P1 P2 , but then QP2 intersects L in Q0 .
If this intersection is P1 we get a contradiction by considering the betweenness relations,
←−→
otherwise Q0 Q = L and thus P2 ∈ L which is a contradiction.

Theorem 4.1.3
Given 4ABC and a line L in the same plane, if L contains no vertex of the triangle
it can’t intersect all three sides.

Proof. By The Plane-Separation postulate L splits the plane into two half-planes and
itself. We have three vertices, therefore two must lie in the same half-plane and since it
is convex, it will also contain the side between them.

Theorem 4.1.4
The Plane-Separation postulate follows from Pasch’s postulate.

Proof. TODO: make this thing more readable: The idea of this proof is to first get our
two half-planes and then prove that they are the right ones, by first proving that they are
convex, then proving the second condition and finally that they cover the hole plane.
Let L be a line contained in a plane E.
Since E contains at least three noncollinear points there is a point P1 ∈
/ L.
Let
H1 = {Q ∈ E : P1 Q doesn’t intersect L}

24
 4.2 Incidence Theorems Based on the Plane-Separation postulate

Let P2 be a point such that P Q intersects L 3.2.5.

H2 = {Q ∈ E : P2 Q doesn’t intersect L}

We will show that H1 is convex (H2 works the same way).

Suppose Q1 , Q2 ∈ H1 then P Q1 , P Q2 don’t intersect L. If Q1 Q2 intersect L, considering
triangle 4P Q1 Q2 by Pasch’s postulate we get a contradiction since L would intersect
one side which by definition it can’t do, thus Q1 Q2 ⊆ H1 .
We will now show that if R1 ∈ H1 and R2 ∈ H2 then R1 R2 intersects L.
Suppose it doesn’t.
We first need to establish R1 ∈ H2 . Consider 4R1 R2 P2 . If R1 ∈
/ H2 then R1 P2 intersect
L, by 4.1.1 R1 R2 or R2 P2 intersect L in both cases it’s an contradiction.
In particular R1 ∈ H2 and P1 R1 ⊆ H1 and P2 R1 ⊆ H2 .
But L intersects P1 P2 thus it must intersect P2 R1 , P1 R1 , R1 which is a contradiction.
The only thing left to show is that H1 ∪ H2 ∪ L = E
Let Q ∈ E. If Q = P1 or Q = P2 we are done. Consider 4QP1 P2 since P1 P2 intersects
with L it must be (by 4.1.1) that L goes through Q (we are done) or it goes throughout
QP1 or QP2 .
By symmetry we can consider only the QP2 case. QP1 doesn’t intersect L, thus Q ∈ H1 .

4.2 Incidence Theorems Based on the Plane-Separation

postulate

The basic idea of this section is to try to define the interior of an angle using the notion
of half-planes, and then use it to define the interior of a triangle.
Note: I skip some more obvious theorems of this section, but I might fill this gaps in the
future.
We will first give the definition and then justify it.

Definition 4.2.1 (Interior of an angle)

Given an angle ∠BAC, the angle lies in a single plane S
we define the interior of it as intersection of two half-planes. They are given by:
• AB, the side containing C.
• AC, the side containing B.
The exterior of an angle is the complement of its interior that lies in S.

25
4 Separation in Planes and Space

Definition 4.2.2
If H1 , H2 are half-planes with common edge L, but are not necessarily in the same
plane we call them sides of L. We call half-planes in the same plane opposite sides.
Comment: this is the definition Moise uses, which I don’t understand the need for.

Theorem 4.2.1
−−→
Given a ray AB and a line L such that
−−→
A ∈ L and AB 6∈ L
−−→
AB − {A} lies in exactly one side of L.

−−→
Proof. First pick a point P on L which isn’t A, then A, B, P are noncollinear, thus AB
lies in exactly one plane.
−−→
Let Q ∈ AB − {A, B}.
If Q is not in the same half-plane as B then QB must intersect L, the intersection can
be A or not A. Both cases lead to simple contradictions.

We are now ready to justify the definition of the interior of an angle.

Theorem 4.2.2 (Interior of an angle is well-defined)

−−→ −→
Given ∠BAC if B 0 ∈ AB − {A} and C 0 ∈ AC − {A} then the interior of ∠B 0 AC 0
is the same as ∠BAC

←−→ ←→ ←→ ←−→
Proof. Since AB 0 = AB, AC = AC 0 the half-planes given don’t depend on the choice of
points on the rays constituting the angle.
←→ ←→
C 0 , C lie on the same side of AB by 4.2.1, and similarly B, B 0 and AC. We are done.

Theorem 4.2.3 (Interior of an angle is convex)

The interior of ∠BAC is convex.

Proof. By Intersection of convex sets is convex.

26
 4.3 Incidence Theorems Continued

Definition 4.2.3 (Interior of a triangle)

We define the interior of a triangle as the intersection of interiors of its angles.

Theorem 4.2.4
The interior of the triangle is convex.

Proof. Intersection of convex sets is convex

4.3 Incidence Theorems Continued

Theorem 4.3.1 (The Crossbar Theorem)
−−→
If D is in the interior of ∠BAC then AD intersects BC − {B, C}.

Proof. We would like to use Pasch’s postulate, but to be able to use it we need a triangle
←→
such that A is between its vertices, thus let F ∈ AB such that F − A − B.
←→ ←→ ←→
The by the postulate tells as that AD intersects BC − {B, C} or F C − {F, C} or C.
−−→
Using our separation theorems we get that AD constitutes that intersection.
←→ ←→
We can eliminate C as an option by observing that if AD = AC then D can’t be in the
interior .
−−→ ←→
Since F C is on different side then AD with respect to AC they can’t intersect2 .
−−→ ←→
So it must be that AD intersects F C − {F, C}.

4.4 Convex Quadrilaterals

Definition 4.4.1 (Quadrilateral)

Given four coplanar points, where each three points aren’t collinear we define

ABCD = AB ∪ BC ∪ CD ∪ AD

we use the notion of sides, angles and diagonals in the obvious waya .
a
But a somewhat good exercise is to define adjacent angles.

2
This could be handled more formally, but I don’t see a point.

27
4 Separation in Planes and Space

Lemma 4.4.1
The vertices of quadrilateral are uniquely determined.

Proof. Same as 3.3.4.

Definition 4.4.2 (Convex Quadrilateral)

This definition hurts the word convex, but is nevertheless standard. . .
We call a quadrilateral convex if for every side (of it) extended to the line, the
quadrilateral doesn’t lie in both half-planes.

Theorem 4.4.1
Quadrilateral is convex iff its diagonals intersect each other.

Proof. By 4.4.2 and 4.4.3.

Theorem 4.4.2
The diagonals of convex quadrilateral intersect each other.

Proof. Lets take ABCD to be our convex quadrilateral. The diagonals are AC, BD
−→
C lies in interior of ∠BAD, thus by The Crossbar Theorem AC intersects BD
−→
Because the quadrilateral is convex, AC − AC lies in different half-plane then BD.
Thus AC intersects BD.

Theorem 4.4.3
If diagonals of a quadrilateral intersect each other, then the quadrilateral is convex.

Proof. Lets take ABCD to be our quadrilateral.

←→
We will show that AB separates the plane in such a way that one half-plane doesn’t
contain any point of ABCD (other lines work the same way).
This is equivalent to showing that C, D are in the same half-plane (because half-planes
are convex).

28
 4.5 Separation of Space by Planes

If they are on the different sides, then CA, BD are on the different sides, and thus can’t
intersect .

Lemma 4.4.2
If each vertex of a quadrilateral lies in the interior of the opposing angle then the
quadrilateral is convex.

Proof. By The Crossbar Theorem the diagonals intersect and by 4.4.1 we are done.

4.5 Separation of Space by Planes

Axioms 4.5.1 (The space-separation postulate)

Each plane E separates the space into two half-spaces/sides such that
• each half-space is convex
• If P, Q belong to different half-spaces then P Q intersect E.
The plane together with two half-spaces cover whole space.

We should now prove the batch of obvious things about half-spaces, theorems equivalent
to the ones for half-planes.

Definition 4.5.2 (Dihedral angle)

If two half-planes share the same edge, but don’t lie in the same plane, we call their
union together with the edge a dihedral angle.
We define the interior of dihedral angle by taking intersection of two half-spaces
(which ones should be obvious).

4.6 The Seven Bridges of Königsberg

This section contains the description and partial solution to the titular problem.

29
5 Angular measure

5.1 Angular measure

In this section we will introduce a notion of angular measure.

Our new structure is [Points, Lines, Planes, d, m] where

m : angles → (0, 180)

Before giving the axioms we need some additional definition

Definition 5.1.1 (linear pair)

We call a pair of angles a linear pair iff they are formed by a line and a single ray.

Definition 5.1.2 (supplementary angles)

Angle are called supplementary if their measures add up to 180.
Notice that this doesn’t tell us how they lie.

Axioms 5.1.3 (Axioms of angular measure)

The axioms are
−−→
M-1 Let AB be on the edge of the half-plane H, for every r ∈ codomain m there
−→
exists exactly one ray AP with P ∈ H s.t. ∠P AB = r
M-2 If D is a point in the interior of ∠BAC then

m∠BAC = m∠BAD + m∠DAC

M-3 Given ∠BAC, if D is in the interior of the angle, then

m∠BAC = m∠BAD + m∠CAD

M-4 If two angles form a linear pair then they are supplementary angles.

31
5 Angular measure

Definition 5.1.4 (congruence of angles)

We say that two angles are congruent iff they have the same measure.

Definition 5.1.5 (Right angle)

The right angle is an angle with a measure of 90.

Theorem 5.1.1
If two angles forming a linear pair are congruent then they are right angles.

Proof. By 5.1.3 the angles are supplementary, and since they are of the same measure,
each must be of measure 90.

Theorem 5.1.2 (angle construction theorem)

−−→ −−→
Given ∠ABC and B 0 C 0 , if H is half-plane which edge contains B 0 C 0 , there is exactly
−−→
on B 0 A0 ⊆ H such that
∠ABC = ∠A0 B 0 C 0

Proof. By 5.1.3, there is a unique angle with measure m∠ABC.

Theorem 5.1.3 (Angle addition/subtraction theorem)

There are two more obvious theorems about adding and subtracting angles. Todo.

Proof. Skipped.

Definition 5.1.6 (Perpendicular ray and lines)

−−→ −→
We say that AB, AC are perpendicular if their union is a right angle. We write
−−→ −→
AB ⊥ AC
−−→ −→
We also say that lines containing AB and AC are perpendicular and use the same
notation.
We use the analogous notation freely and we mix objects in using it.

32
 5.1 Angular measure

Definition 5.1.7 (acute & obtuse)

We call an angle with measure < 90 acute and with measure > 90 obtuse.

Definition 5.1.8 (complementary angles)

Two angles are complementary if sum of their measures is 90.

Definition 5.1.9 (vertical angles)

Two angles are called vertical angles iff their sides form pairs of opposite rays.

Theorem 5.1.4 (Vertical angles theorem)

If two angles form a vertical pair then they are congruent.

Proof. We use the linear pair property repeatedly.

33
6 Congruence Between Triangles

6.1 The Idea of a Congruence

The intuitive idea of congruence in geometry is that two objects are the same if one can
be moved onto another without changing its size or shape.
There are two approaches to congruence, one is to take it as an undefined notion, guided
by certain axioms (synthetic approach) and second is to use other notions to define
(metric approach).

Definition 6.1.1 (congruence of triangles)

Two triangles are congruent if there exists a bijection between their vertices such
that corresponding angles and sides under this bijection are congruent.
If the triangles are 4ABC, 4DEF , and the bijection is A, B, C 7→ E, D, F then we
write
4ABC ∼ = 4EDF

Axioms 6.1.2 (SAS)

SAS postulate (side-angle-side).
Given a correspondence between two triangle, if two sides and the included an-
gle are congruent to corresponding sides and angles then the correspondence is a
congruence.

6.2 The Basic Congruence Theorems

Definition 6.2.1 (isosceles, scalane, equilateral triangle)

We call a triangle
isosceles if it has with two congruent sides
scalene if it is not isosceles
equilateral if all of its sides are congruent.

35
6 Congruence Between Triangles

Theorem 6.2.1 (The Isosceles Triangle Theorem)

If two sides of a triangle are congruent then the angle opposite them are congruent.

Proof. We can use SAS.

Lemma 6.2.1
Every equilateral triangle is equiangular.

Proof. Use The Isosceles Triangle Theorem.

Theorem 6.2.2 (ASA)

Given correspondence between triangles, if two angles and the included side corre-
spond, then the triangles are congruent.
Restatement: Given 4ABC, 4DEF and a correspondence between the vertices
A, B, C 7→ D, E, F , if under this correspondence
• ∠A ∼ = ∠D
• ∠C ∼
= ∠F
• AC ∼
= DF
then
4ABC ∼
= 4DEF

−−→
Proof. By The segment construction there is a point B 0 ∈ DE such that

DB 0 ∼
= AC

By SAS 4ABC ∼
= 4DB 0 F . Therefore

∠DF B 0 ∼= ∠ACB ∼ = ∠DF E

−−→ −−→ ←−→
By 5.1.2 it means F B 0 = F E, and since the intersection with DEe is unique it means
E = B 0 . Finally
4ABC ∼ = 4DB 0 F = 4DEF

36
 6.2 The Basic Congruence Theorems

Theorem 6.2.3
If two angles of the triangle are congruent then so are the sides opposite to angles.
Converse of The Isosceles Triangle Theorem.

Proof. Use ASA.

Lemma 6.2.2
Every equiangular triangle is equilateral

Proof. Using 6.2.3 repeatedly.

Theorem 6.2.4 (SSS)

Given triangles and correspondence between them, if all sides are congruent under
the correspondence, then the triangles are congruent.
Suppose we are given 4ABC, 4DEF and the correspondence A, B, C 7→ D, E, F .

Proof. We start by constructing a copy of 4DEF on the segment AC.

We choose B 0 5.1.2 such that

= ∠F DE and AB 0 ∼
∠B 0 AC ∼ = DE

By SAS
4DEF ∼
= 4AB 0 C

If we show that 4AB 0 C ∼

= ABC we will be done.
Connect B to B 0 . Look, the isosceles triangles! Therefore

∠CBB 0 ∼
= ∠CB 0 B

and
∠ABB 0 ∼
= ∠AB 0 B
but by angle addition this means

∠ABC ∼
= AB 0 C

By SAS we are done.

37
6 Congruence Between Triangles

Definition 6.2.2 (Bisector of an angle)

Bisector of an angle is the ray which split the angle into two congruent angles.

Theorem 6.2.5 (Bisector of an angle is unique)

−−→
Given ∠ABC there is exactly one bisector AD of this angle.

Proof. First we show that at least one bisector exists. Suppose AB ∼

= AC.
Let D be a midpoint of BC. By SSS

4ABD ∼ = 4ACD
−−→
it follows that ∠BAD ∼ = ∠DAC, so AD is indeed the bisector.
−−→
Let AD0 be a different bisector, by The Crossbar Theorem we may choose D0 to lie on
BC. The triangle are 4ABD0 ∼ = 4ACD0 are similar by SAS, so D0 must be a midpoint
of BC. It follows that D = D .0

6.3 Some Remarks on Terminology

This section contains an interesting discussion compering a notion of congruence to a

notion of identity.

6.4 The Independence of the SAS postulate

In this section author presents a model of for our geometry without SAS postulate where
the SAS postulate fails.

6.5 Existence of Perpendiculars

Theorem 6.5.1 (existence of perpendicular)

←→ ←→ ←→
Given a line AC and a point B 6∈ AC there is a line BD such that
←→ ←→
AC ⊥ BD

Proof. We construct an angle ∠CAB 0 such that

= CAB and AB ∼
∠CAB 0 ∼ = AB 0

38
 6.5 Existence of Perpendiculars

←→
Let G be the intersection of BB 0 and AC.
There are two cases.
1. G = A then we are done
2. G 6= A, then 4ABG = ∼ AB 0 G, and since ∠BGA = ∠B 0 GA and the angles are
supplementary, they must be right angles.

39
7 Geometric Inequalities

We start by observation that inequalities between angles and segments can be described
in terms of congruences instead of metric. We can say that one angle is smaller than the
other if the other angle has a sub-angle congruent to it, similarly for segments.

7.1 Geometric Inequalities

Definition 7.1.1 (Interior/Exterior angle of a triangle)

We define interior angle in an obvious way.
Find better definition: Suppose we have triangle 4ABC and there is a point D s.t.
A − C − D, then ∠BCD is an exterior angle.

From the definition it follow that there are 6 exterior angles.

41
8 The Euclidean Program: Congruence
Without Distance

8.1 The Synthetic Postulates

Our presentation of geometry follows that of Birkhoff, as opposed to the classical one, the
one used by Euclid and more recently by Hilbert. The classical treatment of geometry
is called synthetic, while Birkohoff’s is called metric.
We now give a sketch of the synthetic treatment. The structure at the start is [Points, Lines, Planes].
We add

• Notion of betweenness.
• Notion of congruence for segments.
• Notion of congruence for angles.

We pack the two latter ones into one and denote them by the same relation. The
structure now is

[Points, Lines, Planes, Betweenness, Congruence]

The betweenness postulates are:

1. If A − B − C then C − B − A
2. Of any three points on the line, exactly one is between the other two.
3. Any four points on the line can be named in such order that A − B − C − D
4. For any two points A, B there are points C, D such that A − B − C and A − D − B
5. If A − B − C then A, B, C are three different points on the same line.
The congruence postulates are
1. Congruence is an equivalence relation.
2. The segments construction postulate.
3. The segment addition postulate.
4. The segment subtraction postulate.
5. Every segment has exactly one midpoint.

43
8 The Euclidean Program: Congruence Without Distance

6. The angle construction postulate

7. The angle addition postulate
8. The angle subtraction postulate
Another postulate is SAS.

8.2 The Laws of Inequality for Segments

The synthetic treatment of segments begins with the inequalities where the basic idea
was already presented at the beginning of chapter 7.

8.3 Right Angles, Synthetically Considered

We can show from axioms1 that all right angles are congruent2 .

8.4 The Synthetic Form of the Triangular Inequality. Addition

of congruence classes

We are ranting about synthetic treatment. In particular we think its elegant, but there
are a lot of technicalities that are too often ignored.

1
as did Hilbert
2
Euclid took this as an axiom

44
9 Three Geometries

9.1 Introduction

This chapter is very informal an only contains discussion with no proofs.

For simplicity, while the ideas can be extended to three-dimensions at the cost of con-
siderable labour, we will limit ourselves to the plane.
Two lines are called parallel if they lie in the same plane but do not intersect.

Definition 9.1.1 (The Euclidean Parallel Postulate)

Given a line L and point P ∈
/ L there is exactly one line parallel to L which goes
through P .

Definition 9.1.2 (The Lobachevskian Parallel Postulate)

Given a line L and a point P ∈
/ L there are at least two lines containing P that are
parallel to L.

Definition 9.1.3 (The Riemann Parallel Postulate)

No two lines in the same plane are every parallel.

In this book we deal mostly with the first geometry, a little bit with the second and
almost none with the third.

9.2 The Poincaré Model for Lobachevskian Geometry

In this section we assume there exists a model of Euclidean geometry.

45