2.

4 HILBERT'S MODEL FOR EUCLIDEAN GEOMETRY

During the nineteenth century, it became evident to mathematicians interested in geometry that
there was no single axiom set that resulted in a single (i.e., universal) model of geometry. Rather,
there evolved a consensus that the validity of a geometric axiomatic system is dependent on the
consistency, independence, and completeness of the axiom set upon which it is built. To be sure,
it was known that different axiom sets would result in different models, but the study of
geometry was by then no longer restricted by the notion that every resulting model had to be
consistent with the Euclid¬ ean model. Still the Euclidean model for geometry was by far the
most intuitive, and since it was the model with the broadest historical basis, mathematicians took
on the task of constructing an axiom set that would result in the theorems of Euclid but without
the shortcomings mentioned earlier in this chapter. The most well known of these works,
Grundlagen der Geometrie (Foundations of Geometry), which is true to the spirit within which
Euclid worked, was published in 1899 by a German named David Hilbert, arguably the most
prominent mathematician of the era.
Unlike Euclid, Hilbert was well versed in the requirements of modern axiomatics. For
undefined terms, Hilbert chose “point,” “line,” “plane,” “on” (incidence of point to line),
“between” (as a relation concerning three distinct points), and “congruence.” It is a tribute to
Hilbert’s genius (and to his perseverance) that he was able to recognize that all the necessary
terms of Euclidean plane geometry could be defined starting with this simple set of primitives.
For example, Hilbert was able to distinguish between lines and line segments (something Euclid
neglected to do) by using the following definition:

Definition. Line Segment AB. The set of all points that are be¬ tween point A and point B.
Points A and B are called the endpoints of the segment.
Hilbert partitioned his axioms into five groups: axioms of connection (or incidence),
axioms of order, axioms of congruence, the axiom of paral¬ lels (Playfair’s postulate), and
axioms of continuity. The entire axiom set may be found in Appendix B, but for now let’s
consider just the axioms of connection.
Group I Axioms of Connection
1-1. Through any two distinct points A, B, there is always a line m.
1-2. Through any two distinct points A, B} there is not more than one line m.
1-3. On every line there exist at least two distinct points. There exist at least three points which
are not on the same line.
1-4. Through any three points not on the same line, there is one and only one plane.

It would be difficult to fail to notice the contrast between Hilbert’s Postulates 1-1 and 1-2
and Euclid’s first postulate: “To draw a line from any point to any point.’’
Euclid certainly intended to imply the same idea but failed to do so as completely. Hilbert
not only postulates the existence of a line between any two distinct points but also postulates the
uniqueness of that line. (Euclid assumes this throughout but fails to postulate the idea.) In
addition, since Postulate 1-3 describes the existence of points, we are assured of having lines
with which to work, a formality Euclid overlooked.
Since Hilbert’s axiom set is intended to provide the basis for traditional Euclidean
geometry, we should expect that all our intuitive ideas about plane geometry are valid in his
axiomatic system. For example, suppose that / and m are lines. Is it possible that l and m
intersect in more than one point? Since our intuition says no, but no axiom assures that
intersecting lines intersect just once, we should hope that this property can be proved as a
theorem. With this in mind, consider the following.

Theorem. Two distinct lines cannot intersect in more than one point.
Proof. We will proceed indirectly and assume otherwise: Suppose that l and m intersect in two
distinct points called A and B. Under this hypothe¬ sis A and B are distinct points contained by
two distinct lines, a contradiction of Postulate 1-2. Consequently, the assumption that l and m
intersect in more than one point cannot be valid, so that we may conclude that pairs of distinct
lines do not intersect in more than one point. Two points need to be made concerning this proof.
First, the theorem is so obviously true in Euclidean geometry that it might seem to qualify
as an axiom. However, the very fact that the state¬ ment can be proved indicates that it is not
independent of the others. Had Hilbert included it as an axiom, the set of axioms would not be
indepen¬ dent. It is unfortunately true that things are complicated by having to prove obvious
statements. Still the elegance of constructing an independent axiom set outweighs, for
mathematicians of Hilbert’s status, the added bother.
Second, notice the indirect nature of this proof: Assume otherwis-derive contradiction-
conclude theorem. This technique was applied in many of the proofs from the finite geometries
discussed in Chapter 1. If you are still uncomfortable reasoning in this fashion, you should work
hard to overcome this problem, since proofs of this kind are at the core of many of the arguments
in both Euclidean and non-Euclidean geometries. The exercise set at the end of this section
requires a number of indirect proofs such as this.
While Hilbert’s axioms of connection are straightforward, some of the other axioms and
the implications that may be deduced from them are less obvious. In particular, read the
following four axioms of order (those concerning the undefined term “between”):

Group II Axioms of Order

II-1. If point B is between points A and C (we will denote this by A-B-C), then A, B, and C are
distinct points on the same line and C-B-A.
II-2. For any two distinct points A and C there is at least one point B on the line AC such that A-
C-B.
II-3. If A, B, and C are three points on the same line, then exactly one is between the other two.
II-4. Let A, B, and C be three points that are not on the same line, and let m be a line in the plane
containing A, B, and C that does not contain any of the points A, B, or C. Then, if m
contains a point of the segment AB it will also contain a point of segment AC or a point of
segment BC.

The purpose of the axioms of order is to give meaning to the undefined term “between.”
Since “between” is a primitive, we have nothing to guar¬ antee that the term fits our notion of
betweenness except the axioms of the system. While it may not be obvious at first, the
betweenness relation implicitly defined by these axioms is consistent with our intuition.

The axioms of order imply many results we expect in a Euclidean geometry. For example,
Axiom II-2 allows us to conclude that if we can locate two points A and C on a line (which we
can by Axioms 1-3 and 1-1), we can find a third point B on the line that is farther8 from A than
C is. Of course, having assured the existence of B, we may reapply Axiom II-2 to produce a
fourth point, and then a fifth point, and so on. These ideas result in the following theorem.

Theorem. Every line contains an infinity of points.

A similar argument can be used to establish theorems stating that lines do not terminate at
any point and that the points of a line are serial in nature rather than cyclical. This assures us that
the points of a line do not “wrap around,” as would happen with a great circle on a sphere.
Some of the consequences of the axioms of order are self-evident. For example, we can
prove the following obvious theorem.

Theorem. If A and B are points, then there always exists a third point C, such that A-C-B.

This statement is not quite the same as Postulate II-2 (reread Postulate II-2 if this isn’t
clear) but is certainly a property we expect to be valid in the geometry. To show that it is, we
begin by choosing an arbitrary line segment AB (Figure 2.4.1). Axiom 1-3 assures us that there
is a third point D that is not collinear with A and B. Axiom II-2 allows us to locate a point E so
that A-D-E. Axioms 1-1 and 1-2 guarantee the unique line EB, and a reapplication of Axiom II-2
allows us to place point F so that E-B-F. Now consider the three noncollinear points A, E, and B
and the unique line that joins D and F. Axiom II-4 assures us that if DF contains a point ofAE (in
this case point D), then it must also contain a point on EB or AB. Since DF intersects EB at point
F, it cannot intersect at a second point on EB (which axiom excludes this possibility?) so that
there is a point C on AB such that DF  AB = C. Since AB is comprised of the points that are
between A and F, we have now located point C such that A-C-B.
 Axiom II-4, on which the preceding proof heavily relies, was actually stated first by the
German mathematician Moritz Pasch in 1882 and is commonly referred to as Pasch’s axiom.
This axiom fills in a significant gap left by Euclid. Many of the proofs in the Elements rely on
assumptions made from visual cues from diagrams. While the theorems Euclid proved in this
way were true, it was found later that proofs of false statements could be constructed by using
much the same type of reasoning axiom may seem obvious (it is after all an assumption), it or a
postulate equivalent to it is needed to provide valid means of proving several of Euclid’s
theorems.
Hilbert’s third group of axioms, the axioms of congruence, listed here, are self-
explanatory.

Group III Axioms of Congruence

III-1. If A and B are two (distinct) points on line a, and if A' is a point on the same or another
line a', then it is always possible to find a point B' on a given side of the line a' through A'
such that segment AB is congruent to segment A'B'.
III-2. If a segment A'B' and a segment A'B" are congruent to the same segment AB, then the
segment A'B' is also congruent to the segment A"B" or, briefly, if two segments are
congruent to a third segment, they are congruent to each other.
III-3. On the line a, let AB and BC be two segments which except for B have no point in
common. Furthermore, on the same or another line a', let A'B' and B'C' be two segments
which except for B' have no point in common. In that case, ifAB = A'B' and BC = B'C',
then AC = AC'. (This axiom expresses the additivity of segments.)
III-4. If AABC is an angle and if B'C' is a line, then there is exactly one point A' on each side of
B'C' such that AA'B'C' = AABC. Further¬ more, every angle is congruent to itself. (This
is often referred to as the angle construction axiom.)
III-5. If for two triangles AABC and AA'B'C' the congruences AB = A'B', AC = AAC and
ABAC = AB'A'C' are valid, then the congruence AABC = AA'B'C' is also satisfied.

Axiom III-4 says, in a somewhat indirect way, that any angle can be copied to a given
ray, once on each side of the line that contains the ray. Axiom III-5 is in effect the SAS
congruence condition (see Exercise Set 2.4, Problem 16). Euclid lists SAS as a theorem,
however, the proof is flawed because he assumes the ability to move geometric figures around
the plane without distortion. While this use of superposition is reasonable, any proof that makes
use of this technique violates the rules of formal axiomatics unless a postulate is included to
justify doing so. Eventually, a number of geometers legitimatized this technique by formally
postulating the existence of geometric transformations that allow images of geometric figures to
be used in the proof of a great many theorems. Hilbert chose not to travel the “transformation
route,” since an axiom set based on Euclidean transforma¬ tions changes the nature of the
development of Euclidean geometry, even though the results are clearly Euclidean. Since
Euclidean transformations have become an important aspect of geometry, they will be discussed
in detail in Chapter 5.
Hilbert, to his credit, recognized that at least a portion of the SAS congruence condition
for triangles is truly independent of the other tradi¬ tional Euclidean postulates. Consequently,
he concluded that a postulate concerning the congruence of triangles was necessary. Since SAS
is per¬ haps the most obvious of the Euclidean congruence conditions, Hilbert chose that one as
his axiom of triangle congruence. As we shall see in Chapter 3, the other familiar congruence
conditions can be derived as theorems.
By the time Hilbert presented his work in Euclidean geometry, it had already been
established by Felix Klein that the parallel postulate is indepen¬ dent of the other, traditional,
postulates for geometry. In order to help complete the axiom set, Hilbert needed to choose some
form of a parallel postulate. He selected a form of Playfair’s postulate (although in Grundla- gen
der Geometrie he calls it Euclid’s axiom) as his axiom of parallels (Axiom IV-1), which for
convenience is restated here.

IV-1. Let a be any line and A a point not on it. Then there is at most one line in the plane,
determined by a and A, that passes through A and does not intersect a.

Note that this axiom postulates “at most one line’’ rather than “exactly one line parallel to
a,” even though in Euclidean geometry we expect to find exactly one parallel. The reason for this
is, as we will prove in Chapter 3, that the other postulates imply the existence of at least one
parallel so that to postulate exactly one would be somewhat redundant. In order to maintain the
independence of the axioms, Hilbert opted for the weaker version of Playfair’s postulate.
This left but one gaping hole to be filled in Euclid’s geometry the question of continuity of
lines.11 For convenience, the two Continuity postulates are stated here.
Group V Axioms of Continuity
V-l. (Axiom of Archimedes) If AB and CD are any segments, then there exists a number n such
that n copies of CD constructed contiguously from A along AB will pass beyond the point
B.
V-2. (Axiom ofLine Completeness) An extension of a set of points on a line with its order and
congruence relations that would preserve the relations existing among the original elements
as well as the fundamental properties of line order and congruence that follow from
Axioms I through III and V is impossible.

The idea behind Axiom V-l is fairly straightforward, although the statement itself is
confusing at first. The essence of the postulate is that line segments can be measured by units of
an arbitrary size.
During the nineteenth century a number of mathematicians12 studied geometries that
involved spaces that were unbounded but finite. Distance in these and other geometries is
measured in ways much different than in the Euclidean plane. Axiom V-l guaranteed
measurement of the usual (namely, Euclidean) type.
Axiom V-2 does not directly concern us in our discussion of Hilbert’s geometry as an
“upgrade” of Euclid’s work. The completeness postulate is not necessary for the proof of any of
the theorems in Euclidean geometry. By including it, Hilbert allowed himself to make the
connection between his geometry and the formal systems of arithmetic that were being fine-
tuned at the time. Hilbert’s agenda involved establishing a one-to-one correspondence between
the points on any line and the real numbers. This would allow him to declare the geometric
system to be as consistent as the field of real numbers. Hilbert had no doubt that the consistency
of the real numbers could be established, hence that his system would be also. This was a
philosophical as well as a mathematical issue during that era which, alas, was not resolved to the
satisfaction of Hilbert, and of many others.
 Hilbert’s geometry had mathematical significance beyond “patching up the holes” in
Euclid’s work. It was a classic example of the modern axi¬ omatic method, and since it appeared
at about the beginning of this century, it helped set the tone for much of twentieth century
mathematical thought.

EXERCISE SET 2.4

Use Hilbert’s axioms for Euclidean geometry to support informal proofs.

1. Have any of the theorems proved by Euclid in the Elements been shown to be false in
Euclidean geometry? If so, which ones? If not, why did Hilbert feel the need to revamp
Euclid’s work?
2. Use Hilbert’s axioms (you may assume that the full SAS condition has been established) to
prove that the base angles of an isosceles triangle are congruent. {Hint'. Prove that isosceles
AABC is congruent to ACM.)
3. Suppose that W, X, Y, and Z are points with X between W and Y and Y between W and Z.
Prove, using Hilbert’s axioms, that W, X, Y, and Z are distinct, collinear points.
4. In geometry, if we choose a point on a line and consider that point along with all the points
on one side of that point, we have what is usually called a ray. Give a more formal definition
of the term "ray.”
5. How many points are there on a line segment? Provide an argument in support of your
response.
6. Explain what is meant by the following statement: Every line in a plane sepa¬ rates the
plane into exactly two disjoint half-planes. Is this statement true? Would it be true if "line”
were replaced by "line segment”? By "ray”? By circle”?
7. Suppose you are given two distinct points X and Y and a line / that does not contain X or Y.
Describe a criterion that could be used for determining whether the two points are in the
same or different half-planes.
8. Is it possible for a line to intersect all three sides of a triangle? If so, does this contradict
Axiom II-4? If not, explain why not.
9. Prove that if X, Y, and Z are points with Y between X and Z, then (a) AT Pi YZ = Y and (b)
XY U YZ = XZ. (You may assume that line segments include their endpoints.)
10. Prove that two lines that are parallel to the same line are parallel to each other. (Hint:
Proceed indirectly.)
11. What is meant by the term "interior of a triangle”? How can this term be defined formally?
If a line contains a vertex of a triangle and interior points of the triangle, must it intersect the
side opposite the vertex that it contains? Explain why or why not.
12. If a line intersects a side of a rectangle, must it intersect another side of the rectangle?
Explain why or why not.
13. If a line intersects three sides of a rectangle, must it contain a vertex? Explain why or why
not.
14. Hilbert did not define distance in his geometry. We can, however, allow for lengths to be
compared by defining the relation "<” for segments as follows. WX < YZ if and only if there
is a point P between Y and Z such that WX = YP. Using this definition, show that for every
set of points W, X, Y, and Z one of the following statements is true.
(i) WX < YZ
(ii) WX ≅ YZ
(iii) YZ < WX.
15. Show that the “<” relation defined in Problem 14 is transitive.
16. Congruence Axiom III-5 does not postulate, entirely, the SAS congruence con- dition. By a
change in notation it clearly implies that the remaining angles are congruent. We may not,
however, conclude immediately that the remaining sides are congruent. Hilbert omitted
postulating this final congruence because it can be proved as a theorem from the other
axioms. The proof proceeds in the following manner. Suppose the remaining sides BC and
B'C' aronot congruent (Figure 2.4.2). (a) Under this hypothesis there is a unique point D' on
B'C' such that D'C' = BC. Which axiom guarantees this? (b) Next consider AABC and
AA'D'C'. Verify that an application of Axiom III-5 to these triangles allows us to conclude
that LD'A'C = LA and that this result, with the “given,” contra¬ dicts Axiom III-4.