2.2.

A Brief Critique of Euclid 1

2.2. A Brief Critique of Euclid

Note. In this section we discuss some weaknesses of Euclid’s approach to geometry.
We have already mentioned the futility of trying to define every term. We will also
address Euclid’s failure to deal with continuity appropriately and give a “proof”
that every triangle is isosceles by taking advantage of a Euclid’s weakness in dealing
with order relations (and “betweenness”).

Note. We mentioned in Section 1.3. Axiomatic Systems that an axiomatic system

includes undefined terms. Euclid does not take this approach and he has the fol-
lowing definitions of “point,” “straight line” (which is distinguished from a “line”),
and “plane angle.”

Definition. A point is that which has no part.

Definition. A straight line is a line which lies evenly with the points on itself.

Definition. A plane angle is the inclination to one another of two straight lines in
a plane which meet one another and do not lie in a straight line.

Of course this raises as many questions as it answers, since we now focus on the
terms “part,” “lies evenly,” and “inclination.”

Note. The next concern deals with continuity. Many of the results in the Elements
are inspired by compass and straight-edge constructions. In Book I Proposition I
Euclid presents the construction of an equilateral triangle in which additional,
unstated assumptions are needed. As in Figure 2.1(b), he starts with a line segment,
2.2. A Brief Critique of Euclid 2

uses the compass to draw an arc of a circle of radius equal to the length of the
segment and centered at one end of the segment, and then uses the compass to
draw an arc of another circle of the same radius and centered at the other end
of the segment. Of course, this results in a point that is equidistant from both
endpoints of the segment and which can be used to construct the desired isosceles
triangle. This is shown in Figure 2.1(a) for the point “above” the line segment (we
know there is also such a point “below” the line segment). The continuity concern
deals with the intersections of the arcs of the circles. What if the lines have “holes”
in them or if the points on the arcs are distributed in such a way (“like beads on
a string,” as Wiley says on page 40) that the arcs can pass through each other
without intersecting? This is illustrated somewhat in Figure 2.1(b).

Figure 2.1

Note. It might sound odd to worry about these details which certainly violate our
intuitive ideas of circles. In fact, this plays out to a resolution in the 19th century.
Wendell Strong in “Is Continuity of Space Necessary to Euclid’s Geometry?” Bul-
letin of the American Mathematical Society, 4(9), 448–448 (June 1898) (a copy can
be downloaded from projecteuclid.org) discusses what he calls quadratic space. This
2.2. A Brief Critique of Euclid 3

space consists of all points in the Cartesian plane which have quadratic coordinates
(a real number is quadratic if it can be obtained from the integers by a finite num-
ber of rational operations and extractions of square roots). The quadratic space is
everywhere discontinuous, yet any construction that can be performed with a com-
pass and a straightedge can be performed in this space! So in response to Strong’s
question in the title of his paper,“No!” However, two circles with quadratic centers
and quadratic radii which intersect in the continuous Cartesian plane intersect at
quadratic points. By restricting our attention to intersections of lines and circles
(which are themselves constructible from existing [constructible] points and dis-
tances), continuity is not needed! This certainly was not known to Euclid and was
not resolved until the study of field theory in modern algebra arose. Though not
actually developed with these problems in mind, the area of field theory in algebra
ultimately is the tool allowing us to classify Strong’s quadratic numbers. This is
covered in our Introduction to Modern Algebra 2 (MATH 4137/5137) in Section
VI.32. Geometric Constructions and our Modern Algebra 2 (MATH 5420) in Sec-
tion V.1.Appendix. Ruler and Compass Constructions. The rational numbers are
constructible as seen in Section 1.2. Similar Figures (see Figure 1.6). It can be
shown that the set of real constructible numbers C forms a subfield of the field of
real numbers (see Corollary 32.5 in the Introduction to Modern Algebra 2 notes)
and, in particular, the field of constructible real numbers C consists precisely of all
real numbers that we can obtain from Q by taking square roots of positive numbers
a finite number of times and applying a finite number of field operations. From the
Introduction to Modern Algebra 2 we have:
2.2. A Brief Critique of Euclid 4

Theorem 32.6. The field of constructible real numbers consists pre-

cisely of all real numbers that we can obtain from Q by taking square
roots of positive numbers a finite number of times and applying a finite
number of field operations
See also Proposition V.1.16 in Section V.1.Appendix. Ruler and Compass Con-
structions of Modern Algebra 2. Additional details on constructions can be found
in my video “Compass Straightedge Constructions” on YouTube. This study of
constructible numbers ultimately grows out of the study of the three classical com-
pass and straightedge constructibility problems (Doubling the Cube, Squaring the
Circle, and Trisecting an Angle); these problems also serve as the original inspi-
ration for the study of the conic sections. For more detail on the history, see my
online notes for “Introduction to Modern Geometry (History)” (MATH 4157/5157)
on Section 1.8. Three Famous Problems of Greek Geometry and Chapter 3. Conic
Sections.

Note. Interestingly, the complexities of a continuum were first rigorously explored

in the work of Richard Dedekind (October 6, 1831–February 12, 1916). The real
line is a continuum due to the Axiom of Completeness. This is traditionally dealt
with in terms sets of real numbers with upper bounds. Concisely, the Axiom of
Completeness states that every set of real numbers with an upper bound has a least
upper bound. For details, see my online notes for Analysis 1 (MATH 4217/5217)
on Section 1.3. The Completeness Axiom. An alternative approach is the use of
“Dedekind cuts.” This idea was introduced in Dedekind’s 1858 work (first published
in 1872) “Continuity and Irrational Numbers” (a copy can be found online at
2.2. A Brief Critique of Euclid 5

Project Gutenberg). These ideas are informally explained in the setting of an

airplane taking off in my online Calculus 1 (MATH 1910) notes on Appendix A.6.
Theory of the Real Numbers. I also have a video of this material at A.6 Video (of
length (31:05)). On a personal note, I find it amazing that the idea of a line as a
continuum is as old as any other in geometry, but it was not until 1858 that an
appropriate axiomatic system was introduced to deal with this idea!

Note. Another critique of Euclid relates to the concept of “betweenness.” That is,
the order in which points appear on a line is largely left unaddressed; drawings are
used and “. . . it was sometimes for [Euclid] to establish with certainty the location
of one point with respect to others in a given discussion.” See Wylie’s pages 40
and 41. To illustrate this, we consider the following absurd result. This is a widely
circulated idea; a Google search of “all triangles are isosceles” gives about 5000
sites (as of 10/26/2021).

“Theorem.” Every triangle is isosceles.

“Proof.” Let 4ABC by an arbitrary triangle. Let the bisector of ∠BAC, intro-
duce the perpendicular bisector of side BC, and let O be their point of intersection,
as shown in Figure 2.2 below. Let points A0 , B 0 , C 0 be, respectively, the points of
intersection of the perpendiculars from O to the sides BC, CA, and AB. Notice
by our construction that A0 is also the midpoint of side BC.
2.2. A Brief Critique of Euclid 6

Figure 2.2

We now present our argument in the two column style that you may have used in
high school geometry (we also use a notation for angles and triangles from high
school).
1. A0 O = A0 O. 1. Identity.
2. BA0 = A0 C. 2. By construction.
3. ∠OA0 B = ∠OA0 C. 3. Both are right angles.
4. 4OA0 B ∼
= 4OA0 C. 4. Side, Angle, Side.
5. OB = OC. 5. Corresponding parts of congruent 4’s.
6. AO = AO. 6. Identity.
7. ∠C 0 AO = ∠B 0 AO. 7. AO bisects ∠BAC.
8. ∠AC 0 O = ∠AB 0 O. 8. Both are right angles.
9. 4AC 0 O ∼
= 4AB 0 O. 9. Angle, Side, Angle.
10. AC 0 = AB 0 . 10. Corresponding parts of congruent 4’s.
2.2. A Brief Critique of Euclid 7

11. OC 0 = OB 0 . 11. Corresponding parts of congruent 4’s.

12. ∠OC 0 B = ∠OB 0 C. 12. Both are right angles.
13. OB = OC. 13. By Step 5.
14. 4OC 0 B ∼
= 4OB 0 C. 14. Right angle, hypotenuse, leg.
15. C 0 B = B 0 C. 15. Corresponding parts of congruent 4’s.
16. AB = AC 0 + C 0 B. 16. C 0 is between A and B.
17. AB = AB 0 + B 0 C. 17. Steps 10 and 15.
18. AC = AB 0 + B 0 C. 18. B 0 is between A and C.
19. AB = AC. 19. Steps 17 and 18.
Therefore 4ABC is an isosceles triangle. Since 4ABC is an arbitrary triangle,
then all triangles are isosceles, as claimed. Q.E.D. (NOT!)

Note. The problem in the previous “proof” is one of betweenness. The point O of
intersection of the bisector of ∠BAC and the perpendicular bisector of BC in fact
lies outside of 4ABC (as is to be shown in Exercise 2.11.10), so that one of the
points B 0 and C 0 will lie between two vertices of the triangle while the other will
not (it will lie on one of the sides, AC or AB respectively, extended beyond the
vertices of 4ABC). We can see in Figure 2.2 that ∠BAO appears a little smaller
than ∠CAO, so that ∠BAC is not actually bisected by AO.

Note. Since we desire an axiomatic system for Euclidean geometry which will
leave the terms “point” and “line” as undefined (their meaning being given by
the axioms and the theorems which follow from the axioms), we cannot appeal to
figures and drawings! Wylie makes the following comment about the situation (see
2.2. A Brief Critique of Euclid 8

page 42):
“As a matter of fact, from the axioms of Euclid it is impossible to
determine whether the point O is inside or outside the triangle. Worse
yet, it is impossible using Euclid’s axioms, even to give a satisfactory
definition of the inside and outside of a triangle.”
Additional critiques can be aimed at Euclid’s treatment of distances, the measures
of angles, and the idea of congruence which is thought of in terms of superimposing
one object (angle or triangle, for example) on another and requires some concept
of transformations. In the remainder of this chapter, we address these problems
and present an axiomatic system for Euclidean geometry.

Revised: 10/27/2021