Accelerat ing t he world's research.

The role of Euclidean geometry in

high school
Hung-Hsi Wu

The Journal of Mathematical Behavior

Cite this paper Downloaded from Academia.edu 

Get the citation in MLA, APA, or Chicago styles

Related papers Download a PDF Pack of t he best relat ed papers 

Cognit ive development of proof

Walt er Whit eley, Margo Kondrat ieva

Inst ruct ional Sit uat ions and St udent s'Opport unit ies t o Reason in t he High School Geomet ry Class
Pat ricio Herbst

Advanced Mat hemat ics for Secondary Teachers: A Capst one Experience
David Meel
JOURNAL OF MATHEMATICAL BEHAVIOR 15, 221-237 (1996)

The Role of Euclidean Geometry

in High School
HUNG-HSI Wu
University of California

When I was a high school student long ago, the need to study Euclidean geome-
try was taken for granted. In fact, the novelty of learning to zyxwvutsrqponmlkjihgfedcba
prove something was
so overwhelming to me and some of my friends that, years later, we would look
back at Euclidean geometry as the high point of our mathematics education. In
recent years, the value of Euclid has depreciated considerably. A vigorous debate
is now going on about how much, if any, of his work is still relevant. In this
article, I would like to give my perspective on this subject, and in so doing, I will
be taking a philosophical tour, so to speak, of a small portion of mathematics.
This may not be so surprising because after all, in life as in mathematics, one
does need a little philosophy for guidance from time to time.
The bone of contention in the geometry curriculum is of course how many of
the traditional “two-column proofs” should be retained. Let me first briefly
indicate the range of available options in this regard:

1. A traditional text in use in a local high school in Berkeley not too long ago
begins on page 1 with the undefined terms of the axioms, and formal proofs
start on page 22. There is no motivation or explanation of the whys and
hows of an axiomatic system.
2. A recently published text does experimental geometry (i.e., “hands-on ge-
ometry” with no proofs) all the way until the last 130 pages of its 700 pages
of exposition. Moreover, the presentation in these 130 pages fails to give a
clear picture of why undefined terms and axioms are needed, what precisely
the undefined terms are in that context, and what role axioms play in mathe-
matics.
3. The text of a new curriculum discusses only experimental geometry. There
are no proofs, two-column or otherwise.

This work was partially supported by the National Science Foundation. This article is a slightly
expanded version of a lecture presented to the Bay Area Mathematics Project on July 27, 1992. I wish
to thank Alfred Manaster for his constructive criticisms.
Correspondence and requests for reprints should bc sent to Hung-Hsi Wu, Mathematics Dept.,
University of California, Berkeley, CA 94720-3840.

221
222 WU

As a reference point, I should recall that the NCTA4 Standards (1989) recom-
mends that two-column proofs should “receive less emphasis” (p. 159). It cer-
tainly does not recommend their elimination.
To begin our discussion, first allow me to quote something written by a
nonmathematician some time ago:

He studied and nearly mastered the six books of Euclid since he was a member
of Congress.
He began a course of rigid mental discipline with the intent to improve his
faculties, especially his powers of logic and language. Hence his fondness for
Euclid, which he carried with him on the circuit till he could demonstrate with ease
all the propositions in the six books; often studying far into the night, with a candle
near his pillow, while his fellow-lawyers, half a dozen in a room, filled with air
with interminable snoring.

Can you guess who the author is? Of course, it is President Lincoln writing
about himself in his Short Autobiography (Faber, 1983, p. 45). In some sense, all
I am going to do is no more than elaborate on what President Lincoln wrote a
century and a half ago. So let me begin by burning my bridges and declaring at
the outset that there must be proofs in a geometry course. The rest of this article
is spent justifying this claim and outlining what I consider to be a reasonable way
to implement such a recommendation.
So why proofs? Proofs are the guts of mathematics. Producing a proof of a
statement is the basic methodology whereby we can ascertain that the statement
is true. Anyone who wants to know what mathematics is about must therefore
learn how to write down a proof or at least understand what a proof is. The
sciences also use the same methodology to deduce complex phenomena from
first principles. Thus all who want to study science would benefit from learning
about proofs as well. For the others who are outside of science, it comes down to
what Lincoln wrote, that learning how to prove theorems is an excellent way to
sharpen one’s mind. In a larger context, if anyone has any wish at all to find out
how human beings can distinguish right from wrong or true from false, he or she
would find in mathematical proofs the purest form of how this is done. Now I
would like to make some comments about geometric texts that put most or all of
their weight on experimental geometry and intentionally slight the theorem-
proof aspect of geometry. Perhaps the argument is that because hands-on experi-
ments are as efficient at arriving at the truth as abstract arguments, why not
bypass this arduous task of writing down proofs altogether? In case such a
statement does not immediately strike the reader as being silly, let me try to
convince you with a simple example. A standard problem in number theory is to
find integer solutions to equations of the following type (the Fei-mat-Pell equa-
tion):

x2 - 1141y2 = 1.
 HIGH SCHOOL GEOMETRY 223

This is of course the same as looking for positive integers y so that 1 + 1,14 ly2 is
a perfect square. (We exclude the obvious solutions: x = 1, y = 0.) This is a
problem tailor made for experimentation on the calculator. Starting with y = 1,
2, 3, . . . we can work our way up. The case y = 1 is no good because 1,142 is
not a perfect square, for the simple reason that the square of any number must
end in 1, 4, 5, 6, or 9. In fact, nothing works up to 100. For example, with y =
99, we get

I + 1,141 (99)2 = 11,182,942

so that for the same reason it is not a perfect square. Similarly, nothing works up
to y = 100,000. For example, with y = 23,456, we get

1 + 1,141 (23,456)* = 627,759,870,977,

and because it ends in 7, it is not a square. If you try y = 45,678, then

1 + 1,141 (45,678)2 = 2,380,673,319,445,

and

1,542,9432 = 2,380,673,101,249 < 2,380,673,319,445 < 2,380,676,187,136

= 1 ,542,9442.

In fact, for all integers y all the way up to 1025, 1 + 1141~2 is never a perfect
square. In terms of experimentation, one would have given up long before this
and concluded that this particular Fermat-Pell equation has no integer solutions
in x and y. But in fact, we can zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONM
prove that there are an infinite numbers of integer
pairs x and y that satisfy this equation, the smallest being:

y = 30,693,385,322,765,657,197,397,208

and

x = 1,036,782,394,157,223,963,237,125,215.

So much for the power of experimentation. r Now this is not to belittle the
importance of experimentation, because experimentation is essential in mathe-
matics, What I am trying to do is to point out the folly of educating students to
rely solely on experimentation as a way of doing mathematics. Mathematics is
concerned with statements that are true, forever and without exceptions, and

‘I have taken this example from Stark (1978)

 224 wu

there is no other way of arriving at such statements except through the construc-
tion of proofs.
One school of thought claims that secondary school is not the place for
students to learn to write rigorous, formal mathematical proofs, and that the
place to do this is in upper division courses in college. Similar claims have also
been put forth by other curriculum reformers.* With the preceding discussion at
hand, let us approach these claims from a rational perspective. To call a spade a
spade, a so-called “rigorous, formal mathematical proof” is just correct mathe-
matical reasoning, no more and no less. As we are in an age when mathematical
knowledge is at a premium, it does not seem proper that correct mathematical
reasoning should be suddenly declared too profound and too difficult for all high
school students and must be reserved for a few mathematics majors in college. If
this country still has any intention of producing high school students that are first
in the world in mathematics and science by the end of the century,3 the correct
move should be to put more emphasis on the substantive part of mathematics
(i.e., proofs) and not less. In a broader context, mathematics courses are where
the students get their rigorous training in logical reasoning; this is where they
learn how to cut through deceptive trappings to get at the kernel of truth, where
they learn how to distinguish between what is true and what only seems to be true
but is not. They would need all these skills in order to listen to the national debate
and make up their minds about such knotty issues as the national deficit and the
environment, for example. Learning how to write correct proofs is a very impor-
tant component in the acquisition of such skills. (see again the previous quote of
Lincoln). Because the high school students should all be voting to help mold the
destiny of this nation as soon as they reach 18, any proposal that would deprive
them of this training, so vital to their task at hand, must be regarded as ill advised
at best.
Slightly less radical proposals have also been advanced in the Addenda Series
of the NCTM Curriculum and Evaluation Standards (NCTM, 1992a, 1992b). A
fair assessment of their merits and demerits can only be made in the context of
their overall curriculum proposal (NCTM, 1992a) and must therefore wait for a
later occasion. Having said this, I nevertheless feel compelled to address (if only
in an ad hoc manner) a particular point that is strikingly germane to our present
discussion. In deciding to drastically tone down the presence of two-column
proofs to make room for other topics (see the Differentiated Curriculum in
chapter 4 and Appendix II of NCTM, I992a), the idea in these addenda seems to
be “to build the kind of strong geometric intuition that has been shown to be an
important factor for success on the job and in college” (NCTM, 1992b, p. v) at
the expense of properly executed proofs. The strategy to achieve this goal is to

‘A happy exception to this trend is the text of Coxford, Usiskin, and Hirschhom (1991).
‘Recall that this was the clarion call for education reforms by former President George Bush.
 HIGH SCHOOL GEOMETRY 225

approach geometry “from a variety of perspectives: synthetic, coordinate, trans-

formation, and vector” (1992b, p. 4), to allow the students to “explore geometry
in a more eclectic fashion, taking advantage of more connections in the curricu-
lum” (1992a, p. 72), and to be more flexible in the use of less formal “reasoning”
and “justification” where absolute logical validity is of a lesser concern (1992b,
p. 61).
Now there will be no argument from any sensible person, least of all from a
mathematician, against the need to develop a strong mathematical intuition,
geometric or otherwise. There will also be no argument against the fact that the
traditional curriculum, when poorly taught, does not do enough to actively
encourage the development of this kind of intuition. However, as is usually the
case, the prescription by the NCTM Addenda Series is far too strong for this
particular malaise. By presenting geometry right from the beginning from multi-
ple perspectives, such a curriculum guarantees that the students’ understanding
of geometry from any one perspective would not be deep enough to take root;
indeed, the inherent time limitation would see to it that this is the case (see
NCTM, 1992a, 1992b, especially Appendix II, 1992a). In a sense, this is no
different from showing a picture on which several snapshots have been superim-
posed: The untrained eyes would find the picture too blurred to be able to see
what is going on. If there is any objection to Euclid’s presentation of geometry
per se, then valid alternatives abound. One way would be to start with vectors,
state clearly the axioms involved, rederive all the classical Euclidean theorems
from a vectorial point of view, and stay the course. Although there are obvious
gains and losses in this vectorial presentation compared with the Euclidean one,
there is no question that it is feasible. Once the students have gone deeply enough
into vector geometry and have become sufficiently proficient in the use of its
basic techniques, then (and only then) would it be time to discuss the other
approaches.
The reason for going into the details of the preceding objection is that it is
intimately tied up with the next (and my main) objection to the NCTM Addenda
Series, and it is that this curriculum (i.e., the Differentiated Curriculum, NCTM,
1992a) fails to make the students see the most important feature of what consti-
tutes mathematical knowledge, namely, that once the truth of a small set of
statements (axioms) is granted, then the truth of everything else within this body
of knowledge-deduced as they are from the former by precise logical reason-
ing-will leave no room for doubt.4 In saying this, I am also quite mindful of the
fact that mathematics is a multifaceted human activity and is therefore suscepti-
ble to vastly different but equally valid presentations. Thus, a high school mathe-
matics education can be problem based, or it can be theory oriented, or it can be
application oriented. Nevertheless, if such an education does not succeed in

“We ignore the subtleties of GBdel’s theorem.

226 wu

giving the students a firsthand experience of the power and the economy of this
basic axiom-theorem-deductive feature of mathematics, then surely it cannot
claim to have done its job. As Redheffer (1992) so eloquently put it in a some-
what similar context, “If we mathematicians abandon the goal of logical develop-
ment, who will replace us?” (p. 175). Now, as mentioned earlier, the strategy
mapped out in the NCTM Addenda Series with regard to making the students
learn about this aspect of mathematics is basically one involving the use of
reasonings and justifications, and the discussion of proofs is confined to a few
units in the latter part of the 4-year curriculum. Unfortunately, experience shows
that students do not learn correct deductive reasoning by being given small doses
of instructions in a few choice time slots. Rather, they learn it through long
exposure to good models of proofs, through repeated trials and errors, and above
all, through sheer hard work. In other words, as a method of making the students
learn about this most basic feature of mathematics, the strategy given in the
NCTM Addenda Series is likely to fail because its commitment to this cause is
inadequate both in terms of quantity and quality.
Perhaps one more comment about the emphasis of the NCTM Addenda Series
on intuition and informal reasoning would not be out of place. It is a fact that
mathematicians often resort to communicating by this kind of informal reasoning
or heuristic arguments, and that is general survey articles or in colloquium
lectures of the mathematical community, discussion of the intuitive ideas behind
the technicalities usually takes precedence over the technicalities themselves.
The reason for this is that either the parties involved are of such a level of
technical competence that the informal reasoning can be easily converted to a
precise proof, or an intuitive idea is all that the listener or reader needs for his or
her edification, or the details of the technicalities are available in print elsewhere.
In the context of high school mathematics education, however, we have the
diametrically opposite situation of a group of youngsters with almost no technical
competence (in terms of mathematical proofs) and with perhaps no access to any
technical detail beyond that supplied by the texts. In this light, has not the
proposed curriculum from NCTM lost sight of the fact that one of the primary
goals of such an education is precisely to instill this kind of technical compe-
tence? This emphasis on intuition and informality is therefore a move in the right
direction only if it is a supplement to, rather than a replacement of, the teaching
of correct mathematical reasoning; that is, proofs. One can argue with good
cause that such an education should be enlivened with tidbits of great intuitive
appeal or even (according to the prevailing fashions of the times) information of
great fun and relevance. Such an argument would be well taken by anyone of
normal sensibilities. It is quite a different matter, however, to attempt to do so by
taking away part of the core materials. The standard criticism of the traditional
curriculum is that it turns out students who memorize a lot of isolated facts
without knowing why they are true and what they are good for. Given the other
extreme of this current trend of devaluing proofs, what will likely happen is that
 HIGH SCHOOL GEOMETRY 227

a whole generation of students would memorize a lot of ideas without knowing

which of them are true or why they are true. Obviously, a curriculum reform
should aspire to better results than such a sorry scenario.
We now turn to the question of implementation. Let us immediately dispose of
the usual complaints about the standard presentation of proofs in Euclidean
geometry: (a) many of the theorems are too obvious and therefore too boring, and
(b) two-column proofs are too far removed from everyday mathematics. Because
both complaints are most likely the natural adverse reactions to the usual full-
year presentation of two-column proofs of elementary theorems about the trian-
gle and the circle, I will simply confirm the fact that nothing is more deadly than
devoting more than one semester to Euclidean geometry proper. However, before
going into the details of my proposed syllabus, I would like to make some
general comments on (a) and (b). First of all, the “boring” charge should be
looked at from two different angles. On the one hand, and I assume that the
reference here is to the initial theorems right after the axioms, the beginning of
everything does have the habit of being something of a bore. For example, do not
imagine for a second that football players are thrilled by what they have to do in
training camp each year, or baseball players in spring training. However, they do
it anyway, because the boring drills are absolutely essential for their later perfor-
mance. Therefore we should simply tell the students this and then get on with the
mathematics. Moreover, one can easily relieve this perceived boredom by a
selective presentation of some surprising or elegant theorems following the pre-
sentation of the basic results in Euclidean geometry. (For one such example, see
Appendix A.) As to the charge that two-column proofs are passe, I believe this is
the result of the abuse of two-column proofs in previous generations. The reason
for having two-column proofs is that it is the format that makes most clear to a
beginner what a mathematical proof really is: a connected sequence of assertions
each backed up by a reason. The existence of the second column forces the
beginner to recognize the need to put down a reason for each step. Far from being
an anachronism to be laughed at, it is an admirable educational tool. One should
think of the analogy with the training wheel for children who first learn how to
ride a bicycle: It is clearly useful, but it should be abandoned as soon as feasible.
Therefore a reasonable geometry text should begin with the two-column format
when proofs are first introduced, but should make a smooth transition to proofs
in the usual narrative style no later than (say) 4 weeks.5 Of course, using the two-
column format from beginning to end for a full year makes no sense at all.
The question may naturally arise that, if indeed the purpose is to educate
students about proofs, why not pick some topic other than Euclidean geometry?
Because the idea is to start proving theorems from the ground up, a little reflec-
tion would show that the only other topic available for this purpose would be the

%I Appendix A, one finds an example of how to handle proofs in this transitional period: There,
the same proof of a theorem is given in both the two-column format and the usual narrative format.
228 wu

real number system. Anyone who has ever gone through a development of the
real number system, starting from either the Peano axioms or the axioms of a
complete ordered field, would know that this alternative is fraught with perils.
By comparison with Euclidean geometry, it is equally dreary at the beginning
(see, e.g., Appendix B; notice that I have on purpose presented all the proofs in
the two-column format). Moreover, the discussion would soon be dominated by
continuity considerations and then they are definitely out of the reach of the 10th
and 1 lth graders. There is also an extramathematical reason for preferring geom-
etry as an introduction to proofs: In learning to prove something for the first time,
most people find it easier to look at a picture than to close their eyes and think
abstractly. In retrospect then, although the choice of Euclidean geometry as a
starting point for proofs may have been a historical accident, it is nevertheless a
felicitous accident.
The perception that proofs in geometry are too far removed from everyday
mathematics stems to a large extent from a glaring defect in the present-day
mathematics education in high school,6 namely, the fact that outside of geometry
there are essentially no proofs. Even as anomalies in education go, this is cer-
tainly more anomalous than others inasmuch as it presents a totally falsified
picture of mathematics itself. In my limited exposure to secondary school mathe-
matics texts, I have made the astounding observation that many algebra texts do
not give even a semblance of an explanation of the mathematical theorems, let
alone formal proofs. Against this irrational background, no wonder Euclidean
geometry stands out like a sore thumb. The remedy for a problem of such
magnitude cannot be addressed in such a short article. Two general remarks can
be made nevertheless. One is that the sharp subdivision of the traditional curricu-
lum into separate compartments called geometry, algebra, trigonometry, and so
on, should be blurred (although whether such a separation should be abolished
altogether at such an early stage of a student’s mathematical education is open to
a lot of questions on purely pedagogical grounds). At the moment, a simple
problem such as “find the triangle of largest area if its perimeter has a fixed
value” does not fall naturally into any of the preceding headings, and this should
signal some deep-rooted trouble in the present system. It is precisely this atmo-
sphere of compartmentalization that fosters the isolation of Euclidean geometry.
The other remark is more limited in scope and it pertains to the proposed syllabus
in the next paragraph: The suggested topics in the second semester of this year-
long syllabus call for proofs of theorems not in the area of classical Euclidean
geometry. There is room for the discussion of limits, linear transformations,
homomorphisms, quadratic forms, and so on. Needless to say, the depth of the
discussion would depend on the actual classroom situation. In a small way then,
what this syllabus offers is an ad hoc and partial remedy of an enormous failing in
mathematics education in the schools.

“Alas, the same could be said about the teaching of calculus in college at this moment
 HIGHSCHOOLGEOMETRY 229

After this discussion, it is time to set forth and discuss a proposed year-long
syllabus for a course in geometry:

r/a Semester of Experimental Geometry: Basic constructions with ruler and

compass without proofs (e.g., construction of regular polygons, construc-
tion of similar triangles of a given ratio, equal subdivision of a line
segment, tangents to circle from a point); direct verification of basic facts
(e.g., sum of angles of triangles, SAS, ASA, existence of centroid of a
triangle); tessellations of the plane; various manifestations of symmetry.
3/4 Semester of classical Euclidean geometry: Explanation of axiomatic sys-
tem; short exposition of first consequences of the axioms; basic properties
of the triangle, including the concurrence of the three medians, three
altitudes, three angle bisectors and three perpendicular bisectors; congru-
ence; similarity; Pythagoras’ theorem; circle and tangents; inscribed
angles and central angles; selected topics (e.g., 9-point circle, Simson
line, theorems of Ceva and Menelaus, etc., cf. Coxeter & Greitzer,
1967).
Y3 Semester of advanced topics: Curve length, area, and volume; coordi-
nates in two and three dimensions; vectors; transformations in the plane;
isometries; congruence and similarity revisited (cf. Lang & Morrow,
1988).
r/1 Semester of selected topics: For example, conic sections, projective
‘geometry, hyperbolic geometry (cf. Coxeter, 1992; Greenberg, 1980;
Land & Morrow, 1988).

Some of the seemingly cryptic statements require a few words of explanation.

I strongly believe that the students should devote some time at the beginning
getting acquainted with the geometric figures they will be studying and taking in
the geometric beauty nature has to offer. Part of the attraction of Euclidean
geometry is that almost all of its theorems can be pictorially confirmed; for
example, the fact that the three medians of a triangle must meet at a point (the
centroid). Another attractive feature is that many beautiful pictures can be easily
generated with ruler and compass alone; for example, try joining all the diago-
nals of a regular 24-gon. In addition, the obvious beauty of many planar tessella-
tions and geometric configurations resulting from the various manifestations of
symmetry can be easily exhibited. Therefore one should certainly capitalize on
the visual appeal of the subject. Ultimately, because this would represent the first
serious encounter with geometry for most students, they need to develop some
geometric intuition. This initial period allows them to do so through experimen-
tation in an environment free of worries about proofs. It suffices for this experi-
mentation to be done with ruler, compass, and protractor, but if computers
happen to be available, some (but certainly not most) of this experimentation can
230 wu

be carried out on the computer, too. 7 The allocation of ‘/4 of a semester of such
activities must not be regarded as rigidly fixed at the beginning of the year. This
experimentation should be carried out, to varying degrees, all through the year.
When one starts the classical portion of Euclidean geometry, a clear (if neces-
sarily brief) explanation of what an axiomatic system really is would be very
desirable; that is, starting with undefined terms and simple statements (axioms),
one deduces complex ones step by step with the use of logic. This is the best
system mankind has ever devised to ascertain the attainment of truth, and is the
basis of all scientific methods as well.8 The actual presentation of the axioms and
the initial theorems should in no way try to be complete; any complete presenta-
tion involving the betweenness and continuity axioms (among others) together
with the unending progression of obvious and tedious theorems would be peda-
gogically self-defeating. The usual shortcut version, one that makes use of the
real number system from the very beginning and is found in most traditional
texts, is in essence (if not necessarily in detail) adequate. A slightly different
version can also be found in Lang and Morrow (1988). The main point here is to
give the students a demonstration of the spirit of the deductive process as well as
a sample presentation of the way it is done. Once this viewpoint is accepted,
there should be enormous flexibility in the presentation of even the basic geomet-
ric facts. The proof given in Appendix A of the Simson line is a case in point. If
time allows, one would of course prove completely Statements A through D
quoted under “Tools needed for the proof.” If not, however, the students could be
told to accept these on faith for the time being, to read up on these proofs on their
own, or to prove these themselves, provided it is made clear to them that only a
lack of time (rather than any logical difficulties) is responsible for such omis-
sions.9 In most cases, short heuristic arguments given for the omitted materials
would more than suffice to convince the students. In this context, I should
mention in passing that similarity is of course defined with the aid of the real

‘Here is an example of abuse in our present day mad rush toward a rendezvous with the computer
in education. There is software that is supposed to show experimentally why the sum of the angles of
a triangle add up to 180”: Draw a triangle on the computer screen, press a key, and out come the
readings of three numbers representing the measurements of the three angles that, miraculously, add
up to 180”. This is supposed to make the beginners see the light. It seems to me first of all that it
would be much easier to convince the student of this fact by going through the usual simple argument
(draw a line from a vertex parallel to the opposite side and consider the corresponding angles and
alternate angles of parallel lines). Moreover, if a student does become convinced of this fact just by
observing the computer screen without once measuring the angles himself or herself with a protractor,
then 1 would be inclined to cite this as an example of a catastrophic failure in our education of the
young.
W may be mentioned in passing that the great philosopher Spinoza even wrote his treatise on
ethics exactly in this axiom-theorem-proof format. Although that is carrying the axiomatic method a
bit too far, the fact that he did is certainly suggestive.
‘It is interesting to note that the idea of presenting the proofs of selected theorems in this
abbreviated fashion is also endorsed by the NCTM (1992b, Chapter IO), where the method goes
under the name of “local axiomatics.”
 HIGH SCHOOL GEOMETRY zyxwvutsrqponmlkjihgfedcbaZYXWVU
231

number system, but in Euclid’s zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONM

Elements (and in the subsequent axiomatic sys-
tem of Hilbert, 197 1), similarity is defined purely geometrically . The possibility
of doing the latter was in the main due to Eudoxus (408-355 B.C.), and was in
fact mankind’s first breakthrough in understanding the real number system. One
might consider explaining some of this history if time is not a factor.
Two remarks about Euclidean geometry should be made at this point. One is
that even a casual glance through the preceding topics is sufficient to convince
the reader that they are truly basic. Simple facts such as the concurrence of the
three medians of a triangle or the orthogonality of a tangent to the radius of the
circle at the point of tangency, as well as those about congruence and similarity
are taken for granted in any normal discussion of mathematics. ‘0 It seems alto-
gether sensible that this foundational material should be carefully and systemat-
ically explained once and for all. It grieves me to see in some of the more recent
proposed mathematics curricula that this material is slighted or presented in a
haphazard manner in order to make room for more faddish topics related to the
computer. Given time, this madness will no doubt pass, as did all the other once-
fashionable items. A second comment is that people are usually under the misap-
prehension that what one finds in the high school geometry texts is all that there
is in Euclidean geometry. Quite the opposite is true. There was a renaissance of
the geometry of the triangle in the late 19th and early 20th centuries, and the
resulting Lemoine geometry and Brocard geometry contain many mathematical
gems. Anyone interested should look up N. Altshiller-Court’s (1952) College
Geometry in the library for an introduction.
The reason “curve length, area, and volume” are classified as advanced topics
is that these are concepts related to limits, and a careful treatment of the first two
is absolutely essential in making any sense of the number n. In a high school
course on geometry, one naturally does not want to enter into a full-blown
discussion of limits, but one can still steer a middle course and give a clear (even
if nonrigorous) discussion of the meaning of curve lengths and area. It is sad to
note that, among the existing textbooks, very often the discussion of these two
concepts is almost totally worthless. 1l Typically, if students coming out of such a
course were asked what the circumference of a circle of radius r is, the answer
would be “2rv- .” When being asked further what the number IT is, they would
either be silent, or mumble something about 3.14159 . . , or come up with the
statement “the area divided by r*.” In the case of the latter, when pressed for a
definition of the area inside the circle, they would then realize, perhaps for the
first time, that they really do not have the answers. This is not the students’ fault.
It is rather a massive failure of our mathematics education. This failure is all the
more surprising when one realizes how simple it is to teach all these things

“‘The oblique reference to a course that covers these topics as “an archives tour” (NCTM, 1992b,
p. v) probably had something else in mind.
I ‘It goes without saying that I have read only a small number of such texts, but I must again single
out the Coxford, Usiskin, and Hirschhom (1991) text as an exception.
 232 wu

correctly. A model of what is possible can be found in Lang and Morrow (1988);
this book also serves as a good reference for all the topics in this portion of the
syllabus. It goes without saying that every theorem in Lang and Morrow (1988)
that is provable at this level is given a proof. I urge the readers to consult this
book often because it has much to offer.
Finally, let me make a few brief comments about the various selected topics at
the end of this syllabus. It seems to me highly desirable that at the end of every
course, a little time should be reserved for the presentation of something that
makes good use of all that has gone before, and that is also unusually striking or
beautiful. In addition to being emotionally satisfying, such a presentation also
serves as a reward to the students for their effort in learning the materials. The
topics I have chosen, namely, conic sections, projective geometry, and hyper-
bolic geometry, would seem to meet these criteria. What I have in mind about
conic sections is to give a coherent presentation of why the following four defini-
tions of conic sections are equivalent: (a) plane sections of a cone, (b) graphs of a
quadratic polynomial in x and y, (c) loci of a point the ratio of whose distance
from a fixed point to that from a fixed line is constant, and (d) characterizations
in terms of the foci of a conic section. Such a treatment would explain to a
beginner the mysterious appellation of “conic section” for ellipses, parabolas,
and so on. This would also afford an opportunity to demonstrate on a very
elementary level why these curves are interesting and important (e.g., parabolic
lenses of headlights of automobiles, elliptic orbits of planets). Unfortunately, all
this information is spread over many books in bits and pieces at the moment and
no single text seems to go into these details anymore. This is one reason why
such a treatment would be a bonus to a high school student. The topic of
projective geometry is vast, but any student would surely succumb to the charms
of Desargues’ theorem, Pappus’ theorem and Pascal’s theorem. The Coxeter
( 1992) book is somewhat condensed, but its presentation is extremely elegant
and should be accessible with some effort. As for hyperbolic geometry, there is
no better illustration of the difference a single axiom can make in mathematics
than to prove that there are no rectangles (quadrilaterals with four 90” angles) in
hyperbolic geometry, and that two similar triangles are congruent. It follows
from the latter, for example that, once the angle is fixed, there is one and only
equilateral triangle (up to congruence) in hyperbolic geometry! If one is ambi-
tious, one may even try to derive the formula giving the length of a side of an
equilateral triangle in terms of its angle. The Greenberg (1980) book has all this
and more. Nowadays much discussion is given to making mathematics more
attractive and interesting to students. Is there a valid argument that such beautiful
facts about hyperbolic geometry are less interesting or less useful than computer
graphics’?‘* And think of the savings in the school budget!
The root of this whole discussion about the role of Euclidean geometry in the
high school mathematics curriculum is of course the recent drive for curriculum
reform. The origin of this article, on the other hand, lies in my perception that in
 HIGH SCHOOL GEOMETRY 233

this drive, the amount of attention given to pedagogy is not properly balanced by
the equally necessary attention one must give to the underlying mathematics. For
example, take the prevailing stand against proofs in high school mathematics in
general, and in geometry in particular (cf. 2 and 3 at the beginning of this
article); it is clearly grounded on a willful neglect of the nature of mathematics
itself. Mathematics without proofs is still a nice and useful subject for certain
purposes, just as opera without the human voice could be a relaxing and enjoy-
able art form to some people. ‘3 But in either case, the deletion alters the charac-
ter of the original to such an extent that one should really think of it as a creation
of a different genre. Therefore, if the proposed reforms were more balanced in
their focus on both the pedagogy and the mathematics, they would surely have
proceeded quite differently. For example, part of the effort could be devoted to
elevating the subjects other than geometry to a more rational level by incorporat-
ing more proofs than what has been done up to now. It is worth mentioning that
this is not an advocacy for transforming high school mathematics into a unending
sequence of definition-theorem-proofs; no one with any sense would want to
promote intellectual aridity. After all, there is plenty of room for steering a
middle course with a healthy mixture of applications, general discussions of
advanced topics without proofs, and precise theorems and proofs. For example,
we have already referred in an earlier discussion to the possibility of doing curve
length and area without being literally self-contained. Another example would be
to present the previous discussion of the Fermat-Pell equation a? - 1,141 y* = I
to 9th or 10th graders. In a similar vein, one can ask the students in a course of
geometry to compare the area and perimeter of a regular n-gon and to observe
what happens as n approaches infinity, and then explain the answer to them in the
context of the isoperimetric inequality (cf. the discussion of Problem I in Wu,
1994). Advanced topics such as the Fundamental Theorem of Algebra and the
Jordan Curve Theorem can also be made available to the students so long as they
understand that the proofs of these theorems must come later, and so on. In
various and different ways, the books of Lang and Morrow (1988) and Coxford,
Usiskin, and Hirschhorn (1991), wholly or partially, approximate this ideal of a
middle course remarkably well. l4 In any case, one would think that the real task
of a mathematical educator is to find this middle course rather than to impose an
untenable extreme position, such as the abolition of proofs, on the school curric-
ulum. Another example of the lack of attention to mathematics itself is the

i2These theorems about hyperbolic geometry are useful in the contemporary research on Riemann
surfaces and the topology of 3-manifolds.
“Ask Mantovanni, Carmen Dragon, Andre Kostelanetz, and others.
i4The Lang and Morrow book contains much more material than the Coxford, Usiskin, and
Hirschhom book, and is mathematically more advanced. It is my understanding that the latter book is
intended to be used by the best students in the 8th grade, and that most students would use it in the 9th
grade. Therefore a textbook intended for use in the 10th grade should be more along the line of Lang
and Morrow’s book.
234 wu

adverse reaction to the two-column format for proofs. This is nothing more than
the result of focusing only on the formalism without trying to understand the
mathematical motivation behind the formalism itself. The modest goal of this
article is to inject mathematics back into the pedagogical discussion. I have made
some recommendations about the teaching of geometry, and I hope I have been
successful in showing how these recommendations are driven by mathematical
considerations. So long as the goal of a mathematical education is to make the
students learn mathematics, it is my belief that mathematical considerations must
continue to play a vital role.

APPENDIX A: THE SIMSON LINE

Theorem: The feet of the three perpendiculars to the sides of a triangle from a
point of its circumcircle are collinear. This line is called the Simson line of the
point, named erroneously after Robert Simson (1687-1768); it was in fact dis-
covered by William Wallace (1768- 1837) in 1799.

Thus, referring to the figure: let P

be a point on the circumcircle of the
triangle ABC, and let L, M, N be the
points where perpendiculars from P to
side BC, CA, BA (or extensions there-
of) meet these sides, respectively. Then
the claim is that L, M, N always lie on a
straight line, the Simson line of the
point P.

Tools Needed for the Proof

A. The vertices of a quadrilateral lie on the same circle if and only if the sum
of a pair of opposite angles is 180”.
B. The vertices of a quadrilateral lie on the same circle if the angles sub-
tended by an edge at the two opposite vertices are equal.
C. If two inscribed angles on a circle subtend the same arc, then they are
equal.
D. Given angles LAOB and
LCOD with a common vertex
0. Suppose the angles are
equal, AOD is a straight line,
and B and C lie on opposite
sides of AOD. Then BOC is
also a straight line (see figure).
 HIGHSCHOOLGEOMETRY 235

Proof of Theorem (Referring to F’irstFigure on Preceding Page)

1. LAMP+ LANP=90”+90” 1. Hypothesis.

= 180”.
2. Quadrilateral AMPN lies on a 2. By A.
circle.
3. LAMN = LAPN. 3. By 2 and C.
4. LPMC = LPLC = 90”. 4. Hypothesis.
5. Quadrilateral PMLC lies on a 5. By B.
circle.
6. LLMC = LLPC. 6. By 5 and C.
7. LNAP + LAPN = 90”. 7. PNIAN.
8. LBCP + LLPC = 90”. 8. PL I BC.
9. But LNAP + LBAP = 180” = 9. By A and the fact that ABCP lies
LBCP + LBAP. on a circle.
10. Hence LNAP = LBCP. 10. By 9.
11. Hence LAPN = LLPC. 11. Combine 7, 8, and 10.
12. Hence LAMN = LLMC. 12. Combine 3, 6, and 11.
13. AC is a straight line. 13. Hypothesis.
14. Hence NML is a straight line. 14. By 12, 13, and D. Q.E.D.

Same Proof, in Narrative Style

Knowing that AMC is by hypothesis a straight line, we wish to use D to show that
NML must also be a straight line. To this end, we must prove: zyxwvutsrqponmlkjihgfedcb

LAMN = zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFE
LLM C. (1)

We first observe that LAMN = LAPN. This is so because, since PN 1 AN and

PM I AM, we have LAMP + LANP = 90” + 90” = 180”, so that by A, AMPN
lies on a circle. The assertion that LAMN = LAPN then follows from C. We
next observe that LLMC = LLPC. The reason for this is essentially the same:
because PL I LC and PM _L MC, B implies that PMLC lies on a circle, so that
by C again, we have LLMC = LLPC. As a consequence of these two observa-
tions, equation 1 would hold if we can prove instead:

LAPN = LLPC. (2)

Because PN I NA and PL I LC, we have LAPN + LNAP = 90” = LLPC +

LPCL. Thus Equation 2 is in turn a consequence of:

LNAP = LPCL. (3)

But LNAP + LPAB = 180” = LPAB + LBCP, the second equality being
implied by A and the fact that ABCP lies on a circle. Thus LNAP = LBCP,
which is just Equation 3. Q.E.D.
236 wu

APPENDIX B: FIRST CONSEQUENCES

OF THE FIELD AXIOMS

We give a small sample of the immediate deductions from the axioms of a field.
This is of course in the context of defining the real numbers as a complete
ordered field. Thus let zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIH
a, b, and c be arbitrary real numbers. Then they satisfy the
following:

Al. a + b = b + a.
A2. (a + 6) + c = a + (b + c).
A3. There is one and only one number, to be denoted by 0, so that for each
number a, u + 0 = a.
A4. For each number a, there is a number that will be denoted by -a, so that
u + (-a) = 0.
Ml. ab = ba.
M2. a(bc) = (ab)c.
M3. There is one and only one number, to be denoted by 1, which is not
equal to 0, such that for each number a, a . 1 = a.
M4. For each number a different from 0, there is one number, to be denoted
by (l/a), such that a( l/a) = 1.
D. u(b + c) = ub + UC. zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIH

Theorem 1. If a is a number, and if b is a number such that a + b = 0, then

b = -a.

Proof:

1. a + b = 0. 1. Hypothesis.
2. b + a = 0. 2. By Al and Step 1.
3. (b + a) + (-a) = 0 + (-a). 3. By Step 2.
4. b + (a + (- a)) = 0 + (-a). 4. By A2 and Step 3.
5. b + 0 = 0 + (-a). 5. By A4 and Step 4.
6. b + 0 = (-u) + 0. 6. By Al and Step 5.
7. b = -a. 7. By A3 and Step 6. Q.E.D.

Theorem 2. For any number a, - (- a) = a.

Proof:

1. a+(-a)=O. 1. By A4.
2. (-u) + a = 0. 2. By Al and Step 1.
3. a = -(-a). 3. By theorem 1. Q.E.D
 HIGH SCHOOL GEOMETRY 237

Theorem 3: For any number a, a . 0 = 0.

Proof:

1. 1 +o= 1. 1. By A3.
2. a(1 + 0) = a . 1. 2. By Step 1.
3.u~l+u~0=u*l. 3. By D.
4. a + a * 0 = a. 4. by M3 and Step 3.
5. a * 0 = 0. 5. By Step 4 and A3. Q.E.D.

Theorem 4: For any number a, (- 1)a = -a.

Proof:

1. 1 + (-1) = 0. 1. By A4.
2. a(1 + (-1)) = a . 0. 2. By Step 1.
3. a * 1 + a(-1) = a. 3. By D.
4. a + a(-1) = 0. 4. By M3, Theorem 3, and Step 3.
5. a + (-1)u = 0. 5. By Ml and Step 4.
6. (-1)~ = -a. 6. By Theorem 1 and Step 5.
Q.E.D.

REFERENCES
Altshiller-Court, Nathan (1952). College geometry (2nd ed.). New York: Barnes & Noble.
Coxeter, H.S. MacDonald (1992). The real projective plane. New York: Springer-Verlag.
Coxeter, H.S. MacDonald, & Greitzer, S.L. (1967). Geomerry revisited. New York: Random House.
Coxford, Arthur, Usiskin, Zalman, & Hirschhom, Daniel (1991). Geometry. New York: Random
House.
Faber, Richard, L. (1983). Foundations of Euclidean and Non-Euclidean geometry. New York:
Marcel Dekker.
National Council of Teachers of Mathematics (1989). Curriculum and evaluation standards for
school mathematics. Reston, VA: Author.
National Council of Teachers of Mathematics (1992b). Curriculum and evaluation standards for
school mathematics addenda series: Geometry from multiple perspectives. Reston, VA: Au-
thor.
National Council of Teachers of Mathematics (1992a). Curriculum and evaluation standards .for
school marhemotics addenda series: A core curriculum. Reston, VA: Author.
Greenberg, Marvin .I. (1980). Euclidean and non-Euclidean geometry (2nd ed.). San Francisco:
W.H. Freeman.
Hilbert, David (1971). Foundations of geometry (2nd ed.). Chicago: Open Court.
Lang, Serge, & Morrow, Gene (1988). Geometry (2nd ed.). New York: Springer-Verlag.
Redheffer, Raymond (1992). [Review of Handbook of differential equations by D. Zwillinger.]
Bulletin of American Mathematics Society, 26, 17 l- 175.
Stark, Harold M. (1978). An introduction to number theory. Cambridge, MA: MIT Press.
Wu, Hung-Hsi (1994). The role of open ended problems in mathematics education. Journal of
Mathematical Behavior, 13, 115-128.