Content uploaded by Peter Liljedahl
Author content
All content in this area was uploaded by Peter Liljedahl on Nov 11, 2015
Content may be subject to copyright.
COMPUTER MATH SNAPSHOTS
REFLECTIONS ON REFLECTIONS
P
ETER
L
ILJEDAHL
Simon Fraser University
E-mail: liljedahl@sfu.ca
This column will publish short (from just a few paragraphs to ten or
so pages), lively and intriguing computer-related mathematics vign-
ettes. These vignettes or snapshots should illustrate ways in which
computer environments have transformed the practice of mathe-
matics or mathematics pedagogy. They could also include puzzles or
brain-teasers involving the use of computers or computational the-
ory. Snapshots are subject to peer review.
This issue’s snapshot explores some generalizations of the defini-
tion of geometric reflection. Dynamic geometry tools can facilitate
generalizations such as those obtained by relaxing the requirement
that the reflection be through a straight line. The author compares the
families of curves obtained by reflecting thru circular arcs with the
curves generated in response to a physical problem proposed by
Wittgenstein. He suggests that the strategy of generalizing definitions
is a good avenue for bringing students quickly to the activity of doing
mathematics.
Computer Math Snapshots
Editor: Uri Wilensky
Center for Connected Learning and Computer-Based Modeling
Northwestern University, USA
E-mail: uri@northwestern.edu
INTRODUCTION
In this snapshot I use The Geometer’s Sketchpad (KCP Technologies
Inc., 2001) to explore some of the results of changing the definition of
reflection to accommodate reflection across curved lines, in particu-
lar, circular arcs. My motivation for doing such exploration is two-
International Journal of Computers for Mathematical Learning 9: 359–369, 2004.
Ó2004 Kluwer Academic Publishers. Printed in the Netherlands.
fold. First and foremost is my wish to present students with mathe-
matics that is ‘new’, both to the student and to the field. I use this
strategy to help dispel the myths of mathematics as a collection of
facts and skills long since established, and promote the view that
mathematics is an emerging field that involves both ‘learning’ and
‘doing’. My second reason for engaging in this particular exploration
comes from my desire to give students a richer set of experiences with
regards to such taken for granted mathematical concepts as reflec-
tion. In this case my strategy is based on Brown and Walter’s (1974)
method of using ‘‘what-if-not’’ to expand the context of established
mathematics. What I present here are the results of one specific
thread of these rich explorations; a thread that produced a set of
plane curves that are both aesthetically pleasing and hauntingly
familiar; and as such, the basis of further and deeper exploration.
The geometric principles behind the construction of the images are
not difficult. However, in order to keep the article free of clutter I
have taken a minimalist approach to the presentation of the images,
and wherever suitable, opting for aesthetic appeal as opposed to te-
dious details. For the same reasons the derivations of the parametric
equations that describe the curves I have created have likewise been
left out.
WHAT IF NOT?
I begin with the definition of reflection.
A reflection Ruses a fixed line l. Any point on lis its own image. Any other point P
is mapped onto a point P’ such that lis the perpendicular bisector of PP0. As the
name implies, a set of points and their images are reflections of each other; it is as if
line lwere a mirror (Smart, 1998).
Smart is capitalizing on the congruency of the mathematical reflec-
tion of an object across a line and the reflection of an object in plane
mirror. That is, the mathematical definition of reflection creates an
image, which I will refer to as the ‘reflected image’, in the same place
as the conventional understanding of reflection, which I will refer to
as the ‘mirror image’ (see Figure 1). Although these images are in the
same place they are generated in very different ways. The reflected
image (as indicated by the dotted line) is created, as defined above,
through a mathematical reflection across the line of reflection. The
mirror image (as indicated by the solid lines) is determined by the
PETER LILJEDAHL
360
congruency between the angle of incidence a(as measured from the
normal) and the angle of reflection b(also measured from the normal)
along with an extension along the reflected ray by a distance con-
gruent to the incident ray.
If the line is not straight, however, then the notion of a mirror
image still holds while the mathematical concept of a reflected image,
as defined above, no longer has any meaning. This can be fixed if we
simply adjust our definition of mathematical reflection to accom-
modate curved lines. Such a fix is most easily accomplished by
allowing the tangent of a curve to replace (or act as) the line of
reflection. However, this accommodation, although intuitively natu-
ral, destroys the general correspondence between the mirror image
and the reflected image (see Figure 2).
Figure 1.
Figure 2.
REFLECTIONS ON REFLECTIONS 361
Following either image leads to further interesting mathematics.
For example, exploring the mirror image could lead one to the weird
and wonderful world of anamorphic art; a form of art whereby
seemingly incomprehensible pictures are ‘decoded’ by viewing them
through curved mirrors, most often cylinders (Baltrusaitis, 1977). For
this article, however, I follow the path of the reflected image.
REFLECTION ACROSS A CIRCLE
If the line of reflection is an arc as it is in Figure 2 then the original
definition holds (with the understanding that the perpendicular
bisector refers to the tangent of the curve, and not the curve itself).
That is, it holds without the stipulation that ‘‘it is as if line lwere a
mirror’’. But why limit ourselves to an arc? Let’s consider the whole
circle. This will demand some further modifications to our definition.
There are two arcs of the circle that a point Pcan be reflected across
such that the tangent to the arc is the perpendicular bisector of PP0
(see Figure 3).
Although each of these possible images will eventually produce
nice results the particular curves I wish to present here comes from
choosing the closest arc as the line of reflection.
The reflection Ruses a fixed circle c1. Any point on c1 is its own image. Any other
point PðP6¼ OÞis mapped onto a point P0such that a tangent to c1 is the per-
pendicular bisector of PP0and PP0is the shortest segment possible.
These modifications, however, create an ambiguity for the point O.
The solution to this is to leave the mapping of point Oundefined.
O P
P
′
P
′
c1
Figure 3.
PETER LILJEDAHL
362
It should be noted here that this is different from the classic
method of reflecting across a circle called inversion (sometimes
referred to as reflection in a circle). In inversion a point Pis reflected
across a circle of radius rcentred at Oto a point P0such that
OPðÞOP0
ðÞ¼r2(Eaves, 1969).
Because of the basic properties that result from the definition of
reflection used there are two additional circles that are of great
interest (see Figure 4). If the circle of reflection is defined as c1 (a
circle with radius rcentred at O) then the other two circles of interest
are c2 (a circle of radius 2rcentred at O) and c3 (a circle of radius 3r
also centred at O).
In order to better facilitate a discussion of some of the results of
using a circle as a line of reflection I have included (below) a list of
some of the more basic and relevant properties that emerge from my
definition.
1. Any point on c1 is its own image.
2. Any point P, not on c1, will map to a point on a straight line
containing Oand P.
3. Any point on c2 will be mapped to O.
4. Any point on c3 will be mapped to a point on c1.
5. Any point inside c1 will be mapped to a point that is between c1
and c2.
6. Any point between c1 and c3 will be mapped to a point inside of c1.
7. Any point outside of c3 will be mapped to a point outside of c1.
Figure 4.
REFLECTIONS ON REFLECTIONS 363
8. Any point P0inside of c1 can be mapped to from two distinct
points P, one between c1 and c2 and one between c2 and c3.
9. Any point P0between c1 and c2 can be mapped to from two
distinct points P, one inside c1 and one outside c3.
REFLECTION OF A LINE ACROSS A CIRCLE
Although I used The Geometer’s Sketchpad to explore the curves
that resulted from performing the specified reflection on a straight line
the same transformation can be performed simply by attending to
the properties listed above. The software allowed me to define the
transformation of a single point and then move this point along the
straight line to produce the reflection. The software also allowed me to
move the position of the straight line in order to observe the results as
it cuts through the various circles c3, c2, and c1 (see Figures 5 and 8).
The self-crossing nature of the images in Figures 7 and 8 are as a
result of property 3. The straight line cuts c2 in two places, each of
which will map to O.
REFLECTION OF A CIRCLE ACROSS A CIRCLE
Similar approaches can be used to explore the results of reflecting a
circle across a circle. In this more complex case, however, I recom-
c2
c3
c1
a
b
b
′
a
′
Figure 5.
PETER LILJEDAHL
364
mend the use of the software.
1
In order to facilitate an easier dis-
cussion of this process the circle that is acting as the line of reflection
will be referred to as the reflective circle, the circle that is being
reflected will be referred to as the object circle, and the resulting curve
will be referred to as the image (see Figure 9).
For brevity, I introduce here some useful notation. If the reflective
circle is of radius 1 and centred at the origin there would be no further
loss of generality in defining the object circle as having a radius r
centred at (g, 0). Thus, the resulting image can be uniquely described
a
b
a
′
b
′
Figure 6.
a
b
a
′
b
′
Figure 7.
REFLECTIONS ON REFLECTIONS 365
by the parameters rand gof its object circle. The image in Figure 10
can, therefore, be referred to as g= 3.6, r= 1.2 or (3.6, 1.2).
What follows is a collection of curves, labelled with the aforemen-
tioned notation, that result from the reflection of a circle across a circle.
These curves represent only a small portion of the possible images
that can be produced. I left out the trivial results of reflecting an
object circle that is centred at O. The selection of these curves was
guided by my desire to present images produced by placing the centre
of the image circle in each of the four domains (inside c1, between c1
and c2, between c2 and c3, and outside of c3) and then modifying the
a
b
b
′
a
′
Figure 8.
Og
reflective circle object circle
image
Figure 9.
PETER LILJEDAHL
366
radius to create a variety of intersections with each of the circles c1,
c2, and c3.
Although it is onerous to generate these curves without the soft-
ware it is not too difficult, and perhaps an interesting exercise, to
analyse them using only the properties listed above. For example,
upon inspection it quickly becomes apparent that curve 12 is quite
different from all the others; there are three points where it crosses
itself. One of the intersections comes from property 3 (any point on c2
will be mapped to O). The other two intersections come from prop-
erty 9 (any point P0between c1 and c2 can be mapped to from two
distinct points P, one inside c1 and one outside c3).
DEEPER EXPLORATION
In seeing all the curves together I was struck by how aesthetically
pleasing they were, while at the same time intrigued by how haunt-
#1 (0.5, 0.4) #2 (0.5, 1.1) #3 (0.5, 1.8) #4 (0.5, 2.6)
#5 (1.5, 0.4) #6 (1.5, 1.1) #7 (1.5, 1.8) #8 (1.5, 2.8)
#9 (2.5, 0.4) #10 (2.5, 1.1) #11 (2.5, 1.8) #12 (2.5, 2.8)
#13 (3.5, 0.4) #14 (3.5, 1.1) #15 (3.5, 1.8) #16 (3.5, 2.8)
Figure 10.
REFLECTIONS ON REFLECTIONS 367
ingly familiar some of them seemed. In sharing these curves with Nick
Jackiw, he drew my attention to a set of curves that result from one of
Ludwig Wittgenstein’s thought experiments (Cuoco and Goldenberg,
1997; Wittgenstein, 1983). Wittgenstein posed the following problem:
A stick slides smoothly through a pivoting sleeve. As one end of the stick is moved
in a circle, what shape does the stick’s far end describe?
Using The Geometer’s Sketchpad
2
to explore these curves it became
clear that there were, indeed, similarities to some of the curves pre-
sented above (#9–11 and #13–16). To see this I have to generalize
Wittgenstein’s thought experiment to a circle of radius 1 centred at
the origin and a stick of length L pivoting at (p,0). Considering a
general point on the circle to be traced and the resulting image of that
point produces the following parametric equations:
xðtÞ¼ Lðpcos tÞ
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
1þp22pcos t
pþcos t;
yðtÞ¼ Lsin t
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
1þp22pcos t
pþsin t:
ð1Þ
Comparing these with the parametric equations I derived from my
circle images:
xðtÞ¼2ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
1rsin tðÞ
2
r2þg2þ2rg cos t
srcos tg;
pþcos1g
rtpcos1g
r
2ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
1rsin tðÞ
2
r2þg2þ2rg cos t
srcos tg;
pcos1g
r<t<pþcos1g
r;
yðtÞ¼ 2rsin t
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
r2þg2þ2rg cos t
prsin t:
ð2Þ
further similarities (now at a more analytic level) presented them-
selves. In fact, if from (2) r¼1, from (1) L= 2, and from (1) and
(2) p=gthen the equations will be identical. That is, Wittgenstein’s
PETER LILJEDAHL
368
thought experiment is a special case of reflecting a circle across a
circle. I was able to conclude that it is, in fact, possible to generate
any curve produced by Wittgenstein’s experiment through precise
selection of radius and placement of the object circle.
CONCLUSION
The detour I have taken here has very quickly, and very accessibly,
led to the creation of new mathematics. At the very least, such a
detour can only serve to enrich students’ experiences with such ‘taken
for granted’ concepts as reflection. On a bigger scale, such mathe-
matical encounters have the power to change students’ perceptions of
what mathematics is about – and perhaps even create some appeal for
them.
NOTES
1
For anyone interested in the process of doing this on The Geometer’s Sketchpad send
me an E-mail: liljedahl@sfu.ca and I will send the Sketchpad file.
2
See Key Curriculum Press Thought Experiment 192. Available on line at: http://
www.keypress.com/sketchpad/javasketchpad/gallery/pages/wittgenstein.php for a
working applet on this.
REFERENCES
Baltrusaitis, J. (1977). Anamorphic Art. New York: Harry N. Abrams.
Brown, S. and Walter, M. (1990). The Art of Problem Posing, 2nd edition. Hillsdale, NJ:
Lawrence Erlbaum Associates.
Cuoco, A. and Goldenberg, E. (1997). Dynamic geometry as a bridge from Euclidean
geometry to analysis. Geometry turned On: Dynamic Software in Learning, Teaching,
and Research (pp. 33–44). Washington, DC: The Mathematical Association of Amer-
ica.
Eaves, H. (1969). Fundamentals of Geometry. Boston, MA: Allyn and Bacon, Inc.
KCP Technologies Inc. (2001). The Geometer’s Sketchpad. KCP Technologies Inc.
Smart, J. (1998). Modern Geometries, 5th edition. New York: Brooks/Cole Publishing
Company.
Wittgenstein, L. (1983). Remarks on the Foundations of Mathematics. Cambridge, MA:
MIT Press.
REFLECTIONS ON REFLECTIONS 369
