An investigation of student understanding of single-slit diffraction and double-slit

interference
Bradley S. Ambrose, Peter S. Shaffer, Richard N. Steinberg, and Lillian C. McDermott

Citation: American Journal of Physics 67, 146 (1999); doi: 10.1119/1.19210

View online: http://dx.doi.org/10.1119/1.19210
View Table of Contents: http://scitation.aip.org/content/aapt/journal/ajp/67/2?ver=pdfcov
Published by the American Association of Physics Teachers

Articles you may be interested in

Quantitative Investigation of Thermal Expansion Using Single-Slit Diffraction
Phys. Teach. 44, 82 (2006); 10.1119/1.2165434

Photolithographic fabrication of diffraction and interference slit patterns for the undergraduate laboratory
Am. J. Phys. 72, 1328 (2004); 10.1119/1.1775240

Addressing student difficulties in applying a wave model to the interference and diffraction of light
Am. J. Phys. 67, S5 (1999); 10.1119/1.19083

Limitations of ray tracing techniques in optical modeling of silicon solar cells and photodiodes
J. Appl. Phys. 84, 5795 (1998); 10.1063/1.368844

Interference, diffraction, and refraction, via Dirac’s notation

Am. J. Phys. 65, 637 (1997); 10.1119/1.18613

This article is copyrighted as indicated in the article. Reuse of AAPT content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 94.4.77.73
On: Thu, 22 May 2014 20:19:53
 An investigation of student understanding of single-slit diffraction
and double-slit interference
Bradley S. Ambrose, Peter S. Shaffer, Richard N. Steinberg,a) and Lillian C. McDermott
Department of Physics, Box 351560, University of Washington, Seattle, Washington 98195-1560
~Received 9 February 1998; accepted 17 June 1998!
Results from an investigation of student understanding of physical optics indicate that university
students who have studied this topic at the introductory level and beyond often cannot account for
the pattern produced on a screen when light is incident on a single or double slit. Many do not know
whether to apply geometrical or physical optics to a given situation and may inappropriately
combine elements of both. Some specific difficulties that were identified for single and double slits
proved to be sufficiently serious to preclude students from acquiring even a qualitative
understanding of the wave model for light. In addition, we found that students in advanced courses
often had mistaken beliefs about photons, which they incorporated into their interpretation of the
wave model for matter. A major objective of this investigation was to build a research base for the
design of a curriculum to help students develop a functional understanding of introductory optics.
© 1999 American Association of Physics Teachers.

I. INTRODUCTION ences. In contrast, about half of the modern physics students

are physics majors at the sophomore level or above.
In this paper, we report on the ability of introductory and
more advanced students to select and apply an appropriate A. Instructional context
model to account for the pattern produced on a screen when
light is incident on a single or double slit. The motivation for During their first year of physics, students are introduced
this research was our observation that many students who to two different ways of thinking about the behavior of light.
had completed the introductory calculus-based course lacked It is generally assumed at the beginning of the course that the
a consistent conceptual framework for the nature and behav- class is already familiar with the idea that light from an ob-
ior of light. Even the strongest students, who could readily ject travels outward in all directions in straight lines. The
solve standard textbook problems, were frequently unable to students study reflection and refraction, draw ray diagrams,
relate the results to the models for geometrical and physical and solve numerical problems for a variety of simple optical
optics that they had been taught. There was evidence of se- systems. Later, in physical optics, the students learn that
rious difficulties with the wave model beyond the introduc- light is a transverse electromagnetic wave that propagates
tory course. In addition, we found that students in advanced through space. They are taught the concepts and formal rep-
courses often had mistaken beliefs about photons, which they resentations that are used to predict and explain diffraction,
incorporated into their interpretation of the wave model for interference, and polarization.
matter. In a typical modern physics course, an understanding of
The work described is part of a long-term investigation in the wave nature of light is assumed at the outset. When stu-
which the Physics Education Group at the University of dents learn about the photoelectric experiment, their atten-
Washington has been examining student understanding of tion is called to the inconsistency between the results and the
geometrical and physical optics.1–5 A major objective has predictions of physical optics. The discrepancy between
been to establish a research base that can be used as a re- theory and experiment is resolved by the introduction of the
source by physics instructors. Examples of the application of photon. The instructor explains that light appears to behave
this research can be found in curricula developed by our like a wave in certain situations and like a particle in others.
group.6,7 Discussion of the wave–particle duality leads to study of the
wave-like properties of matter and the foundations of quan-
tum mechanics.
II. OVERVIEW OF THE INVESTIGATION
Most of the investigation was conducted at the University B. Research methods
of Washington among students enrolled in one of two The same research methods were used in this investigation
courses: the introductory calculus-based course or a subse- as in others by our group.8 The data consist of interview
quent modern physics course. The study also included stu- transcripts and student responses to written questions.
dents in the algebra-based course, physics majors in a junior-
1. Initial interviews
level quantum mechanics course, and graduate teaching
assistants. Additional data were obtained from an introduc- An initial set of tasks provided the basis for a series of
tory calculus-based physics course at the University of Mary- individual demonstration interviews that were conducted
land. with 46 students from the introductory and modern physics
Although all prospective physics majors at the University courses. The 16 introductory and 30 modern physics students
of Washington enroll in the first-year calculus-based course, who participated were volunteers from several lecture sec-
they constitute less than 5% of the class. Most students in- tions in both courses. The interviews with the introductory
tend to major in engineering, mathematics, or in other sci- students took place during the last week of the quarter, after

146 Am. J. Phys. 67 ~2!, February 1999 © 1999 American Association of Physics Teachers 146
This article is copyrighted as indicated in the article. Reuse of AAPT content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 94.4.77.73
On: Thu, 22 May 2014 20:19:53
 Fig. 2. Diagrams that illustrate ~a! the width of the illuminated region on the
Fig. 1. Apparatus used in individual demonstration interviews. Students are screen for the case of light incident on a wide slit and ~b! the location of the
asked to predict what they would see on the screen: ~1! for the situation first diffraction minimum produced by light incident on a narrow slit.
shown, ~2! for the situation in which the bulb is moved farther and farther
from the mask, and ~3! for the situation in which the slit is made narrower
and narrower.
ing straight lines from the bulb through the slit to the screen.
@See Fig. 2~a!.# The size and shape of the illuminated region
can be determined by considering similar triangles or by us-
all instruction had taken place. In the modern physics course, ing trigonometry. As the bulb is moved farther and farther
the students had reviewed interference and diffraction of from the mask, the student needed to recognize that the rays
light and had covered similar topics in the context of matter through the slit become more nearly parallel. Eventually, the
waves. All earned grades at or above the mean in their re- geometric image approaches the same size as the slit.9
spective courses. The interviews, which lasted from 45 min In the third question, as the slit is made narrower, the
to 1 h, were videotaped and analyzed in detail. illuminated region becomes narrower. Eventually, geometri-
a. Interview tasks. At the beginning of each interview, cal optics is no longer valid and a single-slit diffraction pat-
the student was shown a small unlit bulb, a white paper tern with a bright central maximum and smaller subsidiary
screen, and a cardboard mask containing a slit 1 cm wide and maxima appears on the screen. In an introductory course,
3 cm tall. ~See Fig. 1.! The student was told to assume that Huygens’ principle is used to account for the pattern.10 Light
the room was darkened and that the bulb was lit. The main from the slit is treated as emanating from many coherent,
part of the interview was based on the following three ques- closely spaced, secondary point sources. At the angle u cor-
tions: responding to the first minimum, light from one edge of the
slit is 180° out of phase with light from the center. @See Fig.
~1! What would you see on the screen if the bulb, mask, and 2~b!.# The remaining secondary point sources also cancel in
screen were arranged as shown? pairs. From trigonometry, the condition for the first diffrac-
~2! How would your prediction change if the bulb were tion minimum is a sin u5l ~where a is the slit width!.
moved farther and farther from the mask? If the slit width is decreased further, the first minimum
~3! Assuming the bulb to be very far from the mask ~and moves away from the center of the pattern and the central
very bright!, how would your prediction change if the maximum becomes wider. Diffraction minima occur only
slit were made narrower and narrower? when the slit width is larger than the wavelength. If the slit
In addition to these three main questions, the investigator width is smaller, no minima appear since the waves from all
often posed others. Many of the students were asked about the secondary point sources along the width of the slit cannot
light incident on two slits. In some of the interviews, stu- all cancel at any point on the screen.
dents expressed their ideas about photons, either spontane-
ously or at the suggestion of the investigator. During the 2. Written questions and additional interview tasks
interviews with the modern physics students and subsequent
interviews with the quantum mechanics students, questions The results from the interviews guided the design of writ-
were posed about electron diffraction and interference. In all ten questions that were administered to large numbers of
cases, students were asked to explain their reasoning. students. Analysis of the interview transcripts and the written
The investigator tried to ensure, either tacitly or overtly, responses yielded information on the nature of specific diffi-
that certain simplifying assumptions would be made. If stu- culties, provided a rough measure of their prevalence, and
dents believed that the bulb would not be bright enough to suggested other questions and interview tasks. The written
produce a pattern on the screen, they were asked to suppose questions and additional interview tasks are discussed later
that the bulb was extremely bright. If they seemed to think of in the paper.
the bulb as an extended source, they were told to treat it as a
point source. If they recognized that the light would be com- III. IDENTIFICATION AND ANALYSIS OF
posed of many colors, they were told to imagine a red bulb. STUDENT DIFFICULTIES
The students would have been told to consider only Fraun-
hofer diffraction had any attempted to treat the narrow slit We have organized the student difficulties that we identi-
otherwise. fied into three broad, overlapping categories: misapplication
b. Correct responses to interview tasks. Correct re- of geometrical and physical optics, lack of a qualitative un-
sponses to the three main interview questions could be given derstanding of the wave model, and difficulties with modern
on the basis of material covered in the waves and optics physics concepts. ~Difficulties specific to geometrical optics
portion of the introductory course. A student can answer the are discussed in other papers.1,2! This classification provides
first two questions, which are based on a wide slit, by draw- a convenient structure for discussing a wide variety of stu-

147 Am. J. Phys., Vol. 67, No. 2, February 1999 Ambrose et al. 147
This article is copyrighted as indicated in the article. Reuse of AAPT content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 94.4.77.73
On: Thu, 22 May 2014 20:19:53
 dent errors and for drawing inferences about the nature of the
underlying conceptual and reasoning difficulties.
The errors that are used as illustrations are not intended to
be an exhaustive list. Rather, they are symptoms of the lack
of a coherent conceptual framework for optics. In order to
address this more general ~and far more serious! difficulty, it
is necessary to be aware of common mistaken beliefs that
can hinder the progress of students.
The particular interview and written questions presented in
this paper are representative of a broad range of tasks that we
have used to probe student understanding. The process of
identifying specific student difficulties includes a long se-
quence of interviews, pre- and post-testing, and extended
discussions with students. Questions are developed, modi- Fig. 3. Diagram drawn by a student who incorrectly treated the wide slit
fied, and tested over an extended period on the basis of ex- ~'1 cm! as a single secondary source of light waves. On the basis of her
diagram, she concluded that the illuminated portion of the screen would be
tensive research. In analyzing student responses, we try to larger than the geometric image.
ascertain what students are thinking, rather than interpreting
responses to isolated questions. In this paper, we report only
difficulties that are widespread, persistent, and independent 2. Use of geometrical optics for light through a narrow
of the instructor. The quotes have been selected because they
slit
are typical and not idiosyncratic.
Students also made the converse error to the one discussed
above. Many attempted to apply ideas from geometrical op-
A. Misapplication of geometrical and physical optics tics to account for the pattern produced by light incident on a
narrow slit. One example is provided by responses to a ques-
During the interviews many students failed to consider tion given to about 410 introductory students from several
whether geometrical or physical optics was valid in a given classes with different instructors. The students had com-
situation. They tried to apply one model when the other was pleted all standard lecture and laboratory instruction on
appropriate or to combine ideas from both in a ‘‘hybrid’’ single-slit diffraction. They had also taken an examination
model. that tested on this material. They were shown a single-slit
diffraction pattern with several maxima and minima and
1. Use of ideas from physical optics for light through a asked to predict how covering the right half of the slit ~i.e.,
wide slit narrowing the slit! would affect the distance between the first
minima. The results are given in part 1 of Table I.
Students often tried to apply ideas from physical optics to Mistaken belief that narrowing a slit produces a narrower
predict the appearance of the screen with a 1-cm-wide slit. central diffraction maximum. Despite having performed a
There were a variety of errors. Some students thought that laboratory experiment in which they observed the effect on
diffraction fringes or a fuzzy boundary around the geometric the diffraction pattern of narrowing the slit, about 20% of the
image would be readily visible. These students did not rec- students predicted a decrease in the distance to the first mini-
ognize that the slit was sufficiently wide that, to a very good
approximation, geometrical optics would be valid. Other stu-
dents had serious difficulties that indicated a profound lack Table I. Results from a two-part written question on single-slit diffraction
of understanding of either geometrical or physical optics. As given to introductory students and to physics graduate students. The students
illustrated by the student quotes in this paper, it was evident had completed all standard lecture and laboratory instruction on single-slit
that none of the difficulties arose due to subtleties of the diffraction and had been tested on this material. Percentages have been
apparatus ~e.g., edge diffraction or the type of light source!.11 rounded to the nearest 5%.
The example below is taken from the transcript of a modern
Students in Participants in
physics student. calculus-based graduate teaching
Mistaken belief that all slits, regardless of width or shape, course seminar
can be treated like a single secondary point source of light. (N'410) (N'95)
The student drew a set of curved wavefronts emanating from
the bulb and a second set from the slit. ~See Fig. 3.! She Part 1: Prediction for the effect of
covering half of the slit ~i.e.,
described the light as ‘‘spreading out’’ from the slit, and said
decreasing the slit width!
this ‘‘overcomes the problem of @light# passing through a on the location of the minima.
little hole.’’ In effect, she treated the entire slit as a single, Correct: minima farther from center 70% 90%
secondary point source of light. with correct reasoning 20% 60%
This student also predicted that the illuminated region on Incorrect minima closer to center 20% 5%
the screen would be approximately circular in shape, even if Incorrect minima at same locations 10% ,5%
the slit were rectangular: ‘‘I’m trying to think if it will hap-
Part 2: Comparison of the slit width, a,
pen the same no matter what the shape of the opening is... . It
to the wavelength, l, of the light.
shouldn’t matter what shape it is.’’ When asked how her Correct: a.l 40% 80%
prediction would change as the slit was made narrower or with correct reasoning 10% 55%
wider, she claimed that there would always be a broad, con- Incorrect: a,l 40% 15%
tinuous, bright region. She never predicted the appearance of Incorrect: a5l 10% 5%
maxima and minima.

148 Am. J. Phys., Vol. 67, No. 2, February 1999 Ambrose et al. 148
This article is copyrighted as indicated in the article. Reuse of AAPT content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 94.4.77.73
On: Thu, 22 May 2014 20:19:53
 that the maxima ~not the minima! of a diffraction pattern
occur at angles u. From his explanation, it was clear that the
student was not simply confusing the formulas for diffraction
and interference.
b. Failure to recognize the difference between superposi-
tion in geometrical and in physical optics. It was not only
in single-slit diffraction that we identified a tendency among
students to use a hybrid model. The failure to distinguish
Fig. 4. Sketch by a student who attributed the presence of ‘‘diffraction’’ between geometrical and physical optics also became obvi-
maxima and minima to the interference of light waves emanating only from ous when students tried to account for the interference pat-
the edges of the slit. tern for light incident on a double slit. In particular, many
students did not seem to understand that the superposition of
light must be treated differently in the two cases. This diffi-
mum. They seemed to be applying ideas from geometrical culty became apparent during the analysis of student re-
optics, in which narrowing a wide slit results in a narrower sponses to the following written question that was given to
geometric image. Analysis of the final course grades re- more than 200 introductory students after lecture instruction
vealed that the 20% did not consist only of students at the on interference.
bottom of the class. An equal number of students with grades The students were shown a photograph of the central por-
above and below the mean had this difficulty. tion of a double-slit interference pattern. ~The fringes were of
approximately equal intensity.! The students were asked to
3. Use of a hybrid model with elements of geometrical predict how the pattern would change if the left slit were
and physical optics covered. Only 40% of the students gave the correct answer
that the screen would be almost uniformly illuminated with
Many of the errors made in the interviews and on re- no diffraction minima visible. Almost half of the students
sponses to written questions suggested that students were not made errors based on the ideas described below.
only confused about whether geometrical or physical optics Mistaken belief that each slit alone produces the same
applied but also that they could not separate one model from pattern (only dimmer) as that produced by the pair of slits.
the other. They did not seem to be aware of critical differ- About 25% of the students thought that all the fringes would
ences and applied ideas from both indiscriminately. remain when one slit was closed. Most of these students
a. Inappropriate, simultaneous application of ideas from stated that covering one slit would only make the fringes
both geometrical and physical optics to account for diffrac- become dimmer. They seemed to think that each slit alone
tion. Students often used geometrical optics for light must produce the same pattern as the pair of slits. To account
through the center of a slit and physical optics for light at the for the double-slit pattern, they were effectively adding in-
edges. They drew straight lines to show the path of light tensities from the individual slits without regard to the phase
through the center of the slit and attributed diffraction to difference of the light from the two slits. They were trying to
some sort of interaction between the incident light and the apply ideas from geometrical optics to an interference phe-
edges. Instead of thinking of many point sources along the nomenon.
entire width of the slit, they often drew circular waves that Mistaken belief that each slit alone produces half the pat-
emanated only from the edges. tern produced by the pair of slits. Another 20% of the stu-
Mistaken belief that the central maximum in a diffraction dents treated the double-slit interference pattern as the result
pattern is the geometric image. A firm belief that the central of a juxtaposition of two patterns, each formed by one of the
maximum is due to light through the center of the slit was slits. These students seemed to believe that each slit was
expressed by several students during the interviews. One in- responsible for half of the interference pattern. Some claimed
troductory student stated that for a sufficiently narrow slit, that all the bright fringes on the left half of the pattern would
‘‘some of the light will go through without being diffracted disappear when the left slit was covered. However, only un-
and then the light that strikes the edges will be diffracted der the conditions of geometrical optics is it possible to as-
off.’’ As was the case with this student, the incorrect use of sociate a particular bright region with a particular slit.
geometrical optics for light through the center of the slit was
almost always accompanied by a misapplication of physical
B. Lack of a qualitative understanding of the wave model
optics.
Mistaken belief that a diffraction pattern is produced only The results from the interviews and written questions dis-
by incident light that strikes the slit edges. At least 25% of cussed above indicated that many students did not know
the students who participated in the interviews seemed to whether to use ideas from geometrical or physical optics in a
think that only the light incident on the slit edges contributed particular situation. Even when they recognized that the
to diffraction effects. Many of these students claimed that wave model was appropriate, however, students often did not
light ‘‘bounced’’ off the edges. One modern physics student understand how to apply it in accounting for diffraction or
referred to Huygens’ principle but drew semicircular wave- interference. Some common errors are described below.
fronts emanating from each slit edge. ~See Fig. 4.! He at-
1. Misuse of comparisons between slit size and
tempted to account for a diffraction pattern by attributing the
fringes only to interference of light from the edges of the slit. wavelength (or amplitude) to account for diffraction
Another modern physics student took this idea one step Students usually recognized that the width of the slit af-
further. He said that, since a single slit of width d has two fects the diffraction pattern. However, in many instances
edges that act as point sources, the diffraction pattern should they made inappropriate comparisons between the physical
look identical to that of two point sources a distance d apart. dimensions of the slit and the wavelength or amplitude of the
He wrote the double-slit equation d sin u5ml and decided light. ~Students usually referred to the wavelength, even

149 Am. J. Phys., Vol. 67, No. 2, February 1999 Ambrose et al. 149
This article is copyrighted as indicated in the article. Reuse of AAPT content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 94.4.77.73
On: Thu, 22 May 2014 20:19:53
 though they seemed to know that it is measured along the that the slit width had to be less than the wavelength of the
direction of propagation, which was perpendicular to the light. Many of these students gave explanations similar to
slit.! Many attributed diffraction to the failure of light to those given by the introductory students.
‘‘fit’’ through the slit. Mistaken belief that diffraction and polarization are di-
Mistaken belief that no light passes through a slit if rectly related. In several of the interviews, we included an
a,l and therefore the screen is totally dark. Some students additional task in which we asked students to predict how
who knew the appropriate equation for single-slit diffraction placing a polarizing filter between the small bulb and the slit
believed that light may or may not ‘‘fit’’ through the slit. In would affect what they would see on the screen. We found
the interviews, some students predicted that for the case in that after polarization had been introduced in lecture, stu-
which the slit width is less than the wavelength, no light dents often treated the slit as having the characteristics of a
would get through the slit at all. This claim was made not polarizer. In order to determine the prevalence of this diffi-
only by introductory students. The quote below was taken culty, we posed the following written question to about 400
from an interview with a modern physics student. students in three introductory classes after polarization had
been covered in class. Most of the students had also com-
S: @Pointing to the equation, ‘‘a sin u5l,’’ that he pleted the relevant laboratory experiment.
had previously written down.# If we wanted a to be The students were shown a diagram containing a light
shorter than this wavelength...then this sin u would source, a single narrow vertical slit, and the resulting diffrac-
have to be greater than one, which it can’t be. tion pattern. The pattern had several minima. The students
I: And what do you conclude from that? were told the light is unpolarized. They were asked how, if at
S: That no light can pass through a slit that’s smaller all, the appearance of the screen would change if a polarizer
than the wavelength of light. were placed in front of the slit with its transmission axis
oriented ~a! vertically and ~b! horizontally. Explanations
Mistaken belief that diffraction occurs only if a,l and were required.
that there are diffraction fringes in that case. In addition to The data were similar in all three classes. About 25% of
students who thought that no light would pass through a slit the students correctly predicted that for both orientations of
with width less than the wavelength, there were others who the polarizer the intensity would decrease, but the pattern
seemed to believe that diffraction occurs only if the slit would not otherwise change.12 Analysis of the other re-
width is less than the wavelength ~or amplitude!. sponses indicated that many students were treating the slit as
One example is provided by an interview with a modern if it were a polarizer. They predicted that the pattern remains
physics student who was asked to consider the situation in unchanged when the transmission axis is vertical ~parallel to
which a bright red bulb was placed very far from a single the slit! and becomes significantly dimmer or completely
slit. She correctly used ideas from geometrical optics for the dark when the transmission axis is horizontal. This error was
case of the 1-cm-wide slit but continued to use these same made by about 40% of the students and was the most com-
ideas as she considered narrower and narrower slits. Even mon incorrect answer. The following quote illustrates this
when the investigator asked what would be seen on the type of student response.
screen if the wavelength were just 1 nm larger than the slit ‘‘There would be no pattern because @when the trans-
width, the student used geometrical optics to predict that the mission axis is horizontal, the#...polarizer would block
geometric image of the slit would appear on the screen. one plane of light and the slit would block the other.’’
When asked, however, what she would see on the screen for
a slit width 1 nm smaller than the wavelength, she stated that The second most common error seemed to be based on the
the light ‘‘has to bend in order to fit through’’ the slit. She belief that the relative orientation of the slit and transmission
predicted that a diffraction pattern would appear because axis of the polarizer determines whether physical or geo-
metrical optics applies. About 10% of the students said that
‘‘it’s hard to put something that’s big through something
when the transmission axis is parallel to the slit, there is no
that’s smaller than it is.’’
diffraction. Instead, the geometric image of the slit appears
This type of reasoning was common among introductory
on the screen. Typical student explanations were based on
students as was shown by a written question administered to the incorrect idea that ‘‘the vertical waves passing through
about 410 students. ~These are the same students whose per- the slit do not get diffracted.’’ One student, who seemed to
formance on another question was discussed earlier. They attribute diffraction effects to only the edges of the slit, pre-
had completed traditional lecture and laboratory instruction dicted that the light would not ‘‘spread out’’ because ‘‘there
on single-slit diffraction and had been tested on this mate- would be no interaction with the sides of the slit.’’
rial.! The students were shown a diffraction pattern with sev-
eral minima. They were asked to compare the wavelength to
the slit width. 2. Failure to recognize the critical role of the path length
As can be seen in part 2 of Table I, only 40% correctly (or phase) difference
stated that the slit width would be larger than the wave- During the interviews and on responses to the written
length. About 40% said that it would be smaller than the questions, many students did not seem to recognize the criti-
wavelength. An additional 10% claimed that the slit width cal role of the difference in path length ~or phase! in deter-
was equal to the wavelength. Incorrect explanations included mining the locations of the maxima and minima. Analysis of
‘‘slit width is less than l or else there would not be so much student responses to two written questions provides evidence
diffraction’’ and ‘‘less than l because the light didn’t travel for this generalization. The first, which was administered to
in a straight path.’’ introductory and graduate students, asks about the interfer-
The above question has also been posed to about 95 phys- ence pattern produced by two small objects vibrating up and
ics graduate students who were teaching assistants in the down in unison in a large water tank. The second question,
introductory course. Approximately 15% incorrectly stated which was administered to students in the modern physics

150 Am. J. Phys., Vol. 67, No. 2, February 1999 Ambrose et al. 150
This article is copyrighted as indicated in the article. Reuse of AAPT content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 94.4.77.73
On: Thu, 22 May 2014 20:19:53
 Fig. 5. Top-view diagram of a water
tank with two sources of circular
waves. For each of the three labeled
points ~A, B, and C!, students are asked
to determine whether there is maxi-
mum constructive interference, com-
plete destructive interference, or nei-
ther. They are also asked to determine
the phase difference at each point.

course, asks about the effect on a double-slit interference claimed that some destructive interference occurs since the
pattern of a change in the slit separation. In this case, the waves are not moving in exactly the same direction. They
pattern was produced by electrons. For both written ques- did not recognize that the superposition of waves involves
tions, however, the reasoning required is the same as for phase difference and amplitude, not direction of propagation.
light waves. The first question is described immediately be- Mistaken belief that the path length difference becomes
low and the second later in this paper. negligible at large distances. Some students seemed to be-
On the first of the two questions, the students were shown lieve that the difference in path length ~or phase! is negli-
a top view of a water tank in which two sources of circular gible at all points far from the sources of the waves. For
waves are separated by 2.5l. Three points ~A, B, and C! are example, one student stated that point C would experience
labeled. ~See Fig. 5.! For each point, students are asked to constructive interference because the distance ‘‘from source
determine whether there is maximum constructive interfer- to C is much greater than @the distance# between the
ence, complete destructive interference, or neither. They are sources.’’ Another student gave the same answer and ex-
also asked to determine the phase difference at each point. plained, ‘‘I suppose that 2.5l @the slit separation# is small
Students needed to recognize that at point A the path compared to 400l and 300l, so the sources here act like a
length difference is zero, and at point B it is equal to the single source.’’
source separation ~2.5l!. Thus there is maximum construc-
tive interference at point A and complete destructive interfer-
ence at point B. ~Small differences in amplitude at points far
from the sources are neglected.! For point C, students could 3. Memorization of algebraic formulas without an
use the Pythagorean theorem to calculate the path length understanding of the derivations
from each source and then find the difference. Alternatively,
they could use the approximation that the path length differ- In addition to the investigation that took place at the Uni-
ence, DD, is approximately equal to d sin u, where d is the versity of Washington, we had the opportunity to collect
source separation. The angle, u, for point C is given by some data from students in the introductory calculus-based
sin u'300l/500l50.6. Thus there would be destructive in- course at the University of Maryland. This is one of the test
terference at point C, since DD is equal to an odd half- sites for Tutorials in Introductory Physics, the curriculum
integer multiple of l: DD'(2.5l)(0.6)51.5l. that our group is developing to supplement the lectures and
This written question has been given to about 1200 intro- textbook of the standard introductory course. The structure
ductory students. The results do not seem to depend on of the waves and optics portion of the calculus-based physics
whether or not the students have had lectures on interference. course is similar at the two universities. At both, students
About 35% have given correct answers for points A and B. attend three lectures each week, one small-group 50-min ses-
Only about 10% have attempted to find the difference in path sion, and an associated laboratory course. At the University
length from the sources to point C. The question has also of Washington, all the small-group sessions are tutorials that
been given to about 95 physics graduate students in their first focus on the development of concepts and qualitative reason-
year as teaching assistants. Although more than 80% gave ing. Very little time is spent on numerical problems. For
correct answers for points A and B, their success rate for some classes at the University of Maryland, the small-group
point C was only 55%. Most incorrect responses revealed sessions are tutorials. For others, they are more traditional
that, like the introductory students, they did not try to find recitations, in which the emphasis is on problem solving.
the path length difference for point C. It has been possible to give identical examination ques-
Mistaken belief that the interference of waves at a point is tions in the tutorial and recitation sessions at the University
determined by the path length or the direction from each of Maryland. On qualitative questions similar to those that
source. Students sometimes used path length ~instead of dif- have been described, the students at the University of Mary-
ference in path length! to decide whether the waves from the land who were enrolled in classes with tutorials did signifi-
two sources would interfere constructively or destructively. cantly better than those in classes with traditional lecture
For example, some reasoned that since point B is 300l from sections. We have also been able to compare the perfor-
the sources ~a whole number of wavelengths!, the waves mance of students in the two groups on a quantitative exami-
would interfere constructively. Other students based their nation problem based on material typically covered in the
reasoning on the relative directions of propagation of the lectures, homework, and recitation of an introductory course.
waves from the two sources. Some stated that at point B the On the examination question, the students are told that
waves are moving in the same direction and therefore inter- light of wavelength l5500 nm is incident on a mask con-
fere constructively. Similarly, at points A and C, students taining two slits separated by a distance d530 m m. The dis-

151 Am. J. Phys., Vol. 67, No. 2, February 1999 Ambrose et al. 151
This article is copyrighted as indicated in the article. Reuse of AAPT content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 94.4.77.73
On: Thu, 22 May 2014 20:19:53
 tance between the center of the screen and the first dark
fringe is x51.5 cm. The students are asked to determine the
distance L between the screen and the mask.
To determine the answer, students must recognize that the
first minimum corresponds to a difference in path length
from the two slits, DD, that is equal to l/2. If the screen is
far from the mask, the path length difference is approxi-
mately d sin u or d tan u, where u is the angle measured
from the center of the slits. From geometry, tan u is equal to
x/L. The correct answer, L51.8 m, can be found by solving
the equation l/25d tan u5dx/L.
This problem was administered to 165 students in the tra-
ditional lecture-recitation-laboratory course. Only 15% gave
the correct response. The most common incorrect answer,
L50.9 m, was given by approximately 40% of the students.
Most of these students used the formula (d sin u5ml) that Fig. 6. Diagram drawn by a modern physics student who believed that
they had memorized for the location of an interference maxi- photons travel on straight paths that ‘‘bend’’ near the edges of a slit. She
mum and simply substituted numerical values. In contrast, used the same diagram to predict the appearance of the screen for both a
wide and a narrow slit. In both cases, she predicted a single, broad illumi-
60% of the approximately 115 tutorial students gave correct
nated region wider than the geometric image. Regardless of the width of the
responses. Only 10% of these students gave the incorrect slit, she never predicted that there would be diffraction maxima and minima.
answer that L50.9 m. The tutorial students were much more
likely than the recitation students to consider difference in
path length and to use superposition.
students who had progressed beyond introductory physics. In
most instances, the naive beliefs that they expressed about
C. Difficulties with modern physics concepts the properties of photons could not possibly account for dif-
The primary purpose of the initial interviews with the in- fraction and interference effects. Below are a few examples.
troductory and modern physics students was to elucidate the Mistaken belief that photons move along straight paths
ideas that they had about the wave nature of light. However, that ‘‘bend’’ near slit edges. Many students tried to account
as they attempted to justify their predictions for light incident for the behavior of light by treating photons as point particles
on a single or double slit, 16 of the 30 modern physics stu- that travel along paths that ‘‘bend’’ near the edges of a slit.
dents referred to photons. We used the opportunity to ex- For example, one student stated that the paths of the photons
plore how these students related their ideas about photons to ‘‘would bend around an obstruction so that the light rays...
the behavior of light. In 11 of the interviews, we also had the can spread out.’’ She applied this idea to light passing
opportunity to probe the ideas that the students had about the through the 1-cm slit. She drew a top-view diagram like the
wave properties of electrons. one shown in Fig. 6 and concluded that the illuminated re-
Later in the investigation, we conducted a set of 14 inter- gion on the screen would be considerably wider than would
views with physics majors from a junior-level quantum me- be predicted by geometrical optics. Moreover, she drew on
chanics course. Like the modern physics students in the ini- the same belief for the case of a narrow slit and never pre-
tial interviews, almost all of these students earned final dicted that the pattern on the screen would contain bright and
grades above the mean for their class. In addition to tasks dark fringes. Thus the idea that the path of a photon
that were similar to those posed during the initial interviews, ‘‘bends’’ at the slit edges led this student to incorrect predic-
the quantum mechanics students were asked questions to tions for situations in both geometrical and physical optics.
elicit their beliefs about the de Broglie wavelength of elec- Mistaken belief that photons move along sinusoidal paths.
trons. The results guided the design of written questions that Almost half of the modern physics students who referred to
were posed on examinations in the modern physics and photons ~7 out of 16! tried to account for wave phenomena
quantum mechanics courses. by thinking of photons as point particles that travel along
sinusoidal curves.13 For example, one student drew a dia-
1. Light as photons
gram like the one shown in Fig. 7~a! to show light passing
Twelve of the 16 modern physics students and 6 of the 14 through a narrow slit. He drew a sinusoidal curve with sev-
quantum mechanics students who referred to photons in their eral dots along the curve, stating that ‘‘each point on the
explanations articulated ideas that were unanticipated from wave is a little particle,’’ which he later called a ‘‘photon.’’

Fig. 7. Representative diagrams drawn by various mod-

ern physics students who believed that photons travel
along sinusoidal paths. One student drew the diagram in
~a! while trying to account for diffraction. The student
referred to each dot as a ‘‘photon.’’ Another student
drew the diagram in ~b! while trying to account for
polarization.

152 Am. J. Phys., Vol. 67, No. 2, February 1999 Ambrose et al. 152
This article is copyrighted as indicated in the article. Reuse of AAPT content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 94.4.77.73
On: Thu, 22 May 2014 20:19:53
 He stated that ‘‘some of these particles will crash into this tween the wavelength of light and the de Broglie wave-
wall and the other ones will travel through.’’ Thus, ‘‘part of lengths of particles that have mass, such as the electron.
the amplitude is cut off.’’ He initially thought that his dia- a. Extension of incorrect ideas about photons to elect-
gram illustrated diffraction but later realized that he could rons. During the initial interviews, 5 of the 11 modern phys-
not use it to account for the bright and dark fringes. ics students indicated that they thought that electrons move
Another modern physics student articulated a similar idea along straight paths that ‘‘bend’’ at the slit edges, that elec-
when he stated that a sinusoidal curve represents ‘‘the path trons move along sinusoidal paths, or that two or more elec-
that a particle @of light# would take.’’ To illustrate, he drew a trons are necessary to produce interference effects. We also
top-view diagram of a narrow vertical slit with a sinusoidal posed similar questions to 11 quantum mechanics students in
curve incident on the mask. @See Fig. 7~b!.# He said that the interviews, and 3 of those 11 students made similar er-
‘‘this particle would run into the wall..., so it wouldn’t go rors.
through.’’ He then drew a curve with a smaller amplitude to b. Failure to recognize that the de Broglie wavelength is
represent a wave that passes through the slit. He summarized not an inherent property of an electron (or other particle)
his thoughts by saying ‘‘light that’s oriented more vertically but varies with momentum. In many of the interviews with
is going to pass through,’’ and that ‘‘when your slit kind of quantum mechanics students, there was sufficient time to
gets too narrow...it’s going to block off the photons from probe their understanding of the de Broglie wavelength of
passing through like a polarizer would.’’ electrons and other particles. The results prompted us to de-
Mistaken belief that two or more photons are required for sign written questions that have been given on final course
diffraction or interference minima to occur. Those modern examinations to about 80 modern physics students and about
physics students who predicted that a sufficiently narrow slit 40 quantum mechanics students. The first of these two ques-
would produce a pattern with maxima and minima were also tions is similar to that given in the interviews. Students were
asked how ~if at all! the pattern would differ if the intensity asked to predict how a diffraction ~or interference! pattern
of the incident light were decreased. Several students pre- produced by electrons would change if the velocity ~or ki-
dicted that no interference would occur if the light were suf- netic energy! of the electrons were increased. In the second
ficiently dim. When asked what would appear on a photo- question, students were asked what would happen to the pat-
graphic plate in a single-slit diffraction experiment with tern if the experiment involved particles with greater mass
extremely low intensity light, one student responded as fol- ~e.g., neutrons! but the same kinetic energy as the electrons.
lows: To arrive at a correct answer, students could use the in-
S: If @the intensity# was really really low,... you verse relationship between the de Broglie wavelength and
would just see random distribution of light coming the momentum of the particle. In both cases, the wavelength
in, just like random spots. ... They’re not going to would decrease. Therefore, the interference fringes would
be interfering with each other anymore, because move closer together. For each of the questions, about 75%
there’s so few of them. ... of the students in each class failed to give a correct answer
I: What if we let that experiment go on for a month? with a correct explanation.
What would you think you’d see on the film? On both the examination questions and during the inter-
S: Pretty much lots and lots of dots. ... I guess the views, there were a few students who expressed uncertainty
whole plate would be lit. about the relationship between the de Broglie wavelength
Many of the students were also asked essentially the same and the velocity of an electron. One student articulated his
question for the case of a double slit. Similar difficulties confusion during an interview in the following way: ‘‘I can’t
emerged. For example, while pointing to a minimum in a think if there is a correlation between velocity and the actual
photograph of a double-slit interference pattern, a quantum wavelength of the electron, or if its wavelength is a property
mechanics student explained that if the photograph were re- of the electron itself, regardless of velocity.’’
placed by a photographic plate: Analysis of the remaining written responses revealed two
common types of difficulties that were similar to ideas ex-
‘‘Every time a photon would hit the plate, it would
pressed by students during the interviews. Most of the stu-
make a little spot... . Photons would still hit here, but
dents either failed to recognize that the de Broglie wave-
since only one could pass through, there wouldn’t be
length depends on the electron velocity or applied certain
another one that could destructively interfere with it.
equations for the wavelength of light in a way that was in-
This part @the minimum# would also appear to be
valid for electrons. The students who made both of these
light.’’
errors were evenly distributed above and below the final
The difficulties illustrated above only arose when students mean grades for their classes.4
were asked about diffraction or interference experiments Mistaken belief that diffraction and interference effects
with low intensity light. Many of these students failed to are independent of velocity. On the first written question,
recognize that each photon can be treated as a wave incident about 40% of both the modern physics and quantum mechan-
on the slit, and thus the predictions of physical optics remain ics students predicted that the pattern would not change.
valid for extremely low intensities.14,15 Many stated that a change in electron speed would affect
only the overall intensity of the pattern and not the fringe
locations. One student justified his incorrect answer by stat-
2. Electrons as waves
ing that the ‘‘electrons are just moving faster and l is the
The incorrect beliefs described above were not restricted same, so screen would not change.’’
to photons. Many of the errors made by the modern physics Mistaken belief that all equations that apply to the wave-
and quantum mechanics students revealed that they had the length of light apply to the de Broglie wavelength of elec-
same mistaken beliefs about electrons as they did about pho- trons. About 20% of the students in each class wrote down
tons. They also failed to recognize some key differences be- formulas that are valid for photons but not for electrons.

153 Am. J. Phys., Vol. 67, No. 2, February 1999 Ambrose et al. 153
This article is copyrighted as indicated in the article. Reuse of AAPT content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 94.4.77.73
On: Thu, 22 May 2014 20:19:53
 Some modern physics students incorrectly concluded that the their assistance. Also deeply appreciated has been the coop-
de Broglie wavelength is directly proportional to the electron eration of the instructors at the University of Washington and
velocity by misinterpreting the equation v 5l n . Others, in the University of Maryland in whose classes some of the
both the modern physics and quantum mechanics classes, research questions were asked. In addition, the authors grate-
used the expression for photon energy, E5hc/l, to claim fully acknowledge the ongoing support of the National Sci-
that particles with different mass but with the same kinetic ence Foundation through Grants Nos. DUE 9354501 and
energy have the same de Broglie wavelength. Lacking a DUE 9727648 to the University of Washington and Grant
functional understanding of the wave properties of matter, No. DUE 9652877 to the University of Maryland.
the students tended to focus on memorized formulas and not
a!
on the physics. Present address: Department of Physics, University of Maryland, College
Park, MD 20742-4111.
1
F. M. Goldberg and L. C. McDermott, ‘‘An investigation of student un-
IV. CONCLUSION derstanding of the real image formed by a converging lens or concave
mirror,’’ Am. J. Phys. 55, 108–119 ~1987!; ‘‘Student difficulties in under-
In a standard introductory course, students are presented standing image formation by a plane mirror,’’ Phys. Teach. 11, 472–480
with at least two different conceptual models for the nature ~1986!.
and behavior of light. We found that many students in intro- 2
K. Wosilait, P. R. L. Heron, P. S. Shaffer, and L. C. McDermott, ‘‘Devel-
ductory and more advanced courses do not understand the opment and assessment of a research-based tutorial on light and shadow,’’
basic features of these models nor do they recognize the Am. J. Phys. ~to be published!.
3
conditions under which each should be applied. Even the K. Wosilait, ‘‘Research as a guide for the development of tutorials to
improve student understanding of geometrical and physical optics,’’ Ph.D.
strongest students in the introductory course had serious con-
dissertation, Department of Physics, University of Washington, 1996 ~un-
ceptual and reasoning difficulties. Some of these were preva- published!.
lent in a sophomore-level modern physics course and per- 4
B. S. Ambrose, ‘‘Investigation of student understanding of the wave-like
sisted even among students taking a junior-level course in properties of light and matter,’’ Ph.D. dissertation, Department of Physics,
quantum mechanics. University of Washington, 1998 ~unpublished!.
5
Some of the difficulties that we identified were so severe Other papers that report on our research and curriculum development in
that they had precluded students from attaining the type of physical optics include: K. Wosilait, P. R. L. Heron, P. S. Shaffer, and L.
C. McDermott, ‘‘Addressing student difficulties in applying a wave model
qualitative understanding that should be a minimum outcome to the interference and diffraction of light’’ ~submitted for publication!; B.
of an introductory course. It is impossible to construct a co- S. Ambrose, P. R. L. Heron, S. Vokos, and L. C. McDermott, ‘‘Student
herent model for the treatment of light as a wave without understanding of common representations of light as an electromagnetic
being able both to distinguish and to relate certain basic wave’’ ~submitted for publication!.
ideas ~e.g., wavelength, path length, path length difference, 6
L. C. McDermott, P. S. Shaffer, and the Physics Education Group at the
and phase difference!. Moreover, since many of the critical University of Washington, Tutorials in Introductory Physics, Preliminary
features of the wave model are based on inferences and not Edition ~Prentice Hall, Upper Saddle River, NJ, 1998!. The research re-
ported in this paper has guided the development and assessment of several
on direct observations, students need to be able to interpret tutorials on optics.
various abstract representations. Some difficulties with the 7
Research by our group on geometrical optics has also guided the develop-
formal representations of light as an electromagnetic wave ment of a laboratory-based curriculum intended for the preparation of
are described in another paper, which also includes the dis- precollege teachers, for underprepared students, and for liberal arts majors.
cussion of a tutorial expressly designed to address these See L. C. McDermott and the Physics Education Group at the University
difficulties.16 of Washington, Physics by Inquiry ~Wiley, New York, 1996!, Vols. I and
The results of this study have clear implications not only II.
8
For descriptions of how the Physics Education Group conducts research on
for the teaching of optics in the introductory course but also student understanding of physics, see Refs. 1–5, 13, and 17. Also, see D.
for reform efforts directed toward introducing topics from E. Trowbridge and L. C. McDermott, ‘‘Investigation of student under-
modern physics into that course. We found that after tradi- standing of the concept of velocity in one dimension,’’ Am. J. Phys. 48,
tional instruction, many students could not apply the wave 1020–1028 ~1980!; ‘‘Investigation of student understanding of the concept
model to account for the diffraction or interference of light of acceleration in one dimension,’’ ibid. 49, 242–253 ~1981!; L. C. Mc-
incident on a single or double slit. The ability to do so, Dermott and P. S. Shaffer, ‘‘Research as a guide to curriculum develop-
however, would seem to be a necessary condition for under- ment: An example from introductory electricity. Part I. Investigation of
student understanding,’’ ibid. 60, 994–1003 ~1992!; Erratum to Part I, 61,
standing the wave nature of matter. 81 ~1993!.
In this and other instances, we have found that advanced 9
The term geometric image refers to the bright region on a screen that
study in physics does not necessarily overcome serious dif- would be produced by the rectilinear propagation of light from a source
ficulties with basic material.17 Unless these are explicitly ad- through an aperture to the screen. For a discussion of the differences
dressed in the introductory course, they are likely to persist between this type of image and the real image formed by a converging
and preclude the learning of more advanced topics. There is lens, see F. Goldberg, S. Bendall, and I. Galili, ‘‘Lenses, pinholes, screens,
and the eye,’’ Phys. Teach. 29, 221–224 ~1991!.
a need for instruction that addresses the conceptual and rea- 10
In an introductory course, interactions between the electromagnetic waves
soning difficulties that students have with the wave model and the material of which the slit edges are made are not considered. A
for light. There is evidence that the tutorials that our group is more rigorous approach can be found in S. G. Lipson, H. Lipson, and D. S.
developing help to fulfill this need.3,4 Tannhauser, Optical Physics ~Cambridge U.P., Cambridge, UK, 1995!.
11
Edge diffraction is not typically emphasized in the course, although it is
mentioned briefly in some texts and by some instructors during lecture.
ACKNOWLEDGMENTS 12
The students were not expected to recognize that the axis of polarization of
The investigation that has been described has been a col- the light has an effect on the diffraction pattern. For a discussion of how
the polarization of light can change the diffraction pattern, see T. W.
laborative effort by many members of the Physics Education Mayes and B. F. Melton, ‘‘Fraunhofer diffraction of visible light,’’ Am. J.
Group. Results from the doctoral dissertation of Karen Wosi- Phys. 62, 397–403 ~1994!; T. J. Racey, P. Rochon, and N. Gauthier,
lait strongly influenced the direction of the research. The ‘‘Effect of light polarization on the diffraction pattern of small wires,’’
authors are grateful to Paula Heron and Stamatis Vokos for ibid. 53, 783–786 ~1985!.

154 Am. J. Phys., Vol. 67, No. 2, February 1999 Ambrose et al. 154
This article is copyrighted as indicated in the article. Reuse of AAPT content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 94.4.77.73
On: Thu, 22 May 2014 20:19:53
 13
This difficulty and others related to student beliefs about photons are dis- interference–A demonstration,’’ Am. J. Phys. 40, 1003–1007 ~1972!. An
cussed in R. N. Steinberg, G. E. Oberem, and L. C. McDermott, ‘‘Devel- explanation intended for a first course in modern physics can be found in
opment of a computer-based tutorial on the photoelectric effect,’’ Am. J. P. A. Tipler, Modern Physics ~Worth, New York, 1978!, p. 185.
Phys. 64, 1370–1379 ~1996!. 16
See the second paper in Ref. 5.
14
Another explanation, not usually presented in introductory courses, is 17
In addition to Refs. 2, 3, 4, and 13, see L. C. McDermott, P. S. Shaffer,
based on Feynman’s path integral approach. In this formulation, all paths
and M. D. Somers, ‘‘Research as a guide for teaching introductory me-
between the emitter and a point on the screen are ascribed a phase. The
sum of the contributions of all paths yields the standard intensity pattern. chanics: An illustration in the context of the Atwood’s machine,’’ Am. J.
See R. P. Feynman, QED, The Strange Theory of Light and Matter Phys. 62, 46–55 ~1994!; T. O’Brien Pride, S. Vokos, and L. C. McDer-
~Princeton U.P., Princeton, NJ, 1985!. mott, ‘‘The challenge of matching learning assessments to teaching goals:
15 An example from the work-energy and impulse-momentum theorems,’’
For a description of a demonstration of a low-intensity, double-slit inter-
ference experiment, see S. Parker, ‘‘Single-photon double-slit ibid. 66, 147–157 ~1998!.

ROOM FOR MANEUVER

Bell shook his head and went on, ‘‘@The Dalai Lama# was also very interested in the Big Bang
theory, according to which the world had a start and probably will have an end. This appears to be
somewhat contrary to Buddhist scripture, which emphasizes eternal recurrence; things happen
again and again. I pressed him on that. Of course, I insisted that the Big Bang was a fashion in
science that could change. But if science did become committed to a one-time universe, how could
that be reconciled with Buddhist scriptures? He listened through his interpreter and replied, ‘Well,
it is perhaps not part of the Buddhism to which we are completely committed. We would have to
study our scriptures very carefully, and, usually, there is some room for maneuver.’ ‘Some room
for maneuver’ was the phrase the translator used. I liked that very much.’’

Jeremy Bernstein, ‘‘John Stewart Bell: Quantum Engineer,’’ in Quantum Profiles ~Princeton University Press, Princeton,
1991!, pp. 81–82.

155 Am. J. Phys., Vol. 67, No. 2, February 1999 Ambrose et al. 155
This article is copyrighted as indicated in the article. Reuse of AAPT content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 94.4.77.73
On: Thu, 22 May 2014 20:19:53