Carolyn A. Maher, Arthur B. Powell Combinationatoric and Reasoning

Combinatorics
and Reasoning
Mathematics Education Library

VOLUME 47
Managing Editor
A.J. Bishop, Monash University, Melbourne, Australia
Editorial Board
M.G. Bartolini Bussi, Modena, Italy
J.P. Becker, Illinois, U.S.A.
B. Kaur, Singapore
C. Keitel, Berlin, Germany
F. Leung, Hong Kong, China
G. Leder, Melbourne, Australia
D. Pimm, Edmonton, Canada
K. Ruthven, Cambridge, United Kingdom
A. Sfard, Haifa, Israel
Y. Shimizu, Tennodai, Japan
O. Skovsmose, Aalborg, Denmark
For further volumes:

http://www.springer.com/series/6276
Carolyn A. Maher Arthur B. Powell

Elizabeth B. Uptegrove
(Editors)
Combinatorics
and Reasoning
Representing, Justifying and Building Isomorphisms
123
Editors
Dr. Carolyn A. Maher
Rutgers University
Graduate School of Education
10 Seminary Place
New Brunswick, NJ 08901
USA
carolyn.maher@gse.rutgers.edu
Dr. Arthur B. Powell

Rutgers University
Department of Urban Education
110 Warren Street
Newark, NJ 07102
USA
powellab@andromeda.rutgers.edu
Dr. Elizabeth B. Uptegrove

Felician College
Department of Mathematical Sciences
223 Montross Avenue
Rutherford, NJ 07070
USA
uptegrovee@felician.edu
Series Editor:
Alan Bishop
Monash University
Melbourne 3800
Australia
Alan.Bishop@Education.monash.edu.au
ISBN 978-0-387-98131-4
e-ISBN 978-0-387-98132-1
DOI 10.1007/978-0-387-98132-1
Springer New York Dordrecht Heidelberg London
Library of Congress Control Number: 2010926950
Springer Science+Business Media, LLC 2010
All rights reserved. This work may not be translated or copied in whole or in part without the written
permission of the publisher (Springer Science+Business Media, LLC, 233 Spring Street, New York,
NY 10013, USA), except for brief excerpts in connection with reviews or scholarly analysis. Use in
connection with any form of information storage and retrieval, electronic adaptation, computer software,
or by similar or dissimilar methodology now known or hereafter developed is forbidden.
The use in this publication of trade names, trademarks, service marks, and similar terms, even if they are
not identified as such, is not to be taken as an expression of opinion as to whether or not they are subject
to proprietary rights.
Printed on acid-free paper
Springer is part of Springer Science+Business Media (www.springer.com)
This book is dedicated to the Kenilworth students

who participated in the longitudinal study and
from whom we continue to learn so much. We
thank you for your continuing commitment,
abundant trust, and generous sharing of how
mathematical ideas and ways of reasoning are
built.
Preface
Our research project on mathematical learning focuses on the accomplishments of

a cohort group of learners from first grade though high school and beyond, concentrating on their work on a set of combinatorics tasks. We describe their impressive
mathematical achievements over these years. We illustrate in detail the processes
by which students learn to justify solutions to combinatorics problems that were
challenging for their age and grade level. Based on transcribed video data and learners inscriptions, we provide a careful and detailed analysis of the process by which
mathematical ideas are developed, discussed, modified, expanded, and justified.
Our work underscores the power of attending to basic ideas in building arguments; it shows the importance of providing opportunities for the co-construction of
knowledge by groups of learners; and it demonstrates the value of careful construction of appropriate tasks. Moreover, it documents how reasoning that takes the form
of proof evolves with young children and it discusses the conditions for supporting
student reasoning.
We present in this book strong and compelling evidence that under appropriate conditions and with minimal intervention, learners can develop sophisticated
ideas about proof and justification, generalization, isomorphism, and mathematical
reasoning at an early age and can continue to refine and expand those ideas over
time, developing increasingly sophisticated presentations and representations. We
also describe an extension of this work with groups of undergraduate students, noting similarities and differences between the reasoning of the original cohort group
of younger students and that of the college students.
We include a detailed discussion of all the mathematical tasks, which can be used
in classrooms from elementary school to the graduate college level.
vii
Acknowledgements
We are deeply grateful for the many colleagues who have made this book possible
and would like to acknowledge their contributions.
We thank Fred Rica, Principal of Harding Elementary School, Kenilworth New
Jersey, whose vision made the study possible, and the faculty and administration at
Kenilworth who gave their support and encouragement to students and researchers.
We thank the many graduate students, video crew and videographers, especially
Roger Conant, Ann Heisch, Lynda Smith, and Elena Steencken, as well as the dedicated researchers for their hard work and insights: Alice S. Alston, Robert B. Davis,
John M. Francisco, Barbara H. Glass, Regina D. Kiczek, Judith H. Landis, Amy
M. Martino, Ethel M. Muter, John J. OBrien, Marcia OBrien, Ralph Pantozzi,
Manjit K. Sran, Maria Steffero, Lynn D. Tarlow, and Dina Yankelewitz.
We are particularly grateful to Robert Speiser for his enthusiasm and support, as
well as his invaluable help with task design.
We thank the Sussex County Community College students who participated in
the study.
We thank the staff of the Robert B. Davis Institute for Learning, and in particular,
Marjory F. Palius for research assistance and continued generous help, Robert Sigley
for his overall knowledge and management of the collection as well as his invaluable
technical help, Patricia Crossley for her organization of the data for the studies, and
Manjit K. Sran and Dirck Uptegrove for their wonderful illustrations and artwork.
We are grateful for support for the longitudinal study by: (1) the National
Science Foundation with grants: MDR-9053597 (directed by R. B. Davis and
C. A. Maher) and REC-9814846 (directed by C. A. Maher), and (2) from the New
Jersey Department of Higher Education, the Johnson and Johnson Foundation, the
Exxon Education Foundation, and the AT&T Foundation. Any opinions, findings,
and conclusions or recommendations expressed in this book are those of the authors
and do not necessarily reflect the views of the funding agencies.
ix
Contents
Part I
Introduction, Background, and Methodology
1 The Longitudinal Study . . . . . . . . . . . . . . . . . . . . . . . .

Carolyn A. Maher
2 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Carolyn A. Maher and Elizabeth B. Uptegrove
Part II
Foundations of Proof Building (19891996)
3 Representations as Tools for Building Arguments . . . . . . . . . .

Carolyn A. Maher and Dina Yankelewitz
17
4 Towers: Schemes, Strategies, and Arguments . . . . . . . . . . . .

Carolyn A. Maher, Manjit K. Sran, and Dina Yankelewitz
27
5 Building an Inductive Argument . . . . . . . . . . . . . . . . . . .

45
6 Making Pizzas: Reasoning by Cases and by Recursion . . . . . . .

59
7 Block Towers: From Concrete Objects to Conceptual Imagination

Robert Speiser
73
Part III Making Connections, Extending, and Generalizing

(19972000)
8 Responding to Ankurs Challenge: Co-construction of
Argument Leading to Proof . . . . . . . . . . . . . . . . . . . . . .
Carolyn A. Maher and Ethel M. Muter
9 Block Towers: Co-construction of Proof . . . . . . . . . . . . . . .
Lynn D. Tarlow and Elizabeth B. Uptegrove
10
Representations and Connections . . . . . . . . . . . . . . . . . . .

Ethel M. Muter and Elizabeth B. Uptegrove
89
97
105
xi
xii
Contents
11
Pizzas, Towers, and Binomials . . . . . . . . . . . . . . . . . . . . .

Lynn D. Tarlow
121
12
Representations and Standard Notation . . . . . . . . . . . . . . .

133
13
So Lets Prove It! . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Arthur B. Powell
145
Part IV
Extending the Study, Conclusions, and Implications
14
Doing Mathematics from the Learners Perspectives . . . . . . .

John M. Francisco
157
15
Adults Reasoning Combinatorially . . . . . . . . . . . . . . . . . .

Barbara Glass
171
16
Comparing the Problem Solving of College Students with

Longitudinal Study Students . . . . . . . . . . . . . . . . . . . . .
Barbara Glass
17
185
Closing Observations . . . . . . . . . . . . . . . . . . . . . . . . . .
Arthur B. Powell
201
Appendix A Combinatorics Problems . . . . . . . . . . . . . . . . . . .
205
Appendix B Counting and Combinatorics Dissertations

from the Longitudinal Study . . . . . . . . . . . . . . . . . . . . . .
213
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
215
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
221
Introduction
Carolyn A. Maher, Arthur B. Powell, and Elizabeth B. Uptegrove
The theoretical foundation for the program of research on which this book is based
comes from recognition that individual learning takes place within a community.
The members of that community have access to and are influenced by the ideas
of others. Individual learners are interconnected with other members of the community; engagement with others opens up possibilities for sharing and comparing
representations of ideas and for revising existing schemes and building new ones.
In the activity of problem solving, learners bring forth, communicate, and compare
ideas. They explore whether the ways that others represent ideas correspond with
their own representations, thereby extending their personal repertoires of tools for
dealing with new ideas. In this way further learning takes place and understanding
deepens (Davis & Maher, 1997; Maher, Martino, & Alston, 1993; Maher & Davis,
1990).
The data for this book come from a long-term program of research detailing the
collective building of mathematical ideas, which we call the longitudinal study. In
this book, we explore student work in one of the mathematics strands of the longitudinal study: counting and combinatorics. It investigates how students reasoning
evolved from elementary and high school years to college.
The reasoning of learners is documented by their actions that is, what they do,
say, build, and write as they work on strands of tasks. In studying how participants
make sense of the complexity of problems, we trace the representations they share,
the heuristics they invent and apply, and the modifications they make in building
arguments and in offering justifications for solutions.
The authors of the constituent 17 chapters relate how an ordinary group of school
children manifest over a 12-year period an extraordinary array of mathematical ideas
that they discursively build and how with time their ideas modify and mature as
they reason and justify their ideas. The book reports episodes from a long-term study
of how mathematical ideas and ways of reasoning are built by students over time.
The study has produced over 4,500 h of video, over several sites, involving far more
C.A. Maher (B)

Graduate School of Education, Rutgers University, New Brunswick, NJ, USA
e-mail: carolyn.maher@gse.rutgers.edu
xiii
xiv
Introduction
data than can be presented here. However, we have selected narratives that feature
the voices of several children, as interpreted by a variety of researchers, to weave
together a bigger story about how students can educate us about the multifaceted
nature of mathematical development. In an important sense, the really big story is
still being written as our work in preserving and further analyzing those 4,500 h of
video through the Video Mosaic Collaborative1 continues to reveal new narratives.
We invite readers to view the videos at http://www.video-mosaic.org/. Along with
the narratives offered in the book chapters that follow, these videos enable readers
to trace in detail the development of counting/combinatorics ideas and ways of reasoning of learners over more than a decade. Thus, while only some of the childrens
voices appear in this book, we are indebted to all of them for sharing their developing mathematical ideas over time and in divergent contexts, which we continue
to study and consider how these childrens extraordinary mathematical reasoning
may inspire the fields of mathematics education, teacher education, and the learning
sciences.
To structure a story that emerges from the chapters, the editors have divided this
book into four parts. The two chapters of the first part, respectively, provide historical background of the research study from which the details of the later chapters
emerge and describe the design of the study. The first chapter describes the study and
the purpose of the research, how the study began, and the conditions under which the
research was conducted. It also briefly describes the mathematical ideas and ways of
reasoning that emerged from the study. (The details are presented in later chapters.)
The second chapter presents the method of the study, its design, including selection
of participants, data collection, and analysis, as well as the strand of tasks on which
participants were invited to work. The chapter also discusses the importance of the
task design for helping learners to develop ways of reasoning.
The second part of the book contains five chapters. These chapters chronicle
the work of the studys participants over a 7-year period from grades 28, tracing
the development of their mathematical ideas, heuristics, and forms of reasoning.
In particular, the reader will learn how the participating children represented their
ideas; developed schemes and strategies; reasoned in specific ways; built inductive arguments; reasoned by cases and by recursion; and connected numbers in
Pascals Triangle to results of previous problems. The authors of Chapter 3 discuss how young children use representations to express their mathematical ideas
while building a solution to a particular counting problem (the shirts and jeans
1
The Video Mosaic Collaborative is a research and development project sponsored by the National
Science Foundation (award DRL-0822204) directed by C.A. Maher, G. Agnew, C.E. Hmelo-Silver,
and M.F. Palius that is leveraging the Rutgers Community Repository to preserve the unique video
collection amassed by The Robert B. Davis Institute for Learning at Rutgers University through
two decades of research with over four millions dollars of grant funding from the NSF (awards
MDR-9053597, REC-9814846, REC-0309062 and DRL-0723475). In addition to preserving the
video collection, new tools are being developed for conducting design research and an empirical
study that use the videos in the context of teacher education. The editors gratefully acknowledge
this considerable support from the National Science Foundation and wish to clarify that all views
expressed in this book are those of the authors are not necessarily those of the NSF.
Introduction
xv
problem, described fully in Appendix A, along with all combinatorics problems discussed herein). They show how children structure their representations in response
to requests to justify their problem solution and build convincing arguments to
early counting problems. The authors of Chapter 4 and 5 discuss students work
on different versions of the towers problems (which involve determining how many
towers can be built of various heights when selecting from cubes of various numbers of colors). They show the emergence of different forms of reasoning (cases,
contradiction, recursion, and induction) and how, motivated by the need to find
the sample space for a basic probability exploration, students revisit the inductive
argument for building towers. Chapter 6 discusses how participants collaboratively
build representations that help them use reasoning by cases and by recursion to
develop justifications for their solutions to classes of pizza problems. (Pizza problems involve determining how many pizzas it is possible to make when selecting
from various numbers of toppings and under various other constraints.) Completing
this part of the book, Chapter 7 presents the results of an interview with 13-yearold Stephanie, who discusses the relationship between the towers problems and the
binomial expansion, including how the towers answers can be found in Pascals
Triangle.
The six chapters of the books third part closely examine the mathematical work
of the research participants during their high school years. It shows how the students
built important connections using sophisticated mathematical reasoning. In these
chapters, the story revolves around the students proof making, use of representations, acquisition of standard notation, and forging of conceptual connections among
isomorphic problems. Specifically, Chapter 8 shows that as they revisit their representations and arguments, students refine representations and clarify arguments. In
Chapter 9, students working in groups on towers problems are seen to find and generalize formulas, using methods including controlling for variables, justification by
cases, and induction. Chapter 10 shows how a tenth-grade students binary notation helped his group form connections among the pizza and towers problems, the
binomial expansion, and Pascals Triangle.
Chapter 11 details how representations are a source for making connections in
solutions to pizza and tower problems, resulting in the students mapping the structure of the solution of these problems to Pascals Triangle and how their increasingly
sophisticated use of representations led to further development of mathematical reasoning and justification. Chapter 12 discusses how students moved from personal
to standard notations in order to express in general form their understanding of
solutions to the pizza and towers problems and to extend their understanding in
creating an isomorphism from the numerical results in those problems to Pascals
Triangle. The chapter also shows how the students understanding of extensions of
the pizza and tower problems led to their understanding of the addition rule for
Pascals Triangle. The final chapter of Part III, Chapter 13, reveals how as high
school seniors, days before graduation, the students used their understanding of
relationships between the pizza and tower problems and Pascals Triangle to solve a
third isomorphic problem the Taxicab Problem. (This problem involves finding the
number of routes from the starting point the taxicab stand to various points on
xvi
Introduction
a rectangular grid.) They recognized the isomorphism, used it to make conjectures

about the new problem, saw the need to prove their conjectures, and provided a convincing argument. This chapter concludes by examining some of the extraordinary
mathematical accomplishments of the cohort group of students.
The last part of the book, consisting of four chapters, takes stock and looks forward. Chapter 14 examines the epistemological growth of the students, viewed from
their own perspectives. Students reflections on their learning over the years challenges common views about student engagement in learning, and gives insight into
how students view their own sense making in doing mathematics.
Chapter 15 examines a different student population college undergraduates
and their work with the set of combinatorics problems. The chapter shows that when
adult college students are asked to justify ideas and make convincing arguments, an
understanding of mathematical reasoning, proof, and generalization can emerge. In
Chapter 16, Glass compares the strategies developed by children and older learners for solving the combinatorics problems and discusses the implications for adult
learning.
In closing, Chapter 17 presents the epistemological and methodological contributions of the book. We argue that students must be actively and purposely engaged in
their learning so as to take ownership of and be proud of their accomplishments.
Mathematics educators and teachers need to create opportunities for students to
engage in ways similar to those described in this book. We have shown that in a
program of carefully selected tasks, with minimal intervention by educators who
pay careful attention to students arguments and justifications, students can perform
mathematically at high levels. In addition to developing mathematical competency,
students who participated in the study gained confidence and a sense of empowerment and were successful in their career choices. They learned to trust their own
mathematical ability and they did not rely on outside authority for validation. This
confidence, sense of empowerment and propensity to reason carefully has been carried over outside their mathematical work; these students found that the knowledge
and ways of working that they gained through their participation in the longitudinal
study continues to help them in many other areas of study and employment.
Contributors
John M. Francisco
Secondary Mathematics Education, Department of Teacher Education &
Curriculum Studies, University of Massachusetts Amherst, Amherst, MA 01003,
USA, jmfranci@educ.umass.edu
Barbara Glass
Sussex County Community College, Newton, New Jersey, USA,
bglass@sussex.edu
Carolyn A. Maher Graduate School of Education, Rutgers University, New
Brunswick, NJ, USA, carolyn.maher@gse.rutgers.edu
Ethel M. Muter
5280 Antioch Ridge Drive, Haymarket, VA 20169, USA, emmuter@gmail.com
Arthur B. Powell Department of Urban Education, Rutgers University, Newark,
NJ, USA, powellab@andromeda.rutgers.edu
Robert Speiser
799 E 3800 N, Provo, UT 84604, USA, pinyonsmoke@gmail.com
Manjit K. Sran
Mathematics Department, Monroe Township High school, 1629 Perrineville Road,
Monroe Township, NJ 08831, USA, msran@monroe.k12.nj.us; College of
Business and Management, DeVry University, 630 U.S. Highway One, North
Brunswick, NJ 08902, USA, msran@devry.edu
Lynn D. Tarlow
Department of Secondary Education, The City College of the City University of
New York, New York, NY 10031, USA, ltarlow@ccny.cuny.edu
Elizabeth B. Uptegrove Department of Mathematical Sciences, Felician College,
Rutherford, NJ, USA, uptegrovee@felician.edu
Dina Yankelewitz
Department of General Studies, The Richard Stockton College of New Jersey,
Pomona, NJ, USA, dyankelewitz@gmail.com
xvii
Part I
Introduction, Background,
and Methodology
Chapter 1
The Longitudinal Study

Carolyn A. Maher
1.1 Theoretical View

Where do new ideas come from? Our view is that building new ideas is a process;
new ideas come from old ideas that are revisited, reviewed, extended, and connected
(Davis, 1984; Maher & Davis, 1995). Building new ideas also involves the retrieval
and modification of representations of existing ideas. The representations that a
learner builds for a mathematical idea or procedure can take different forms physical objects or actions on objects, words, and symbols, for example. As the learners
experience increases, old representations become elaborated, extended, and linked
to new ones (Maher, 2008; Davis & Maher, 1997).
The problem tasks that are posed to learners are critical to their learning
(Francisco & Maher, 2005); they should be well defined, open-ended, and open
to extension and generalization. The connections that the learner makes when analyzing and developing solutions to these problems provide further opportunity for
growth in knowledge. Students are encouraged to revisit earlier problems because
requirements to justify and generalize solutions can help students to see underlying
mathematical structure. It is a widely accepted view that when learners understand the fundamental structure of a subject, the gap between elementary and
advanced knowledge is reduced (Bruner, 1960). There is increasing evidence that
learners, under certain conditions, can build meaningful, mathematical relationships
and understand the structure of mathematical problems at an early age. For example, a study of Norwegian children indicated that even as young as Grade 3, learners
are able to unearth the underlying structure of the mathematics of problem tasks
(Torkildsen, 2006).
A central component of the learning process is encouraging students to communicate their ideas. Sfard (2001) suggests that students learn to think mathematically
C.A. Maher (B)

C.A. Maher et al. (eds.), Combinatorics and Reasoning, Mathematics

Education Library 47, DOI 10.1007/978-0-387-98132-1_1,

C Springer Science+Business Media, LLC 2010
C.A. Maher
by participating in discourse about ideas arguing, asking questions, and anticipating feedback. We have emphasized that justifying ideas in problem solving is an
essential component of mathematical reasoning (Maher, 2002, 2005, 2008; Maher &
Martino, 1996a; Martino & Maher, 1999). Learners, in communicating their ideas,
share personal mental images representations. When students make their representations public, they have an opportunity to talk further about them, compare them,
and later revisit them. Similarities and differences in ideas naturally emerge. When
learners try to convince others that their answers are correct, they can reorganize
and reformulate their representations so as to make convincing arguments. In summary, students learn mathematics by engaging in the process of building their own
personal representations, communicating them as ideas, and then providing support
for those ideas by reorganizing and restructuring representations. Our view is that
this process is a necessary prerequisite both for developing the idea of mathematical
proof and for making suitable connections between problems of equivalent structure
by building isomorphisms.
In this book, we discuss how a group of students developed new and increasing
sophisticated mathematical ideas by revisiting, reviewing, extending, and connecting old ideas that they had begun developing in first grade. They developed and
modified representations that became increasingly elaborated and extended. They
participated in serious mathematical discourse. And ultimately they built a strong
and durable understanding of the solutions to a set of mathematical tasks. Our longitudinal work is important because it reveals the processes that these learners used
to build structural understanding of solutions to mathematical tasks.
1.2 Background of the Study

The longitudinal study began in 1987 in Kenilworth, New Jersey. This was during a time when behaviorism mainly governed mathematics instruction. It was a
time before the reform movement in the United States emphasizing conceptual
understanding had made its entry. The K-8 Harding Elementary School in the
working-class community of Kenilworth, New Jersey, was typical of others at that
time. Half-hour sessions were devoted to mathematics, and mathematics instruction
was mainly rote. The rule was drill and practice for carrying out memorized procedures. For the most part, even the brightest students from the school did not excel
when they moved on to high school mathematics classes, only doing average work.
Most members of the community and most teachers had rather low expectations for
student advancement.
But Fred Rica, principal of the Harding Elementary School, had higher expectations for the students in his school. Formerly an elementary grade classroom teacher
in Kenilworth, Fred Rica knew his staff and students well. Like other concerned educators, he knew when the system was not serving its student population. He turned
to Rutgers University for help with instruction, first in language and literacy and
then in mathematics. It was shortly after this professional development work that
Fred Rica and Carolyn Maher created a partnership between the Kenilworth Public
Schools and Rutgers University. It should be noted that the RutgersKenilworth

partnership, with its focus on students building meaning of mathematical ideas and
working collaboratively with each other, began long before the National Council of
Teachers of Mathematics published its reform standards.
Initially, the project began as a teacher development intervention in mathematics. The Rutgers University team of researchers and graduate students worked for
3 years to help teachers build an understanding of the mathematics they were
expected to teach and to learn to be attentive to the developing understanding of
their students. (See Davis & Maher, 1993; OBrien, 1994, for a detailed study of the
teacher development project.)
The project could not have survived the early years without the full support and
active participation of the Kenilworth school administration. In particular, principal
Rica actively participated in the teacher-training sessions, encouraged teachers to
become involved, and made sure that students who were involved in the study were
available to the researchers. Original financial support for the partnership came from
the Kenilworth school district and through volunteer efforts of the Rutgers team. The
Kenilworth school district continued to fund the study for several years as a component of its mathematics teacher development mission. The Rutgers research group
received outside funding for the research from two National Science Foundation
grants. The first grant awarded to Principal Investigators Robert B. Davis and
Carolyn A. Maher was when the students were in Grade 4; the second grant
awarded to Principal Investigator Carolyn A. Maher was when students were in high
school.
1.2.1 Teacher Development Component

It is not surprising that the teachers at the Harding Elementary School were not
prepared to teach mathematics with understanding. What is surprising was the
expectation of principal Fred Rica that the teachers were capable, with some professional development and classroom support, of understanding the mathematics they
were expected to teach. In fact, this view was remarkable for its time.
The teacher development team was made up of mathematics education doctoral
students who had considerable experience in schools; its first members were Alice
S. Alston and Judith H. Landis. The team worked closely with Fred Rica and
his teachers to establish a program of activities that involved not only videotaped
teacher workshops and classroom sessions, but also study of those workshops and
sessions. The Rutgers team worked directly with students and with their teachers,
first observing classroom sessions and later collaborating with the teachers in the
design and implementation of lessons. Alice Alston also worked in the classrooms
alongside the teachers.
Principal Rica obtained school funding to support teachers summer work to
revise the existing curriculum. Two years of summer professional development
assisted by John OBrien and Alice Alston resulted in a movement from a drill
and kill approach to one in which students building of mathematical understanding
C.A. Maher
was central. Curriculum revisions included the use of more engaging and thoughtful
lessons for the students and the introduction of manipulatives that allowed students
to build models of their solutions.
Some Kenilworth teachers who participated in the teacher development programs
also became involved in classroom action research. As teachers were introduced
to new resources and tools, they developed new units and piloted them during the
school year. Through course work opportunities at Rutgers, some teachers studied
the mathematical learning of their own students (Landis & Maher, 1989; Landis,
1990; Maher 1988; OBrien, 1994).
1.2.2 Intervention Design

The Rutgers team was interested in what mathematical concepts students could
learn with minimal intervention from teachers. Classrooms were organized so that
children might work together and collaborate on problem tasks. Children were
encouraged to use each other as resources in their investigations, to construct models of solutions with available tools, and to revisit tasks and discuss their strategies
and solutions. An important observation during the first 3 years was that students
produced arguments that took on a variety of forms of reasoning to support their
solutions to the problems. By Grade 4, it became increasingly clear to researchers
that students reasoning, in a natural way, took the form of proof. Children began
their investigations by searching for patterns, organizing solutions, searching for
completeness, deriving strategies for keeping track and checking, and then reorganizing justifications into arguments that were proof-like in structure. Using each
other as resources, children freely shared ideas, questioned each other, argued about
the reasonableness of ideas, and became comfortable in sharing and communicating
with each other.
What encouraged both the school staff and the university collaborators was the
enthusiastic feedback from students. The children enjoyed talking about their ideas;
they engaged with each other with energy and enthusiasm, becoming increasingly
more comfortable making their ideas public. Their way of working underscored a
demand for sense making, which then evolved as a cultural norm.
This book explores student work for one of the mathematics strands of the longitudinal study: counting and combinatorics. It investigates how students reasoning
evolved over the course of the longitudinal study that continued from elementary
and high school years to college.
1.3 Longitudinal Study: Grades 13

In order to study the effectiveness of the intervention, the Rutgers team decided
to follow a class of students throughout their elementary grades as they worked
on mathematical investigations that were not part of the school curriculum. The
study began with a class of 18 first-grade students from the Harding School. These
children, randomly assigned to one of three first grades, became the initial focus
group; they were together for Grades 13 as part of the school design. Throughout
the study, students engaged in strands of thoughtful mathematics activities designed
by the researchers. Although the mathematical investigations were not part of the
curriculum, the concepts that were introduced would later become part of the regular
school mathematics curriculum.
1.4 Longitudinal Study, Grades 48

After Grade 3, the students were distributed among different classrooms, according to school policy. However, the principal worked with Rutgers researchers to
facilitate maintenance of a focus group of 12 students for research purposes. When
families moved and new families entered the district, the composition of the focus
group changed, but an attempt was made to maintain a group of comparable background and interest. Although some students stayed with the study from the start
(and are still in touch today), some students moved from the district and new students joined. During middle school, the school arranged for the cohort group to
continue working with researchers during school hours, 46 times a year in two
90-min sessions and one 45-min session each time.
1.5 Longitudinal Study: High School Years

In 1996 the high school in Kenilworth was closed, as the school district became
part of a regional system. The community joined forces to protest the merger and
succeeded after 1 year. Hence, the first year of high school (ninth grade) proved
disruptive for the students, although some math problem-solving sessions were
conducted with small groups of students during that year in local homes, usually on Saturdays. After Kenilworth de-regionalized and the students returned to
Kenilworth for the remaining 3 years of high school, groups of students resumed
participation in the longitudinal study in informal, after-school sessions that were
held during the year, usually on Friday. While students no longer met with
researchers during regular class hours, 14 students (some from the original group
of first graders and others who had joined the study at various times during middle
school and high school) made time in their schedules to meet after school about
46 times a year for problem-solving sessions that lasted 12 h or longer. This
group included ten students who had been with the study since Grade 1, two students who had joined the study in Grade 6, and two who joined in high school
(Grade 11).
1.6 Longitudinal Study: Beyond High School

All students in the focus group applied to Rutgers University, and all were
accepted a remarkable achievement for the district. However, not all students
attended Rutgers; they attended a variety of universities, public and private; besides
C.A. Maher
Rutgers, these included Cornell University, Kean University, St. Johns University,
and the University of Pennsylvania. Majors included accounting, American studies, animal science, computer science, criminal justice, economics, engineering,
English, and mathematics. All are now either employed or in graduate school.
Some of the students have continued to meet occasionally with researchers during and after college. They do not generally work on problems (although sometimes
old problems are revisited), but they talk about how being in the study has affected
them, their academic careers, and their future plans.
In the next chapter, we detail how the study was conducted and we discuss
selected problems that formed the cornerstone of the student investigations over
the years.
Chapter 2
Methodology
Carolyn A. Maher and Elizabeth B. Uptegrove
2.1 Introduction
In this chapter, we discuss how data were collected and analyzed, and we briefly
describe some results, which will be more fully explored in later chapters. We
summarize student work on fundamental problems and note how this work led to
exceptional growth in the students mathematical understanding.
Researchers (professors at the Rutgers University Graduate School of Education
and their students) conducted all problem-solving sessions with the students; the
sessions were always videotaped with one or more cameras. Researchers observed,
described, and coded the videotape data, and they kept written and electronic files
of the emerging theoretic, analytic, and interpretative ideas about the students
mathematical behaviors. Researchers paid careful attention to childrens use of
inscriptions, the connections they made between and among codes, and their emerging and extended ideas and ways of reasoning. Critical events in childrens reasoning
were flagged and transcribed and transcripts were coded according to the research
questions. The connected series of events that formed a trace led to the emergence
of a narrative (Maher & Martino, 1996a; Powell, Francisco, & Maher, 2003).
The videotapes, researcher notes, and student notes did not capture every interaction or every case of student learning. Some students sat silently during discussions;
but they had quietly absorbed a problem or quietly developed a solution that came
to light some time later in a different situation. Therefore, although we can make
inferences about what is observed, we cannot assume that a student who is quiet
does not understand.
By videotaping children as they worked together on mathematical tasks over long
periods of time, we were able to trace the origin and development of their mathematical ideas. We observed what children said to one another and showed to one
another. We used videotapes and transcripts to study the meanings that children gave
C.A. Maher (B)



10
C.A. Maher and E.B. Uptegrove
to mathematical situations and to note the different representations they made public. A detailed analysis of data made it possible to trace the origin and evolution of
childrens arguments. Our data indicate how children expressed their ideas through
spoken and written language, through the physical models they built, through the
drawings and diagrams they made, and through the mathematical notations they
invented.
2.2 Theoretical Perspectives

Guiding our work is the view that children come to mathematical investigations
with theories they can modify and refine. We observe them do so in settings that
combine personal exploration and suitable social interaction. The theories we consider can include criteria to decide (1) what, at some given moment, needs to be
investigated, (2) how to conduct such an investigation, (3) what key features need
to be explored in detail, (4) when useful progress has been made, and, given such
progress, (5) if further investigation might be needed. We have found that theories
of this kind often empower striking and effective ways for children to work conceptually with mathematical ideas, often using concrete objects as specific anchors for
their thinking.
2.3 Selected Problems

Mathematics arose from the need to count, measure, and calculate, but the discipline
evolved to include abstraction, logical reasoning, and the search for and analysis of
patterns. Good mathematical problems are therefore those which give rise to the
need for abstraction, systematization, and pattern recognition. A focus of the study
was therefore to select problems that would give rise to these needs.
Another focus of the longitudinal study was on doing problems that were not part
of the regular curriculum, because it was important for the students to come to the
problems fresh, without pre-taught algorithms. A major strand of the longitudinal
study therefore consisted of problems in combinatorics, because in working on these
problems, students can find the need to organize their work systematically, look for
patterns, and generalize their findings; also, counting problems were at the time outside the regular elementary school curriculum and therefore unfamiliar to students.
In addition, these problems lend themselves to the use of multiple personal representations that can be shared. Freudenthal (1991) cites the study of combinatorics
as a most important matter for reinvention (p. 53), specifically because combinatorics can be learned through paradigmatic examples and because problems in
combinatorics give rise to the need for convincing proof, including mathematical
induction.
Another purpose of the longitudinal study was to provide an environment in
which certain socio-mathematical norms could be established to elicit in children
Methodology
11
sense making, argumentation, and justification in mathematics. As Yackel and Cobb

(1996) and Cobb, Wood, Yackel, and McNeal (1992) suggest, an appropriate social
context must be created to encourage students to try to convince others of the truth
of the mathematical ideas that they build. The longitudinal study was structured
to investigate and track the nature of the schemes that students developed and the
methods that the students used to build and retrieve representations to solve mathematical tasks. In addition, the study attempted to trace how students shared ideas
and how these ideas were adapted and assimilated by other students.
Appendix A describes all the combinatorics problems that students worked on
over the years. We summarize here some example problems, along with brief
accounts of strategies and representations used by students and forms of reasoning
that developed.
2.3.1 Shirts and Jeans

Students worked on the shirts and jeans problem at the end of second grade and
again at the beginning of third grade (1989 and 1990):
Stephen has a white shirt, a blue shirt, and a yellow shirt. He has a pair of blue jeans and
a pair of white jeans. How many different outfits can he make?
During second grade, most students drew pictures of outfits; some drew lines
between shirts and jeans, and others made lists of outfits. Notational choices
influenced the way they reasoned about the data. For example, Stephanie used bluewhite to stand for the white shirt/blue jeans outfit, and also for the blue shirt/white
jeans outfit. Contextual issues also played a role in the problem solving. For example, Dana discarded the white jeans/yellow shirt outfit on grounds that the resulting
outfit did not match and was thus not fashionable. That different students got different answers was not problematic for the children; in second grade, students seemed
comfortable with the notion that answers varied between three and seven outfits.
They willingly shared their interpretations and strategies and talked to each other
about their findings. In third grade, when the children were again presented with
this problem, they did not remember how they had solved the problem earlier, nor
did they remember their earlier answers. Of particular interest is that evidence of further elaboration of earlier strategies emerged. Students used and built on strategies
of their second-grade partners. For example, Stephanie indicated different outfits by
drawing lines between drawings of shirts and jeans, as Dana had done in second
grade.
By third grade, techniques for checking and for keeping track, such as controlling
for variables, were complete. Earlier ideas and strategies were refined to produce
complete, elegant solutions rather quickly. Students built on their heuristics to solve
more complex extensions of the problem to include belts and hats as parts of outfits.
What was especially significant for the researchers was the evidence of how students built on earlier ideas and, without intervention or approval from researchers,
12
continued their problem solving, driven by earlier heuristics and sense making to
produce correct solutions that they could justify.
2.3.2 Towers
Early in third grade (1990), students were given the four-tall tower problem for the
first time:
Your group has two colors of Unifix cubes. Work together and make as many different towers
four cubes tall as is possible when selecting from two colors. See if you and your partner
can plan a good way to find all the towers four cubes tall.
The definition of a tower is an ordered sequence of Unifix cubes, snapped

together. Each cube can also be called a block. Each tower has a bottom and a top.
The height of a tower is the number of its cubes. We say two towers are the same if
their colors match, block by block, from top to bottom. Unifix cubes are interlocking
cubes that come in various colors (typically blue, red, yellow, white, and green).
In fourth grade (1991), students worked on the five-tall tower problem. Then in
fifth grade (1992), they revisited the four-tall version. In 10th and 11th grades, they
were asked to provide a justification for the n-tall tower problem. Students discussed
variations and generalizations of the solution and they used their organization of
towers by cases and knowledge of the binomial expansion to build an understanding
of how Pascals triangle grows.
Their work on the towers problems also illustrates how their representations
changed over the years. At first, they used Unifix cubes to build towers. Eventually,
they turned to drawings and codes, for example, using letters R and Y to mean red
and yellow cubes. In some cases, a more general code emerged; some students
would use X and O or 0 and 1 to indicate any two colors. More details on these
emerging strategies are given in Sections 2.2 and 2.3.
2.3.3 Pizzas
In order to introduce a variation of the tower problem and to investigate how students
reasoned with an isomorphic problem, the researchers introduced the set of pizza
problems. When the students were first given the problem in fifth grade, they interpreted the task as allowing different toppings on each half of the pizza, an alternative
that they knew that was available in some pizza restaurants. In response to their interest in counting the varieties allowing toppings on half a pizza, the researcher asked
them to solve it with only two toppings available. This pizza with halves problem is
as follows:
Kenilworth Pizza has asked up to help them design a form to keep track of certain pizza
sales. Their standard plain pizza contains cheese. On this cheese pizza, one or two toppings
could be added to either half of the plain pie or the whole pie. How many choices do customers have if they could choose from two different toppings (sausage and pepperoni) that
Methodology
13
could be placed on either the whole cheese pizza or half a cheese pizza? List all possibilities. Show your plan for determining the choices. Convince us that you have accounted for
all possibilities and there could be no more.
The strategy that the students developed for the solution established the heuristic
that was applied later when there were five toppings available, again, allowing some
or no toppings on half the pizza. The final problem, the five-topping pizza problem,
proved a trivial special case for the pizza with halves problem that they first solved
successfully:
The local pizza shop offers a plain cheese pizza. On this cheese pizza, you can place up to
five different toppings. How many pizzas is it possible to make?
Pizza with Halves was the first of several variations of the pizza problem that the
students worked on over the years. It illustrates a basic philosophy of the study we
did not start students off with easier problems and then progress to the more difficult ones. Instead, students began with the more difficult versions of the problems,
which required them to tackle several challenges at once organization (making
sure no pizzas were repeated and none were omitted), notation (how to distinguish
between pepperoni and peppers, for example), and forming a valid argument how
to convince the researchers (and themselves) that they had the right answer.
Looking at students answers to the pizza problems over the years, we see growth
in organization and in representations. At first, students drew fairly accurate renditions of pizzas; they drew circles to indicate pizzas, and inside those circles were
wavy lines to indicate sausages and smaller circles to indicate pepperoni, for example. When they had to answer a question involving half pizzas, they drew lines
down the middle of their pizza circles to show both halves, and they listed all
the pizzas using full words (for example, whole plain, half sausage half plain).
Eventually, they turned to codes, starting with single letters or combinations of letters (to distinguish between peppers and pepperoni, for example) and then moving to
more abstract symbols such as 0s and 1s. These representations and organizational
strategies are discussed more fully in Chapter 6.
In 11th grade, some students investigated Pascals triangle and Pascals identity
(the addition rule for Pascals triangle). Using the metaphor of the pizza problem,
they explained how the triangle grows by explaining how the number of possible
pizzas grows as new toppings become available. In an extraordinary session lasting over 2 h one evening in 1999, students generated a slightly nonstandard but
mathematically correct equation for Pascals identity using standard combinatorial
notation:
N
X

+
N
X+1

=
N+1
X+1
A detailed description of the students work on Pascals identity is given in

Chapter 12.
14
2.3.4 Taxicab
The group that had generated Pascals identity was introduced to the taxicab
problem in 12th grade (2000):
All trips originate at the taxi stand, in the upper-left corner of a grid. The problem is to
find the shortest route to three specific points on the grid and to determine the number of
shortest routes to each point.
Their work on this problem is discussed in greater detail in Chapter 13. It is interesting to note that by this time, the students, without prompting, solved the general
problem, in addition to answering the specific questions. For any point on the grid,
they showed why the general answer was correct, and they demonstrated the connection to isomorphic problems (towers and pizzas) and to the binomial expansion.
What is also interesting from this session is that the students took on the roles of
eliciting justifications from each other. Their pursuit of explanations that made sense
and that connected to earlier tasks was quite remarkable.
2.4 Concluding Remarks

The purpose of the longitudinal study was not to teach the students particular topics
in combinatorics or other areas of mathematics. Instead, the aim was to establish
a culture where the correctness of an answer came from the sensemaking of the
students, rather than from the authority of the researcher. We asked students questions about what was convincing, what made sense, and how they developed their
answers. In justifying their answers, students usually exceeded our expectations.
We were impressed by the seriousness with which students approached the problems and the collegiality of their work, as well as by the forms of reasoning they
developed. In the early years of the study, children began to use inductive reasoning, to organize work by cases, and to think about justification through contradiction.
By middle school, these forms of reasoning were more sharply defined, and other
forms of reasoning emerged, such as controlling for variables. In high school, students began the process of building isomorphisms, using their own notation as well
as standard notation to describe how some problems were related to each other and
ultimately to Pascals triangle.
In the following chapters, we provide details on the specific problems, the specific strategies and representations used by the students, and the specific results
they generated. In the next chapter, we discuss the students earliest work on
combinatorics problems, the second- and third-grade work on shirts and jeans.
Part II
Foundations of Proof Building (19891996)
Chapter 3
Representations as Tools for Building

Arguments
Carolyn A. Maher and Dina Yankelewitz
Date and Grade:

Tasks:
Participants:
Researchers:
May 30, 1990 and October 11, 1990; Grades 2 and 3

Shirts and Jeans
Dana, Jaime, Michael, and Stephanie
Carolyn A. Maher and Amy M. Martino
3.1 Introduction
In this chapter, we discuss the childrens earliest work on combinatorics problems,
the second- and third-grade efforts on the shirts and jeans problem:
Stephen has a white shirt, a blue shirt, and a yellow shirt. He has a pair of blue jeans and
a pair of white jeans. How many different outfits can he make? Convince us that you have
them all.
In examining their problem solving, we focus on the childrens early use of personal representations. When introducing mathematics to children, it is important to
invite them to use their personal representations to express their ideas and ways of
reasoning (NCTM, 2000). These representations are the basic elements that children draw upon to express their ideas as they begin to engage in more abstract
and logical reasoning. Childrens representations and how they are connected and
related are foundational building blocks toward more sophisticated processes that
lead to the creation of new mathematical ideas. When children are provided with
opportunities to reinvent mathematics, they are in a better position later on to
recognize their own need for abstraction, generalization, and logical reasoning (see
Chapter 13).
C.A. Maher (B)



17
18
C.A. Maher and D. Yankelewitz
3.2 Representation as a Tool for Problem Solving

Davis (1984) has written extensively about the role of representation in mathematical thinking. According to Davis, representations are mental models that allow for
the association between the properties of a mathematical idea and the idea itself.
These ideas are not stored in the mind in words or pictures, and so when we explore
what these intangible representations are, we are only approximating their true
nature.
Davis stated that doing mathematics involves a series of steps, similar to those
of a computer executing a program, through which the student must cycle one or
more times. First, in an attempt to make sense of the problem, the student builds
a representation for the input data. Second, the student searches his memory for
knowledge that will assist in solving the problem. Finally, the student maps the data
representation with the knowledge representation. When the mapping seems accurate enough to tackle the problem at hand, the student uses techniques associated
with the knowledge representation to solve the problem.
Students use representations that they build to make sense of and attribute meaning to the mathematics that they are doing. They use mathematical tools, which,
according to Davis and Maher (1997) include mathematical notation, spoken and
written language, physical models, drawings, and diagrams.
According to Maher and Martino (1996a), students who are encouraged to build
and use multiple representations as they work on problems become sense-makers
and active members of the mathematical community. The use of different tools
to build and express ideas allows students to make connections between different representations and understandings and to better understand the mathematics
that they are learning. In addition, when students build and express multiple external representations, this allows observers (such as teachers, researchers, and fellow
classmates) to better understand the students ideas. Using representations to make
sense of problems and using representations to communicate ideas are therefore the
building blocks of effective argumentation.
3.3 Early Counting Task Strand Shirts and Jeans

As noted in the previous chapter, the researchers in the longitudinal study aimed to
provide the students with mathematical problems for which they had no algorithms
and which would afford them opportunities to find patterns, be systematic, and generalize findings. Combinatorics problems were well suited to these goals. In the
sections that follow, we will consider the specific mathematical ideas, fundamental
both to combinatorics in particular and to mathematics in general, that are elicited
by the tasks that were used in the longitudinal study.
The shirts and jeans task (above) introduces the fundamental counting principle,
a key idea in combinatorics.
Representations as Tools for Building Arguments
19
Fig. 3.1 A diagram and an

organized list for displaying
the shirts and jeans solution
In solving this problem, students may abstract the mathematics underlying the
real-world situation; they may come to realize that the number of combinations of
shirts and jeans is equivalent to the product of the number of shirts and the number of jeans. This type of multiplication is generally the most difficult for children
to model, comprehend, and apply (Cathcart, Pothier, Vance, & Bezuk, 2006). This
problem also introduces the need for notation or symbols to represent the real-world
items described in the problem. So this problem creates a need for a bridge between
the real-world situation presented and the mathematical ideas that will provide a
solution. The problem also provides students with a chance to realize that an organization of the facts (by means of a diagram or an organized list) can help them to find
a solution; this need for structure is fundamental to mathematics. The problem also
requires students to think about how to justify their solution to others and convince
others that they have found all the combinations. It has the potential to give rise to
the need for direct or indirect arguments (Fig. 3.1).
There are six combinations. In the figure, the letter B indicates a blue item,
W indicates a white item, and Y indicates a yellow item. The shirt colors are
listed on the left and the jean colors are on the right. The blue shirt can be combined
with either the white jeans or the blue jeans to form an outfit; so two (and only two)
outfits can be made using the blue shirt. The same is true for the white shirt as well
as for the yellow shirt. Therefore, there are 2 3 or 6 possible combinations. All
possible combinations are accounted for, since any other attempted combinations
will be duplicates of the ones listed above. For example, there cannot be a third
outfit formed using the blue shirt, because only white and blue jeans are available.
The same is true for the other color shirts.
3.3.1 Second-Grade Problem Solving

The students worked on the shirts and jeans problem in the second and third
grades. The strategies of three students, Dana, Stephanie, and Michael (see Fig. 3.2),
are analyzed and discussed in Martino (1992), Maher and Martino (1992a), and
20
Fig. 3.2 Dana, Stephanie,

and Michael, grade 2
Maher and Martino (1996). In addition, selections of the video were included
in the Private Universe Project in Mathematics (Harvard-Smithsonian Center for
Astrophysics, 2000), and the mathematical thinking and representations of these
students are discussed there.
In the second grade, all three students drew pictures of shirts and jeans to represent the items in the problem, and they used the pictures in their attempts to
find different combinations of the shirts and jeans. Stephanie drew three shirts and
labeled them w for white, y for yellow, and b for blue. She drew two pairs
of jeans, similarly labeled b and w. She then began to make a list, writing the
letter symbolizing the shirt directly above the letter that represented the jeans that
together comprised an outfit. She then numbered the combinations that she found.
However, when recording the fifth combination, she erased the w that she initially
wrote to make a combination of a white shirt and blue jeans, and, in its place, wrote
a y to show the combination of a yellow shirt and blue jeans (see Fig. 3.3). She
then told the researcher that she had found five combinations and she was convinced
that she had found them all.
Fig. 3.3 Stephanies grade 2

written work
21
Fig. 3.4 Danas grade 2

written work
Dana, early on, expressed verbally her understanding of the structure of the solution when she indicated that each of the three shirts could be combined with each
pair of jeans. She said, He can make all three of these shirts with that outfit
(Martino, 1992, p. 47). It can be concluded from this statement and the subsequent
problem-solving steps that she took that she had built a scheme that closely matched
the problem solution. As Martino (1992) notes, From her explanation it can be
inferred that Dana possessed a key strategy for exhausting all possible combinations (p. 48). Dana used a strategy of connecting her representations of shirts and
jeans that she had drawn with lines as is shown in Fig. 3.4.
An interesting decision that Dana made was not to draw a line between the yellow
shirt and white jeans because a yellow shirt and white jeans do not go together.
She then used Stephanies strategy of listing and numbering the combinations, so
Dana also arrived at a solution of five combinations. However, Stephanies solution
lacked the combination of a yellow shirt and blue jeans, while Danas was missing
the combination of a yellow shirt and white jeans (the combination that did not
match, according to Dana). We can conclude that she was aware of all possible
outfits but her sense of fashion resulted in her rejecting the yellow shirt and white
jeans.
Michaels strategy differed significantly from that of his classmates. He drew
diagrams of the different color shirts and jeans, but said that he had arrived at three
combinations: a white shirt with white jeans, a blue shirt with blue jeans, and a
yellow shirt with yellow jeans (see Fig. 3.5).
Although Stephanie and Dana pointed out that the shirts and jeans did not have
to be the same color, Michael did not make any changes to his own solution.
3.3.2 Third-Grade Problem Solving

In the third grade, the students were again given the shirts and jeans task. Stephanie
and Dana again worked together, and they immediately began to draw diagrams to
represent each item in the problem. Stephanie then suggested that they draw lines to
22
Fig. 3.5 Michaels grade 2

written work
show each combination, and they arrived at a total of six combinations. When questioned about why they drew lines to show the combinations, Stephanie explained
that that was to ensure that they do not make any duplicate combinations. Figure 3.6
shows Stephanies drawing with numbered lines to keep track of the outfits.
Dana, in her grade 3 drawing, again used a tree representation to form all shirts
and jeans outfits as indicated in Fig. 3.7.
Fig. 3.6 Stephanies grade 3

written work
23
Fig. 3.7 Danas grade 3

written work
Fig. 3.8 Michaels grade 3

written work
Michael worked on this task with another student, Jaime. This time, Michael
used the strategy of connecting lines between the shirts and jeans to represent the
possible outfits, but, unlike Stephanie and Dana, he drew lines between the words
in the problem, rather than between drawings of the shirts and jeans as indicated in
Fig. 3.8. For example, he drew a line between the word white and the word blue,
signifying an outfit of a white shirt and blue jeans. Using this strategy, Michael also
arrived at a solution of six combinations. He used a strategy similar to that used
by Stephanie in the second grade: he listed the combinations by writing the letter
representing the color shirt above the letter representing the color jeans.
3.4 Cognitive Implications and Differences Observed

In the second grade, none of the three students arrived at the correct number of
outfits, although the way they solved the problem gave evidence that they were
24
building schemes that could account for some or all of the outfits. All three students
drew pictures in order to model the shirts and jeans, and all three used notation (first
letter abbreviations) to indicate the colors of the shirts and jeans. Their representations of the problem included letters and line diagrams. It is important to note
that Dana showed evidence of building the scheme for controlling for variables in
the second grade. If not for her sense of style, Dana would have arrived at the correct answer of six outfits in the second grade. In the third grade, all three students
arrived at a correct solution and none recalled that their correct solution was different
from the solution they found in grade 2. This time, Stephanie offered a justification,
explaining that the lines that they drew between the shirts and the jeans ensured that
they accounted for all possibilities and also that they had not counted any combination more than once. Stephanie also used a system of counting that enabled her to
keep track of her outfits. The childrens representations show that they are aware of
the components of the problem and how they are put together.
3.5 Discussion
This task prepares the groundwork for future tasks in combinatorics. It invites students to bring forth personal representations and it offers opportunities for sense
making, so that students can begin to discuss how they arrived at their solutions and
how they know their solutions are correct. Also, different ways of reasoning can be
explored while students can learn how to formulate organizational schemes that can
help them solve other problems in mathematics. In addition, the real-world setting
of the problem shows the direct connection between the mathematical ideas and the
world that students know. It also gives students a chance to think about the influence
that real-world considerations have on mathematics, as can be seen from Danas
sense of fashion and her insistence that one combination of shirts and jeans cannot
constitute an outfit. Thinking about real-world considerations (which are important and which can be ignored) is a necessary step on the journey to mathematical
understanding and abstraction.
It is interesting to note the variety of strategies and approaches that were used in
this problem. For example, in second grade, the three children used three different
strategies to solve the problem. Although none of the strategies produced a correct
answer, it should be emphasized that arriving at the correct answer was not the goal.
In fact, when asked as third graders what answer they gave in grade 2, these students
all responded that they found six outfits.
For the researchers, the goal of the sessions was not for the children to give the
correct answer; we were confident that they would eventually succeed. Our primary
goal was to engage the children in thoughtful problem solving that could trigger
growth in schema. We wanted solutions to be meaningfully constructed by the students. As the following chapters will reveal, a pattern of working with the students
in the longitudinal study was to revisit problems and solutions in cycles so that earlier ideas could be built upon and new representations could be revealed. As we
25
studied the progress made by the students over the years, we gained insight into
their developing ideas and ways of reasoning. We observed the heuristics they used
and the schemes that were later retrieved and modified.
Over the months, there is evidence of durable learning. Also, there is evidence
that children learned from each other as revealed in elements of one anothers strategies reappearing in their second attempt at solving the problems. More importantly,
perhaps, is that each student incorporated strategies in unique ways. For example,
Michaels use of lines and notation to show combinations (see Fig. 3.8) differed
significantly from the way that Stephanie and Dana used lines and notation. This
suggests the importance of encouraging the use of students personal representations
in building solutions.
These episodes also demonstrate some benefits of group work. The contribution
of each student to the cumulative body of knowledge enables students to arrive at
their own solutions while simultaneously benefiting from the knowledge of others.
As the data indicate, the children built durable schemes to solve the mathematical tasks that they were given. They used personal representations, as was seen in
Danas and Stephanies second-grade work, to make meaning of the problem situation and to make an organized, systematic attempt to solve the problem. In addition,
the data show that the representations and arguments that were originally built by
individual students were used to effectively communicate to others the schemes
upon which they had been built. This can be inferred from the students work in
the third grade, when these schemes and representations were assimilated into the
solution strategies and representations of classmates.
In this chapter, we have seen that, at an early age and at the beginning of their
investigations, students worked to make sense of the mathematics; they readily
communicated their ideas and built on others ideas in forming their solutions. In
Chapter 4 , we follow two of these students Dana and Stephanie along with a
third student, Milin, as they work on more problems designed to help them explore
fundamental ideas in algebra and combinatorics.
Chapter 4
Towers: Schemes, Strategies, and Arguments

Date and Grade:

Tasks:
Participants:
Researchers:
19901992; Grades 3 and 4

Towers
Dana, Jeff, Michelle I., Milin, and Stephanie
Carolyn A. Maher and Amy M. Martino
4.1 Introduction
In the previous chapter, we examined the representations, strategies, and
problem-solving schemes used by four second- and third-grade students to build
their solution to the shirts and jean problem (which was to determine how many
outfits could be formed from three different shirts and two different pairs of jeans
and to provide a convincing argument of the solution). In their effort to make sense
of the components of the problem and to monitor their work, the students developed
various notations to represent the data and illustrated the use of certain strategies.
In this chapter, we examine how those students and others in the longitudinal study
build on those representations and strategies in their work on some towers problems. (A towers problem involves determining how many towers can be built of a
given height from a specified number of colors of Unifix cubes, small plastic cubes
that can be stacked together. Because Unifix cubes have a vertical orientation they
have a top and a bottom so do towers. An n-tall tower is one that was built from n
Unifix cubes. Appendix A provides an analysis of solutions to the towers problems.)
In this chapter, we examine the representations and strategies such as looking for
patterns, guess and check, and controlling for variables that were used by students
as they worked on the towers task. We trace students use of heuristics and ways of
reasoning that were exhibited in their earlier problem solving with shirts and jeans.
Finally, we trace the growth in students mathematical reasoning as their arguments
and solutions took on proof-like forms.
C.A. Maher (B)


27
28
C.A. Maher et al.
As students were introduced to new problems and worked to make sense of the
problem tasks, we observed growth in their knowledge as evidenced by the models
they built, the identification of new and more elaborate patterns, and the structure of
the arguments they provided in support of their solutions (Maher, 2002). Older ideas
were elaborated and expanded upon. Students active engagement in the problem
solving gave opportunity to build new ideas and methods of argumentation. As they
attempted to resolve issues that could not be solved with their existing schemes, new
schemes were built to accommodate the conditions of the problems. The structure of
the towers problem served as an assimilation paradigm (Davis, 1984) for students
later work with problems of similar structure, providing the students with a conceptualization that we see used in later years to tackle more complex combinatorial
problems.
4.2 Stephanie
We discuss here Stephanies emerging strategies as she worked on the towers
problem in the third and fourth grades.
4.2.1 Stephanie Grade 3, Class Session

The third-grade students in the study were asked to find all possible combinations
of four-tall towers that can be made when selecting from two colors (in this case,
red and blue). The strategies of a number of these students are documented and
discussed in Martino (1992). We present here a discussion of Stephanies strategies
as she worked on the task in the third grade (see Fig. 4.1).
Fig. 4.1 Dana (left) and

Stephanie (right), grade 3
tower exploration
29
Fig. 4.2 Stephanies four-tall

opposites
RED
BLUE
Stephanie and Dana began by working independently. Stephanie built ten towers.
She began by making a four-tall tower and then its opposite, that is, a new tower
of the same height with the second color in the corresponding position. Her first five
towers included two sets of opposites, such as the tower with four blue cubes and
its opposite, the tower with four red cubes (see Fig. 4.2).
Dana also initially built ten towers, including two pairs of opposites. Then the
two girls decided to combine efforts, and Stephanie took each of Danas towers in
turn and checked it against her own to see if it was a duplicate.
STEPHANIE:
DANA:
STEPHANIE:
Everything we make, we have to check. Everything we make. . .

Lets make a deal. Everything we make, we have to check.
All right. Ill always make it and youll always check it.
Okay, you make it and Ill check it.
When a duplicate was found, it was dismantled and returned to the pile of cubes.
After this process, Dana and Stephanie now had 14 tower combinations. Stephanie
suggested that Dana build new towers while she checked each new tower against
the existing ones to ensure that it was not a duplicate. They finally eliminated all
duplicates, and after attempting to find more combinations but not succeeding, they
concluded that there were only 16 combinations, since they had checked many times
and could not find new towers.
This activity was marked by a number of emerging strategies. First, Stephanie
and Dana used trial and error to find as many towers as they could. In addition, both
thought of finding a tower and its opposite in an attempt to generate as many towers
as possible, but neither used this strategy extensively or consistently. Further, the
two decided to compare results and eliminate duplicates, and ultimately used this
strategy of elimination to find the remaining tower combinations.
Stephanie and Danas attempt to prevent duplication of combinations as they
worked on the towers task is reminiscent of their strategy for solving the shirts and
jeans task in the second and third grades (see Chapter 3 for an in-depth discussion).
As they worked on the solution to that task, they used lines to ensure that they
counted each combination of clothing once and only once. They explained to the
researcher:
30
RESEARCHER:
STEPHANIE:
C.A. Maher et al.
What are these lines that you drew? You drew lines between the
shirts and the pants.
So that we could make sure; so instead of we didnt do that
again and say, Oh, that would be seven, eight, nine, 10. We
just drew lines so that we can count our lines and say, Oh we
cant do that again, we cant do that again.
As they worked on the towers task, Stephanie and Dana again were careful to
check each combination against the others to ensure that there were no duplicates.
RESEARCHER:
STEPHANIE:
How could you be sure that you havent made any of them twice
or that one of you got them all? Is there a way that you could be
sure?
Well, there is a way. You could take one, like say we could take
this one, this red with the blue on the bottom and we could go,
we could compare it to every one. And the ones that match that
dont match, put back; and the ones that do match, eliminate.
Stephanie and Dana were then asked to predict how many three-tall towers they
could build. Stephanie first predicted that there would be the same number 16;
and other groups predicted that there would be more three-tall towers than fourtall towers. Upon experimentation, they found that removing one cube from each of
their four-tall towers resulted in duplicates, or pairs, and they concluded that there
were only eight combinations of three-tall towers (see Fig. 4.3).
During an interview the next day, Stephanie explained why there were fewer
three-tall towers than four-tall towers.
RESEARCHER:
STEPHANIE:
RESEARCHER:
STEPHANIE:
What do you think you learned from what you did?

Well, we learned that . . . you might think thered be more
because there are less blocks so theres more combinations you
can make. Theres less because once you take one block off,
say you have red, red, red, red, and you have red, red, red, blue.
Once you take red, one red away and one blue away, theyre the
same.
Oh . . . So then you dont have more. You haveHave less.
Fig. 4.3 From four-tall to

three-tall
RED
BLUE
31
Stephanie and Dana began to make conjectures and inferences based on their
previous knowledge; both went on to suggest that there would be more five-tall
towers than four-tall towers.
4.2.2 Stephanie: Grade 4, Class Session

In the fourth grade, on February 6, 1992, the students were asked to find all possible
five-tall towers, selecting from two colors (in this case, yellow and red). Stephanie
and Dana began to make towers along with their opposites (see Fig. 4.4); and they
checked their work as they progressed to prevent duplication.
DANA:
STEPHANIE:
DANA:
STEPHANIE:
And then I got another idea.

Well, tell me it so I can do the opposite.
Im going to do the red this, thatShow me. Oh, okay, and Ill do the red and Ill do it with the red
at the top.
At one point, they realized that an individual tower could be turned upside down
to create a new tower. Dana called this new tower cousin (see Fig. 4.5). They used
this strategy to find more possible arrangements. After forming as many towers as
they could using this strategy of trial and error, they arrived at 32 different towers,
arranged in pairs with a tower and its opposite and a tower and its cousin.
Dana also considered different ways of arranging specified sets of towers that she
referred to as families. An example of Danas family is the elevator pattern consisting of exactly one red cube (see Fig. 4.6). In her discussion with the researcher,
Dana justified that there could only be five towers in this family because it only
goes up to five blocks. Her reasoning indicates an argument by contradiction of the
Fig. 4.4 Stephanie and

Danas group work
32
C.A. Maher et al.
Fig. 4.5 Danas tower and

cousin
RED
YELLOW
Fig. 4.6 Danas family of

one red cube and four yellow
cubes
RED
YELLOW
given condition that the towers should be five-tall. If another red were added, the
result would be a six-tall tower.
RESEARCHER:
DANA:
RESEARCHER:
DANA:
Are there any other members of this family?

No.
Why not?
Because it only goes up to five blocks.
Stephanie and Dana located other families of towers with exactly two red cubes
(see Fig. 4.7). Stephanie explained to the researcher:
With two [red cubes] together, you can make four. With one [yellow cube] in between, you
can make three. With two [yellow cubes] in between, you can make two. With three [yellow
cubes] in between, you can make one. But you cant make four in between or five in between
or . . . anything else because you can only use five blocks.
33
Fig. 4.7 Families of five-tall

towers with exactly two red
cubes
RED
YELLOW
By the close of this whole class activity, Stephanie and Dana had begun to explore
an exhaustive method of finding the combinations of towers that were five cubes tall.
This marked Stephanies first use of a partial argument by cases as she worked on
the towers task.
4.2.3 Stephanie: Grade 4, Interviews

Stephanies work on problems involving towers continued throughout the fourth
and fifth grades (see Maher & Martino, 1996a, 1996b, 1997) and again in grade 8
(see Chapter 7). Stephanies growth in understanding of the idea of a mathematical
proof is further documented by Martino and Maher (1999) and Maher and Speiser
(1997b). Data from these episodes are presented here with attention to the emergent
strategies that Stephanie used while working on the tower tasks.
In an interview following the class session described above, Stephanie, using
red and blue cubes, extended her family organizations of opposites, cousins, and
elevators, to include a new organization, the staircase pattern (see Fig. 4.8). She
discovered that introducing additional patterns sometimes resulted in duplicate towers that needed to be eliminated by checking. She said, Yeah, we kept we kept
finding different patterns, but we didnt check it with the other patterns.
Fig. 4.8 Stephanies use of

different patterns resulting in
duplicates
RED
BLUE
34
C.A. Maher et al.
The interviewer asked Stephanie if there was a way she could be sure of how
many towers of a specific type could be made.
STEPHANIE:
RESEARCHER:
STEPHANIE:
RESEARCHER:
STEPHANIE:
RESEARCHER:
STEPHANIE:
RESEARCHER:
STEPHANIE:
I guess . . . a very lucky guess.

Is there anything else possible for towers with exactly one blue?
No.
Why are you convinced?
Because if there are towers of five, you can only build that many
[with one blue cube]. You cant really be convinced for everything because theres no absolute way . . . you cant go and say
Im right.
[referring to the set of towers with one blue cube that Stephanie
had shown] Well, this is an absolute way.
Yeah, this is one of the absolute ways.
This absolute way is when you looked at only one blue and I
wonder if you could find absolute ways for looking at maybe
two blues, three blues, or four blues.
You could. Yeah, it is possible to have a certain number and get
it right.
With this exchange, Stephanie demonstrated that the elevator pattern provided a
convincing argument for justifying the number of towers with exactly one (or four)
of a color. She also seemed to consider that other organizations, such as exactly two
of a color, could be convincing. She began to consider families of towers as belonging to cases that could be justified individually to create the mutually exclusive and
exhaustive set of cases for building an argument for finding all five-tall towers.
In the latter part of the session, Stephanie used letters O and B to represent two
colors. She made a grid with rows and columns to represent different six-tall towers.
Notice, in Figs. 4.9 and 4.10, that Stephanie kept the entries in two rows constant, the
top two rows in Fig. 4.9 and the bottom two rows in Fig. 4.10. Notice, also, in both
figures Stephanie applied her elevator pattern while holding both rows constant.
Fig. 4.9 Towers with top two

rows constant
35
Fig. 4.10 Towers with

bottom two rows constant
BLACK
WHITE
Fig. 4.11 Stephanies global organization
In a subsequent interview, Stephanie shared with her classmates the strategy of

controlling for variables, that is, keeping the color of a cube in a particular position
constant. This method of controlling for variables was useful to her in keeping track
of larger number of towers (Maher & Martino, 1996a).
During an individual interview on March 6, 1992, Stephanie presented a complete argument by cases. She was able to produce a global organization for four-tall
towers. In her justification, she focused on number of white cubes yielding five categories of towers: towers with no white cubes, towers with exactly one white cube,
towers with exactly two white cubes, towers with exactly three white cubes, and
towers with exactly four white cubes (see Fig. 4.11).
4.3 Milin
4.3.1 Milin: Grade 4, Class Session
During the February 6 class session, Milin worked with Michael on the five-tall
towers task. Milins work has been referred to in earlier publications (Alston &
Maher, 2003; Maher & Martino, 1996a) and was analyzed in greater detail by Sran
(2010). Together with Michael, Milin began by using the strategy of building a tower
using trial and error and then making an opposite for each tower. Michael and Milin
36
C.A. Maher et al.
Fig. 4.12 Milins cousin and

opposite ways of making
pairs
RED
YELLOW
Fig. 4.13 Milins cases of

red cubes separated by one,
two, and three yellow cubes
RED
YELLOW
Fig. 4.14 Class discussion and sharing of solutions
always paired their towers. Milin sometimes used the opposite strategy to make
a pair and other times he utilized the cousin pairing, by inverting a tower. At the
conclusion of the group work, the boys found all 32 towers. The strategies used by
the two students on February 6, 1992, included: trial and error, building an opposite
tower to complete a pair by switching the color of each cube, building an opposite
tower to complete a pair by inverting the original tower, and monitoring work by
checking for duplicates by comparing to previous towers (see Fig. 4.12).
Milin noted that there were three possible combinations of towers in which the
red cubes were separated by one yellow cube, two in which the red cubes were
separated by two yellow cubes, and one in which the red cubes were separated by
three yellow cubes (see Fig. 4.13).
During the sharing session (see Fig. 4.14), the children were attentive as they
listened to the findings and strategies used by their classmates.
37
4.3.2 Milin: Grade 4, Interviews

In an interview on February 7, the day after the class session, Milin made sets of
towers using the elevator method for moving a cube of one color to each floor of
the tower and by moving two cubes of one color the same way. He then found the
remaining combinations by trial and error, and by grouping towers together with
their opposites. Although Milin believed that he had found all combinations, he was
only able to provide a convincing argument for his elevator patterns and his solid
towers (see Fig. 4.15).
Later during this interview, Milin began to consider simpler cases, and he said
that there were four towers that could be built that were two cubes tall, and two
that could be built that were one cube tall. Milin continued exploring simpler cases
after this interview and brought the cubes home to further explore his idea. During
the second interview 2 weeks later (on February 21), Milin reported that there were
16 four-tall towers. Later on in the interview, Milin showed towers that were one-,
two-, and three-cubes tall, and he recorded the number of combinations that were
possible for each (see Fig. 4.16).
Fig. 4.15 Milins partial
organization by cases and
opposites
RED
Fig. 4.16 Milins one-, two-,

and three-cube tall towers
YELLOW
38
C.A. Maher et al.
Fig. 4.17 Milins inductive

reasoning with families
BLACK
LIGHT BLUE
Then, in a third interview (on March 6), he showed that larger towers could be
built from smaller ones. For example, one can build four two-tall towers from the
two one-tall towers (a blue cube or a black cube) by placing either a blue or a black
cube on the blue cube and then placing either a blue or a black cube on the black
cube. Milin showed that groups of larger towers could be included in the family of
the smaller tower from which it was built (see Fig. 4.17).
Later during the March 6 interview, Milin suggested that his rule for generating
taller towers from shorter towers breaks down after five-tall towers. Toward the end
of this interview, he retracted this claim, and he suggested that there were 64 possible
combinations of six-tall towers. When asked if his pattern would hold for towers
taller than five, he said it should, indicating, We followed the pattern till five. Why
cant it follow the pattern to six?
4.3.3 Small Group Interview: March 10, 1992 Grade 4

Three weeks later, in a small group interview, arguments were presented by a group
of four children for accounting for all possible towers, three-tall, selecting from two
colors. This group sharing is referred to as The Gang of Four (see Fig. 4.18). It
was conducted so that the children could share their strategies and arguments for
Fig. 4.18 Milin, Michelle, Jeff, and Stephanie (left to right)
39
building towers of a variety of heights in earlier investigations. A simpler version of

the problem was chosen deliberately for this session, as the evaluation was intended
to identify the forms of reasoning and methods of justification that the children used
to convince themselves and one another of the validity of their solutions (Maher &
Martino, 1996b).
The session began with the researcher asking the students how many sixtall towers could be built. Milin answered, probably 64. He was asked to
explain why, and he described his inductive rule: multiply the previous answer by
two.
MILIN:
RESEARCHER:
MILIN:
RESEARCHER:
MILIN:
Well, because there was a pattern.

Whats that?
You just times them by two.
Times what by two?
The towers by two, because one is two, and then we figured out
two is two, and then, I mean four, and then-
Milin used inductive reasoning to justify his solution, extending the problem
beyond the three-tall case given to the group. He said that there were two one-tall
towers, four two-tall towers, and eight three-tall towers. He was asked to re-explain
how he progressed from four to eight towers. In this clip, he noted that a cube of
each color could be added to the top of the shorter tower to build the taller tower.
RESEARCHER:
MILIN:
RESEARCHER:
MILIN:
RESEARCHER:
MILIN:
RESEARCHER:
MILIN:
RESEARCHER:
MILIN:
Why eight? Thats what Jeffrey asked about.

I know.
Go ahead. Let Milin persuade Jeff.
If you do that, you just have to add for each one of those you
have to addEach one of what? These four?
Yeah. You have to add one more color for each one.
Which way are you adding it? Where are you putting that one
more color, Milin?
No, two more colors for each one. SeeSo this one with red on the bottom and blue on the top.
You could put another blue or another red.
Later, Milin explained the logic behind the leap from two-tall towers to three-tall
towers using inductive reasoning. He was able to demonstrate his doubling rule with
each individual tower. In his own words, he noted that there were two possible cubes
to be added to each three-tall tower to make a four-tall tower: This was for three, so
you could add two for each one of the three. Milin explained this doubling rule by
drawing two three-tall towers from a two-tall tower by adding a different color cube
on top. He took his first two-tall tower with a blue cube on the bottom floor and a
red cube on the top floor and generated two three-tall towers by first adding a red
cube on the third floor and then adding a blue cube on the third floor (see Fig. 4.19).
It is interesting to notice how Milin chose to draw the towers rather than use actual
cubes as he had during his individual interview.
40
C.A. Maher et al.
Fig. 4.19 Milins

representation showing his
doubling rule
Fig. 4.20 Stephanies

representation of an argument
by cases
In this session, Stephanie presented an argument by cases to account for building

all possible towers, three-tall. She represented the towers in a grid using letters B
and R, for blue and red cubes as indicated in Fig. 4.20. Details of the session are
described in Maher and Martino (1996b). She showed that there was only one way to
form a tower without any blues. Then she showed that there were three combinations
of two red cubes and one blue cube using the staircase pattern.
She then used an argument by contradiction to show that this pattern could be
used to show that there were only three possible combinations. Stephanie said,
Well, theres no, theres no more of these because if you had to go down another one youd
have to have another block on the bottom. But then you have with three blues well, not
with three blues. Ill go like this.
Stephanie used the staircase pattern to argue by contradiction that there could not
be a fourth arrangement of two red cubes and one blue cube. What is of interest here
is that Stephanie felt the need to prove that her argument by cases was complete and
convincing, even though no one had challenged her answer. Stephanie continued her
argument by cases by describing all the possible combinations of two blue cubes and
one red cube. Stephanies organization was interesting in that she separated the case
of two blue cubes into sub cases: two adjacent blue cubes and two non-adjacent
blue cubes. When her classmates pointed out that these two cases could be grouped
into one broader group, Stephanie insisted on continuing her explanation as she had
originally presented it. The entire conversation follows, starting with Stephanies
description of the all red tower.
STEPHANIE:
RESEARCHER:
STEPHANIE:
RESEARCHER:
All right, first you have without any blues, which is red, red, red.
Okay, no blues.
Then you have with one blue
Okay.
STEPHANIE:
MILIN:
STEPHANIE:
RESEARCHER:
JEFF:
MILIN:
JEFF:
STEPHANIE:
RESEARCHER:
STEPHANIE:
RESEARCHER:
STEPHANIE:
RESEARCHER:
STEPHANIE:
MILIN:
RESEARCHER:
MICHELLE:
STEPHANIE:
RESEARCHER:
STEPHANIE:
RESEARCHER:
STEPHANIE:
41
Blue, red, red; or red, blue, red; or red, red, blue. All right. You
could put blue, blue, red; you could put red, blue, and blue.
You could put blue, red, and blue. You could put . . .
Yeah, but thats not what I am doing. Im doing it so that theyre
stuck together.
Okay.
There should be one there could be one with one red and then
you could break it up and theres one with two reds and theres
one with three reds and then . . .
Ah, but see you did the same thing, but theres the blue.
See, theres all reds and theres three reds, two reds. There
should be one with one red. And then you change it to blue.
Well, thats not how I do it.
Lets hear how Steph well hear that other way; thats
interesting. Okay, now, so what youve done so far is
One blue, two blue.
Okay, no blues
One blue, two blue.
One blue, and two blues, but Milin just said you dont have all
two blues, and you said that why is that?
All right, so show me another two blues. With them stuck
together, because thats what I am doing.
In that case, no.
Okay, so now what are you doing, Stephanie?
What if you just had two blues and they werent stuck together,
you could
But thats what Im doing. Im doing the blues stuck together.
Okay.
Then we have three blues, which you can only make one of.
Then you want two blues stuck apart not stuck apart; took
apart.
Separated?
Yeah, separated. And you can go blue, red, blue right here.
Although Stephanie insisted on explaining her method of using two categories

of towers with two blue cubes during this session, she later indicated (in a written
assessment) that she understood the arguments of Milin and Michelle. At that time,
Stephanie organized her cases as her classmates had suggested, producing a more
elegant proof by cases (Maher & Martino, 1996a).
Toward the end of the session, the students used Milins argument by induction as
a stepping-stone to generalize the solution to the towers problem. Their progression
of this understanding is documented by Maher and Martino (1997, 2000) and Maher
(1998).
42
C.A. Maher et al.
Stephanie, during interviews preceding The Gang of Four, noticed a relationship between the height of towers and the total number produced and conjectured
a doubling rule. During the Gang of Four session, she made reference to her
doubling pattern and offered that there would be 1,024 ten-tall towers that could be
built selecting from cubes of two colors. However, during that session, she chose
to justify her solution to the three-tall tower problem with an argument by cases
(Maher & Martino, 2000). In Chapter 5, while working on another problem, Guess
My Tower, we see Stephanie learn why the doubling rule works as she investigates
Milins inductive argument.
4.4 Summary of Strategies and Justifications

Figure 4.21 outlines the strategies, representations, and forms of justification used
by Stephanie and Milin during the five sessions on the towers problems. Both
Stephanie and Milin began by using trial and error and justifying their solution
empirically. They both progressed to more sophisticated strategies and forms of justification. Stephanie looked for patterns and controlled for variables to eventually
formulate her justification using cases. Milin considered simpler cases and then recognized the recursive nature of the problem, arriving at his inductive justification.
Both Milin and Stephanie arrived at a complete justification of their solution during
The Gang of Four session. In addition, both students chose not to use the Unifix
cubes to represent their towers but instead used notations in a grid (Stephanie) and
drawings (Milin) to represent the different tower combinations.
STEPHANIE
Strategies
Class
Session
Interview 1
Interview 2
Interview 3
Gang of
Four
Tools
MILIN
Justification
Trial and error;

Unifix cubes Patterns,
Opposites; cousins,
partial cases
Staircase
Partial cases
Drawings
Pattern recognition
and symbols
Controlled for
variables
Staircase;
Controlled for
variables
Staircase,
Controlled for
variables; Pattern
recognition
Strategies
Drawings
and symbols
Drawings
and symbols
Emergent
cases
Trial and error;

Opposites;
Staircase
Patterns; partial
cases; simpler
problem
Considered
simpler cases
Inductive pattern
recognition
Grid with
symbols
Case
argument
Inductive pattern
recognition
Partial cases
Tools
Justification
Unifix cubes Partial cases
Drawing and Inductive argument

symbols
Unifix cubes Partial Cases

Unifix cubes Partial induction
Unifix cubes Emergent induction
Fig. 4.21 Strategies, representations, and justifications used by Stephanie and Milin
4.5 Discussion
The Gang of Four session evidenced particular structures and modes of reasoning
by Milin and Stephanie in their justification of their solutions to the towers task. The
students built and refined their representations over a period of time in which they
43
had the chance to reflect upon the task, recognize emergent patterns, and choose
schemes that best matched the representation that they had formed. Stephanie used
symbols within a matrix to organize the towers by cases; Milin used drawings of
towers to explain how they grew. The call for justification of the three-tall tower
task enabled Stephanie and Milin to make public the schemes that they had built
earlier.
There are some similarities in Stephanies early use of representations for both
the towers task and shirts and jeans tasks. In the second grade, Stephanie listed the
outfit combinations by using the initials of each color and recording the combinations in a vertical format (see Fig. 3.2). When working on the towers task, she again
used initials for the colors of cubes using a grid organization to show the different
towers. Stephanie also used the heuristic of controlling for variables as she organized her tower combinations, a strategy that her partner Dana had used in the shirts
and jeans task.
Milins strategies of considering simpler cases and pattern recognition were powerful tools in his building of an inductive argument. As will be seen in Chapter 5,
both students schemes proved durable as Stephanie and her classmates folded back
to reflect on their earlier work to make sense of more complex combinatorial tasks
in later grades.
Importantly, these data show the advantage to revisiting tasks, group discussions
about ideas, and sharing strategies. All of these components play a key role in the
formulation and refinement of justifications. Stephanie and Milin, after having had
multiple opportunities to think about and justify their ideas, presented a compelling
argument to classmates during the group evaluation setting. As is evidenced in later
years, unique aspects of the discussions that continued among this community of
learners further triggered the development of more complex cognitive structures,
triggered by the students need to produce justifications for combinatorial tasks of
ever-increasing complexity.
In Chapter 5, we follow Stephanie and other classmates as they continue to work
on understanding Milins inductive argument for building towers as they retrieve
earlier frames and cognitive structures revealed during the Gang of Four work.
Chapter 5
Building an Inductive Argument

Date and Grade:

Task:
Participants:
Researchers:
February 26, 1993; Grade 5

Guess My Tower
Bobby (Bobby is called Robert in later chapters), Matt, Michelle
I., Michelle R., Milin, and Stephanie
Carolyn Maher, Alice Alston, and Amy Martino
5.1 Introduction
In the previous chapters, we followed the strategies, schemes, and arguments built
by second-, third-, and fourth-grade students as they worked on combinatorial tasks.
In this chapter, we trace how Stephanie and her classmates tried to make sense of
the inductive method of generating towers. This strategy was originated by Milin,
but it was eventually adopted by many other students. We attempt to identify the
moments at which individual students gained ownership of the inductive argument
and explained their new understanding to others.
5.2 Early Ideas

In third and fourth grades, the children continued to build powerful strategies and
schemes as they worked on the tower problems (see Chapter 4). To support their
solutions, students followed two different approaches. Stephanie and others made
extensive use of argument by cases. Milin, over a series of task-based interviews,
built an inductive argument and he was able to use an inductive argument to show
how to generate the number of combinations of towers of any height. Stephanie also
C.A. Maher (B)



45
46
C.A. Maher et al.
observed this doubling rule (that the number of towers of height n was double the
number of towers of height n 1). Stephanie conjectured a doubling rule from her
successful case-building justifications of towers of different heights.
5.2.1 Stephanies Individual Interview: May 15, 1992

On May 15, 1992, Stephanie participated in another interview. When asked whether
she had thought about the problem since the small-group session, she showed
the interviewer a sheet of paper on which she recorded the doubling method of
finding towers of a specific height. The interviewer introduced the idea of using
a tree diagram to show a recursive pattern that could be used to generate towers
and to organize the tower combinations, similar to Milins inductive scheme. The
researcher showed Stephanie the first two levels of the tree diagram and then asked
Stephanie to extend the tree to include three-tall and four-tall towers. Stephanie
responded by producing a partial extension of the tree organization as indicated in
Fig. 5.1.
5.2.2 Written Assessments for Stephanie and Milin: June 15, 1992
At the end of the fourth grade on June 15, 1992, the children participated in a written
assessment (see Fig. 5.2).
The children worked in pairs on this assessment to provide a convincing
justification of the towers task. Stephanie and Milin were partners and provided
individual written work of their solution. (See Fig. 5.3 for Stephanies written work
and Fig. 5.4 for the work of Milin.) In her letter, Stephanie gave an elegant argument
by cases to show that she found all the towers, and then used a doubling pattern to
predict taller towers, offering a general method. She said, All you have to [do] is
find the no. [number] for the problem before and mulity [sic] by 2.
Notice that the representations included ideas from both Stephanie and Milin,
with the generation of numbers of towers as the height increases. Milin made a grid
of towers three-tall using letters B and G to represent the two colors. Then he
paired the two sets of opposites and two sets of cousins (see Fig. 5.4). He also
demonstrated an understanding of the doubling rule in his written work. Notice, too,
Fig. 5.1 Stephanies partial

extension of the tree diagram
Fig. 5.2 Assessment June 15, 1992
Fig. 5.3 Stephanies June 15, 1992, written assessment
47
48
C.A. Maher et al.
Fig. 5.4 Milins June 15, 1992, written assessment
on page 4 of both childrens work that they were considering the building of threetall towers selecting from three colors. Stephanie showed some of those towers;
Milin wrote in the lower corner 3 1 = 3 and beneath it, he wrote 3 3 =.
5.2.3 Written Assessments for Stephanie and Milin:

October 25, 1992
At the beginning of the fifth grade, individual assessments were given on October
25. The children worked alone and produced individual written work (Fig. 5.5).
In a follow-up individual written assessment at the beginning of fifth grade,
Stephanie again used an argument by cases, and she checked her solution by using
the doubling method (see Fig. 5.6). Its interesting that again Stephanie relied on
Fig. 5.5 Assessment October

25, 1992
49
Fig. 5.6 Stephanies October

25 written assessment
her case argument to justify her solution. (Refer to Maher & Martino, 1996a, for
further details.)
Milin, in his October 25 letter, again explained his doubling rule and drew all
eight three-tall towers in a grid. He then also included the one- and two-tall towers
(see Fig. 5.7).
Fig. 5.7 Milins written response to October 25, 1992, assessment
50
C.A. Maher et al.
5.3 Investigating Inductive Reasoning

Further activities with towers led several students to make advances in understanding of inductive reasoning previously introduced by Milin to some classmates. We
discuss below the traveling of ideas within a small community of students, initiated
first, by Milin; then, from Milin to Michelle I.; then from Michelle I. to Matt; then
from Matt to Stephanie, Bobby, and Michelle R.; and then from Stephanie to the
entire group. These ideas were triggered by Milins inductive argument.
5.3.1 Stephanie and Matts Beginning Exploration

Later in the fifth grade year (on February 26, 1993), the children worked on a problem called Guess My Tower, in which they were asked to decide what kind of
tower was most likely to be selected at random from a box containing first all
possible three-tall towers and then all possible four-tall towers (see Appendix A).
The problem called for a revisiting of tower building, enabling the researchers to
monitor the durability of students strategies and arguments. Since solving the Guess
My Tower problem required the building of a sample space for all possible events, it
required that the students revisit the question of finding the total number of four-tall
towers. We have seen that earlier, Stephanie had already built and justified by cases
all possible four-tall towers and used the doubling method for determining the total
number of towers. She was exposed to Milins inductive argument during previous
sessions on March 10 and June 15.
Stephanie and Matt worked together on this problem first, using paper and pencil
and then by building actual towers using Unifix cubes. They found all eight towers
(see Fig. 5.8).
The researcher asked Stephanie and Matt to predict how many four-tall towers
they would find. Stephanie remembered the pattern that she had noticed the previous
year. She said, Oh, I remember the way that you could make sure how many. It was
Fig. 5.8 Stephanie and

Matts towers
51
whatever number you got from the last one, you multiply by two, and then you get
the number of how many there will be for the next one. Stephanie predicted, using
the doubling rule, 16 four-tall towers, but they were only able, using a trial and error
strategy, to find 12 towers. Stephanie insisted that there should be 16 based upon her
confidence in the doubling rule. She shared with Matt the doubling pattern.
STEPHANIE:
MATT:
STEPHANIE:
Well, a couple of us figured out a theory because we used to see

a pattern forming. If you multiply the last problem by two, you
get the answer for the next problem. But you have to get all the
answers. See, this didnt work out because we dont have all the
answers here.
I thought we did.
No. I mean all the answers, all the answers we can get . . . I dont
know what happened! Because I am positive it works. I have my
papers at home that say it works. I know that you had to multiply
it [the total number of towers of a given height] by something.
Maybe it wasnt two because I know it worked. Maybe it was
adding two.
Stephanie and Matt continued to attempt to find more four-tall towers, but they
could not find more than 12. Stephanie continued to assert that the total would have
to be 16.
STEPHANIE:
MATT:
STEPHANIE:
I dont know how it worked. I know it worked. I just dont know

how to prove it because Im stumped.
Steph! Maybe it didnt work!
Oh no. No. Because Im pretty sure it would . . . I think we goofed
because Im still sticking with my two thing. Im convinced that I
goofed, that I messed up because I know that . . .
Stephanies memory of the doubling pattern and the fact that it had been used to
verify her solution to the towers problem in the past was sufficient for her to remain
convinced of its validity even when faced with a discrepancy in their own results.
5.3.2 Milins Explanation and Michelles Aha!

As Stephanie and Matt worked to find more towers, Milin attempted to convince
his partner Michelle I. that the inductive method of generating towers was a valid
way of accounting for the total number of towers of a given height. Recall that
both Milin and Michelle I. participated in the Gang of Four session almost a year
earlier on March 10, 1992, in which Stephanie gave her argument by cases and Milin
introduced his inductive argument. (Refer to Chapter 4 for details.) Michelle I. was
introduced to Milins argument the previous year in the group discussion and gave
evidence of some understanding of the doubling pattern introduced in the session.
In fact, building from Stephanies organization of towers by cases, Michelle placed
2 above each representation of a tower entry from each of the three-tall towers, to
52
C.A. Maher et al.
Fig. 5.9 Milin explaining inductive reasoning to Michelle I.
Fig. 5.10 Michelle I. continues Milins argument to four-tall towers
indicate that two new four-tall towers can be generated from each. In this session,
however, Michelle told Milin that she did not understand his explanation of how
he generated new towers using an inductive argument. Milin then explained his
reasoning for a second time (see Fig. 5.9).
Once Milin finished explaining his reasoning for building three-tall towers by
adding a red cube or a yellow cube to the tower from the previous stage, Michelle I.
interrupted Milins explanation, extending his reasoning to towers four cubes tall.
She explained the method to the researcher, commenting, This is a lot simpler,
from the last time we explained it. Michelles explanation demonstrated that she
had come to her own understanding of Milins method and was able to extrapolate
the number of towers four-tall using her solution to three-tall towers (Fig. 5.10).
Michelle also shared her understanding of the inductive method with Stephanie
and Matt (see Fig. 5.11).
5.3.3 Matts Explanation and Stephanies Aha!

Matt seemed immediately to grasp Michelles explanation of Milins inductive generation of towers. As Michelle I. continued to generate taller towers, Matt joined in
the explanation, commenting and so its a family tree (see Fig. 5.12). Matt then
proposed that Milins method might be connected with the doubling pattern that
53
Fig. 5.11 Michelle I. shares her understanding with Stephanie and Matt
Fig. 5.12 Michelle I. and Matts family tree
Stephanie had been using. The researcher suggested that Matt be given an opportunity to explain to the group how Milin and Michelle I. had generated towers.
Matt eagerly complied, showing that one could find the total number of towers of
any height (see Fig. 5.13), by using a tree diagram to build towers in an organized
fashion, explaining.
Fig. 5.13 Matt explaining to

Bobby and Michelle R. the
family tree
54
C.A. Maher et al.
Stephanie and Matt moved on to talk to Bobby and Michelle R., who showed
them that they, also, had found 16 towers. Stephanie commented, When we multiplied it out we got 16. . .. But we werent able to find that many. We were only able
to find . . . like 12. After Matt, joining Michelle I., had demonstrated his understanding of the inductive pattern, the researcher asked Stephanie to explain Matts
reasoning to Michelle R. who still was not convinced about the inductive method.
RESEARCHER:
STEPHANIE:
RESEARCHER:
I want to know how you are going to get to two-high.

Okay. Once theres no more, theres absolutely, positively no
more, you cant build any more with one. So you go to the next
number. And the next number is two. Okay? So you have four
of two.
Thats a big jump for me, Stephanie. Youre jumping too fast
from four to two. I dont know how they came. I dont know
how they grew.
Stephanie, who did not provide an explanation of how the towers were growing
from one- to two-tall, was interrupted by Matt who nudged her aside, saying, move
over and began to explain how the towers were growing. As Matt explained to
Michelle R. and Bobby, an attentive Stephanie looked on.
MATT:
RESEARCHER:
MATT:
RESEARCHER:
MATT:
All right. Now, from here, you add an opposite, an opposite, an

opposite or the same color on. So then you add the yellow and
the red on to the last one. So you have . . .
What do you think of that, Michelle and Bobby? . . . Do you
understand what he is saying?
. . . So you have a red on the bottom . . .. You add a red or a
yellow on top. You have the same yellow on the bottom, but
you add a red or a yellow on top. Then you have . . .
Is that all you can do?
Thats it. And then, well, for this you have red, yellow, yellow,
red, like that.
As Matt finishes explaining the generation of three-tall towers from the two-tall
towers, Stephanie joins Matt and explains how two four-tall towers can be generated
from each three-tall tower. A confident and elated Stephanie (see Fig. 5.14) offers
to explain their method to the entire class declaring
Yes! I knew it! I knew it! I knew it! . . . I told him all along, I was right . . .
Stephanie demonstrated her understanding using the cubes to build four-tall

towers and, counting the towers, commented
STEPHANIE:
RESEARCHER:
STEPHANIE:
I understand. Im just very happy that my rule worked.

Your rule worked. But what . . . you know what I think is really
valuable . . . for people to understand is to know why that rule
worked.
Well, I know what it is now. I, I figured it out! But Im just
happy that it worked.
55
Fig. 5.14 I knew it!
RESEARCHER:
STEPHANIE:
Because you see how you can forget a rule, but if you know why
it worked . . .
Yeah, yeah.
5.3.4 Stephanies Sharing Milins Family Tree

Finally, during a whole class discussion, Stephanie confidently explained the
reasoning behind her doubling pattern to her classmates as shown in Fig. 5.15.
STEPHANIE:
I have one red, okay? And I have a yellow and from each of these
you can make two because all you have to do is you add on . . .
you can add on a red to a red and a yellow to a red . . . and for
the yellow you can add on a red to the yellow and a yellow to the
yellow, okay?
Fig. 5.15 Stephanie explains

to the group
56
C.A. Maher et al.
MICHELLE I.:
STEPHANIE:
ALL:
STEPHANIE:
So you dont have to look for duplicates.

Then each one of these has two, like, okay? If this is like a family
tree . . . the mother, the parents . . . the mother, the parents . . . and
then six kids, okay? Well, no. Actually eight kids . . . then they
have eight kids and each one of them has two kids. And this one,
you can add one red, one yellow, one, yellow, one red, one red,
one yellow . . . .
And on, and on, and on, and on.
Cause each one of them is different . . . you keep adding on. And
then here you can add the exact same pattern.
5.4 Discussion
Argument
Other
Participants
Primary
Participant
Interaction
Type
Date
Event
Throughout the session, the students had many opportunities to reconstruct earlier
ideas and share them with others. They had occasion to revisit earlier ideas and
they were encouraged to explain and re-explain their arguments. Communication of
ideas, encouraged by the researcher, was the students responsibility. Clearly the
students took ownership and learned together and from each other. Figure 5.16
Michelle I.,
Stephanie, Jeff
Inductive
Milin, Stephanie,
Jeff
Milin, Michelle I.,
Jeff
Extension of Stephanies cases to doubling
3/10/1992
SG (E)
Milin
3/10/1992
SG (E)
Michelle I.
3/10/1992
SG (E)
Stephanie
4
5
6
7
8
5/15/1992
6/15/1992
RT (I)
WA
Stephanie
Milin
6/15/1992
10/25/1992
10/25/1992
WA
WA
WA
Stephanie
Milin
Stephanie
9
10
11
2/26/1993
2/26/1993
2/26/1993
PT
PT, RT
SG
Milin
Michelle I.
Michelle I.
12
2/26/1993
SG
Matt
13
2/26/1993
SG
Stephanie
14
15
2/26/1993
PT
Stephanie
Connects inductive pattern to doubling
2/26/1993
WC
Stephanie
Presents inductive pattern to group
Michelle I.
Milin
Milin, Matt,
Stephanie
Stephanie, Bobby,
Michelle R.
Michelle R. and
Bobby
Argument by cases, noting of the doubling

pattern
Doubling pattern; organization by cases
Doubling rule
Cases; doubling; 3 colors
Induction, doubling,
Organization by cases, verification using
the doubling pattern
Explains inductive pattern
Illustrates inductive pattern
Presents inductive pattern
Presents inductive pattern
Leads presentation of inductive pattern
Fig. 5.16 The building of an inductive argument (PT: partner talk; SG: small-group discussion; RT: researcher talk; WC: whole class discussion; I: interview; E: evaluation; WA: written
assessment)
57
shows a trace by which the members of the group built an understanding of the
inductive argument for the growth of the towers.
Gaining ownership of a mathematical idea involves a process by which the
learner takes responsibility for knowing. Faced with a conflict in ones understanding, a learner can work with others to express what may be clear or unclear in
their understanding so that an obstacle can be resolved. Faced with a conflict, students can be motivated to find resolution. Value added is that there is personal
gain and confidence in ones ability as a problem solver and achieve understanding. Children, given the opportunity to share ideas, can contribute to each others
growth in understanding.
In this chapter we have shown how sharing and discussion helped students attain
ownership of mathematical ideas. In the next chapter, we show how students made
use of the strategies developed here in solving new problems in counting.
Chapter 6
Making Pizzas: Reasoning by Cases

and by Recursion
Date and Grade:

Tasks:
Participants:
Researcher:
March and April 1993; Grade 5

Pizzas with halves; Pizzas
Amy-Lynn, Ankur, Bobby (Robert), Brian, Jeff, Matt, Mike,
Michelle I., Michelle R., Milin, Romina, and Stephanie.
Carolyn Maher
6.1 Introduction
In previous chapters, we followed the students in Kenilworth as they worked on the
shirts and jeans task and the towers problems in the second through fifth grades. In
this chapter, we discuss five sessions during which these students worked to make
sense of the pizza problems. These problems presented new challenges and required
the students to adapt the representations and solution strategies that they had previously formed to meet the needs of these tasks. (Portions of the data analyzed here
are described in Bellisio (1999), Muter (1999), and Tarlow (2004).)
As we trace the students problem-solving attempts, we will identify the forms of
justification and reasoning that were used by the students, the methods of notation
that they used, and the heuristics and strategies that were developed as the students
worked to resolve the complexity of the problems.
During the first four sessions, the students worked on the pizza with halves
task:
A local pizza shop has asked us to help them design a form to keep track of certain pizza
sales. Their standard plain pizza contains cheese. On this cheese pizza, one or two toppings could be added to either half of the plain pizza or the whole pie. How many choices
do customers have if they could choose from two different toppings (sausage and pepperoni)
C.A. Maher (B)



59
60
C.A. Maher et al.

that could be placed on either the whole pizza or half of a cheese pizza? List all possibilities. Show your plan for determining these choices. Convince us that you have accounted
for all possibilities and there could be no more.
During the fifth session, the students were presented with three related tasks; they
used the strategies that had been developed for the more complex pizza with halves
task to tackle these variations.
6.2 First Session: Initial Pizza Explorations

During the first session, the students worked in two groups as they made their first
pizza explorations. We will trace the discussion of the two groups separately, though
they worked simultaneously on the task.
6.2.1 Group 1
Five students (Jeff, Matt, Michelle I., Milin, and Stephanie) worked together.
They started by drawing pictures to represent the pizza combinations. They settled their differences in notation as they began to talk about what their drawings
represented.
STEPHANIE:
MATT:
MICHELLE I.:
MATT:
JEFF:
MICHELLE I.:
STEPHANIE:
MATT:
MICHELLE I.:
MATT:
JEFF:
STEPHANIE:
MATT:
MICHELLE I.:
P equals cheese.
S is sausage.
Why dont you just put P for plain?
P is plain.
P is pepperoni.
Plain is cheese.
Hold on! Nobody is explaining this to me. What is PE?
Pepperoni.
Or just regular P would be pepperoni.
Okay, P is pepperoni.
One pie, all cheese. [Jeff draws a circle and writes C in it.]
What is half a pie?
HP?
No, were just drawing.
Just as Stephanie and Dana had discussed their method of notation with each
other as they worked on the shirts and jeans task in the second grade (see Chapter 3),
the fifth-grade students were careful to make sure that their notations matched before
they began to discuss their findings. This group of students used notation as well as
drawings to identify their pizza combinations. Matts suggestion to use a notation
of HP to signify half a pie signaled the need to clarify what would be symbolized
using letters and what would be drawn.
Making Pizzas: Reasoning by Cases and by Recursion
61
In addition to discussing their method of notation, they discussed which pizza

combinations were distinct and which were identical and could only be listed
once.
MICHELLE I.:
JEFF:
MICHELLE I.:
JEFF:
STEPHANIE:
Then we can switch the side so that the sausageNo, its not going to matter which side. Its going to be the same
pie no matter what side sausage is on
No, its a different side, though. Its the same pie but its a
different side.
Nobodys going to care if its this or if its like this. [He points to
two circles, one with c/s and the other with s/c.]
Nobodys going to call up and say I want the pepperoni on the
left side of the pizza.
Stephanies real-world contextualization of the problem settled the argument,

and all the students agreed that the two combinations (sausage/cheese and
cheese/sausage) were not different in practice.
The students worked to find and justify their pizza combinations. After some
discussion, four of the five agreed that there were ten possible pizzas, while Milin
was still unsure that they had found all combinations.
The group of students discussed their ideas with each other and looked
at their own work. Matt found that he had named one pizza S and another
one CS but both represented the same pizza, since all pizzas contained cheese
(see Fig. 6.1).
Matt and Stephanie then discussed how they should organize their list of pizzas. Matt made a chart with C, P, and S as headings, listed all the possible pizza
combinations for each column, and then eliminated the duplicate combinations as
indicated in Fig. 6.2. Stephanie then showed him a different way of organizing
pizzas:
You just show how many different combinations you can make with like two little things,
three little things, four little things . . . So you put the one toppings here, all the ones with
one topping on top. And then the ones with two toppings you put those in the second row,
and the ones with three toppings, you put them in the third row.
Fig. 6.1 Matts first picture

of half pizzas
62
C.A. Maher et al.
Fig. 6.2 Matts organized list

of pizza combinations
Stephanies method of organization is of note here. When Stephanie worked on

the towers problem (see Chapters 4 and 5), she organized the towers by cases. She
first considered towers containing no cubes of a color, then one of that color, then
two of that color, three of that color, and so on. Now when she began to organize
her pizza combinations, she did the same, first considering pizzas with only one
topping, then those with two toppings, and finally those with three toppings.
At the close of this session, the five students were convinced of their solution;
they had begun to systematize their combinations and discuss their differences in
organization. We will now contrast the work of the second group of students and
identify key differences in their strategies and representations.
6.2.2 Group 2
Amy-Lynn, Ankur, Bobby (Robert), Brian, Mike, Michelle R., and Romina were
in the second group. As the students began their investigation, Ankur, assisted by
Brian, took a leading role. He verbally listed six possibilities; when asked to show
that he had all the possibilities, he directed the others to write down the choices as he
dictated them. Five of the others worked with Ankur and also contributed by calling
out possible combinations. Each student wrote the list of pizzas that was compiled
by the group (see Fig. 6.3).
Ankurs list, like the lists of the other members of the group, used words to
describe the pizza combinations. However, his drawings at the end of the list showed
Fig. 6.3 Ankurs written

work on March 1, 1993
63
his move toward using letters to symbolize toppings. Of note as well is his division
of the pizzas into halves as well as fourths, which may have been a factor that
resulted in the students confusion about the total number of combinations.
Mike worked alone. From the start, he used drawings of circles to represent pizzas, with small circles to represent pepperoni and wavy lines to represent sausage
(see Fig. 6.4). He did not use any letters to symbolize his toppings. This method
may have contributed to his confusion and duplication of pizzas.
After a few minutes, the students decided to discuss the possibility of having
more than one topping on half of a pizza. They then started to make lists together,
as the students listed pizzas aloud and wrote them down, and Mike drew another set
of pizza pictures, as Amy-Lynn labeled them (see Fig. 6.5).
After this phase, some students had a list of 11 pizzas, while Brian conjectured that there were 12. Amy-Lynn found that there was a duplicate in her list of
11 pizzas and pointed out to Mike that he also had one repeated combination.
ROMINA:
ANKUR:
AMY-LYNN:
BRIAN:
Its the same thing if you turn it around.

No its not because of, half of it is mixed and the other half of it is
pepperoni.
We have that. We have half plain, half pepperoni sausage.
Mixed.
Fig. 6.4 Mikes initial

written work (on two pages)
on March 1, 1993
Fig. 6.5 Mike and Amy-Lynns version of the groups work (two pages)
64
ANKUR:
MIKE:
C.A. Maher et al.
Thats mixed.
That isnt mixed though [pointing to the ninth pizza].
In earlier years, as students worked on the shirts and jeans task and the towers
problem, they were careful to eliminate duplicate combinations. As the students
worked on the pizza with halves task, they did the same, and like the other group
they had lively discussions about what was considered a duplicate. Their differences
in their definition of mixed pizzas, however, were not resolved until the next session.
The students carefully color-coded the pizza representations, but the first session
was over before they had a chance to analyze or organize their list.
6.3 Second Session: Further Pizza Explorations

The groups remained static between the first two sessions, and the students returned
the next day and continued working on the problem. However, the students from
the two groups had discussed the problem between the sessions and shared their
solutions with each other.
6.3.1 Group 1
Michelle I. began the discussion by commenting that the other group had found
more than ten combinations, and that this groups solution from the previous session
was therefore probably incomplete. Stephanie insisted that their original solution
was correct. She said, I mean, maybe theyre wrong. Did you think of that?
This group of students began to make charts to organize their pizza combinations.
Matt was first to complete his chart, which showed the ten combinations categorized
according to the number of distinct pizza toppings on each pizza as indicated in
Fig. 6.6.
Michelles chart, as indicated in Fig. 6.7, showed a total of 13 combinations.
Milin told Michelle I. that she had duplicates in her list.
MILIN:
MICHELLE I.:
MILIN:
Shes wrong. C P . . . Look, C.

Yeah, cheese, pepperoni, sausage.
He has it, but this is cheese pepperoni. That means cheese
pepperoni versus cheese, that doesnt count.
Fig. 6.6 Matts list of pizza

combinations
65
Fig. 6.7 Michelle I.s list of

pizza combinations
MICHELLE I.:
JEFF:
MATT:
MILIN:
MATT:
Yes it does because cheese and pepperoni are mixed together. On

the other side is just cheese.
But that would be the same as that, though. Think about it. Watch.
It has to be halves.
It already has cheese.
Were only working with halves, not quarters.
Stephanie then asked the group why they wrote a C for cheese at some times and
not at others. Jeff insisted that cheese was not a topping.
JEFF:
STEPHANIE:
JEFF:
STEPHANIE:
JEFF:
STEPHANIE:
MICHELLE I.:
JEFF:
Cheese is another way of saying plain. You get a plain, cheese

pizza. [Jeff writes C = plain.]
Then why do we put cheese here and here and here?
If we put plain that would get confused with that.
Why couldnt we jut put like that?
What does that mean?
Half of the pizza.
One side sausage and one side plain.
You could do that. It doesnt matter, but we decided to put cheese
there.
Now that the students had finally settled their differences about notation for a
second time, they concluded that there were ten combinations. However, they first
discussed whether a pizza with sausage and pepperoni mixed together was the same
as or different from the pizza that had sausage on one half and pepperoni on the
other.
The researcher asked the students to explain their solution and justify it. She
reviewed their drawings and lists and asked Matt and Milin why and how their
lists differed from one anothers. They reported that Matt had organized the pizzas
according to the number of toppings, while Milin had three categories: whole pizzas,
half pizzas, and mixed pizzas.
66
C.A. Maher et al.
6.3.2 Group 2
As part of the research protocol, researchers had photocopied the students work
from the previous day, kept the originals, and returned the copies to the students.
When the students returned the next day, they were disappointed that the photocopies did not show their color-coding. They were joined by Researcher 1, who
asked them to think about organizing their work before taking the trouble to draw
elaborate diagrams. They decided to check whether all students had the same list
of pizzas. As pizzas were checked off, Mike made a list of the accepted pizzas.
The students realized that some of their pizza combinations were identical, such as
the half pepperoni half sausage plain and half pepperoni-plain half sausage. A
question arose as to whether a half pepperoni/half sausage and pepperoni pizza was
allowed, and the group decided that it was. They again came up with a list of ten
pizzas.
As they tried to justify this answer, Amy-Lynn made a list of the pizza combinations. The students were convinced that there were only four whole pizza
combinations, but were not sure that they had found all half pizza combinations.
Ankur suggested that Amy-Lynns list be revised to show the different categories of
pizzas. Michelle wrote this new list, first writing the four whole pizza combinations,
and Ankur pointed out that there were only three half pizzas that did not contain any
mixture of two toppings. They then listed the pizzas that contained topping mixtures on one of the two halves. Michelle labeled her cases whole and halves.
The researcher then asked them to point out the difference between the first three
half pizzas and the remaining four.
RESEARCHER:
BRIAN:
BOBBY:
Okay, what was special about these first three halves? They
were different from the others.
They werent mixed.
They were in the halves. If you want to use plain on one side,
theres only two possible ones, plain on one side and pepperoni
has to be on the other side or sausage.
Brian then justified Bobbys claim using an argument by contradiction:

Because plain is considered like a topping so with plain, only two other toppings, because
that is all they give you . . . so you cant use plain again.
They were not yet prepared to state that they could justify the entire solution,
although Brian argued that one case, the whole pizzas case, was fully accounted
for. After further discussion, they concluded that all the toppings were used in all
possible combinations for each case, and so they reported that they felt they could
justify their solution.
67
6.4 Third Session: Getting the Right Answer

At the third session, the researcher discussed with the students the different kinds
of arguments that they had used. When they separated the pizzas into categories of
whole, half, and mixed, they were using an argument by cases. In previous investigations with towers, they had also used an inductive argument. They argued by
contradiction when they had stated that no more whole pizzas were possible
because all toppings were used. Researcher 1 also discussed the importance of using
a notation that clearly represented their ideas and communicated the meaning of the
ideas clearly to others. The session ended with the following discussion:
ANKUR:
RESEARCHER:
JEFF:
RESEARCHER:
CHORUS:
RESEARCHER:
CHORUS:
RESEARCHER:
JEFF:
RESEARCHER:
Did we get that problem right?

What do you think? Is this a question? Ankur? Did you get
the problem right? How many of you believe that you got the
problem right?
Yeah, can you tell us the answer to it?
What is the answer to it?
Ten. Ten.
Have you proved it?
Yes!
Then why are you asking me?
Cause maybe, youre the oneIts up to you. You shouldnt have to ask me.
6.5 Fourth Session: Giving the Solution

At this session, the children were asked to write a letter describing their work on the
pizza with halves problem to Drs. Davis and Alston. Their letters give the solution,
but do not provide a justification for the answer. As an example, Mikes letter is
shown in Fig. 6.8.
Amy-Lynns written explanation is of note in the method she described. She
wrote, We moved the toppings around to make new ones. When we made one,
we made the opposite. She then listed the three cases that her group had delineated
(whole, half, mixed) and listed the combinations that could be formed for each case.
Although Amy-Lynn did not explain what she meant by her terminology of
opposites, this word is reminiscent of the students explorations with towers in
the third, fourth, and fifth grades. (See Chapters 4 and 5.) During those activities,
the students worked on finding tower combinations, and they consistently searched
for towers and their opposites, which were the towers with colors reversed. AmyLynns reference to this strategy indicates that the students used similar heuristics
as they tackled two problems that, on the surface, seem to need very different
problem-solving tactics.
68
C.A. Maher et al.
Fig. 6.8 Mikes letter on

March 5, 1993
6.6 Fifth Session: Additional Justifications

This session, lasting 2 h and 30 min, provided an opportunity for students to work on
four different pizza problems and to expand their repertoire of justifications. They
continued the pizza with halves problem, and they also worked on variations of
the four-topping pizza problem, including thin and thick crust and the four-topping
pizza with halves problem.
6.6.1 Problem 1: Pizzas with Halves

At the beginning of this session, the students spent about 40 min discussing the
three pizzas with halves cases. The whole pizzas case proved to be problematic,
because this case needed to be defined more clearly. Brian and Ankur provided a
precise definition:
ANKUR:
BRIAN:
ANKUR:
. . . cause if you take one slice, theyre all gonna be the same with
the mix.
If you take one slice and you take another slice and you compare it, it
would have sausage and pepperoni on both.
They would both have the same.
69
Fig. 6.9 Final decision on

cases for the pizzas with
halves problem
This definition clearly marked the difference between whole pizzas and other
pizzas. The students continued by placing the pepperoni/sausage whole pizza in
a subcategory of its own and also placing the three mixed pizzas as a subcategory
under the half pizzas category. Their final set of cases is given by the table in
Fig. 6.9.
6.6.2 Problem 2: The Four-Topping Pizza Problem

The students then considered the Four-Topping Pizza Problem:
Kenilworth Pizza has asked up to help design a form to keep track of certain pizza choices.
They offer a cheese pizza with tomato sauce. A customer can then select from the following
toppings: peppers, sausage, mushrooms, and pepperoni. How many choices for pizza does
a customer have? List all the possible choices. Find a way to convince each other that you
have accounted for all possibilities.
It took about 15 min for the students to find 16 pizzas by randomly generating
combinations of toppings. Ankur suggested an organizational strategy:
ANKUR:
RESEARCHER:
ANKUR:
RESEARCHER:
ANKUR:
Okay. You start with the first one, thats P. And you mix it with
the second one. Thats P slash S. And then you start with the first
one again, skip the second one and go to the next one. Thats M;
P slash M. Then you start with P again and mix it with the fourth
one, PE. And then you start with the S since thats the . . . cause
you cant use plain. We start with S and mix it with M.
Wheres that?
S M. Then we start with S and PE, right here. And we start with
M and PE. S and P is right here, the first one. [He points to P/S.]
Okay. So why is it you cant go M with P?
Because you already have it. P M. [He points to P/M.]
This marked the first use of a recursive strategy on pizza problems. When questioned by the researcher about their solution, Brian explained that they were sure
70
C.A. Maher et al.
that they had found all possibilities because we have an order. The students were
very confident about their solution. As Brian and Ankur told the members of another
group:
ANKUR:
BRIAN:
ANKUR:
BRIAN:
Sixteen. And we can prove it.

Sixteen.
And you guys cant prove it.
And we can prove it.
6.6.3 Problem 3: Another Pizza Problem

Following the four-topping pizza problem, the students considered a problem called
Another Pizza Problem:
Kenilworth Pizza was so pleased with your help on the first problem that they have asked
us to continue our work. Remember that they offer a cheese pizza with tomato sauce. A
customer can then select from the following toppings: peppers, sausage, mushrooms, and
pepperoni. Kenilworth Pizza now wants to offer a choice of crusts: regular (thin) or Sicilian
(thick). How many choices does a customer have? List all the possible choices. Find a way
to convince each other that you have accounted for all possible choices.
Within less than a minute, without writing anything, students were proclaiming
the answer of 32 pizza combinations. Mike provided the explanation:
Well, since theres sixteen to make with those toppings, you put a Sicilian crust on it. Thats
sixteen. Plus then you put a regular on it, and thats 32. Sixteen and sixteen.
Thus Mike provided a succinct explanation of the students reasoning by

recursion.
6.6.4 Problem 4: The Final Pizza Problem

The session concluded with the Final Pizza Problem, which included components
of all the previous problems:
At customer request, Kenilworth Pizza has agreed to fill orders with different choices for
each half of a pizza. Remember that they offer a cheese pizza with tomato sauce. A customer
can then select from the following toppings: peppers, sausage, mushrooms, and pepperoni.
There is a choice of crusts: regular (thin) or Sicilian (thick). How many different choices
for pizza does a customer have? List all the possible choices. Find a way to convince each
other that you have found all possible choices.
The entire group worked on this problem for about 50 min. Ankur, Brian, Jeff,
and Romina jumped from answer to answer without seriously considering reasons
why the numbers would be an answer to the question. When Matt proposed a method
based on Ankurs previous method, they began to focus on finding a justified solution. Matt began with the 16 pizzas that were the answer to the first pizza problem.
Then he made half pizzas, using cheese on one side with each of the remaining 15
71
Fig. 6.10 Brian, Romina,

Jeff, and Ankurs written
work on April 2, 1993
Fig. 6.11 Matts written

work on April 2, 1993
combinations on the other half of the pizza. The next pizza had peppers on one half
and each of the remaining 14 combinations on the other half. Matt continued this
recursive procedure to arrive at an answer of the sum of 116, times 2 (to account for
the two different choices of crust). Refer to Fig. 6.10 for Matts work. Although his
procedure was correct, Matts answer was incorrect, due to an arithmetic mistake.
Matt was unable to convince the other students to consider his procedure which
is represented in Fig. 6.11.
After some of the students left, a subgroup consisting of Ankur, Brian, Matt,
Milin, Mike, and Stephanie continued to work on the problem. Romina proposed a
new categorization procedure:
First youre gonna have your wholes and then youre gonna have half with one topping on
one side and two toppings on the other side. Then youre gonna have one topping, three
toppings. Four top- I mean, one topping with four toppings. And then youre gonna go to
72
C.A. Maher et al.

[sighs] here. Ill write it down. Youre gonna have a whole first. And then youre gonna
have all your wholes, which equals up to, I think, five. And then youre gonna have half of it
with one topping, half of it with two. Then youre gonna keep on doing that until you come
to four. Then youre gonna go two; one, two; oh, two three.
After some discussion, the students were ready to listen to Matts argument. This
time they agreed that his method would count all the possible combinations and not
include any duplicates, and they accepted the recursive method as a valid strategy
of justifying the solution to this complex task. The children, pleased with their work
and solutions, expressed jubilation at the end of the session.
6.7 Discussion
During these five sessions, the students used two kinds of justifications: proof by
cases and recursive arguments. These forms of reasoning had been seen in the
schemes displayed in students earlier work on the towers problem. Now we see
how the students retrieve, build upon, and extend earlier schemes to reason about
pizza problems, despite differences in surface features. As the students folded back,
activated, and drew on previously built cognitive structures, they thought about new
ways of applying these strategies and they worked on settling differences in understanding and notation. They were animated and engaged in these interactions. In
addition, we see students adopting some strategies of others as they worked on new
problems; we see the students sharing ideas with their group members and attempting to convince them of the validity of their methods. In this way, ideas traveled
across the learning community and were applied and adapted to new challenges.
The episode shows the continual development of cognitive structures that were built
earlier.
In Chapter 7, we examine in detail the work of one of the members of this group,
Stephanie. We observe how, 3 years later, she expanded on the explanations discussed here and in Chapter 5 and explored connections between these problems and
Pascals triangle, showing further elaboration of her cognitive structures.
Chapter 7
Block Towers: From Concrete Objects

to Conceptual Imagination
Robert Speiser
Date and Grade:

Tasks:
Participant:
Researchers:
March 13 and 27, April 17, 1996; Grade 8

Binomials and towers
Stephanie
Carolyn A. Maher and Robert Speiser
7.1 Introduction
In previous chapters, we looked at the development of various forms of reasoning
in students working in a classroom in small group settings. In this chapter, we focus
on an individual student we examine Stephanies development of combinatorial
reasoning. In previous chapters, we saw how Stephanie, working with others and
on her own, made sense of the towers and pizza problems. In this chapter we see
how Stephanie extended that work. In her examination of patterns and symbolic representations of the coefficients in the binomial expansion, using ideas from earlier
explorations with towers in grades 35, she examined several fundamental recursive
processes, including the addition rule in Pascals Triangle.
This chapter centers on how children can build fundamental mathematical understanding, over time, through extended task-based explorations. They create models,
invent notation, and justify, reorganize, and extend previous ideas and understandings to address new challenges. By the time of the interviews that we report here,
we had been observing Stephanie for 8 years. Her work in combinatorics began in
grade 2, with the shirts and jeans problem (refer to Chapter 3). Even at this early
stage, she would validate or reject her own ideas and the ideas of others, based on
whether they made sense to her or not. Stephanie would monitor and often refer to
ideas and conclusions of other group members and would often integrate the ideas
R. Speiser (B)
799 E 3800 N, Provo, UT 84604, USA
e-mail: pinyonsmoke@gmail.com


73
74
R. Speiser
of partners into her work and discourse. This constant, extended process of evaluation and revision helped her to keep track of data and to reconsider, strengthen, and
extend her explanations (Davis, Maher, & Martino, 1992; Maher & Martino, 1991,
1992a, 1992b).
In grade 3, Stephanie was introduced to investigations with block towers (see
Chapter 4) that enabled her to build visual patterns of her ideas, such as the local
organization within specific cases, based on ideas like together, separated, how
much separated. She recorded tower arrangements first by drawing pictures of
towers and placing a single letter on each cube to represent its color, and later by
inventing a notation of letters to represent the colored cubes. Stephanies working
knowledge about towers, gained over long periods of time through very concrete
explorations, led as, we shall see, to powerful and personally meaningful new ways
to work with mathematical ideas.
7.2 Theoretical Perspectives

We believe that children come to mathematical investigations with theories that can
be built upon, modified, and refined. In turn, childrens theories and their ways of
working with these theories help us, as researchers, to constitute our own conceptions of childrens emerging work and thought, and so affect the way we build the
discourse, day by day, that we will share with them. In the task-based interviews
that we report, we, too, will seek to build a theory. Our emphasis on building theory
informs directly how we structure research interviews. Initially, one interviewer will
engage the child in a specific exploration, seeking to estimate the working theory
that might guide the childs thinking. Later, in the same interview or in a subsequent
follow-up interview, key ideas noted so far will be pursued primarily by the child,
who initiates, and then increasingly directs, the discourse. In such interviews, we
frequently begin with very concrete discussions, followed by what might be called
a teaching phase intended to investigate deeper connections. In such interviews,
children will sometimes make powerful connections early and so break the flow we
might naively have imagined. We have come to view such unique outcomes as
potential opportunities to gain important insights from the children that we study.
Therefore, when a childs connection appears to break the flow, the interviewer, on
principle, will invite more detailed explanation.
In Mindstorms (1980), Seymour Papert reflects on how he built his personal
mathematical understanding an understanding that inspired his later work based
on his personal experience, as a young child, playing with gears. In a similar way,
some of Stephanies key mathematical understandings can be traced to her activities, in the early grades, when she used block towers to investigate conceptually
important counting problems.
The specific arguments that Stephanie investigates below were first developed
and explained in Speisers paper (1997), where block towers underpin a concrete
microworld for productive exploration. These arguments, shaped specifically within
Block Towers: From Concrete Objects to Conceptual Imagination
75
the given microworld, were triggered by the early Gang of Four investigations
(Maher & Martino, 1996a, 1996b, chap. 4), which first describe the reasoning and
argument that enabled Stephanie and three other children, at age 9, to discover the
idea of mathematical proof, as they built and then debated strategies for counting block towers. Building from this work with towers (and inspired by the young
Papert) we seek precise, particular descriptions (1) of how Stephanie actually does
strong mathematics based on towers, and (2) of what, specifically, might constitute
its strength.
7.3 Setting
Stephanie participated in the longitudinal study starting in first grade. Stephanie and
her classmates were challenged in their mathematics classrooms to build solutions to
problems and construct models of their solutions. This setting, which for Stephanie
continued to grade 7, encouraged differences in thinking that were discussed and
negotiated. In fall 1995, Stephanie moved to another community and transferred to
a parochial girls school. Her mathematics program for grade 8 was a conventional
algebra course. Stephanie continued to participate in the longitudinal study through
a series of individual task-based interviews. A subset of these interviews provides
the data for this chapter.
7.4 Guiding Questions

The following questions guide our analysis in order to consider, systematically, the
ways in which Stephanies past experience is drawn upon: (1) How does Stephanie
work with towers in building images and understandings for higher mathematical
ideas? (2) What is the role of past experience in building new ideas? (3) How are
her ideas modified, extended, and refined over time?
Data come from two of eight individual task-based interviews of Stephanie. The
interviews were videotaped with two cameras, positioned to capture in detail what
was said, written and built and to include less tangible data such as tone of voice,
speech tempo, and where people look while they converse and work. Transcripts and
analyses of the interviews were made and verified by a team that included several
graduate students in addition to both authors. Stephanies written work prepared
for the interview, and several observers notes, provide further data. The teaching
experiment was conducted over a 6-month interval (November 8, 1995, to May 1,
1996). Each interview, approximately one and one-half hours in length, would typically begin with inquiries about the mathematics that Stephanie currently studied in
eighth-grade algebra, both to open opportunities to talk about that mathematics and
to explore her thinking about fundamental mathematical ideas.
76
R. Speiser
7.5 Results
To introduce each data segment, we provide a brief discussion of the mathematics
Stephanie was invited to explore. On this basis, we can more clearly understand
each segment as a momentary snapshot of Stephanies emerging understanding.
The correspondence between binomials and towers. During the March 13, 1996,
interview, Stephanie, unprompted, made a connection to towers, by examining her
symbolic representation of the expansions of (a + b)2 and (a + b)3 .
STEPHANIE:
RESEARCHER:
STEPHANIE:
RESEARCHER:
STEPHANIE:
RESEARCHER:
STEPHANIE:
So theres a cubed [a3 ].

Thats 1.
And theres three a squared b [3a2 b] and theres three a b
squared [3ab2 ] and theres b cubed [b3 ]. [Interviewer writes 1 3
3 1 under 1 2 1 as Stephanie speaks.] Isnt that the same thing?
What do you mean?
As the towers.
Why?
It just is.
Stephanie asserts (in her own way) that each three-high tower gives a noncommutative monomial of degree 3 in a and b, and she has indicated that these
non-commutative monomials, indexed by the corresponding towers, collect to give
the coefficients for the commutative monomials that appear in her expansions of
(a + b)2 and (a + b)3 . Our interpretation, therefore, is that Stephanie visualizes
towers (referring to mental models she does not have plastic cubes at this point)
to help her organize the terms that she collects. We believe that Stephanie reasons
about polynomials based on her mental images of towers.
Working at home before the interview, Stephanie had written out the first six
powers of the binomial a + b, and brought her written calculations to the interview. The interviewer covered Stephanies paper, guessed the coefficients for the
sixth-power expansion, and wrote down the terms in full. Her coefficients were the
same as Stephanies, although one monomial was slightly different. Several minutes
further in the conversation, Stephanie gives further evidence, that she proceeds by
visualizing towers and then reasoning based on her mental images.
RESEARCHER:
STEPHANIE:
RESEARCHER:
STEPHANIE:
RESEARCHER:
STEPHANIE:
RESEARCHER:
STEPHANIE:
RESEARCHER:
STEPHANIE:
So you have two factors of a. Right?

Um hm.
You have one of those. One thing with two factors of a. One
thing with two as in it.
Um hm.
I dont want to think of as. I want to think of red.
Okay [laughing].
Can you switch that a minute?
Yeah.
So now I have one thing with two reds. What thing can I be
thinking of with two reds?
Thats a tower thats two high.
RESEARCHER:
STEPHANIE:
RESEARCHER:
STEPHANIE:
RESEARCHER:
STEPHANIE:
RESEARCHER:
STEPHANIE:
RESEARCHER:
STEPHANIE:
RESEARCHER:
STEPHANIE:
RESEARCHER:
STEPHANIE:
77
Okay. And here Im talking about two things.

Um hm.
One is.
Red and . . .
one is . . .
. . . one is yellow.
Is that possible in two high?
Yeah.
Having one red and one yellow? There are two of them?
Yeah.
Which two?
Cause the one is the red could be on the top or the bottom, with
the yellow the same thing.
What about b squared?
Um. Two yellow.
In a March 27, 1996, interview, Stephanie is invited to explain to a second

interviewer (unfamiliar with her recent work) what had happened in the March 13
interview described above. Here Stephanie begins with towers, then reviews the
binomial coefficient notation C(n, r), working through a sequence of examples with
increasing n. Stephanie remarks that r is a variable, which she understands can
range from 0 to n. This observation shifts the level of abstraction upward from specific towers (as above) to patterns of formal symbols as in Pascals Triangle. Hence,
at this point, n, the height, and r, the number of red blocks for given n, will both
vary. This richer context triggers, with encouragement from Interviewer 1, a confident, detailed, and carefully presented recapitulation by Stephanie of the recursive
construction of the towers of height n from the towers of height n 1, as it had been
introduced by classmate Milin in grade 4 and revisited in grade 5 (see Chapters 4
and 5).
During a previous interview, on March 13, Stephanie also referred to Pascals
Triangle, in particular to its addition rule, to make similar predictions, but she had
done so in a conceptually quite different domain: to predict, in effect, the numbers
of n-tall towers in each given case (of r red blocks, say, for given r) for new values of
the height. Stephanies choice to center, in the present interview, directly on binomials strongly suggests that Stephanie now grasps the isomorphism between Pascals
Triangle, which she had built, at first, to summarize her towers cases, and the array
of coefficients for her polynomial expansions of the powers (a + b)n , for variable n.
On this basis, further interviews were planned, with towers available to serve as
concrete anchors to establish formal facts about the C(n, r), viewed either formally
as binomial coefficients or as counts of combinations, or, more concretely, as the
numbers of specific kinds of towers.
Fermats recursion. One goal for the March 27 exploration was to offer Stephanie
the tools shed need to construct a formula, originally due to Fermat (Weil, 1984),
that expresses the relationship between two successive binomial coefficients. In
symbols, here is Fermats formula:
78
R. Speiser

C(n, r + 1) =
nr
r+1

C(n, r)
(7.1)
This equation, applied repeatedly beginning with the simple case r = 0, leads
directly to the standard formula for C(n, r).
To make sense of this formula, it seems especially helpful in this setting to interpret Equation (7.1) in terms of towers. For concreteness, take red and yellow for the
colors of the blocks available. On the right side, C(n, r) counts the towers of height
n that have exactly r red blocks, hence n r yellow blocks. Call these the original
towers. On the left side, C(n, r + 1) counts the towers of height n with exactly r +1
red blocks. Call these the new towers. In concrete terms, Equation (7.1) tells us that
the number of new towers can be found by multiplying the number of old towers by
the number of yellow blocks in each and then dividing by the number of red blocks
in a new tower.
In the data below, the interviewers will fix n, the height of a tower, and then
vary r, beginning either with r = 0 or r = 1, for which C(n, r) is either known to
Stephanie or easily determined by inspection. For each r, the interviewers will then
invite Stephanie to construct new towers from a given set of original towers and
explore with her what she has found. The construction process Stephanie explores
will work for any height n and any r < n.
For concreteness, we explain this process when n = 4 and r = 1. In this case, we
have four original towers, each with a red block in one of four available positions.
From each given original tower, we can build new towers by replacing one of its
three yellow blocks with a red block. For each of the four original towers, we can
therefore build three new ones. Working in this way (well say by day) we obtain
four groups of three new towers. The 12 towers constructed in this way clearly
include each possible new tower. For example, consider Fig. 7.1 as an example:
Working by day, begin with one original tower four blocks high with one red block
(shaded). We obtain three new towers, each with two red blocks, by replacing in
turn each of the three yellow blocks in the original tower with a red block.
The total we have just obtained instantiates the product (n r)C(n, r) on the right
side of Equation (7.1), but this is a key point the new towers we just built are not
distinct. In fact, each tower appears exactly twice among the 12, as the denominator
r + 1 predicts. To understand how duplicates emerge, consider a particular new
tower. This tower has exactly two red blocks. Each of these two red blocks can be
replaced (working, we shall say, by night) with a yellow block, producing one of
two original towers. This construction, which reverses what we did by day, shows
Day
Fig. 7.1 Working by day:

replace each yellow block
(unshaded) by a red block
(shaded)
79
that each new tower will appear exactly twice among the 12 we had constructed.
In particular, there will be exactly six towers of height 4 that have exactly two red
blocks. Because towers correspond to combinations, we have used the known result
C(4, 1) = 4 to show that we have C(4, 2) = 6. For example, consider Fig. 7.2. By
night, begin with a new tower four blocks high (n = 4) with two red blocks. We
obtain two original towers, each with one red block, by replacing one of the two red
blocks in the given new tower with a yellow block.
Stephanie began to explore the construction shown in Fig. 7.1 during the March
27 interview, first with three-tall towers. Her blocks were blue and green. Continuing
to four-tall towers, she next built the four towers with exactly one green block and
guessed initially (but incorrectly, perhaps based on her experience with three-tall
towers) that from each such original four-tall tower she could obtain two new ones.
RESEARCHER:
STEPHANIE:
RESEARCHER:
STEPHANIE:
RESEARCHER:
STEPHANIE:
RESEARCHER:
INTERVIEWER 2:
RESEARCHER:
STEPHANIE:
I wonder why you get two of them.

I dont know. Maybe cause its bigger.
What would that have to do with it?
I dont cause you have more room to build on.
Tell me, can you explain to me?
Oh, well, maybe its because like you already have one [green
block] thats taking up space, so you only have three places
to move it.
I gotcha, okay.
Okay.
So what would you predict if you were building towers five
high?
Youd have four.
Here we see Stephanie revise and then explain her observations, starting from a
set of four-tall towers that she had physically built. On this basis, she extends her
observations to a set of five-tall towers she has just imagined. So far, she knows that
duplicates appear in the construction she discusses, but has not yet explored in detail
how or why they do.
Revisiting the same construction in the next interview session (April 17, 1996),
Stephanie considers duplicates directly. This time (as in the earlier examples) her
colors will be red and yellow, and the variable r will count the red blocks in a tower.
After reviewing, for four-tall towers, the construction of new towers from original
towers, with r = 0, 1, 2, and 3 in succession, the researcher invites Stephanie to
Fig. 7.2 By night, replace

each of the red blocks
(shaded) by a yellow block
(unshaded)
Night
80
R. Speiser
predict how many duplicates she would expect for towers of height 5 and helps her
build the towers that she needed as they proceed. They begin with a tower with five
yellow blocks (the case r = 0).
STEPHANIE:
RESEARCHER:
STEPHANIE:
RESEARCHER:
STEPHANIE:
RESEARCHER:
STEPHANIE:
RESEARCHER:
STEPHANIE:
RESEARCHER:
STEPHANIE:
RESEARCHER:
STEPHANIE:
RESEARCHER:
STEPHANIE:
RESEARCHER:
STEPHANIE:
The first ones one.

Theres one of those
times five.
Why five?
Cause theres five positions.
Okay.
Divided by one, cause they come in groups of one.
Um, hm.
Five.
Okay. So thats five things taken one at a time.
Yes. The second one
Why dont you write that down? Five things taken, equals five
things taken one at a time. [Stephanie writes.]
Okay. For the second one, um, theres four spaces. But theres
out of five so its five times four and theyll come up in groups
of I dont know, um, thats what we dont know though.
All right. So. Lets can we make these five? Just, here.
Well, maybe they might come in groups of two?
One. Lets think about at least one of these.
They might come in groups of two, I guess.
Here, just as they begin to build the five five-tall towers with one red block,
Stephanie repeats, it seems, the mistaken guess that she had made earlier for fourtall towers. At this point, the interviewer arranges the five towers they have built in
front of Stephanie and offers Stephanie the tower with its one red block in the top
position.
RESEARCHER:
STEPHANIE:
RESEARCHER:
STEPHANIE:
RESEARCHER:
STEPHANIE:
RESEARCHER:
STEPHANIE:
RESEARCHER:
STEPHANIE:
RESEARCHER:
STEPHANIE:
Okay. So what youre saying here move some of this aside

um, okay. Lets think of that one.
Okay.
There are five.
[Builds a tower with a second red block just below the first.]
You have one like that. [She builds another tower with a second
red block two spaces below the first.] One like that.
Well, can you predict before you do it?
Yeah, theres going to be four from each.
Four from each.
Yeah.
Okay. So and whats the each? How many make up each?
How, wh, what do you mean?
Youre saying, its four from this.
Yeah, four from
RESEARCHER:
STEPHANIE:
RESEARCHER:
STEPHANIE:
RESEARCHER:
STEPHANIE:
RESEARCHER:
STEPHANIE:
RESEARCHER:
STEPHANIE:
RESEARCHER:
STEPHANIE:
RESEARCHER:
STEPHANIE:
RESEARCHER:
STEPHANIE:
RESEARCHER:
STEPHANIE:
81
What doesOne.
-each mean in this case?
Oh! Like theres going to be four from this one. Four from that
one. Four from that one. Four from that one. Four from that one.
Okay. So how many eaches?
Theres five.
Five eaches. Okay.
Yeah.
All right. So that, you say, five times four.
Yes. I have that. I just dont know what theRight.
-bottom part it
So and by the groups, you mean. The groupings you mean.
Groups like one after weve put them all out. Like how many
groups, theyre going to come inI dont know. Im
duplicates?
Im wondering. When you say you divide by
Oh! Cause thats the number of duplicates that there are.
Again, working by day, with towers on the table, Stephanie corrects her guess, but
then a new step tries to go further. She has just built 20 towers in five sets of four.
Now she proposes to restructure her set of 20 towers into a groups of duplicates, or
at least to find the number of such groups. In effect, she has proposed the key step
by herself: to find how many duplicates each new tower has.
RESEARCHER:
STEPHANIE:
RESEARCHER:
STEPHANIE:
But how do you know beforehand? Do you think theres a way?

[Building towers.] Oops.
So if this, um, is going to be a pattern to this the five times
four what do you think you would divide by?
Five times four what do you think Id um maybe two.
Working backward from the known entry, 10, in Pascals Triangle, Stephanie
confirms that she indeed will need to divide 20 by 2. The explorations then continue
with ten towers of height 5 with two red blocks. Working by day, Stephanie predicts
that 30 new towers can be built beginning with her ten originals. This time each new
tower will have three duplicates.
STEPHANIE:
RESEARCHER:
STEPHANIE:
Ten. So it would be ten times three and you divide by three.

[Writes as she speaks.]
And it worked?
Yeah. And the next one, there is two spaces to put it and you
have ten. So theres ten times two, and you divide by two?
[Continues writing.] And the last one theres one space to put
it its five times one divided by five equals one.
82
R. Speiser
In this exploration, in effect, Stephanie has explained how the corresponding row
(1, 5, 10, 10, 5, 1) of Pascals Triangle emerges numerically from the pattern of
Equation (7.1), which she has not yet seen. At each step, she connects the product
(n r)C(n, r) directly to the operation of replacing one yellow block with a red
block.
We note, however, that Stephanie has not yet identified the denominator,
r + 1, with the number of red blocks in a new tower. Instead, she seems to follow a numerical pattern that she has observed empirically. At this point, Stephanie
does not yet seem able to explain why the number of duplicates that she observes
must necessarily be r + 1. Nonetheless, we remain astonished, after 12 years, by the
depth and strength of the connections Stephanie has made, based on her familiarity
with towers.
The addition rule in Pascals Triangle. By March 1996, as noted above, Stephanie
already knew the additive pattern that relates successive rows of Pascals Triangle.
In symbols, this addition rule can be expressed as follows:
C(n 1, r) = C(n, r 1) C(n, r)
(7.2)
According to this formula, each row of Pascals Triangle can be computed from
the row before it, by adding each pair of successive entries in the row above. To
connect this formula to combinations, and in this way make sense of it, we will read
each term as a count of towers. Specifically, the first term on the right counts the
towers of height n that have exactly r 1 red blocks, while the second term counts
towers of the same height, but with one additional red block. So interpreted, the
right side of Equation (7.2) at least suggests that every tower of height n + 1 that
has exactly r red blocks can be constructed from suitable shorter towers of height
n, either by placing a red block on top of a tower of height n with r 1 red blocks
or by placing a yellow block on top of a tower of height n with exactly r red blocks.
A special case is shown in Fig. 7.3: In the top row, we begin with two sets of towers
of height 3 (n = 3): one tower with no red block (r = 0) and three towers with one
red block (r = 1, shaded). To accomplish the recursion, attach a red block (R) on
Fig. 7.3 A specific example

of the addition rule for
Pascals Triangle
83
top of the single tower in the first set, and a yellow block (Y) on top of each tower
in the second set, to produce four towers of height 4 with one red block.
Again we work both day and night. By day, attaching blocks as shown in Fig. 7.3,
it is not difficult to see that the resulting new towers of height n + 1 must be distinct.
Then (by night), if we remove the top block of each possible tower of height n + 1
that has exactly r red blocks, its clear that all such towers have been counted on the
right side of Equation (7.2).
In the data segment soon to follow (later in the April 17, 1997, session)
Interviewer 1, drawing several rows of Pascals Triangle, writes down the numbers 1 and 3 that correspond to the towers shown in the top row of Fig. 7.3. She is
just about to write the number 4 below them, and then draw two diagonal lines, to
associate the numbers 1 and 3 to the 4.
RESEARCHER:
STEPHANIE:
RESEARCHER:
STEPHANIE:
RESEARCHER:
STEPHANIE:
RESEARCHER:
STEPHANIE:
RESEARCHER:
STEPHANIE:
RESEARCHER:
STEPHANIE:
RESEARCHER:
STEPHANIE:
RESEARCHER:
STEPHANIE:
Okay. Um. Lets explore, um which one should we explore?

[Draws lines as above.] Lets do this one.
Um, hm.
Do you know what this one means? If you had to build this one,
what would that tower look like?
That one?
What would that one look like? What would those two look
like?
[A pause, while Stephanie builds towers.] I think that one would
be like this. [Stephanie has built the tower of height 3 with all
yellow blocks, and she indicates the one that Interviewer 1 has
drawn] and that one.
Three high, no red.
Like this. [She has just built the first two towers of height three
with one red, as in Fig. 7.3.]
Okay. Three high, exactly one red.
Yes.
Okay.
Oh! Wait! [Builds the remaining tower.]
Okay. Makes you dizzy after a while, doesnt it? Cause I think
I see exactly one also. Even when you make it, I just believe
youre gonna do it. Okay. Now. When were doing this [points
to the 1, the 3, and the 4 that she has written].
Um, hm.
Whats different about these and this tower here [taps the
number 4 in Pascals Triangle] that I call four? There.
Well its four high.
For a few lines, Interviewer 1 and Stephanie review the towers of height 3 that
Stephanie has built and has physically in front of her. They easily agree that the
number 4 that Interviewer 1 has written beneath the entries 1 and 3 should count
the towers 4 blocks high that have exactly one red block. These towers have not yet
been built, and they will not be built in the conversation that will follow.
84
RESEARCHER:
STEPHANIE:
RESEARCHER:
STEPHANIE:
RESEARCHER:
STEPHANIE:
RESEARCHER:
STEPHANIE:
RESEARCHER:
STEPHANIE:
RESEARCHER:
STEPHANIE:
RESEARCHER:
R. Speiser
I want to know from here

Uh, hm.
what you do to these [the towers of height 3]Well.
-to get me, to get meWell, Id build them higher.
Well, dont do it yet. Just think about it for a minute. Remember
what theyre going to look like.
Yeah.
Theres going to be exactly one red.
This would go here [she moves the all-yellow tower of height
3] and there would be red.
No. No. We start with these [points to the number 4 again]. I
dont want you to touch these [indicates the towers of height 3].
I want you to tell me what youre gonna do to these so that when
youre all doneUm, hm.
-you end up with exactly one red. But youve got to make them
all four tall.
This point is delicate. Stephanie knows (empirical experience!) the four-high

towers with exactly one red block, and she can easily imagine them. But does she
understand, working by day, how those towers can be built from the four towers of
height 3 she has in front of her?
STEPHANIE:
RESEARCHER:
STEPHANIE:
RESEARCHER:
STEPHANIE:
RESEARCHER:
STEPHANIE:
RESEARCHER:
STEPHANIE:
RESEARCHER:
STEPHANIE:
RESEARCHER:
STEPHANIE:
RESEARCHER:
STEPHANIE:
Im going to put a yellow here [points to the first tower on the

right in the top line of Fig. 7.3],
Okay.
Im going to put a yellow here [points to the next tower],
Right.
Im going to put a yellow there [points to the third tower in
the same group] and Im gonna put a red there [points to the
all-yellow tower].
Okay. So how many ways how many do you end up with?
Four.
Four. So from the one three tall with no reds
Um, hm.
And the three three-tall with one red, right?
Yes.
You end up four four-tall with one red.
Yes.
Isnt that neat?
Yeah.
In this and later interviews, Stephanie first masters this way of working in continued conversations with Interviewer 1, but she then goes on, in later sessions with
her peers, to teach the line of reasoning she begins exploring here to others.
85
Fig. 7.4 Stephanies tower

exploration, grade 8
A conceptual reflection. In these data, we see Stephanie refining and revising new
ideas that she has built from raw materials she draws from prior experience with
towers and combinations. This prior experience includes a variety of proofs (first by
cases, later by induction), expressed concretely with block towers and more formally
through language and notation that she and peers have personally developed and
refined throughout their long collaboration.
This process of revision and refinement, which we emphasize throughout, might
be most clearly visible across the data we see here as a progressive movement from
sets of towers that Stephanie built physically (see Fig. 7.4) toward sets of towers
that Stephanie comes to imagine. These imagined towers (such as the final set of
four above) are not simply visualized as static images from prior tasks; indeed, the
new towers have been constructed based on new conceptual ideas that Stephanie has
begun to build, in real time, as the interviews proceeded.
7.6 Discussion
In an earlier paper, based on just a fraction of the data we considered here, we used
the metaphor of text to state the following conclusions (Maher & Speiser, 1997b,
p. 131):
Images, patterns, and relationships have become mathematical objects that Stephanie sees
and works with mentally to build abstractions. Our conversations with her elicited both
spoken and written texts. These texts, together with our interpretations, anchor an analytic
narrative of the development of certain mathematical ideas. Such texts (which we propose
to view as work in progress) extend through time and serve as records of particular events
upon which later texts can comment. Further, they can serve as raw material from which
new texts can be composed.
We revise our texts, and so does Stephanie, as our experiments proceed through detailed
interactions with each other. Hence, as Stephanies developing judgment enters the discussion, her presentations offer raw materials that help to focus and direct the researchers
later task designs and explorations. Our agenda for the interviews, seen as an emerging text,
continues to be rewritten, reconsidered, and revised, often in direct response to goals that
Stephanie pursues.
86
R. Speiser
After the interviews we have considered here, events that neither Stephanie,
nor her peers, nor the researchers could foresee in 1996, would offer opportunities
for everyone involved to deepen, reconsider, refine, and extend their previous perspectives and conclusions. In the discussion we have just presented, perhaps most
striking is the heightened prominence we see of personal, conceptual imagination to
address new problems and, in the process, to give form to new and powerful ideas.
In the next chapters, we follow other students from the longitudinal study who
also build on previous explorations to make sense of Pascals Triangle and rules for
its generation.
Part III
Making Connections, Extending,

and Generalizing (19972000)
Chapter 8
Responding to Ankurs Challenge:

Co-construction of Argument Leading to Proof
Carolyn A. Maher and Ethel M. Muter
Date and Grade:

Tasks:
Participants:
Researcher:
January 9, 1998; Grade 10

Towers
Ankur, Brian, Jeff, Mike, and Romina
Carolyn Maher
8.1 Introduction
In previous chapters, we saw elementary students work in classrooms on counting
problems presented by researchers. In this chapter, we observe a group of five
high school students working under different circumstances. When the students in
the longitudinal study entered high school, they no longer worked on problems in
class. Instead, the students who remained in the study worked on problems in afterschool sessions scheduled by the researchers and for which the students rearranged
their after-school schedules in order to attend. In addition, this session is unique
because the students worked on a problem proposed by a fellow student. In this session, Ankur and Mike were invited to propose and solve their own problem. Ankur
proposed a new towers problem, which became known as Ankurs Challenge:
Find all possible towers that are four cubes tall, selecting from cubes available in three
different colors, so that the resulting towers contain at least one of each color.
Mike and Ankurs approach was to start with the total number of four-tall towers
built from three colors and then subtract the number that did not fulfill the stated
criteria. They started by writing combinations of towers, using the numbers 1, 2, and
3 to represent colors red, blue, and yellow, respectively, and using a 0 to represent
the duplicated color. At first, they omitted some towers in their count, but later in
the session they discovered the missing set.
C.A. Maher (B)



89
90
C.A. Maher and E.M. Muter
As Mike and Ankur proposed a variety of numerical answers (72, 54, 45, and
36) to the problem, Jeff, Brian, and Romina joined the discussion. Jeff suggested,
based on a preliminary listing, that the answer to Ankurs challenge was 36. (His list
had 37 towers, but he thought that it included a duplicate.) Although they worked
separately, the two groups periodically exchanged comments and suggestions. The
first firm finding on which all agreed was that the total number of four-tall towers
that can be built when selecting from three colors is 34 (81) towers.
8.2 Rominas Presentation of Proof

At this point the dynamics of the working groups changed. Jeff joined with Mike
and Ankur, who were working on developing a justification based on cases. Brian
sat quietly, sketching out ideas on a sheet of paper. Romina worked on her own, at
times thinking out loud, as illustrated here:
You know, it might be 36. Cause Im working with sixes now. And okay, you put them, like
you pair em up. Cause youre only gonna have . . .
At this point, Romina put up her hands and indicated that she needed to collect
her thoughts. She continued:
Let me think first, organize my thoughts a little. Were gonna have them together. Together
like over here [indicating her list]. These are together. These are together. These are
together. Like two of the same color together. And then, in like a pattern, like, well put
them somewhere and then well switch them around, so Im up to 24 now and Im going to
put them the same way here and here. So then thats 30. And I put the same ones here and
here. Here, here. I didnt put them, and then theres your 36.
Figure 8.1 gives Rominas first list of possibilities.

Romina explained that two cubes of the same color had to be in each tower and
that she was using that fact to create a pattern. Using X, O, and 1 to represent the
three colors, Romina listed the six towers that could be created using a single pattern
(that in which the repeated color occupied the first two positions in the tower), and
then she moved on to the next possible arrangement. This was the beginning of a
justification based on cases. She explained that by using this method she had found
24 towers so far and that there were two additional groupings that she had yet to
complete. Counting all of the towers in her list would mean that there were 36
possible towers as the answer to Ankurs Challenge. At this point, Romina began to
generalize her solution. Instead of listing all of the combinations, she listed only the
cubes that were duplicated in each tower.
As Romina tried to get the attention of Jeff, Mike, and Ankur in order to explain
her solution to them, Brian interrupted and pointed out that she had duplicated
one of the rows in her list. Undaunted, Romina reached for a clean piece of paper
and began to redraw her table. Brian watched carefully as she worked, offering
helpful suggestions as she created this version of her solution. She drew boxes
showing the duplicated cubes in five of the six possible positions. With Brians
Responding to Ankurs Challenge
91
Fig. 8.1 Rominas first

attempt to write out the
towers of Ankurs Challenge
assistance, Romina reviewed what she had already written and included the sixth
row (Fig. 8.2).
She was then ready to present her thoughts to Jeff, who had moved over to see
what she and Brian were doing. She counted her rows to verify that she had the
required six and proceeded to explain what she had done to Jeff. She said:
You know were gonna have two of the same color. Right? Two of the same color, which
stands for putting these and these, right? And youre gonna have them in, and then the
rest you fill up, right? And then youre gonna have the . . . And theres only two other ones
you could have. So this, you have this one, you have to multiply it by two. Well, one, two,
Fig. 8.2 Rominas

generalization
92

three, four, five, six. . . . You multiply this by two [indicating each row in turn], multiply
this by two, multiply this by two, by two, by two, and by two. And then, one, two, three,
four, five, six.
This attempt to convey the idea that each row represented a four-tall tower that
had to contain two cubes of one color with one of second color and one of the third
color was not convincing to Jeff. Romina did not explain the reason for multiplying
by two.
Romina and Brian decided that Romina would write her ideas more neatly, and
she produced a third version of her diagram. As Ankur and Mike began talking to
the researcher about their solution, Jeff and Romina interrupted with the information
that they had found a solution of 36, and they could prove that it was correct. They
asked Mike to pay attention to their explanation and, although he agreed, it became
clear that he was still thinking about his own ideas for finding the complement to
Ankurs Challenge.
Rominas presentation improved as she related her ideas this second time. The
discussion follows.
ROMINA:
ANKUR:
ROMINA:
ANKUR:
ROMINA:
JEFF:
BRIAN:
So you have to organize them so that you dont have any doubles.
So either you can have them next to each other. You can have them
separated by one. You have them on the ends, in the middle, two and
fourth spot, and third and fourth spot. So thats six. Okay. Now you,
in the other spots, you can have an o and an x. Those are colors. Like
these are three different colors an o and an x and an x and an o.
Right?
Yes.
So thats six.
Yes.
So you have to multiply each of these six by two.
But you couldnt have like x x because that wouldnt fit the requirement. So you multiply each one by two. So that would give you 12.
Correct? Cause that means you could have like this . . .
Like the x is in the first, the o in the first spot.
This time, Romina explained the code and her organization, iterating all the possible positions for the duplicated color. She explained how she placed the remaining
two colors in her diagram and said why she multiplied each row by two (to account
for the fact that the two non-duplicated colors could switch positions). Jeff and Brian
added clarification of the reason for multiplication by two.
Ankur accepted Rominas solution. Mike, however, remained fixed on his earlier
strategy to justify that the remaining towers, those that formed the complement of
Ankurs Challenge, numbered exactly 45. He was not willing to accept the number 45 as the difference between 81, the total number of towers that they had
agreed on at the outset, and the 36 combinations that Romina, Brian, and Jeff had
found. Although Ankur and Jeff claimed that Rominas proof should be sufficient to
prove that 45 was the solution to the complement problem, Mike wanted an explicit
proof:
93
Fig. 8.3 Rominas

chalkboard version of her
proof of Ankurs Challenge
ANKUR:
JEFF:
MIKE:
The only way you could prove you were right is to prove the other side.
We proved the other.
Thats not enough for me. I want to prove the other.
At the end of the session, Mike asked Romina to restate her proof, admitting that
he had not been paying attention the first time that she had presented it to the group.
Romina graciously complied, further refining both her explanation of the proof and
the diagram she utilized as she spoke while writing it on the chalkboard. Refer to
Fig. 8.3.
At the groups next meeting, Romina brought a written copy of her proof of
Ankurs Challenge. Notice her further refinement in showing 36 possibilities in her
written work shown in Fig. 8.4.
Fig. 8.4 Rominas second

version of her proof of
Ankurs Challenge
94
In Rominas written explanation, she indicated a realization that there were three
different colors for selecting the blocks and four positions on a tower to place a
block. She indicated, also, that there would be a double of one of the colors and a
single of the two remaining colors. She now used a different notation yellow (Y)
and blue (B) for the single blocks and red (R) for the duplicate block. Romina wrote
that placing the red cubes in all possible positions would produce six towers. She
pointed out that there were only two possibilities for the remaining colors, thereby
producing 6 2 or 12 towers with double red blocks. She then concluded that by
considering the remaining 3 colors, 3 12 or 36 towers would yield all possible
combinations, thereby producing an elegant justification for the solution to Ankurs
Challenge.
8.3 Discussion
In their work on Ankurs Challenge, Romina, Brian, Jeff, Ankur, and Mike demonstrated how they worked as problem solvers. Ankur posed an interesting problem,
which he and Mike partnered to solve. They started by applying previous knowledge (how to find the number of towers when selecting from two colors) to a
new situation (three colors are now available). They successfully found the total
number and proceeded to use a subtraction strategy. They listed exceptions by
case; when the notation they originally chose proved inadequate, they introduced a
new notation.
Later, when provided with an answer derived from a different approach, Mike
continued to work on Ankur and his initial strategy. Although he accepted the direct
approach explained by Romina from the other group, he was unsatisfied that Ankurs
and his approach, seemingly reasonable, did not work.
Romina and her group used a direct approach. They brought together ideas and
notations from the past as they constructed a solution for Ankurs Challenge. They
used a variation of the binary coding scheme that Mike had introduced the previous month. They profited from the strategy that Ankur had presented in the fifth
grade of fixing one of the variables and then considering the possibilities that satisfied that case. As Romina explained her strategy to Brian, she formulated different
ways to express her thoughts; Brian assisted by pointing out additional cases. As
Romina tried to communicate her ideas so that Jeff would follow, she revised her
representations.
Mike, ready to hear about the solution of the other group, asked Romina to repeat
her explanation. In response, she presented her work at the blackboard with further refinement of their representations. Her final written summary provided another
opportunity for detail, refinement, and generalization. In summary, we can say that
Romina and her group profited by using their personal representations, communicating them as ideas, and then providing support for those ideas by reorganizing and
restructuring representations. Further, in each iteration of the argument, Romina
made refinements and clarified her reasoning. This suggests advantages for students
95
when afforded more than one opportunity to explain and write about their ideas.
Each explanation has the potential to contribute to a deeper understanding and for
multiple ways to represent ideas.
Romina, 1 week later, shared a written solution to the problem, indicating an
interest in refining her explanation and demonstrating her motivation for further
thought and reflection.
In the following chapters, we return to the pizza and towers problems and see how
groups of students continue to refine their representations, clarify their reasoning,
and extend and generalize their understanding of mathematical ideas by revisiting
old ideas, communicating their findings, and listening to the findings of others.
Chapter 9
Block Towers: Co-construction of Proof

Lynn D. Tarlow and Elizabeth B. Uptegrove
Date and Grade:

Tasks:
Participants:
Researchers:
November 13, 1998; Grade 11

Towers
Ali, Angela, Magda, Michelle, Robert, and Sherly
Carolyn Maher, Alice Alston, Susan Pirie
9.1 Introduction
In previous chapters, we observed elementary school students working to make
sense of the towers problems by building representations, formulating conjectures,
and defending their solutions in discussions with classmates and researchers. In this
chapter, we observe a cohort of high school juniors as they engage in explorations
and constructions in the towers problem. During this session, the students found and
generalized formulas for solutions to the original towers problem (building towers
when selecting from two colors of Unifix cubes) and extensions (with more than
two colors of cubes), using methods including controlling for variables, justification
by cases, and inductive reasoning.
9.2 Building Towers

In the 2-h session, students worked in pairs on tower problems. They came up
with a general rule for the number of possible towers of height n when selecting
from x colors (xn ) and an explanation of that result based on an inductive argument
based on generating all possible towers of a given height. Their arguments contained
reasoning by cases, induction, and reasoning by contradiction. In addition, Robert
produced an equation for the number of towers having exactly two cubes of one
color (when selecting from two colors), for a tower of any height.
E.B. Uptegrove (B)
Department of Mathematical Sciences, Felician College, Rutherford, NJ, USA
e-mail: uptegrovee@felician.edu


97
98
L.D. Tarlow and E.B. Uptegrove
9.2.1 Angela and Magda

Neither Angela nor Magda had previous experience with the towers problem, as
they had both joined the longitudinal study in sixth grade. In this, their first experience with towers, they found all 16 towers, four-tall, selecting from two colors.
Interestingly, they used strategies similar to those developed by the fourth and fifth
graders that participated earlier in the study (see Chapter 4). The girls organized
their work by cases: (1) one blue, (2) two blues, (3) three blues, and (4) four of
the same color. The two single-color cases consisted of one tower each; the oneblue-cube and three-blue-cube cases exhibited a local organization; they built those
towers by moving the single cube of one color into each of the four possible positions. When asked how they knew that they had all the towers with one blue cube,
they described their organization:
MAGDA:
ANGELA:
The blue is in each position each time.

Yeah, each possible position because theres only four spots.
Initially, they had no support for accounting for the towers in the two-blue-cube
case; they explained that they were unable to find any more. However, after they
found four of the towers for this case, and they were asked how they knew they
had them all, Angela alluded to a preliminary organization using controlling for
variables strategy, that is, holding the top and bottom cubes constant. She said:
Well, I mean, I dont know how to explain it, theres just like no other possibilities for it.
I mean, theres only four places, you have them, like you know, yellow on top, blue on the
bottom, and the blue on top, yellow on bottom, then blue on top and bottom, and yellow on
top and bottom.
As she was saying this, Angela found the two towers for this case that were
missed. There are two towers with yellow on the top and blue on the bottom and
two towers with blue on the top and yellow on the bottom; they had originally found
only one of each of those pairs.
When asked to determine the number of three-tall towers, Angela and Magda
moved from building towers to drawing them, again with an organization by cases.
The eight towers that they found were organized in three cases: (1) one blue cube, (2)
two blue cubes, and (3) all one color. Each case was locally organized, as shown in
Fig. 9.1. After thinking about their findings, they developed a general rule; according to what they called Angelas Law of Towers, the number of n-tall towers when
you have x colors to choose from is xn . Thus Angela and Magda not only provided a solution to the specific four-tall towers problem posed, but they also posed
a generalization from towers with two colors to towers with x colors.
Case 1
Fig. 9.1 Angela and Magdas

list of three-tall towers
Case 2
B Y
Case 3
99
9.2.2 Sherly and Ali

Sherly and Ali were new to the study, participating for the first time in grade 11.
Their tower building strategy was also similar to that of the third- and fourth-grade
students who, when first encountering the towers problem, used the strategy of
grouping towers in pairs of opposites. (When the towers are placed side by side,
the cubes in corresponding positions are opposite colors.) They found all 16 four-tall
towers using this strategy. When they were asked to predict the number of three-tall
and then five-tall towers, they made a prediction based on patterns. They conjectured that since the number of four-tall towers is 16 (4 times 4), there would be 9
three-tall towers (3 times 3) and 25 five-tall towers (5 times 5). After realizing that
there are only 8 three-tall towers, they revised their conjecture for five-tall towers
to 24, which follows a pattern of multiples of 8. During the whole group discussion (mentioned below), they were exposed to the conjecture of the other groups
(that there would be 32 five-tall towers), but they were not convinced. As part of
their work on five-tall towers, Sherly and Ali were asked to investigate the case of
five-tall towers with exactly two blue cubes. They organized this group of towers
by cases: (1) two blue cubes together; (2) two blue cubes separated by one yellow
cube; (3) two blue cubes separated by two yellow cubes; and (4) two blue cubes
separated by three yellow cubes; using this organization they found all ten towers
that satisfied this condition. Refer to Fig. 9.2 for a diagram of the two blues separated by one yellow case. During the discussion of this work, Sherly provided a
proof by contradiction: When a researcher asked if the two blue cubes could be separated by four yellow cubes, Sherly noted that this could not happen, because there
would be six [cubes in the tower] then. We see that even though Ali and Sherly had
not previously been exposed to the way younger students in the longitudinal study
worked, they had by the end of the session adopted some of the methods that had
been developed earlier by their classmates.
Fig. 9.2 Sherly and Alis
organization for two blues
separated by one yellow
9.2.3 Michelle and Robert

Michelle and Robert were both in the longitudinal study from the first grade.
Although Robert and Michelle sat along side each other, they used different organizational strategies. Michelle built towers randomly; she said it was without any
set plan, and she rearranged the towers into pairs that she called twos. (These
100
Fig. 9.3 Roberts three cases
of four-tall towers with
exactly two blue cubes

Case 1: blue on top
Case 2: blue on second

position from top
Case 3: blue on third

position from top
were pairs of opposites like those built by Sherly and Ali, described above.) Robert
immediately built and organized the towers by cases, focusing on the blue cube; his
five cases were zero, one, two, three, and four blue cubes.
When asked to show that he had all the possible four-tall towers with exactly
two blue cubes, Robert provided a justification based on controlling for variables.
He showed how he held the upper blue cube in a fixed position beginning at the
top, while he moved the lower blue cube down one position each time he built a
new tower. When the lower blue cube had been moved down to all of the possible positions, the upper blue cube was moved down one position. The process was
repeated until it was not possible to move either blue cube down. With these three
cases, Robert demonstrated that there were six towers in the two-blue case. Refer to
Fig. 9.3 for a diagram of Roberts organization.
Robert and Michelle next worked on three-tall towers. Robert showed that the
three-tall towers could be built by removing the top cube from each four-tall tower
(giving two identical sets of 8 towers) and then removing duplicates. This reasoning
foreshadows Roberts use of inductive reasoning later in the session. Robert and
Michelle went on to chart their results for the total number of towers two-tall through
five-tall. See Fig. 9.4.
Their entry for two-tall towers is incorrect (the actual number is 4), but the rest
of the numbers are correct. They found a pattern that they believed applied to towers
of height 3 and taller, and Robert used that pattern to predict that there would be 64
six-tall towers. Their rule for the number of towers was 2h , where h is the height of
the tower.
The researcher then asked them to consider the question of how many five-tall
towers would have exactly two blue cubes. Robert and Michelle built those towers,
using Roberts strategy of controlling for variables by holding the top cube fixed and
then moving it successively lower in the tower. Robert and Michelle did more than
Height
Fig. 9.4 Robert and

Michelles chart for number
of towers for height 26
2
3
4
5
6
Number of towers
2
8
16
32
64 (prediction)
Fig. 9.5 Roberts results for

the number of towers with
two blue cubes
Height
2
3
4
5
6
7
8
101
# of towers with exactly 2

blue cubes
1
3
6
10
15
21 (prediction)
28 (prediction)
Difference
2
3
4
5
6
7
was asked, finding that for towers with heights two, three, four, five, and six, there
are one, three, six, ten, and fifteen towers with exactly two blue cubes, respectively.
Robert predicted, based upon the pattern plus two, plus three, plus four, plus five
(see Fig. 9.5), that for seven-tall and eight-tall towers, there would be 21 and 28
towers with exactly two blue cubes.
Robert then looked for an explicit formula for the number of towers with exactly
two blue cubes for any height. With his explanation for why a formula would be
useful, we observe Robert thinking like a mathematician:
What if someone just, they just did this problem for the first time, and they just came up
with like how many two yellow for fifteen, and they wanted to find out for fourteen without
doing it? How would they do that? . . . We are not going to sit down and write out one plus
two plus three plus four plus five . . .
Roberts table of results indicates that he developed a rule: multiply the height by
half the height minus point five. This corresponds to the formula for the number
of combinations when selecting two objects from a set of h objects:
h ( h 1)
h!
h =
=
=h
2
2 ( h 2 ) ! 2!
(( h 2 ) .5)
There was an extended discussion of this formula; the researchers wanted to

know where 0.5 came from and how Robert thought of it. He said, I dont know, it
just seems to work. Again thinking like a mathematician, Robert wondered aloud if
the formula would work for towers build from three colors. (He conjectured that the
2s in the formula would be replaced by 3s.) Although Robert did not pursue this
thought, this led to a discussion of the general towers formula. A researcher asked
the students to illustrate how powers of 2 gave the numbers of towers for various
heights. In the following discussion, we see how Robert, initially confused about
the meaning of 21 , comes to an understanding with the help of Michelle and goes
on to describe the use of the general rule.
102
RESEARCHER:
ROBERT:
MICHELLE:
ROBERT:
RESEARCHER:
ROBERT:
RESEARCHER:
MICHELLE:
ROBERT:
RESEARCHER:
ROBERT:
RESEARCHER:
ROBERT:
Show me what it would be for two to the first. What would the
towers be?
Uh, two blues. This and this. [Robert indicates one tower with
two blue cubes, and one with two yellow cubes].
Its just ones, so it would just be one, one.
Yeah, total combinations likeTwo to the first would be, what would they be, show me.
Right here.
No.
No, no. That isnt two to the first; this is two to the first.
[Michelle holds one single blue cube and one single yellow
cube.]
Ah.
Okay, so two to the first would these guys. Two to the second,
according to your theory would be how much?
Oh, right here. Four. We have them right here. [Robert indicates
the four two-tall towers that were already built.] Theres four.
How did you get from here [one-tall towers] to here [two-tall
towers]?
We just built on top of them, I guess.
Robert followed up by demonstrating how each one-tall tower would generate two two-tall towers: one with a blue cube on top and another with a yellow
cube on top, for a total of four two-tall towers. Refer to Fig. 9.6. Later, when the
researcher asked Robert to show that his three-tall tower list was complete, he introduced an inductive argument based on the procedure he used to generate the two-tall
towers:
ROBERT:
RESEARCHER:
ROBERT:
RESEARCHER:
And then here [the set of four two-tall towers] you have all you
could have on the bottom, and you are just adding to the top,
I guess. I guess you just take this [a two-tall tower] and add a
blue and a yellow to the top. All the way through, and then
you know what I am saying?
Does that make sense?
Yes, it does.
Does it keeping going?
Fig. 9.6 Robert generates

four two-tall towers from the
two one-tall towers
ROBERT:
103
Yeah, I guess it would keep going on forever. Thats why that

thing works. Because you are just adding an extra set of two.
9.2.4 Group Work

Robert and Michelle were invited to discuss their inductive argument with Magda
and Angela. They set up towers as shown in Fig. 9.6.
A portion of their discussion follows.
MAGDA:
ROBERT:
ANGELA:
ROBERT:
ANGELA:
ROBERT:
MICHELLE:
ANGELA:
You kind of add that one on top of that color?

We just took this [one of the two-tall towers] and added a yellow
and blue, and took this [another two-tall tower] and added a yellow
and a blue and took that and added a yellow and blue like for all of
them. [Robert indicates each two-tall tower in turn.]
Oh.
Do that for all of them and they grow another row. You just keep
going. Add yellow and blue to each one. Thats it. It looks nice.
Thats very lovely
She said it branches.
It branches.
Yes it does.
The researcher asked if this process would work with three different colors.
Angela responded that there would be three branches and more towers (There
would be three little thingies. . . . It would be a lot bigger), and she added, Would
you like to see our theory? It is x to the n. Robert responded, Yeah, we have that
same theory. He directed Michelle to change their formula from 2h (the total number of towers of height h when there are two colors) to xh , where x is the number of
color choices and h is the height of the tower.
Four students (Robert, Michelle, Magda, and Angela) worked as a group to write
up this result; Robert dictated, Michelle did the writing, and Magda and Angela
observed. Robert said they had to carefully explain their notation so that any reader
would understand what they had written and so that all the students would be able
to explain it. Refer to Fig. 9.7 for their write-up.
Xh
X = # of colors
h = height
Fig. 9.7 The groups general

towers result
Tells all combinations
104
9.3 Discussion
In this chapter, we discussed a session in which students working in small groups
and in a larger group gave specific and general answers, with justifications, to
mathematical questions. They found formulas to answer specific towers questions;
they generalized the formulas to handle general towers questions; they often gave
convincing justifications for their answers, using methods including controlling for
variables, justification by cases, and induction; and they explained most of the
patterns they found by showing how the patterns made sense in the context of the
problems. Although Robert sometimes took the lead in moving further and creating understandable explanations, all students (even those separated and new to the
study) worked diligently on the problems, took care to explain their thinking, and
made convincing arguments and justifications.
In the next chapter, we observe another cohort of 10th-grade students working on
the towers and pizza problems. They too found ways to make sense of the problems,
generalize answers, and provide convincing justifications for those answers. In addition, their work led them to a deeper understanding of the binomial coefficients and
Pascals Triangle.
Chapter 10
Representations and Connections

Ethel M. Muter and Elizabeth B. Uptegrove
Dates and Grade:

Tasks:
Participants:
Researchers:
December 1997 through March 1998; Grade 10

Towers and pizzas
Carolyn Maher and Robert Speiser
10.1 Introduction
In the previous chapter, we viewed a cohort of high school students from the longitudinal study as they explored the towers problems. In this chapter, we observe a
different cohort of students also exploring the towers and pizza problems. In the five
sessions discussed here, spanning December 1997 through March 1998, the five students in this cohort were reintroduced to the towers and pizza problems, which they
last explored in elementary school as described earlier in Chapters 5 and 6. They
found general solutions to those problems and a way to organize their solution lists
to prove that all solutions were present. They recognized that those problems were
related to each other, to the binomial coefficients, and to Pascals Triangle. They
made use of their understanding of the structure of those problems to form preliminary ideas about the meaning of Pascals Identity. We show how their development
and use of a sophisticated general representation scheme helped them make these
connections and generalize their knowledge.
10.2 Session 1: A Common Notation

On December 12, 1997, when they were in the tenth grade, Ankur, Brian, Jeff,
Mike, and Romina began to meet with the researchers as a group for 1- to 2-h
after school sessions that continued throughout high school. At their first meeting,
E.B. Uptegrove (B)


105
106
E.M. Muter and E.B. Uptegrove
the researchers proposed the pizza problem to them as they snacked on pizza.
When first presented with the four-topping problem in the fourth grade, the students
had enthusiastically tackled the problem by randomly generating combinations. As
described in detail in Chapter 5, in approximately 15 min they found all 16 possibilities by using an alphabetical coding scheme to represent each pizza as it was
generated.
In tenth grade, working on the three-, four-, and five-topping pizza problem, Jeff
and Romina utilized an alphabetical coding scheme similar to the one that they used
in fourth grade (p for pepperoni, m for mushroom, etc.). Ankur and Brian used a
numerical coding scheme (1 through n for the n different toppings). Mike, however,
worked alone. He selected a unique binary number coding scheme to keep track of
the combinations. The binary representation became a useful tool in many of their
future discussions of combinatorial problems.
As the first four students discussed the problem, they realized that they needed a
common notation and adopted the alphabetical model. They kept the plain pizza separated, and so they found 7 possible pizzas for the three-topping case (plus plain) and
15 pizzas for the four-topping case (plus plain). When they found 30 five-topping
pizzas (plus plain), they realized it did not fit the pattern. They hypothesized that
a doubling rule might be involved and they decided to rethink their solution. At
this point, Mike re-entered the discussion in order to introduce his binary coding
scheme. He proposed that pizzas be represented by binary numbers; a four-topping
pizza would be represented by a four-digit binary number, with a 1 in the kth digit
representing the presence of the kth topping and a 0 representing the absence of
the kth topping. For example, all one-topping pizzas are represented by all fourdigit binary numbers with exactly one 1: 0001, 0010, 0100, and 1000. This coding
scheme was more easily generalizable than the letter code scheme: to add another
topping, just add another binary digit. After listening to Mikes explanation of the
binary code, the group made a connection between the two representations (letter
codes and binary notation: use the letters that stand for toppings as column headers
for the list of binary digits). Refer to Fig. 10.1 for their table; O stands for onion,
M for mushroom, P for pepperoni, and S for sausage.
Mikes understanding of the binary system and the way it could be used to
describe the solution to the pizza problem gave him an insight into a generalization
about the number of pizzas that can be created from n toppings; Mike hypothesized that the answer to the n-topping pizza problem is 2n . The group discussed the
numerical solution for some time there was some confusion about whether the
coefficient might be n 1 or n + 1, or whether the answer might be 2n 1, possibly
Fig. 10.1 Students table

linking topping codes and
binary notation [annotation
added]
onion pizza
mushroom pizza
10
107
due to the uncertainty about how to count the plain pizza, but ultimately all agreed
that the solution was 2n .
As they were wrapping up the session, the researcher asked if the problem
reminded them of other problems, and Brian mentioned towers: Every thing we
ever do always is like the tower problem. In order to investigate the possible relationship between the two problems, the students worked on the three-tall tower
problem and concluded that the answer was the same as for the three-topping pizza
problem; there are eight three-tall towers, just as there are eight possible threetopping pizzas. Because they were focused on relating the pizza toppings to the
cubes colors, they concluded that the problems were similar but not identical.
Ankur noted, for example, that a red-yellow tower is different from a yellow-red
tower, but a pepperpepperoni pizza is the same as a pepperonipepper pizza.
10.3 Session 2: Towers and Pizzas

One week later, the students returned and resumed their discussion of a possible
relationship between pizza and towers problems. Although they were asked to consider only the two-color towers problems, they kept returning to the question of how
to count the possible number of towers when there were cubes of three or more colors. Looking at this issue led them to the realization that when the height of the
towers is the only variable under consideration, the towers problem is identical to
the pizza problem. This time, they mapped the height of the tower to the number of
pizza topping choices (contrary to the previous week, when they attempted to map
number of colors to number of pizza topping choices). A portion of their discussion
follows.
JEFF:
MIKE:
JEFF:
MIKE:
JEFF:
ANKUR:
JEFF:
ANKUR:
JEFF:
ROMINA:
ANKUR:
If the only variable were changing is height, it stays the same.

It would be the same as the pizza.
What would that be like changing on the pizza, though?
You could change the height, the number of toppings.
Changing the height would be like changing the number of toppings.
Yes.
Changing the color would be like, what?
Say what you just said again.
All right. When we change the height of the box, from like two to
three, its like changing the topping on the pizza from a possible two
toppings to three toppings.
Okay.
Okay.
Brian and Mike returned to the question of changing the number of colors available for building towers, and Mike proposed that when there are three colors to
choose from, there are nine possible two-tall towers (3 times 3). After a 10-min
discussion, the other four students agreed. In the course of the discussion, they
attempted to clarify the meaning of the base and exponent in each problem. In doing
108
this, they were able to answer a question that they had not answered in the previous
session, that of the meaning of the 2 in the formula 2n . In the previous session, they
had determined that n represented the number of toppings, and in this session, they
determined that the 2 stands for the two toppings choices: on or off the pizza:
BRIAN:
MIKE:
Two has to stand for something.

It stands for something; n was the number of toppings and 2 is what
you could either have 0 or 1. You either have a topping or not.
Later, Romina presented the groups findings to the researchers:

ROMINA:
MIKE:
JEFF:
JEFF:
ROMINA:
JEFF:
Okay. For the pizza problem, the 2n [meaning 2n ], the two represents
either topping or no topping. Right?
Theres two different possibilities for each.
Thats why theres two. We didnt know, I dont think we explained
that last time, why it was two.
Topping or no topping, and thats what the two is. Now the n, Romina.
Is toppings.
The number of toppings.
Mike used binary notation again in this session, this time using it to represent the
two colors of the towers problem. He noted that instead of relating binary digits to
the presence or absence of pizza toppings, he could relate them to colors of cubes:
Zero is blue and one is red. Figure 10.2 shows Mikes table of solutions for both
the two-topping pizza problem and the two-tall towers problem. The column headers
1 and 2 represent the two pizza topping choices and the two levels of the tower. The
1 and 0 represent topping/no topping and blue cube/red cube.
Mike explained that if the labels at the top of the chart stood for pizza toppings
(m for mushrooms and p for pepperoni), the zeros and ones would represent the
presence or absence of the topping. If the labels stood for positions in towers, the
zeros and ones would represent the color of a cube; e.g., one would represent blue
and zero would represent red.
During the second half of this session, at the researchers request, the students
explored geometric interpretations of the binomial expansion. Figure 10.3 represents their drawing of (a+b)2 . They went on to spend over half an hour working on
drawings and three-dimensional models for (a+b)3 , although no model was entirely
satisfactory to them. These investigations can be seen as preparation for their later
work describing the isomorphic relationship among the towers and pizza problems
and the binomial expansion.
Fig. 10.2 Mikes listing of

two-tall towers and
two-topping pizzas
1
0
1
0
0
1
1
0
1
0
1
0
0
1
1
0
10
109
Fig. 10.3 Geometric

interpretation of (a + b)2
a2
ab
ba
b2
In this session, the students gave the researchers clear and convincing explanations of the isomorphism between the pizza and towers problems and of the meaning
of the components of the formula, and it appeared that they had a firm grasp of
the underlying structure of the problems. But a few months later (in sessions discussed later in this chapter), we see them return to their focus on relating number
of pizza toppings to number of colors (instead of to the height of the towers), once
again deciding that the problems were similar but not identical. This illustrates how
important it is to revisit problems and re-examine solutions, to solidify and expand
students understanding.
10.4 Session 3: Towers and the Binomial Expansion

When the students met in January 1998 after the holiday break, they returned to the
topic of towers. The researchers gave them a problem from fourth grade: When you
are choosing from red and yellow cubes, how many five-tall towers can you build
containing exactly two red cubes? They immediately answered ten, and they were
then challenged to provide an explanation. Mike and Ankur provided a justification
in approximately 2 min, using Mikes binary coding scheme, with 0 representing
a yellow block and 1 representing a red block. As Ankur explained their solution,
their organization improved; they begin to control for variables by holding the red
cube fixed in the top position and then moving the second red block into successively
lower positions until it reached the bottom position. This process was repeated, holding the red cube fixed in the second and then third and fourth positions. Figure 10.4
shows their original list followed by the re-organized list of all ten towers. This
Fig. 10.4 Mike and Ankurs

two lists of five-tall towers
with exactly two red cubes
1
1
0
0
0
0
1
1
0
0
1
1
0
0
0
1
0
1
0
0
1
0
0
1
0
1
0
1
0
0
0
0
0
1
1
0
0
1
1
0
1
0
0
0
1
0
1
1
0
0
0
1
0
1
0
1
0
0
1
0
0
1
0
0
1
0
0
1
1
0
0
0
1
0
1
0
1
0
1
0
1
0
0
0
1
0
0
1
0
1
0
0
0
1
1
0
1
0
0
1
110
illustrates the importance of revisiting and re-explaining answers; in the process

of explaining their solution, they organized the list so as to make it clear that all
possibilities were accounted for and none were missing.
The other group (Jeff, Romina, and Brian) worked on their solution for approximately 25 min. Their solution depended on first finding all possible five-tall towers;
they recalled from previous work that there are 32 such towers. They built a justification based on cases; their cases were (1) all red: 1 tower, (2) one red and four
yellow: 5 towers, (3) two red and three yellow: 10 towers, (4) three red and two
yellow: 10 towers, (5) four red and one yellow: 5 towers, and (6) all yellow: 1 tower.
While Brian, Jeff, and Romina were working on their solution, and Ankur
and Mike were done, Ankur proposed a problem that became known as Ankurs
Challenge. The groups work on this problem was discussed in Chapter 8.
At the end of this session, the researcher introduced some of the notations of
combinatorics. She told the students that asking how many five-tall towers have
exactly two red cubes is the same as asking how many combinations there are when
selecting two of five objects. She showed four different ways to write this, as shown
in Fig. 10.5.
She concluded with a discussion of the binomial expansion and Pascals Triangle.
Following up on the previous sessions investigation of the binomial expansion, she
wrote the expansion of (a + b) to powers 03, drew Pascals Triangle, and then asked
the students to think about the relationship. (Refer to Fig. 10.6 for the researchers
drawings).
In the following excerpt, the researcher hinted about the relationships that the
students were in the process of discovering.
RESEARCHER:
ANKUR:
RESEARCHER:
The question is, whats the relationship here? How could you
model it? How could you show this relationship? And why does
it work? Thats the question. So thats sort of the direction. Are
you interested in knowing that? I think you have the bits and
pieces to put it together.
Some of the pieces are really small.
Theyre bigger than you think. Youve been working on this for
a long time.
5
5C2 C5,2 2
Fig. 10.5 Notation for selecting two of five objects
Fig. 10.6 The binomial

expansion and Pascals
Triangle
C25
10
ROMINA:
RESEARCHER:
ANKUR:
BRIAN:
111
Is this what we did today though?

Youve been dealing with some of this today. So think about it.
So are all of the things we learned for the past 8 years sort of
combined into one thing?
Imagine that.
Immediately following that discussion, the researcher asked the students to make
concrete the numbers in Pascals Triangle, by thinking about them in a very real
way (linking them to towers problems).
RESEARCHER:
ANKUR:
RESEARCHER:
ANKUR:
RESEARCHER:
ANKUR:
RESEARCHER:
ANKUR:
RESEARCHER:
ANKUR:
RESEARCHER:
When you first came in here today, you produced that number
ten. [She refers to the first 10 in row 5 of Pascals Triangle 1
5 10 10 5 1.] Right?
Yes.
And what problem were you solving?
Two were red and three something else.
Okay. So you can think of that ten in a very real way, if you
want to, right?
Yeah.
Can you think of those other numbers in a real way? Does that
help?
The 1 is, in 1 4 6 4 1, the 1 represents all red. The other 1
represents all yellow I guess . . .
All red and all yellow for what?
Of four high.
So this is four high. [The researcher points to row 4 of Pascals
Triangle.] And these are all red. [The researcher points to the
first 1 in that row.]
This marks the first time the towers problem was explicitly linked with Pascals
Triangle, when row 4 of Pascals Triangle was connected to the four-tall towers problem. Before the session ended, the researcher asked the students to think
about the meaning behind the addition rule for Pascals Triangle in the specific
case of how the 6 in row 4 was generated from the two 3s in row 3. Although
the students did not offer an answer at this session, it is noted here as the first
time they were asked to think about Pascals Identity. In this session we see three
instances where the students were invited to think about how all the individual problems could be related, both to each other and to abstract mathematical entities,
but without any explicit instruction about how to make a connection between the
problems.
10.5 Session 4: Pizzas, Towers, and Pascals Triangle

In this session, three students (Ankur, Jeff, and Romina), in a first meeting with visiting researcher Robert Speiser, explored the relationships among the pizza problem,
112
the towers problem, and Pascals Triangle, and (for the first time) they discussed
Pascals Identity in terms of operations on physical objects (adding cubes to towers).
When Professor Speiser asked the students about their recent work, the students
did not mention a relationship between the towers and pizza problems. Instead, as
they had initially done back in December, they maintained that the problems were
different. Romina, recalling Ankurs earlier reasoning, said that red-blue cubes on
a tower are different from blue-red cubes, but sausagepepperoni is the same as
pepperonisausage. Ankur added that a five-topping pizza problem is like a fivecolor towers problem. Jeff agreed, saying that a tower could have two of the same
color but that a pizza could not have pepperonipepperoni. Although all three had
participated in the earlier, correct, discussion of the relationship between the towers and pizza problems, they recalled now only their own original ideas. They said
nothing about the height of the tower being connected to the number of toppings, or
how on-the-pizza/not-on-the-pizza could be made to correspond to blue/red cubes
via the binary representations 0 and 1.
The researcher reminded the group about the combinatorics notation that had
been introduced a month earlier and she reminded them how the notation was related
to the five-tall towers problem. She went on to demonstrate the binomial expansion
and to ask explicit questions: What are the relationships, if any, among (a+b)5 , the
five-tall towers problem, the five-topping pizza problem, and the fifth row of Pascals
Triangle? This question is significant in terms of the students later work, as it represents the first time the students were asked to think about a four-way link, among
the binomial expansion, the two combinatorics problems, and Pascals Triangle. The
students were able to make the connection, evoking and expanding Ankurs explanation from the previous month of how the four-tall towers problem could be found in
row 4 of Pascals Triangle. (This also anticipated their night session explanation of
entries in Pascals Triangle in terms of pizzas.) In the following excerpt, the students
linked the binomial expansion to the towers problem.
RESEARCHER:
ROMINA:
ANKUR:
RESEARCHER:
ANKUR:
RESEARCHER:
ANKUR:
What are the as and the bs here?

Colors. . . .
a and b is red and blue. . . .
What do you mean by red and blue?
a is red and b is blue. Thats [red-blue tower] a b. So b a would
be a blue red.
So how, if you have them in front of you, how would they look
different?
Red and blue, reds on top, and blues on the bottom. Blues on
top and reds on the bottom.
In the following episode, Ankur explains how to find the answers to the five-tall
towers problem in row 5 of Pascals Triangle, and then Jeff and Romina locate the
pizza answers in row 6. (Row 6 of Pascals Triangle contains the numbers 1 6 15 20
15 6 1.)
ANKUR:
JEFF:
This [1] is no red.

Yeah.
10
ANKUR:
JEFF:
ANKUR:
ROMINA:
ANKUR:
JEFF:
ANKUR:
JEFF:
ROMINA:
JEFF:
ROMINA:
ANKUR:
JEFF:
ROMINA:
ANKUR:
ROMINA:
ANKUR:
JEFF:
ANKUR:
JEFF:
ROMINA:
JEFF:
ROMINA:
JEFF:
ROMINA:
JEFF:
ROMINA:
ANKUR:
113
So theres one with no red. Theres six with one red. . . . Theres fifteen
with two reds. Twenty with three reds. Six with five reds.
And one with noAnd one with noNo.
No. Six reds.
One with six reds. . . . All right. Now. What does that have to do with
pizza?
Just relate the tower problem to the pizza problem.
Well, were saying that this [1] is a pizza with just plain.
Yeah. Thatll be the plain pizza.
Plain. This [6] is with all your six toppings.
Thats with one topping.
You cant exactly relate these numbers to the pizza problem.
Well, well try really quick.
Yeah. You can. Cause this [1] is plain, just plain pizza.
And what will the other 1 represent?
With everything on it.
Okay.
So this is plain.
Okay. Six withWith one of each. Fifteen is withTwo toppings.
Just two toppings out of your six. Twenty is with three toppings.
Fifteen is with the four toppings. Six is with the five toppings.
Five toppings.
And the other one isAnd the one is with all of them.
Like the supreme.
Is that good?
Cool. Were on fire today.
Thus Ankur, Jeff, and Romina used the two combinatorics problems they knew in
order to explain the numbers in Pascals Triangle. This was the first time they were
observed connecting the pizza problem to Pascals Triangle. Although Ankur had
initially been reluctant to attempt a definition of the relationship between Pascals
Triangle and the pizza problem (You cant exactly relate these numbers to the pizza
problem), he still participated in the discussion and at the end expressed satisfaction
with their work. (Were on fire today.) We noted earlier that the students received
no special concrete rewards for participation in the study. Ankurs remarks illustrate
our belief that the intellectual enjoyment involved with solving difficult problems
was a factor in the students continuing involvement in the study.
After explaining the link between specific numbers in Pascals Triangle and the
pizza and towers problems, the students described an instance of the addition rule in
terms of towers problems. They explained the instance of Pascals Identity shown in
114
Fig. 10.7 One instance of

Pascals Identity which
students linked to towers
problems
All red
Add blue
Add red
10
10
15
20
15
All blue
Fig. 10.7. They described the two 10s in row 5 and of the 20 in row 6 as counting
classes of five-tall towers, and they described the process by which the 20 (counting a different class of six-tall towers) could be generated from the two 10s. The
transcript below gives portions of their discussion.
RESEARCHER:
What are those tens counting? And what does the twenty count?
JEFF:
ANKUR:
ROMINA:
ANKUR:
The tens showThe tens show two of one color.

And three of another.
One color and two of another color. . . . Thats why its ten and
ten. But then, at the top of each one, you can put eitherYou could either put a red or like blue.
The first ten in that row of five high has two reds and three
blues? Were counting reds?
Yes.
And the second ten hasThree reds and-three reds and two blues. Now coming down here, the twenty
is supposed to count the ones that have three of each.
Three red. Three reds and three blues.
Right.
So how do the two tens add to give the twenty?
Because in these ten, where theres three reds and two blues,
you want to make it three reds and three blues. So you put a
blue on top of each one.
JEFF:
RESEARCHER:
ANKUR:
RESEARCHER:
ANKUR:
RESEARCHER:
ANKUR:
JEFF:
RESEARCHER:
ANKUR:
This is the first instance where a connection was made between Pascals Identity
and a specific concrete combinatorics problem.
10.6 Session 5: Towers, Pizzas, and Pascals Triangle

Ankur, Jeff, Mike, and Romina attended this session, 4 weeks after session 3. Much
of the session was devoted to attempts by Ankur, Jeff, and Romina to explain to
Mike specific instances of Pascals Identity in terms of towers and the binomial
expansion. This session also provided another example of the importance of revisiting problems and re-explaining solutions. In the February session, described above,
the students had successfully explained the addition rule shown in Fig. 10.7 in
10
Fig. 10.8 Another instance

of Pascals Identity discussed
in terms of towers problems
115
1
1
3
4
1
4
terms of towers: to create the 20 six-tall towers that have three red and three blue
cubes, add a blue cube to the 10 five-tall towers with two blue cubes and add a red
cube to the 10 five-tall towers with two red cubes. But this time, when asked to
explain the similar case shown in Fig. 10.8, they did not recall the previous (correct)
explanation. Instead, they tried to find a solution that would involve disassembling
and re-assembling existing towers, an approach that surprised and confounded the
researcher.
They correctly mapped the numbers 1 and 3 into tower groups (1 = white-whitewhite and 3 = blue-white-white, white-blue-white, and white-white-blue, as shown
in Fig. 10.9).
But they tried to explain the 4 by breaking apart the tower representing 1 tower
and distributing its cubes among the other three towers. After the researcher questioned this method, Mike gave a different explanation for how to represent the 4.
JEFF:
RESEARCHER:
JEFF:
ANKUR:
RESEARCHER:
ANKUR:
ROMINA:
MIKE:
ANKUR:
Weve got this [the white-white-white tower, representing 1].

And were saying how this goes together. [Jeff has assembled
the three towers each with one blue cube to represent the 3.
Refer to Fig. 10.9.] Were saying- [Jeff starts to dismantle the
white-white-white cube.]
No. No. Dont take that apart. BecauseWell, thats why I made this. So I could.
We made another one so we can take that one apart. . . . And
show you.
You mean, you mean, you mean you get the four by taking
something apart?
Youre not taking it apart.
Youre not taking it apart; youre just seeing how they go
together. . . .
You dont really have to take it apart to show this, cause look.
Each one, the reason why they combine, each one of these four
blocks [towers] is going to have something added to them to
equal the same thing.
Yeah.
Fig. 10.9 Students map

specific towers to 1 and 3 in
row 3 of Pascals Triangle
116
MIKE:
ANKUR:
MIKE:
ANKUR:
MIKE:
These blocks [towers] are going to have, theyre going to have a

white block added to them. [Mike indicates the three three-tall
towers with one blue cube.]
Theyre going to have a b added to them.
And this ones [the white-white-white] going to have a, a blue
added to it.
An a added to it.
And theyre going to equal the same thing. Thats why youre
going to have the four. [Refer to Fig. 10.10 for a diagram of
Mike and Ankurs suggestions.]
The other students accepted Mikes explanation and apparently this time they
comprehended the process, as evidenced by their later work in the night session (see
Chapter 12) and subsequent interviews. In addition, Ankur reiterated a link noted
in the February session, observing that the as and bs in the binomial expansion
could be connected to the blue and white cubes, respectively, in the towers. The
same connection was explained a year later during the night session.
Next, the researcher asked the students to relate the tower problems to binary
notation and the pizza problem; she said, If you had to make up a pizza problem
to model this row [row 2 of Pascals Triangle], whats the pizza problem? Ankur
reiterated the position that he had taken in two previous sessions, that a peppers and
pepperoni pizza is the same as a pepperoni and peppers pizza; it appeared that he
had not yet firmly established that there was a connection between the two choices
of cube colors for each cube in a tower and the two choices for each topping on
or off the pizza. Although the group noted that the nth row of Pascals Triangle
could be linked to the n-topping pizza problem, they did not propose an explanation about how to use the numbers in the nth row to enumerate pizzas. When the
researcher asked for clarification and Ankur insisted that there was no relationship
between colors and pizza toppings, Mike interrupted with his own explanation. It is
Fig. 10.10 Mike and Ankur

illustrate 1 + 3 = 4 by adding
cubes to towers
10
117
interesting to note the similarity between this episode and the earlier one in which
the students discussed how to connect pizza problems to Pascals Triangle. In the
earlier episode, Ankur initially denied a connection. Also, in that episode, when the
connection was established, Ankur, with Jeff and Romina, quickly caught on and
proceeded as enthusiastic participants in the exploration and explanation process. In
this episode, also, we see a student express satisfaction with the groups intellectual
achievement. (Romina said, Oh, wow!) Figure 10.11 illustrates the four two-tall
towers the group made as part of this process.
RESEARCHER:
ANKUR:
MIKE:
ANKUR:
ROMINA:
MIKE:
ANKUR:
MIKE:
ANKUR:
MIKE:
JEFF:
ANKUR:
ROMINA:
MIKE:
ROMINA:
MIKE:
ROMINA:
ANKUR:
Now wait. Now Im lost again. What, what, what was this?
. . . [The Researcher indicates the single white and blue cubes
representing row 2 of Pascals Triangle.]
The colors dont, dont look at the colors.
No. No. No.
Just look at this [Pascals Triangle]. . . . But the colors dont
specifically represent anything.
Yeah.
Yes. It does.
No, it dont.
Topping. [Mike points to the blue cube.] Or no topping. [Mike
points to the white cube]. Just say like that. And if you look at
it like this, you know.
So all of the whites are no topping?
Yeah. [Mike takes the white-white-white tower.] Then this is a
plain pizza with a choice. If you had a choice of three toppings.
All right.
Okay.
Okay.
This [the blue-white-blue tower] would be a pizzaOh. With the one. Ooh.
-with two different toppings, without the other, third topping.
Thats what I was asking.
Okay.
Pizza with
Second Topping
Plain Pizza
Fig. 10.11 The students link

between two-tall towers and
two-topping pizzas
Pizza with
First Topping
Pizza with
Both Toppings
118
JEFF:
ROMINA:
ANKUR:
ROMINA:
JEFF:
MIKE:
ANKUR:
JEFF:
ANKUR:
JEFF:
ANKUR:
ROMINA:
RESEARCHER:
ANKUR:
ROMINA:
RESEARCHER:
ROMINA:
. . . Well, yeah. Well, if youre just saying that this [the whitewhite-white tower] is the pizza with three no toppings, its plain.
Its just a plain pizza.
All right. All right. So thats [blue-blue tower] two toppings.
Yeah.
Yeah. All right. So.
Thats [white-white tower] . . . a choice of two, but you want it
plain.
You have a choice of two toppings.
Yeah, so this is, this [blue-blue tower] is choice of two using
two. This [blue-white tower] is choice of two using one.
Two using one.
This [white-blue tower] is choice of two using the other one.
Thats using the other one. And thats [white-white tower] using
nothing.
Yeah.
And thats all the possibilities?
Yes.
Yeah.
You like that?
Oh, wow!
During this episode, the other three members of the group immediately accepted
and built upon Mikes brief remarks. All he had to say was topping and no topping, and all three of the others began immediately to form connections between
specific individual towers and specific pizzas. This represented the fifth discussion
of the pizza problem in 4 months, and at least three members of the group apparently
began this discussion without a clear idea of the essential feature of the problem
(topping versus no topping), as opposed to a surface feature (the fact that the toppings could be selected in any order). But it appears that this discussion helped
them finally to make sense of the isomorphic relationship, because the pizza problem was the one that the group selected during the night session a year later, to
explain Pascals Identity.
10.7 Discussion
During these five tenth-grade problem-solving sessions, the students worked independently, sometimes spontaneously splitting themselves into subgroups, sometimes working individually, but always sharing their ideas with the other members
of the group. By sharing, the students were able to incorporate others ideas into
their own understanding of the justifications. An example is Mikes introduction of
the binary notation code. Mike watched Jeff and Romina work for a while, listened
as the other four students exchanged ideas, and then began focusing on his own
paper. He later presented his conception of how the binary system mapped onto the
10
119
solution of the pizza problem. Mike recalled an episode from an eighth-grade class
and applied his previously constructed knowledge to a totally new situation. His
introduction of a coding system, the zeroes and ones of the binary system, to the
justification being built by the group of five students was an important contribution.
It became the students notation of choice for future problems.
Over the course of these sessions, we observed the students investigating problems that had been explored in earlier years, retrieving earlier ideas and images
as they built solutions and justifications. These ideas and images sometimes reappeared just as they were formed in the prior occurrences. In the third session of the
sophomore year, we see Mike and Ankurs swift production of a justification for the
number of five-tall towers with exactly two red cubes. They reproduced a justification by cases that had originally been built in their fifth-grade classroom, but using
Mikes binary notation. In addition, they offered a second justification, utilizing a
strategy that depended on controlling for variable that was first introduced by Ankur
while solving a pizza problem in grade five. His ownership of this strategy allowed
him to adapt it for use in the isomorphic block tower problem.
For the same problem, Jeff retrieved a strategy used during that same fifth-grade
session. Mike and Ankur had enthusiastically participated in the whole classroom
discussion, which culminated in a proof by cases. Although Jeff was in the room,
he was not focused on the classroom discussion; he was looking at patterns in the
towers that he built. Jeffs partners, Romina and Brian, also had more difficulty in
providing a justification during the session in their sophomore year. In grade five,
while Mike and Ankur were active participants in the classroom discussion and Jeff
was quietly pursuing his own line of thinking, Brian and Romina were in another
classroom. Although they worked on the five-tall towers problem, their class did
not offer a convincing justification for the answer to that problem. The difficulties
experienced by these students as they worked on the block tower problems as tenth
graders might be explained by the absence of some earlier experiences. They constructed the images and representations to the block tower problem for the first time
in this tenth-grade experience.
In the attempt to think about the potential connection between the pizza and
block tower problems, the students came to discuss many powerful mathematical
concepts. While they were able at an early point to determine that the answer to the
n-topping pizza problem is 2n , they came to this number by recognizing the pattern
of {2, 4, 8, 16, 32, . . .}. They determined that n represented the number of toppings,
but did not provide a satisfactory explanation for the base 2 until they began the
discussing the possible relationship between pizzas and block towers; then that they
came to see that the base 2 represented the presence or absence of a topping. This
realization came, not from working on the pizza problem, but instead as a result of
their search for an answer to the three-color four-tall tower problem. Thus we see
that the opportunity to work on open-ended problems and follow paths determined
by the interest of the moment can lead to greater understanding of other problems. In
this case, the opportunity to investigate an isomorphic problem provided the students
with the tools necessary to complete the formulation of the imperfectly developed
earlier idea.
120
In summary, the students investigated isomorphic problems in combinatorics and

used them to explore how Pascals Triangle grows and to make sense of Pascals
Identity. Between December 1997 and March 1998, they first found general solutions to the pizza and towers problems, using letter and number codes and binary
notation to enumerate the pizzas and towers. Then they organized their lists of solutions, organizing the pizza problem solutions according to number of toppings and
the towers problems solutions according to the number of cubes of one color. These
lists not only provided a way to show that all cases were present, but they also
provided the means to associate those cases with the numbers in Pascals Triangle.
In discussions with the researcher and other researchers, the students described the
isomorphic relationship between the pizza and towers problems. Their extensive
repertoire of representations proved essential; in this process, they made use of
words, written inscriptions, and concrete materials (as when Mike held up a blue
cube and a white cube and said topping and no topping). The opportunity to
revisit problems also proved crucial, as students often needed to have two, three, or
more discussions on the same topic before critical ideas were firmly established.
In the next chapter, we observe another cohort of students who also make sense
of the relationships among towers, pizzas, and Pascals Triangle. They brought
their own experience, their own representations, and their own ideas to the problem, but they too used personal representations, communicated findings, and made
generalizations that showed their increased understanding of these problems in
combinatorics.
Chapter 11
Pizzas, Towers, and Binomials

Lynn D. Tarlow
Date and Grade:

Tasks:
Participants:
Researcher:
March 1, 1999; Grade 11

Pizzas, Towers, and Pascals Triangle
Amy-Lynn, Angela, Magda, Michelle, Robert, Shelly, Sherly,
and Stephanie
Carolyn Maher, Alice Alston, Regina Kiczek, Ralph Pantozzi
11.1 Introduction
In the previous chapter, we observed a cohort of tenth-grade students as they
investigated the connections among the pizza problems, the towers problems, the
binomial coefficients, and Pascals Triangle, leading to their increased understanding of the meaning of the numbers in Pascals Triangle and how the triangle grows.
In this chapter, we see how another cohort of students, composed of two subgroups with different backgrounds (from students new to the study to students who
had been in the study from the start) worked together, sharing their ideas, comparing
representations, and discussing relationships among problems. Through their collaborative work, they too came to discover generalized rules for the pizza and towers
problems, see how both problems were related to Pascals Triangle, and explain the
meaning of Pascals Identity through the use of two different metaphors: solving the
pizza problems and solving the towers problems.
The eight students at this session were organized into two groups of four. Each
group worked independently. The four students at Table A (Robert, Stephanie,
Shelly, and Amy-Lynn) had participated in the tower and pizza investigations in
grades 35 through five. Only Robert had participated in the previous 11th-grade
tower investigation discussed in Chapter 9. Of the students at Table B (Angela,
Magda, Michelle, and Sherly), only Michelle had explored the pizza problems in
L.D. Tarlow (B)

Department of Secondary Education, The City College of the City University of New York,
New York, NY 10031, USA
e-mail: ltarlow@ccny.cuny.edu

121
122
L.D. Tarlow
the early grades, but all had been present at the previous 11th-grade tower session.
This time, the students were given the four- and five-topping pizza problems:
A local pizza shop has asked us to help design a form to keep track of certain pizza choices.
They offer a plain pizza that is cheese and tomato sauce. A customer can then select from
the following toppings: pepper, sausage, mushrooms, and pepperoni. How many different
choices for pizza does a customer have? List all the choices. Find a way to convince each
other that you have accounted for all possible choices. Suppose a fifth topping, anchovies,
were available. How many different choices for pizza does a customer now have? Why?
11.2 Table A: A Connection Between Pizzas and Towers

As the students at Table A began to talk about listing pizzas, Shelly complained that
she just did this in school, combinatorics stuff, but she was not able to remember
a formula, although she thought it involved factorials. Her remarks showed that she
remembered the form but not the meaning; she said:
I dont know if its factorial or combination. I dont know if you would just do like five
factorial plus four factorial plus three factorial plus two factorial plus one factorial. . . . I
cant remember. That was the last section we did. Its so pathetic.
Shelly was correct in that the solution can be seen as a sum (but of five combinations, not of five factorials), although the solution is often seen as a power of 2: two
choices (on or off) for each of four toppings give 16 possible pizzas:

4
4
4
4
4
+
+
+
+
= 24 = 16
0
1
2
3
4
Nevertheless, Amy-Lynn agreed that factorials were appropriate, and so Shelly
took the sum of 1! to 5! and got an answer of 153. Although no one questioned
this use of factorials, the group decided that the answer needed to be verified; and
so they started to list the possible pizzas. Stephanie and Shelly discussed possible
strategies:
STEPHANIE:
SHELLY:
STEPHANIE:
Do we just want to, um, plot out the pizzas, like with shirts and
pants or towers? Do you know what Im talking about?
Yeah. The tree diagram type thing.
Yeah, kind of like that. Or is there an easier way to do it that Im
just not thinking of?
Stephanie, Shelly, and Amy-Lynn then proceeded to use tree diagrams to represent possible pizzas, using letter codes for topping combinations. As an example,
refer to Fig. 11.1 for a diagram of Shellys tree.
All of the students except Robert started out by including plain as a topping to
be combined with other toppings. After a discussion about real pizzas, they decided
that this was unnecessary. As Shelly said, A pepperoni is a plain with pepperoni.
They kept the plain pizza on their lists, though, and just crossed out the duplicates
that resulted. They compared answers and, unprompted, decided to create a new
11
123
pl
Fig. 11.1 Shellys tree of all

16 pizzas when there are four
toppings to choose from
pe
pep
s
s
pe
m
m
pe
pe
s
pl = plain
pe = peppers
s = sausage
m = mushrooms
pe = peperoni
list according to number of toppings. Enumerating this way, the students found that
with four toppings available, there were 16 possible pizzas: one plain, four with
one topping, six with two toppings, four with three toppings, and one with four
toppings (1 4 6 4 1). They recognized these numbers as following a pattern and
belonging in Pascals Triangle, but they also realized that they needed more of an
explanation of the answer. A portion of the discussion between Shelly and Stephanie
follows.
SHELLY:
STEPHANIE:
SHELLY:
STEPHANIE:
SHELLY:
STEPHANIE:
SHELLY:
STEPHANIE:
SHELLY:
STEPHANIE:
SHELLY:
One, four, wait a minute. One four six four one, so the next one
will be one. This is theThe triangle.
The triangle.
Yeah. So the next one is one five ten ten five one.
Were done. [Shelly laughs.]
But what does that mean? [All three girls laugh.]
I dont know. . . .
But what it, like, what does one four six four one. That means
nothing to me.
It means nothing to me either, but its the pattern we saw.
So we have a pattern, but how do we apply it to getting sixteen
pizzas?
That would be the problem.
Stephanie and Shelly followed up with a discussion of how the numbers in row
4 could be matched to the answers to the four-topping pizza problem.
STEPHANIE:
SHELLY:
STEPHANIE:
SHELLY:
STEPHANIE:
So, well, okay, lets figure this is saying that we have one plain
pizza.
Uh-huh.
And then we have four pizzas with two toppings?
With one. Because its the plain and then with the one topping.
Okay. So we have four pizzas with one topping. And we have four
pizzas with two toppings. Oh, no, we have six pizzas with two
toppings, four pizzas with three toppings.
124
SHELLY:
STEPHANIE:
L.D. Tarlow
And one pizza with four toppings.

Okay.
Then they used the next row of the triangle to determine the number of possible
pizzas with five available toppings, 32 (the numbers 1 5 10 10 5 1 representing the
numbers of pizzas with zero through five toppings in order).
After the group found the relationship between pizzas and Pascals Triangle,
Researcher 1 asked the students to explain the addition rule for generating new rows
of Pascals Triangle from existing rows (Pascals Identity) in terms of pizzas. The
students had some trouble figuring out how that would work. Stephanie said, I just
cant get past the fact that you cant make a pizza out of other pizzas. I think maybe
if it was applied to something else I could look at it differently. Robert suggested
that they try relating the towers problem to Pascals Triangle. Robert started making
drawing in order to follow this line of thinking, but the others went back to thinking
about pizzas.
Stephanie and Shelly tried to use pizzas to try to explain one particular instance
of the addition rule on Pascals Triangle: 1 + 3 = 4. See Fig. 11.2.
Stephanie continued to maintain that using pizzas to explain Pascals Identity
did not make sense, because, like, one is no topping, so adding one to three doesnt
materialize another topping. She understood that, in general the nth row of Pascals
Triangle represented all the possible pizzas that could be made with n toppings available, but she could not see how one row generated the next one. She called over
Researcher 4 and he suggested that she think of having made all eight three-topping
pizzas and then finding that another topping had become available. Then Stephanie
realized that each existing pizza could either acquire the new topping or stay the
same. Therefore, the addition rule could be explained by thinking of each existing
pizza either acquiring a new topping or staying the same. After Stephanie was satisfied that she understood how the addition rule worked, the researcher moved on
to another topic with Robert. Robert had drawn rows 05 of Pascals Triangle, and
to the right of each row he wrote the sum of the numbers and the sum expressed
as a power of 2. (Refer to Fig. 11.3.) He gave the general rule: that the number of
possible combinations for pizza toppings is given by 2 to the number of choices.
Researcher 4 asked Robert if he had thought about what role the number 2 played.
Robert replied that he remembered that with towers, the total number of tower combinations was two to the something, and the pizza situation was the same thing.
Amy-Lynn was listening, and she concurred: That was with a lot of the problems,
they went by two and it had something to do with powers.
1
1
1
1
Fig. 11.2 Two instances of

Pascals Identity
3
4
2
3
6
1
4
11
125
1
Fig. 11.3 Roberts diagram

of Pascals Triangle and
powers of 2
20 = 1
1
1
1
4
5
10
23 = 8
1
6
10
21 = 2
22 = 4
1
1
24 = 16
1
25 = 32
The question of the role of the number 2 was temporarily deferred, as Researcher
3 stopped by, and Stephanie demonstrated that she had made sense of the relationship between pizza problems and Pascals Identity by explaining two instances of
the addition rule, first 1 + 3 = 4. (Refer to Fig. 11.2.)
You already have three pizzas with one topping. And the plain pizza becomes the pizza with
the new topping. Okay, so this becomes, instead of one plain pizza, this is one pizza with
one topping. Cause this ones getting like the pepperoni thrown into it. And that produces
the four pizzas with one topping.
Then she explained how 3 + 3 = 6. (Refer to Fig. 11.2.)

Now you already have three pizzas with two toppings. So these three pizzas with one topping get an extra topping added on. So these become three pizzas with two toppings. And
then three pizzas with two toppings plus three pizzas with two toppings equal six pizzas.
Stephanie also noted that each pizza moved to two places in the row below; in
one move the pizza remains the same and in the other move the pizza gets the new
topping. For example, a pizza with peppers could be moved to the left and remain a
pizza with peppers, or it could move to the right and become a pizza with peppers
and mushrooms. Amy-Lynn then returned to the unanswered question about the role
of the 2 in 2n , Roberts expression for the number of possible pizzas when n toppings
are available: Maybe thats where he got the two to the n; maybe that is where the
two comes from. Shelly and Stephanie agreed:
SHELLY:
STEPHANIE:
RESEARCHER:
STEPHANIE:
RESEARCHER:
STEPHANIE:
RESEARCHER:
ROBERT:
That makes sense. Yeah thats where each of the twos come
from.
That was good, that was really good
This one [the first 1 in row 3] only goes here [to the 3 in row 4]?
Does it go here [to the 1 in row 4] too?
Yeah, it drops down as a plain pizza.
Because this one is the plain, I see. Okay, so your drop-down
idea is that it stays the same.
It stays the same once. And it changes once. Where I guess Amy
got the two.
Very interesting. Do you agree with this?
Uh-huh.
Once the formula question was settled, the researcher asked the students to imagine what the numbers on Pascals Triangle might mean with respect to towers.
126
L.D. Tarlow
Stephanie and Robert said that the numbers in row 3 (1 3 3 1) were for threetall towers, and Robert added that the height was the same as the number of
available pizza toppings. Researcher 3 asked, So, the ultimate question now is,
what if you have n toppings? Stephanie and Robert responded: 2 to the n. The
other students agreed; they wrote up their work as shown in Fig. 11.4. Stephanie
then described how part of the row in Pascals Triangle that would represent
n available toppings would look, one, n and then whatever, and then, n, one
(1 n . . . n 1).
As a group, Amy-Lynn, Shelly, and Stephanie explained what each of the numbers in the 1 3 3 1 row of Pascals Triangle would represent in terms of towers, using
the colors blue and red. Researcher 1 asked: So why does one plus three give you
four? You have towers three tall. Now you have towers four tall. Shelly responded,
Cause youre just adding the extra block on. When asked if they could visualize
the three towers with one red block, Stephanie and Amy-Lynn responded together,
One with a red at the top, one with a red in the middle, and one with the red on the
bottom.
The group was asked to make a picture of Unifix cubes to illustrate the addition
1 + 3 = 4 (the same process illustrated in Fig. 11.2 that they had explained earlier in
terms of pizzas). Stephanie drew diagrams of towers, using b for blue Unifix cubes
and R for red, as shown in Fig. 11.5. Stephanie and Shelly then explained: Each of
the three-tall towers gets a cube added on top to become a four-tall tower. The tower
with no red cubes gets a red cube (R), and the three towers that already have one
red cube each get a blue cube (b). These towers represent the case of four four-tall
towers with exactly one red cube in each tower.
Shelly said that it was easier to explain the two thing with the towers because
there are only two colors. Stephanie agreed and added that with all the pizza toppings, it throws you off; you expect eight hundred pizzas. Researcher 1 asked if
there were another way to think about the pizzas. Robert supplied the explanation of
the isomorphism: the number of available pizza toppings corresponds to the height
Fig. 11.4 Students

justification for their
generalization of 2n
11
Fig. 11.5 Stephanies

illustration using towers for
1+3=4
127
b
b
b
R
b
b
R
b
b
b
b
R
b
b
b
R
b
R
b
b
b
b
R
b
b
b
b
R
of the tower, and the two colors would indicate whether or not a topping was on the
pizza. Researcher 1 pointed to the towers that Stephanie had drawn and asked what
the b in a tower would mean if you were thinking of pizzas. Stephanie answered
that it would mean that you either had or did not have the topping. Researcher 1
then asked the group what a four-tall tower would represent in terms of pizzas.
Robert, Stephanie, and Shelly explained together that each of the four cubes indicated whether or not you would choose each of the four toppings. Stephanie added
that it would be, for example, mushrooms going all the way across, and each tower
that had an R in that position would indicate that pizza did not have mushrooms on
it. The other toppings would be represented in the same way.
11.3 Table B: Connection Between Pizzas and Pascals Identity

Angela, Michelle, Magda, and Sherly started work on the four-topping pizza problem by making individual tree diagrams. This work was interrupted with a question
about order: Angela asked whether a sausage and pepperoni pizza is the same as a
pepperoni and sausage pizza. Although the consensus was that this did represent the
same pizza, they decided to leave duplicates in the tree diagram and remove them
later.
After they prepared their tree diagrams, Magda noted that she originally had 24
combinations on her tree diagram for the four-topping pizza, but after crossing out
duplicates, she was left with only one pizza that had exactly four toppings; this led
the group to realize that they would have to cross out a lot of pizzas. Angela found
15 different pizzas that can be made when there are four topping choices. In order to
confirm this answer, the group agreed that each member would work on a different
case: Sherly would do the one-topping case, Angela would do the two-topping case,
Michelle would do three toppings, and Magda would do four toppings. Magda had
only one pizza; for the other three cases, Angela and Sherly made lists and Michelle
used a tree diagram. They concluded that there were 15 total combinations; four
with one topping, six with two toppings, four with three toppings, and one with four
toppings, confirming Angelas earlier answer. When Researcher 2 stopped by to ask
about their solution, Michelle pointed out that they had not counted the plain pizza;
therefore, they now had 16 pizzas.
128
L.D. Tarlow
Next the group continued their work for the five-topping pizza question, with
anchovies as the new topping. Again they distributed working on the different cases
among the group, with Magda adding the five-topping case to her original fourtopping case, and the others keeping the same cases. Using this procedure, they
found all 32 five-topping pizzas. Angela realized that the number of possible pizzas
doubled with the addition of one topping, and so she conjectured that for the threetopping case, there would be eight possible pizzas. She confirmed this by finding
those eight pizzas.
Researcher 2 asked the students to explain their work. Sherly and Magda
explained that they had found combinations by substituting the new topping,
anchovies, for each of the other toppings in their previous combinations. In this
way, they would not have to cross out answers.
Seeing the doubling pattern with the pizzas, Magda recalled that there was also
a doubling pattern in the towers problem, which they had not been able to explain
before. Angela recalled the formula for towers from the previous session (we came
up with that whole like x to the n thing). She did not observe that the problems
were structurally similar, though, noting that order seemed to make a difference with
towers (a tower with a red cube on top of a yellow cube is a different tower from
one with a yellow cube on top of a red cube), whereas a pepperoni and mushroom
pizza is the same pizza as a mushroom and pepperoni pizza.
Researcher 2 asked the group to continue to explore finding pizzas with other
topping choices. They gave the numbers for two choices (four) and one choice (two),
for which Angela listed the possible pizzas. They predicted that with six choices, the
number of possible pizzas would be 64.
Then Magda had what was described as a breakthrough: She noticed that the
numbers they were finding were also seen in Pascals Triangle. A discussion with
Researcher 2, Sherly, and Angela ensued. Magda at first thought she was mistaken,
but the two other students convinced her that she had seen the pattern of Pascals
Triangle (see Fig. 11.6).
Fig. 11.6 Magda and Sherly

discuss Pascals Triangle
11
MAGDA:
RESEARCHER:
MAGDA:
RESEARCHER:
MAGDA:
SHERLY:
ANGELA:
MAGDA:
RESEARCHER:
ANGELA:
MAGDA:
ANGELA:
SHERLY:
MAGDA:
ANGELA:
129
I am thinking. One, three, three, one. I dont know, maybe it has

something to do with this.
Why dont you put it on a new piece of paper cause you are
about to give out I dont know if that works though. Does it?
Lets see.
No, it doesnt work because we dont have the four in here, or
something. You know that, like Pascals.
Pascals, yeah.
Yeah But it doesnt work, so scratch out my idea.
No, wait just a minute. Explain what you see and what doesnt
work.
Wait, wait. How does it work?
But it doesnt work for this one [row 5].
Six, four and four. Yes it does. What are you talking about?
Yeah.
Oh wow, it does. Maybe my idea works.
Yeah, cause then if you have zero toppings, there is only one
[pizza]. Magdas smart. Who would have figured?
The diagram of Pascals Triangle that Magda drew is shown in Fig. 11.7.
The researcher then asked them to explain how the addition rule for Pascals
Triangle (Pascals Identity) could be explained in terms of pizzas. After a lengthy
discussion, Angela said that the 6 in row 4 (representing six pizzas with exactly two
toppings) would be generated from the two 3s in row 3 (representing three pizzas
with one topping and three pizzas with two toppings): add the new toppings (peppers) to the three pizzas that had one topping. She said, Just go, peppers/sausage,
peppers/pepperoni, peppers/mushroom, right? There you go. The three other pizzas that already had two toppings each (sausage/pepperoni, sausage/mushroom, and
pepperoni/mushroom) would be unchanged and become part of the two-topping
group at the next level.
Fig. 11.7 Magda illustrates

how Pascals Triangle is
related to the pizza problem
130
L.D. Tarlow
At Magdas suggestion, the group tested this rule by explaining other additions on
Pascals Triangle using pizzas. When Magda was asked to explain the relationship
between pizzas and Pascals Triangle, she responded:
Okay, so we have the three 1-toppings ones, which was the sausage, pepperoni, and mushroom. So those are the three combinations for the 1-topping ones. Because we are moving
to four toppings, we needed to add an extra topping. So we just added peppers because this
stands for two topping pizzas, so we just added the peppers.
The group noted that the other three represented pizzas that already had two
toppings, so those three pizzas did not change.
In summary, the group found a pattern and a rule for the general pizza problem.
Their organization helped them to see a relationship between the pizza problem and
the numbers in Pascals Triangle. They were able to explain the meaning of Pascals
Identity in terms of generating successive groups of pizzas, with more choices for
toppings. Finally, they noticed that the pizza and towers problems had the same
answers.
11.4 Discussion
All of the students in this session solved the pizza problem and justified their
solution using proof by cases. They connected their topping combinations to the
numbers on Pascals Triangle and used answers to the pizza problem in order to
explain the addition rule for Pascals Triangle. They also noted the doubling pattern
as the number of available toppings increased. In addition, the students at Table A,
who had participated in the longitudinal study and explored the tower and pizza
problems in the earlier grades, explained their reasoning for the doubling rule using
both pizzas and towers. Furthermore, they explained the addition rule for Pascals
Triangle using towers as well as pizzas. Finally, the Table A group constructed a
three-way isomorphism between the tower problem, the pizza problem, and the
numbers on Pascals Triangle. We also note that although the students at Table
A made an attempt to use a partially remembered formula, they did not accept the
formulaic answer but went on to use their own methods to verify the answer; in
the process, they abandoned that formula and instead generated their own correct
formula.
These 11th-grade students exhibited advanced reasoning skills as they thought
about the problems, justified their solutions, and made connections. The students
who had participated in the longitudinal study in the early grades demonstrated that
they had benefited from exposure to thoughtful mathematical experiences over a
long period of time. They retrieved representations that had been built years before
and then modified those representations in order to build and extend their mathematical knowledge and make connections to other mathematical ideas. The students
who came to the study later on also demonstrated the ability to think deeply about
the problems and justify their solutions.
11
131
The students were presented with challenging problems and given the responsibility to solve them. They did not work in isolation; rather, they were active
participants in a learning environment where ideas were shared and discussed. In
the course of making sense of their observations and of what their peers were saying and doing, they moved back and forth between their representations, which had
become less concrete and more abstract and symbolic. They developed, modified,
refined, and extended representations, which enabled them to solve increasingly
more complex combinatorics problems. They built an understanding of fundamental
mathematical ideas and used those ideas to justify solutions to problems. Using their
personal notations, they extended their reasoning and made connections to other
ideas in combinatorics. They were offered opportunities to work on rich, demanding
problems, to think carefully about their ideas, and to discuss their ideas with their
peers and with the researchers. Given the opportunity to think and reason together,
the students constructed deep and powerful mathematical ideas.
In this chapter and the preceding chapters, we observed students make sense of
Pascals Triangle based on personal experience with combinatorics problems and
making use of personal notation. In the next chapter, we see how one group of
students made sense of standard mathematical notation by building on their personal
notation and knowledge.
Chapter 12
Representations and Standard Notation

Date and Grade:

Tasks:
Participants:
Researchers:
May 18, 1999; Grade 11

Towers, Pizzas, and Pascals Triangle
Carolyn Maher and Regina Kiczek
12.1 Introduction
In the preceding chapters in this section, we considered how students made sense of
Pascals Triangle and isomorphic combinatorics problems using their own increasingly sophisticated and abstract representations. In this chapter, we see how one
group built on those ideas in order to derive, explain, and record Pascals Identity
(the addition rule for Pascals Triangle) using standard mathematical notation. This
remarkable demonstration of how students can come to make sense of complex
mathematical ideas was captured during the session that came to be referred to as
the Night Session, since it took place on a weekday evening from 7:30 to 10:00 PM
(Uptegrove, 2004).
In Chapter 10, we described how, during their sophomore and junior years of
high school, Ankur, Brian, Jeff, Mike, and Romina made use of the pizza and towers problems to develop complex mathematical notions: recognize isomorphisms,
generalize findings, and represent ideas using their own personal representations,
which had become increasingly sophisticated and symbolic over the years. In the
session described in this chapter, they made use of standard combinatorial notation to communicate, clearly and concisely, the ideas about Pascals Triangle and
Pascals Identity that they had previously developed. They derived Pascals Identity,
wrote it in standard notation, and explained the meaning of the standard notation in
terms of general versions of the pizza and towers problems.
E.B. Uptegrove (B)



133
134
E.B. Uptegrove
In the following sections we discuss the strategies used by the students to make
sense of the standard notation. We note how the use of increasingly sophisticated
notation accompanied the students building of general notions about the meaning
of Pascals Identity. We show how their organizational strategies proved key in their
making sense of the standard notation and of the relationships between the combinatorial problems, Pascals Triangle, and Pascals Identity. Further, we show that they
found in the standard notation an essential tool for expressing their understanding
of Pascals Identity in general form.
12.2 Summary of Earlier Student Work

As discussed in Chapter 10, during sophomore- and junior-year problem-solving
sessions, Ankur, Brian, Jeff, Mike, and Romina revisited and extended previous
work on two familiar combinatorial problems. In the first session of their sophomore year of high school, they were reintroduced to the towers and pizza problems.
In subsequent sessions, they found a way to organize their solution lists to prove
that all solutions were present; they recognized that those problems were related to
each other, to the binomial coefficients, and to Pascals Triangle; they found general
solutions to those problems; and they used those problems to form preliminary ideas
about the meaning of Pascals Identity.
The key organizational decision to organize pizzas by number of toppings and
towers by number of cubes of a given color not only helped the students show that
The coefficients of the binomial expansion:

1
10
10
The numbers also represent 5-tall towers with ...

1
10
10
0 red cubes
1 red cubes
5 red cubes
2 red cubes
4 red cubes
3 red cubes
The also represent pizzas with ...

1
10
10
0 toppings
5 toppings
1 topping
2 toppings
Fig. 12.1 Row 5 of Pascals Triangle
3 toppings
4 toppings
12
135
they had all the solutions, but also helped them see the relationship between the two
problems and Pascals Triangle. It enabled the students to realize, for example, that
the fifth row of Pascals Triangle (1 5 10 10 5 1) contains not only the coefficients of
the binomial expansion but also the solution to both the five-topping pizza problem
and the five-tall towers problem. Refer to Fig. 12.1.
When the students were introduced to the use of standard notation to describe the
binomial coefficients, they were able to make use of the connection they had already
formed between the binomial coefficients and the towers and pizza problems. They
applied the standard notation to the pizza and towers problems; knowing how to
generate the answers to the pizza and towers problems in their own notation enabled
them to use the standard notation to describe the general pizza and towers answers
and finally to express the general rule for building Pascals Triangle.
12.3 The Night Session

In the night session Ankur, Brian, Jeff, Mike, and Romina returned to the investigations of Pascals Triangle that they had begun in their sophomore year. By the end
of this session, they had written Pascals Identity in standard notation and provided
a sound explanation of its meaning. They did this by looking at general forms of the
towers and pizza problems, referring back to their previous explanations of specific
instances of Pascals Identity in terms of towers and pizzas, and making use of the
binary notation that had been introduced by Mike some 18 months earlier.
At the beginning of the session, the first three students to arrive (Jeff, Mike, and
Romina) talked about that days class work, which had been to find the coefficients
of the expansion of (a+b) n . Jeff brought up what they called choose notation, the
notation to denote combinations, using the nCr function on their calculators.
In this episode, when Jeff was trying to explain how to find a particular coefficient
of the expansion of (a+b)10 , Romina spontaneously introduced the towers problem
with the words ten high and two reds. Jeff and Mike elaborated that this meant
building towers ten cubes tall, selecting from two colors, and counting how many
there are containing exactly two cubes of one of the colors (red).
JEFF:
ROMINA:
JEFF:
. . . If we were looking for a plus b to the tenth say, . . . it was

1 a to the tenth and then 10 a to the ninth b to the first, right?
. . . [The next coefficient] was 45, but we were working on how
to figure it out. We knew it was the choose thing, whatever that
means. . . . What was it? Ten choose two? . . . Like, uh, was it
N-C-R? [Jeff is referring to buttons on his calculator.] Two, is
that how you do it? [Jeff writes 10 nCr 2.] Right? . . . And that
equals 45, and thats the answer. . . . Were not really sure how
all this works but its like . . .. If you have ten different. What is
it? Ten different things . . .
Ten high. Ten high.
Ten high. How many.
136
ROMINA:
JEFF:
RESEARCHER:
JEFF:
MIKE:
JEFF:
E.B. Uptegrove
How many would have two reds, only two reds.

How many would have two, two reds.
One more time.
If you have towers with ten high and two colors.
How many different places can you put two reds in there?
And like a would be one color and b would be the other color.
Their original explanation of ten choose two was so brief; it would have been
difficult for anyone not familiar with their work to understand the references ten
high and how many would have two reds. But the elaboration (although it still
assumed knowledge of the towers problem) shows that they knew that the coefficients of the binomial expansion to the tenth power were related to the ten-tall
towers problem.
A few minutes later when the researcher asked the group to discuss the choose
notation that Jeff had mentioned, Mike drew a few rows of Pascals Triangle on the
board and explained that any row could be expressed in choose notation; for example, the row 1 3 3 1 could be called 3 choose 0 through 3 choose 3. When the
researcher asked the students to talk about the addition rule for Pascals Triangle in
that notation, Romina suggested a new vehicle, the pizza problem, even though she
had just used the towers problem in the previous explanation. Mike used the pizza
problem to explain a particular case of addition: think of the nth row of Pascals
Triangle as representing all the possible pizzas that can be made when there are n
toppings to choose from, and think of generating new rows of Pascals Triangle as
making new pizza toppings available. Then the pizzas represented by the first three
in row 3 are the one-topping pizzas (when there are three toppings to choose from).
You can either add the new topping to those three pizzas (making them two-topping
pizzas) or let them remain one-topping pizzas. If you add the new topping, you now
group them with the second three in that row (the pizzas that already have two toppings), resulting in six pizzas with two toppings. If you do not add the new topping,
those three pizzas are added to the one pizza that had no toppings (and that had the
new topping added to it), giving four pizzas with one topping. Figure 12.2 illustrates
Mikes explanation. A portion of their discussion is given below.
Fig. 12.2 Examples of Pascals Identity
12
MIKE:
JEFF:
MIKE:
JEFF:
MIKE:
JEFF:
MIKE:
JEFF:
MIKE:
137
Lets go to this one. This would be like three different places, I guess.
[Mike indicates row 3, which is 1 3 3 1.] . . .
That would be a plus b to the third.
All right, lets say you have like, heres a number, all right? [Mike writes
000.] Zero means no toppings. One would be a topping. So first category
is everything with no toppings. [Mike points to the first 1 in row 3.] And
thats your number for that one. [Mike points to 000.] Thats like, like
binary numbers or something. Next would be- [Mike writes 001, 010,
then 101.] Theres all the, the ones that have one topping.
Right, you got to write that 0 at the end. You messed up. Last one should
be a hundred, not a hundred and one.
I knew that. [Mike changes 101 to 100.] Theres all the ones that have one
topping. . . .. Theres your 3 choose 1 and theres three different combinations you could put that. . . .. But, um, when you have a new when you
add another place, another topping. [Mike draws dashes to the right of the
four numbers already there. Refer to Fig. 12.3.]
That could be one or the other, one or the other, one or the other.
So, it could be one or the other. It could be a zero or one, a zero or one,
zero or one. [Mike writes 0 and 1 above each dash.] So all these threes
would either move up a step onto the next category and have two toppings.
[Mike points to the 6 in row 4.] Or they might stay behind and still only
have one if they have the zero. [Mike points to the 4 in row 4.] So three
get a topping, go to this one [Mike points to 6.] and three wont, will stay.
[Mike points to 4.] These three [Mike points to the first 3 in row 3.] with
one topping wont get one so, you know, you can put them in the same
category as this one.
Thats their four? Yeah.
Thats four. . . .. And you know, the three that had two toppings wont get
any. [Mike draws a line from the second 3 in row 3 to the 6 in row 4. Refer
to Fig. 12.2.] And you could put them in together with the ones that did
get something. Thats why you would add.
After Mike explained the specific instances of 3+3=6 and 1+3=4 in terms of pizzas, the researcher (R1) rewrote row 3 of Pascals Triangle in standard combinatorial
notation and asked the students to write other rows in that form and show an example
of the addition rule. Figure 12.4 shows what they did. Their discussion follows.
000 01
00101
01001
100 01
Fig. 12.3 Binary listing of 3 choose 0 and 3 choose 1
138
E.B. Uptegrove
Fig. 12.4 Showing 3 + 3 = 6 in combinatorial notation
RESEARCHER:
MIKE:
RESEARCHER:
JEFF:
ANKUR:
MIKE:
JEFF:
ANKUR:
Show me that 3 plus 3 is 6. Which ones would it be? . . .

This one and that one. [Mike points from 3 choose 1 and 3
choose 2 to 4 choose 2.] . . .
Okay, so youre saying 3 choose 1 plus 3 choose 2 equals 4
choose 2. Right? Okay. So whats 4 choose 2 plus 4 choose 3?
. . . 4 choose 2 plus 4 choose 3? That would be, that would be 55 choose. . . .
5 choose 3.
Why is he 5 choose 3?
Because its always the one on the right. [Ankur means that the
choose number of the sum is the same as the choose number
of the rightmost addend.]
Mike observed that the bottom number indicated the number of toppings actually
used, so that when a topping was added, the bottom number changed and when a
topping was not added, the bottom number did not change.
The researcher asked the group to continue by writing a general (nth) row of
Pascals Triangle and to use that row to discuss the meaning of the addition rule.
Figure 12.5 shows the two general rows that Jeff drew. In spite of the researchers
suggestion to use lowercase n to indicate row number and r to indicate a number
in the middle of the row (following standard usage), Jeff used uppercase letters
N and X.
Brian arrived after the group had been working for almost an hour; first Jeff
explained to Brian how Fig. 12.5 related to that days work in their regular math
class relating to Pascals Triangle:
Were explaining the general addition, the addition rule using chooses to fill out the triangle,
and this here would be N choose X plus 1 and then N choose X plus 2 and so on to whatever
N equals.
Then the group was asked to write the addition rule in general form. Figure 12.7
shows Jeff working at the board as they discussed the problem. This discussion
follows.
Fig. 12.5 Rows N1 and N

of Pascals Triangle
N 1 . . . N 1 . . . N 1
0
N 1
X
N N N N N N N
0 X 2 X 1 X X+ 1 X +2 N

12

N N N + 1
+
=
X X+ 1 X + 1
139
n n 1 n 1
=
+
r r 1 r
Fig. 12.6 Pascals Identity in students notation and as shown in textbooks
RESEARCHER:
ANKUR:
MIKE:
JEFF:
MIKE:
JEFF:
MIKE:
JEFF:
MIKE:
ANKUR:
MIKE:
JEFF:
RESEARCHER:
JEFF:
BRIAN:
JEFF:
BRIAN:
JEFF:
ROMINA:
BRIAN:
JEFF:
BRIAN:
JEFF:
BRIAN:
JEFF:
Can you write it as an equation? Just like you wrote three plus
three equals six.
N plus, just that plus that. [Ankur points to the entries N choose
X and N choose X + 1 in Fig. 12.5.]
N choose X.
N choose X plus N choose X plus one. [Jeff writes on the board
as he speaks. Refer to Fig. 12.6.]
Equals that. . . .
Plus one equals that right there. [Jeff points to N+1 choose X+1.]
. . . Then, well, thats, thats because this would be gaining an X
and going into the X plus 1. [Jeff points to N choose X.]
Yeah.
And this would be losing an X. [Jeff points to N choose X+1.]
No, no, not losing, not getting anything.
Staying the same.
And the top numbers have changed because you have more.
Because youre adding; you have more things [to choose from].
Say it so Brian can follow it because he wasnt here for the
earlier pizza discussion.
What, what were doing is the next line of the triangle.
Remember how today in class the other triangle was one, two.
Yeah.
Three, that whole row there. Well, thats the increase in N and
then the X plus one. . . . Say were doing pizzas.
All right.
If you add another topping onto it.
You know how we get the triangle and how we go 1 2 1 and add
those two together?
Yeah.
We were explaining why you add.
All right, keep going.
Because [Jeff points to N choose X+1.] . . . If it gets a topping,
thats why it goes up to the X plus 1. [When a new topping
is available, the second (choose) number in the expression is
increased by 1.] And in this one, its staying the same, right?
[Jeff points to N+1 choose X.] And thats why its going there.
Make sense?
Yes. It actually does.
So, so that would be the general addition rule in this case.
140
E.B. Uptegrove
Fig. 12.7 Two versions of

Pascals Identity
The students version of Pascals Identity is equivalent to a standard textbook

version of this equation, with n equivalent to N+1 and r equivalent to X+1. (See
Fig. 12.6) Following the production of the equation in combinatorics notation, the
students were asked to convert that notation to factorial notation. They did so; their
work is shown in Fig. 12.7.
Earlier in this session, Mike had explained 3 + 3 = 6 (Fig. 12.3) by stating that
the threes are from the three-toppings row of Pascals Triangle: the first three represents the three one-topping pizzas that become two-topping pizzas, and the second
three represents the three two-topping pizzas that remain two-topping pizzas; the
six represents the six two-topping pizzas that can be made when there is a fourth
topping available. Now the students generalized this rule using the standard notation, which they called choose notation: N choose X gives the number of pizzas
that have exactly X toppings when there are N toppings to select from and N choose
X+1 gives the number of pizzas that have exactly X+1 toppings. Moving to the next
row down in Pascals Triangle means that you increase the number of available toppings by one, and so N increases by one. Adding the new topping to the first addend
(N choose X) and not adding the new topping to the second addend (N choose X+1)
gives a group of pizzas with X+1 toppings when there are N+1 toppings to select
from.
The students had described instances of Pascals Identity in earlier sessions. For
example, in their sophomore year, Ankur, Jeff, and Romina had provided a specific
explanation using towers: In order to create a six-tall tower with exactly three red
cubes, add a red cube to the five-tall towers that have exactly two red cubes and
add a blue cube to the five-tall towers that have already have three red cubes. (This
was discussed in more detail in Chapter 10.) But this night session explanation was
the first explanation of Pascals Identity using standard notation to state a general
result.
12.4 Durability of Understanding

Three years after the night session, in individual interviews, Mike, Romina, and
Ankur were asked to recall this work. In that time, all three took math classes in
college, but none studied combinatorics. All three were able when prompted to write
12
141
the formula, and all three were able to provide a cogent explanation of the addition
rule. As an example, we discuss here the interview with Mike.
Over the years in which he worked on the combinatorial problems, Mike progressed from drawings and codes through his personal (binary) representation
system to the standard combinatorial notation. From the time he introduced his ideas
about binary notation to his fellow students to his most recent interview over 5 years
later, he demonstrated the ability to make sense of the problems and of the notation,
both through the use of his chosen notation and through the use of the combinatorics
tasks. Mike took the lead in devising representations, finding connections, and making sense of the tasks. He recognized structural similarities between problems; he
moved between different representations with ease; and he extended, generalized,
and reorganized his knowledge when he discussed it with others.
In this interview, which took place when Mike was in his second year of college,
the researcher (R1) showed Mike a diagram of Pascals Triangle and asked him to
recall how the group used the pizza problem to think about Pascals Triangle. Mike
spent a little time regenerating the meaning of specific entries in Pascals Triangle
and then, in response to a question about a general rule, he reconstructed the formula that had been developed during the night session. A portion of their discussion
follows. Mike began with the two-topping pizza problem.
MIKE:
RESEARCHER:
MIKE:
Okay. If you had no toppings, that would be one pizza.

Okay. So where is that on the triangle?
Well, Im going to just draw it. . . . And then well find it. . . .
If youre using just one topping, you can make two possible
pizzas with that. And then if you have all the toppings, thats
one. Right. And then automatically I see that relates to this
row. [Mike points to row 2 of Pascals Triangle (1 2 1).] And
Im pretty sure it would go down, this is like a third topping
and a fourth topping. [Mike points to rows 3 and 4 of Pascals
Triangle.] Now I think the way I thought about it is, like, the row
on the outside [leftmost entry in a given row] would be your
plain pizza. And theres only one way to make a plain pizza.
And . . . the next one over would be how many pizzas you could
make using only one topping, and then so on until you get to the
last row [the rightmost entry in a given row] which is all your
toppings. And, once again, you can only make one pizza out of
that. . . .
The researcher then suggested that Mike work on a general rule:

RESEARCHER:
MIKE:
RESEARCHER:
MIKE:
And at that session, I asked them to write an equation to show,

for instance, how that might happen from one row to the next.
So can you just do that, write. . . .
Like a general equation?
Well, that was what I was going for ultimately. . . .
To give an amount for any spot in the row.
142
RESEARCHER:
MIKE:
E.B. Uptegrove
Right. . . .
All right, so I guess well give, you know, the row a name.
Call that r. And I guess the spot in the row, like, you know,
zero topping, one topping. Call that, n sounds fine. [There is a
pause; then Mike writes the left part of the equation shown in
Fig. 12.7.] Im just going to like work this out in my head and
see if it actually works. [A few seconds later, Mike adds the
right part of the equation.]
This equation, shown below, is equivalent to the textbook version and to the
night session equation, although he used different variables. (Textbooks usually use
n choose r instead of r choose n, and the sum is given on the right side.) We
can see that Mike did not rely on symbol manipulation. He linked the numbers to
a problem task that made sense to him, and then he expressed the relationships in
that task in symbolic form. We conclude that Mike was reconstructing the substance
rather than merely remembering the form.

r
r
r1
+
=
n
n+1
n+1
12.5 Discussion
Exploring previously unexamined complexities of the towers and pizzas problems
was a mathematically challenging task of the sort recommended by Davis and
Maher (1990) to foster students ability to engage in real mathematics developing
their own mathematical theories, for example. Conditions important for the development of new mathematical ideas were in place: these students had ample time for
exploration of mathematical ideas and the opportunity to express their own ideas.
The students existing representations were taxed by new questions about how to
relate these problems to each other, to Pascals Triangle, and to the binomial coefficients and about how to represent a general instance of Pascals Identity. Hence,
there was a need to reorganize existing knowledge and to use new tools for dealing
with these new ideas. We have shown that these students did make use of a new
tool standard mathematical notation for dealing with their ideas about Pascals
Identity.
When they first started working on the pizza and towers problems, Ankur, Brian,
Jeff, Mike, and Romina built towers and drew pictures of pizzas. As early as middle
school, they began instead to use symbolic notation. (For example, they used letter
and number codes to stand for the objects they were investigating.) Besides continuing the use of codes during high school, the students also found increasing use
for the standard notations of mathematical discourse. For example, the binary notation that they began to use in high school was more powerful than the letter codes
because it was easily extended (adding a cube to the tower or a topping choice to
12
143
a pizza corresponded to adding a binary digit) and it was applicable to both pizza
and towers problems, thus making it easier for the students to identify the similar
structures of the two problems. Using binary notation helped the students focus on
the isomorphic structural aspects of the combinatorial problems (the duality of the
choices) rather than the surface features (the different pizza toppings, for example).
Binary notation was also an easily generalizable notation in three ways: (1) adding
an extra digit corresponded to making a new pizza topping available and to increasing the towers height by one block; (2) adding a 1 corresponded to adding that
newly available topping and to adding a block of the designated color to the tower;
and (3) adding a 0 corresponded to not adding the new topping and adding a block
of the other color. This idea that it was not necessary to know the current number (or
names) of pizza toppings or the current height of the tower in order to describe what
happened next was important in the students production of the general equation for
Pascals Identity.
During the course of discussions over 4 months of their sophomore year, these
students first noted that the pizza and towers problems had the same answer in specific cases. Then they linked specific answers to the pizza and towers problems to
specific entries in Pascals Triangle. Finally, they described the links among binomial coefficients, pizza toppings, and towers. (Blue block = a = topping on the
pizza; white block = b = topping off the pizza.) During the night session, they built
on their knowledge of these links in order to produce the general form of Pascals
Identity. We claim that their ability to map corresponding mathematical structures
among these three representations is a strong indication of their mathematical competency and it indicates more competency than, for example, simply being able to
reproduce or use a memorized formula.
The way these students organized their answers to the pizza and towers problems was a key organizational element that helped them to form connections among
those problems and Pascals Triangle. They also made extensive use of their personal representations at the beginning of the process. But once those connections
were formed, the students began to make general statements about Pascals Triangle
and Pascals Identity, and they had less use for personal representations. Finally,
although they were able to articulate general information about Pascals Triangle
and Pascals Identity, they did not represent the generalizations symbolically until
the night session.
After they had made the association between Pascals Triangle and the combinatorial problems, the students demonstrated an ability to describe any selected
entries in Pascals Triangle in terms of the combinatorial problems. For example,
they described the numbers in row 6 as representing six-tall towers with zero through
six red cubes, respectively. The fact that they could explain any instance suggested
that they had an idea of the general rule; but without the standard notation, they
could express their general ideas most easily by referring to specific examples. By
the time of the night session, these students seemed to know general rules about
generating Pascals Triangle, but they lacked the notation to express these rules in a
concise way. They were at the point where they needed standard notation in order
to proceed further.
144
E.B. Uptegrove
We suggest that these findings point to one way that teachers can follow the
recommendation by the NCTM (2000) to use sound professional judgment when
deciding when and how to help students move toward conventional representation
(p. 284). Teachers should aim to help students to develop a powerful organization,
one that lends itself to a mapping onto formal notation. In that way, the formal
notation can be seen as the solution to a problem that arises during the students
own investigations: the problem of how to express in a general way the findings that
the students have developed on their own.
In this chapter, we have seen how this group of students learned about the relationships among well-known combinatorics problems and Pascals Triangle. In the
following chapter, we observe the same students working on new problems in combinatorics and using what they learned about the pizza and towers problems and
their relation to Pascals Triangle in order to make sense of that unfamiliar problem
the Taxicab Problem.
Chapter 13
So Lets Prove It!

Arthur B. Powell
Date and Grade:

Tasks:
Participants:
Researchers:
May 5, 2000, Grade 12

The Taxicab Problem
Brian, Jeff, Mike, and Romina
Carolyn A. Maher and Arthur B. Powell
13.1 Introduction
In previous chapters, we observed students throughout middle school and high
school working on and making sense of two isomorphic problems in combinatorics the towers problems and the pizza problems. In this chapter, we see how
students just finishing high-school work on another isomorphic problem, demonstrating the application of techniques and ways of thinking that they developed
throughout their previous years in the study. We further address the challenge that
Davis (1992a) proposes to mathematics education researchers to investigate the
emergence among learners of what lies at the core of mathematics: mathematical
ideas. Here, a cohort of four high-school seniors Brian, Jeff, Mike, and Romina
elaborates mathematical ideas and reasoning through work on the Taxicab Problem.
They display criteria and techniques for justifying claims and an awareness of the
power of generalizing, particularly as an aid to respond to special cases.
13.1.1 The Task

The problem-solving session was held in a classroom during the late afternoon,
after school hours. During the session, which lasted about 1 h and 40 min, the four
A.B. Powell (B)

Department of Urban Education, Rutgers University, Newark, NJ, USA
e-mail: powellab@andromeda.rutgers.edu


145
146
A.B. Powell
students collaborated on a culminating, performance-assessment task of the research

strand on combinatorics the Taxicab Problem:
A taxi driver is given a specific territory of a town, shown below. All trips originate at the
taxi stand. One very slow night, the driver is dispatched only three times; each time, she
picks up passengers at one of the intersections indicated on the map. To pass the time, she
considers all the possible routes she could have taken to each pick-up point and wonders if
she could have chosen a shorter route.
What is the shortest route from a taxi stand to each of three different destination points?
How do you know it is the shortest? Is there more than one shortest route to each point? If
not, why not? If so, how many? Justify your answer.
13.2 Justifying Claims

It is a non-trivial cognitive task for students to recognize which statements or claims
in their mathematical discourse require justification or proof. This is particularly
true if the students deem the claim to be obvious or if the students are in the
midst of group problem solving with intellectual peers. On May 5, 2000, in the late
afternoon, after school, and just a few weeks shy of their high-school graduation,
Brian, Jeff, Romina, and Mike are seated around three sides of a trapezoidal-shaped
table, on top of which are four black felt-tip markers, sheets of blank paper, and
a problem statement. The statement is of a problem in which one is to determine
in a given rectangular grid the number of different shortest paths between pairs of
specified colored, endpoints (black and blue, black and red, and black and green).
A researcher asks the four seniors to read the Taxicab Problem and to see whether
they understand it. Jeff asks aloud whether one has to stay on the grid lines and
whether they represent streets. The researcher responds, Exactly. Each student
has taken a marker. Among themselves, they observe that from the black endpoint
13
So Lets Prove It!
147
or taxi stand, five and seven are respectively the number of blocks it takes to reach
the blue and red endpoints or pick-up points. Moreover, some assert that different routes to each point have the same length as long as one doesnt go beyond
the particular pick-up point. Especially noteworthy from cognitive and pedagogical viewpoints, Brian says to his colleagues, So, lets prove it! After a few quiet
moments, a discussion ensues as to how they know that their claim is true.
13.2.1 Generalizations, Isomorphisms, and Transitivity

After further individual and collective work and discussions, Brian, Jeff, Mike, and
Romina decide that to determine the number of paths between three specialized pairs
of endpoints they need to generalize the problem. This moment is a watershed event
in their mathematical work on the Taxicab Problem. Through their various heuristic actions, the students generate data that they consider reliable. They reflect on
numerical patterns in their array of data, observe that it resembles Pascals Triangle,
and conjecture that Pascals arithmetic array underlies the mathematical structure
of the problem. How do they justify this conjecture? They embark on building an
isomorphism between the Towers Problem and the Taxicab Problem since from
previous experience they know that Pascals Triangle underlies the mathematical
structure of the Towers Problem. The students strategy can be interpreted as justifying their conjecture by transitivity: (a) the mathematical structure of Pascals
Triangle is equivalent to that of the Towers Problem and (b) the mathematical structure of the Towers Problem is equivalent to that of the Taxicab Problem; implying
that (c) the mathematical structure of Pascals Triangle is equivalent to that of the
Taxicab Problem. The students knew that (a) is true and demonstrate (b) to justify
and conclude (c).
13.2.2 Reasoning and Justifying

In the following sections, we discuss students methods of reasoning and ways of
justifying their statements.
13.2.2.1 Realizing the Need to Discursively Build a Justification
Two and a half minutes after receiving the task, Romina begins the first student-tostudent interaction. It centers on a question about a relation that she notices about
which Romina invites her colleagues to comment.
ROMINA:
BRIAN:
ROMINA:
BRIAN:
MIKE:
Isnt it like anyway you goPretty much, because lookAs long as you dont go like past it. [Facing Brians direction.]
The first one- No.
Well what if you go to the last one-
148
BRIAN:
ROMINA:
BRIAN:
JEFF:
ROMINA:
JEFF:
BRIAN:
JEFF:
ROMINA:
JEFF:
ROMINA:
BRIAN:
A.B. Powell
You can go all the way down and go over and go down three and go
over two. [Tracing the routes above the problem sheet with a black
marker in his right hand.]
Isnt it- Dont they all come out to be the same amount of blocks?
Five.
Five?
Five? I got seven.
Uh, which one- Yeah, we were both looking at the red one.
Im looking at blue. [Mike is tapping his pen on the grid along
intersection points.]
Yeah.
Oh, okay.
All right. I mean pretty much.
As long as you dont go like past it youre fine. So its the same thing.
So, lets prove it.
Rominas interrogative, isnt it like any way you go, they [the lengths of routes]
all come out the same . . . as long as you dont go past it [the pick-up point]?
suggests that she is aware of a relation among efficient (as long as you dont go past
it) paths or routes between the taxi stand and the red pick-up point. She observes
that as long as one does not go beyond the red pick-up point that the numbers of
blocks traversed or lengths of routes to red equal each other. Specialized to the red
pick-up point, she expresses three awarenesses about relations among objects: (a)
an efficient route will be a shortest route, (b) there can exist more than one shortest
route, and, her central observation, (c) efficient routes have the same length. These
three ideas are important and fundamental for progressing toward a resolution of the
problem task.
At first, Brian disagrees (The first one- No, cause-) and then, examining routes
to the blue pick-up point, attempts to understand Rominas remark (You can go
all the way down and go over and go down three and go over two). Afterward,
Jeff and Romina try to understand Brians assertion, five, for the number of
blocks traversed by shortest routes between the taxi stand and the red pick-up point.
Ultimately, Brian sees that they are speaking about routes to the red point (Yeah,
we were both looking at the red one.). While, they understand that he is referring
to the blue pick-up point (Im looking at blue.). Taking up Rominas observation
for the red pick-up point along with his own for the blue point, Brian suggests, So,
lets prove it.
Brians proposal is not immediately entertained. However, after about 1.5 min,
Jeff poses a question that places Brians proposal back onto the agenda, and the
students discuss how they know that Rominas unchallenged assertion is true.
JEFF:
MIKE:
ROMINA:
MIKE:
So why- why is it the same every time?

Youre going left and right.
Ours is a four by one, right? Its the only way to go.
Its the only way you can go. Yeah, its a four by one, unless you go
backwards a couple of times.
13
So Lets Prove It!
ROMINA:
MIKE:
BRIAN:
MIKE:
JEFF:
MIKE:
ROMINA:
MIKE:
ROMINA:
MIKE:
JEFF:
MIKE:
JEFF:
ROMINA:
MIKE:
ROMINA:
JEFF:
ROMINA:
JEFF:
MIKE:
ROMINA:
BRIAN:
JEFF:
MIKE:
JEFF:
MIKE:
ROMINA:
MIKE:
JEFF:
ROMINA:
MIKE:
JEFF:
MIKE:
JEFF:
149
You cant go, wellI know that would be dumb.

[inaudible] the shortest route only if you go forward.
But the only- You cant go diagonal so you have to go up and down.
So if the thing is down this many and
Over that many, its the same
Its the sameIts the same area
No matter how you do it, no matter how you do it its- you have toyou cant get around doing that. [Pointing and gesturing around his
grid]
All right.
You cant get around going four down and right one cause -.
All right, yeah. All right.
You cant go over there. You cant get around doing that.
Yeah.
What if I were to go like to the red when I go one, two, three, four[Pointing at her problem sheet.]
But theyre not asking for that.
Five, six, seven.
Five, six, seven. Its the same thing.
Like how- how am I going to- like how would IIts the same thing.
Its the same.
-devise an area for that? Like this- this area up here? [Motioning with
her pen on her grid, indicating the area of the rectangular space whose
vertices are taxi stand and the red pick-up point.]
Like plus and [Inaudible].
Well, its not area.
Its not area. Its just aIts the perimeter. Its like each one being one.
One, two, three, four, five, six, seven. [Pointing at Rominas paper
and counting the length of a route to the red destination point.] [Jeff
scratches his head.]
All right.
Theres no way you can get around going- [gesturing with his hands]
Going seven blocks.
No, yeah, I understand.
Across that many and down that many because you cant go diagonally. Cant- [gesturing with his hands over his problem sheet across
to the left and then down]
Yeah.
Cant get around it, so- [gesturing with his hands]
I mean, thats the most sensible way I think to say that. Right? And
they want to know how many though.
150
A.B. Powell
Justifying Rominas observation, reiterated by Jeff, or, equivalently, entertaining

Brians proposal becomes a shared project of the participants. When Jeff poses his
question (why is it the same every time?) and the others understand his it to
mean the set of efficient routes to a pick-up point. Mikes immediate response,
coming just 4 s after Jeff finishes uttering his question, is in contrast to the silence
that Brians proposal received almost 2 min earlier. The ensuing discursive exchange
hints that the issue of the why Rominas observation was true in general remained a
concern of the participants and that they are only now prepared to tackle it.
Mike explains that to reach a pick-up point, the shortest distances will always
require one to move a fixed number of units down (south) and a fixed number across
(east) and observes that within the grid one cannot travel diagonally. Brian reminds
the others that only going forward will produce a shortest route. Mike generalizes
his awareness to all routes. Jeff signals that he is convinced, saying, I mean, thats
the most sensible way I think to say that. In the process of the groups discourse,
Jeff and Mike help Romina to see that area is not an operative idea in this task.
In the above conversational exchange, the participants engage in socially emergent cognition (Powell, 2006), providing discursive evidence to several ideas: (1)
movement within the given portion of the taxicab plane goes left or right and up
or down; (2) diagonal movements are not permissible; (3) the taxi stand and each
pick-up point together define a rectangle in which the pair of points are located
at opposite ends of a diagonal, and the problem task involves moving along the
perimeter but does not concern the extent of space that a rectangle occupies; (4) the
number of units down plus the number of units across are objects related by addition
to produce the length of a shortest path; (5) any route to the blue pick-up point will
involve four blocks down and one block across; and (6) each horizontal and vertical
line segment of the grid can be considered as one unit in length.
By the end of the exchange, Jeff, who in the form of a question reintroduced
Brians proposal that they justify the idea that the length of efficient routes from the
taxi stand to a pick-up point are equivalent, expresses satisfaction with Brian and
Mikes argumentation (I mean, thats the most sensible way I think to say that.),
checks whether the others agree (Right?), and reminds his colleagues that they
can now turn their attention of the crux of their task: And they want to know how
many though.
The discursive exchanges in the three episodes quoted above are critical. They
present the major occasion in which the participants ferret out the nature of the problem space and build fundamental ideas essential for investigating the problem task.
The participants establish what are the basic objects of taxicab geometry (points
and line segments or routes); basic awareness of the Taxicab Problem (there can
be more than one shortest route to an intersection point in the taxicab plane); and
implicitly note a distinguishing feature between Euclidean and taxicab geometries
(how distance is measured). This distinction emerges when Mike observes that in
the context of the problem task, one cannot travel diagonally, he touches upon the
fundamental distinction between the metric of Euclidean geometry and that of taxicab geometry. Moving forward with the ideas they have built that were illustrated
in the three episodes, the students shift their focus to delve further into the problem
task and generate considerably more data.
13
So Lets Prove It!
151
13.2.2.2 Generalizing to Specialize

The students take a decisive turn in their investigation: they generalize the problem. Instead of determining the number of shortest paths between each of the three
specialized pairs of endpoints, the work to uncover a pattern among the numerical
values that represent the number of shortest paths between the taxi stand and any
point on the grid. They first examine points in close proximity to the taxi stand. This
decision proves to be a watershed event in their mathematical work on the Taxicab
Problem.
13.2.2.3 Building Isomorphisms to Justify

The transcription of the problem-solving session contains 1,869 turns of speech.
The portion of the transcript that relates to the students building an isomorphism to
justify their solution transpires over many turns of speech, spanning from turn 159
to turn 1,320. Space does not permit us to present a full illustration of the development of the ideas and reasoning that comprise the students work toward justifying
their solution. They have continual discursive interactions with the aim of building an isomorphism between a rule for generating the entries of Pascals Triangle
and the number of shortest routes to points on the taxicab grid. Early in their work,
they manifest embryonic thinking about an isomorphism. Romina wonders aloud,
cant we do towers on this? (This groups previous work on the towers problem
is discussed in Chapters 4, 5, 10, and 12.) Her public query catalyzes a negotiatory
interlocution among Mike, Jeff, and her. Jeff, responding immediately to Romina,
says, thats what Im saying, and invites her to think with him about the dyadic
choice that one has at intersections of the taxicab grid. Furthermore, he wonders
whether one can find the number of shortest routes to a pick-up point by adding
up the different choices one encounters in route to the point. Romina proposes that
since the length of a shortest route to the red pick-up point is 10, then ten could
be like the number of blocks we have in the tower. Rominas query concerning
the application of towers to the present problem task prompts Mikes engagement
with the idea, as well. As if advising his colleagues and himself, he reacts in part by
saying, think of the possibilities of doing this and then doing that. While uttering
these words, he points at an intersection; from that intersection gestures first downward (doing this), returns to the point, and then motions rightward (doing that).
Similar to Jeffs words and gestures, Mikes actions also acknowledge cognitively
and corporally the dyadic-choice aspect of the problem task. Through their negotiatory interactions, Mike, Jeff, and Romina raised the prospect of as well as provided
insights for building an isomorphism between the Taxicab and Towers Problems.
The prospect and work of building such an isomorphism reemerges several more
times in the participants interlocution and, each time, they further elaborate their
insights and advance more isomorphic propositions. Eventually, the building of isomorphisms dominates their conversational exchanges. Approximately 35 min after
Romina first broached the possibility of relating attributes of the Towers Problem
to the problem at hand, the participants reengage with the idea. Romina speculates
that between the two problems one can relate like lines over to like the color and
152
A.B. Powell
then the lines down to the number of blocks. What is essential here is Rominas
apparent awareness that each of the two different directions of travel in the Taxicab
Problem needs to be associated with different objects in the Towers Problem.
Romina uses this insight later in the session. She transfers the data that she and
her colleagues have generated from a transparency of a 1-cm grid to plain paper.
Their data are equivalent to binomial coefficients. She identifies one unit of horizontal distance with one Unifix cube of color A and one unit of vertical distance
with one Unifix cube of color B:
Like doesnt the two- theres- that I mean, thats one- that means its one of A color, one
of B color [pointing to the 2 in Pascals Triangle]. Heres oneits either one- either way
you go. Its one of across and one down [pointing to a number on the transparency grid and
motions with her pen to go across and down]. And for three that means theres two A color
and one B color [pointing to a 3 in Pascals Triangle], so here its two across, one down or
the other way [tracing across and down on the transparency grid] you can get three is two
down [pointing to the grid].
Furthering the building of their isomorphism, Mike offers another propositional

foundation. Pointing at their data on the transparency grid and referring to its diagonals as rows, he notes that each row of the data refers to the number of shortest
routes to particular points of a particular length. For instance, pointing to the array
1 4 6 4 1 of their transparency, he observes that each number refers to an intersection point whose shortest route is four. Moreover, he remarks that one could
name a diagonal by, for example, six since everything [each intersection point]
in the row [diagonal] has shortest route of six. In terms of an isomorphism, Mikes
observation points to two different ideas (1) it relates diagonals of information in
their data to rows of numbers in Pascals Triangle and (2) it notes that intersection
points whose shortest routes have the same length can have different numbers of
shortest routes.
Later in responding to a researchers question, the participants develop a proposition that relates how they know that a particular intersection in the taxicab grid
corresponds to a number in Pascals Triangle. They focus their attention on their
inscriptions in Fig. 13.1, which shows empirical data of shortest routes between the
taxi stand and nearby intersection points. In array A, the green numbers (lighter
shade inside sqiares) show empirical data of shortest routes between the taxi stand
and nearby intersection points. Jeff wrote the 1 s on the side in blue (darker shade)
to augment the appearance of the numerical array as Pascals Triangle. From the
participant perspective, to the left of Jeffs numbers, Romina wrote in green (lighter
shade) the numbers 1, 2, and 3 to indicate the row numbers of the triangular array.
Array B shows their drawing of Pascals Triangle. The first five rows contain empirical data; the remaining two rows contain assumed data values based on the addition
rule for Pascals Triangle.
Mike and Romina discuss correspondences between the two inscriptions.
Referring to a point on their grid that is five units east and two units south, Romina
associates the length of its shortest route, which is seven, to a row of her Pascals
Triangle by counting down seven rows and saying, five of one thing and two of
another thing. Mike inquires about her meaning for five and two. Both Romina
13
So Lets Prove It!
153
Fig. 13.1 Participants data arrays A and B
and Brian respond, five across and two down. She then associates the combinatorial numbers in the seventh row of her Pascals Triangle to the idea of five of
one thing and two of another thing, specifying that, left to right from her perspective, the first 21 represents two of one color, while the second 21 is five of one
color, presuming the same color. Using this special case, Romina hints at a general
proposition for an isomorphism between the Taxicab and Towers Problems.
13.3 Conclusion
The narrative of these four students working on the Taxicab Problem has three
sections. The first concerns their recognition of the need to justify an observation
that they made immediately after reading the problem statement. The observation maybe simple but their recognition of the significance of the observation
and that it needed to be justified before progressing on with resolving the problem is rather sophisticated. This sophistication in their mathematical work speaks
to the sociomathematical norms (Yackel & Cobb, 1996) that they have developed through their longitudinal experience working on open-ended problems. This
sociomathematical norm is subtle and akin to the way mathematicians work.
The second section of the narrative pertains to their decision to seek a general
solution to the problem and that such a solution would be easier than trying to count
the number of shortest routes between each of the three pair of given endpoints.
This is an instance of what can be called generalizing to specialize. That is, finding
a general solution of a problem situation in order to answer more specific questions
of the problem. Often the general case is easier to solve than special cases.
Finally, the third section of the narrative revolves around not only with the recognition that claims needs to be justified but also with a particular proof strategy that
154
A.B. Powell
emerged in the students attempt to justify their resolution of a generalized form of

the Taxicab Problem.
Important sociomathematical norms (Yackel & Cobb, 1996) are evident in the
students mathematical interactions in the first and third threads: claims need to be
justified and a problems solution needs to be connected or linked to attributes of
the problem. These norms emerge from the mathematical interactions of students
who have a collective history of problem solving through occasional interactions
over their school years with researchers from Rutgers University who increasingly
over the years left the students to structure their own mathematical investigations in
response to given tasks.
In this chapter and the previous chapters of this section, we have given the
researchers perspectives on the students mathematical work. In Chapter 14, we
examine this work from the point of view of the students.
Part IV
Extending the Study, Conclusions,

and Implications
Chapter 14
Doing Mathematics from the Learners

Perspectives
John M. Francisco
Date and Grade:

Tasks:
Participants:
Researcher/Instructor:
19992000; high school

Clinical interviews
John Francisco
14.1 Introduction
The previous chapters focused on aspects of the cognitive development of the
students in the longitudinal study. The present chapter looks into the epistemological
growth of the students. During the longitudinal study, individual clinical interviews
were conducted with the students with the goal of capturing the mathematical beliefs
that the students might have developed in connection with their experiences in the
longitudinal study. This chapter reports on the analysis of five such interviews.
The results provide insights into the students views about mathematics and about
how it should be learned and taught. The findings challenge the widespread view
that students below college hold nave epistemological views; support studies that
show that students who experience constructivist learning environments tend to
develop sophisticated epistemological beliefs and highlight the important of past
mathematical experiences in framing individuals mathematical beliefs.
Research on students views about mathematics can be placed within the field of
personal epistemological beliefs. This is a field traditionally concerned with describing individuals views about the nature of knowledge and knowing. A substantial
amount of research has been conducted within the field since the pioneering work
of Perry (1970) with Harvard college students. Even more research has been associated with this field since the epistemological construct was expanded to include
individuals views about learning, teaching, and intelligence through the work of
J.M. Francisco (B)
Secondary Mathematics Education, Department of Teacher Education & Curriculum Studies,
University of Massachusetts Amherst, Amherst, MA 01003, USA
e-mail: jmfranci@educ.umass.edu


157
158
J.M. Francisco
Schommer (2002) and Schommer and Walker (1995) and some continental scholars
(De Corte, Opt Eynde, & Verschaffel, 2002; Leder et al., 2002). Even though the
expansion has not been free of controversy, it is recognized that this expansion has
brought the research on the field closer to classrooms practice.
Students epistemological beliefs have been examined in relation to a variety of
constructs. Students beliefs have been studied in relation to the students home and
school environments (Hammer & Elby, 2002) and their teachers epistemological
beliefs (Hofer, 1994; Lyons, 1990; Pirie & Kieren, 1992; Roth & Roychoudhury,
1994). There has been also extensive research that has examined students beliefs
within disciplines (Carey & Smith, 1993; Ceci, 1989; Lampert, 1990; Konold,
Pollatsek, Well, Lohmeier, & Lipson, 1993) as well as across disciplines (Case,
1992; Sternberg, 1989). However, a comprehensive review of the field (Pintrich &
Hofer, 1997, 2002) suggested that a number of challenges remain to be addressed.
One particular challenge is the need for more research on the epistemological
beliefs of students below college level. Except for a few cases (Schoenfeld, 1989;
Pehkonnen, 2002), most research in the field has remained at the college level.
The review notes that there have been few studies involving students below college
and even fewer below high school. The review further points out that lack of such
research has resulted in students below college being assigned nave epistemological beliefs only because research findings show that entering college students tend
to hold such views. Another challenge is the lack of studies that have examined the
epistemological beliefs of students who have experienced a constructivist learning
environments. The few existing studies (e.g., Hofer, 1994) have been exploratory in
nature.
The present study grew out of a natural interest on the part of the researchers to
examine the epistemological growth that participating students in the longitudinal
study might have experienced in connection with the particular conditions in which
they were asked to do mathematics. The researchers were particularly interested
in the students beliefs about (1) success and failure in mathematics, (2) knowing mathematics, (3) learning and teaching mathematics, and (4) how the practices
that they assigned to doing or learning mathematics compared with those in other
disciplines. However, the researchers also sought to make a contribution toward
deepening the research communitys understanding of the epistemological beliefs
of students below college, particularly at the high school level, and of students who
experienced constructivist mathematical environments. The researchers viewed the
longitudinal study as a learning experiment, rather than a teaching experiment,
through which they tried to understand how students construct mathematical ideas
while working on open-ended mathematical tasks in particular conditions. However,
there were no preconceived ideas about what students were to learn or how they
were supposed to learn. Students constructed mathematical ideas and reasoning
were results, not preconceived goals, of the research. This was consistent with a
constructivist approach to learning in the sense that participants had plenty of opportunities to construct and accordingly revise their ideas without any guidance from
the researchers, but rather within their own community of learners.
The students were interviewed about their experiences in the longitudinal study,
and from these interviews, inferences are made about their mathematics beliefs.
14
Doing Mathematics from the Learners Perspectives
159
Their conversations provide insights on their mathematical beliefs and challenge the
widespread view that students below college level hold nave views in contrast to
studies that show students who experience constructivist leaning environments tend
to develop sophisticated epistemological beliefs. However, the results also highlight
the importance of past mathematical experiences in the development of individuals
mathematical beliefs.
This study used a phenomenological approach to the students experiences in
the longitudinal study. Researchers avoided imposing any interpretive framework
on the students (Wilson, 1977) and sought to infer the students epistemological
views among the meanings that the students assigned to their experiences in the
longitudinal study (Creswell, 1998). Overall, the approach was similar to Perrys
(1970) idea of inferring individuals epistemological beliefs from their reflections
on educational experiences.
Data for the present study consisted of 1-h videotaped individual interviews with
the five participating students about their experiences in the longitudinal study. The
four males and one female Ankur, Brian, Jeff, Mike, and Romina agreed to
be interviewed and videotaped. However, it was the students long experience in
the longitudinal study, starting in first grade, that constituted the major criterion
for selecting the students to take part in of the interviews. Their long-term participation in the longitudinal study satisfies the criterion sampling method (Miles &
Huberman, 1994) recommended for phenomenological studies.
The interviews used a semi-structured interview protocol. There was a clear goal
(i.e., capturing the students views about mathematics as a discipline with particular practices and criteria of validity), but the interview proceeded by eliciting and
building on students reflections on their experiences in the longitudinal study to
ascertain the students epistemological views. Typically, the interviews started with
the open question, What are your memories of the longitudinal study? Then the
researchers tried to steer the interviews toward obtaining insights on the students
views on mathematics.
14.2 Findings
The search for answers to the research questions generated five major themes about
personal success and failure in mathematics, mathematics as sense making, mathematics as a discovery activity, mathematics as an activity involving discourse,
and the relationship between mathematics and other disciplines. These themes are
described below, along with supporting statements from the students. (Emphasis
was added to quotes.)
14.2.1 Personal Success/Failure in Mathematics

All of the students described themselves as confident and good in mathematics.
Ankur even said, Im well above average in mathematics, and modesty prevented
160
J.M. Francisco
Mike from describing himself as being better than a normal kid. There were
differences, however, on how the students construed mathematical success. Mike
and Ankur emphasized personal interest and hard work as the ultimate sources of
success. They argued that those who like mathematics can be successful because
they are willing to work harder in mathematics than those who do not like it.
The other students stressed the importance of previous mathematical experiences
and training. In particular, they singled out particular aspects of experiences in the
longitudinal study such as collaborative work and opportunities to come up with
ideas, as opposed to merely receiving them from teachers or experts, as having contributed to their success and confidence in mathematics. Romina further suggested
that confidence and success is built over time:
In fourth grade, I didnt know who you were. Now were comfortable with you. Youve
been our teachers for 10 years. Thats what youve been to us, so now its easier, and we
know whats expected of us, what we have to do. Before we would wait for you to give us a
little start or a little push and point us in the direction. Now you hand us a problem and you
just kind of leave, and we just do it ourselves. We just start experimenting and see what we
can give you.
She also suggested that lack of success can be a function of how success is
defined. She explained that, although she generally felt confident in her abilities,
she might not feel confident in situations where she is asked to engage in rule-based
mathematics, as in textbooks, as opposed to ways that are personally meaningful:
They might throw out, Oh, do you know this rule? Im like, No, but if you sit me down,
maybe I know it. I know it in my own way, not in their way. Everything I explain is in
my own words, not in anyone elses words. Its not from some mathematician from a 1,000
years ago, because I dont know that. I didnt know what the pyramid [Pascals Triangle]
was called. I just know everything in my own way. Everything has Rominas definition to it.
There were no suggestions that the students considered success as a quality or

trait that people are born with. On the contrary, a closer analysis of the students
reflections suggests that the students converge on recognizing the importance of
past mathematical experiences in promoting mathematical confidence and ability
either directly or indirectly through motivation. Rominas last statement also suggests that standardized testing has the potential to portray otherwise bright students
as mathematically weak only because the students do not do mathematics in the
ways prescribed in textbooks or by experts.
14.2.2 Knowing Mathematics as Sense Making

The students reflections on their experiences in the longitudinal study emphasized
the importance of understanding as opposed to memorization of concepts. For example, Mike reported gaining increased conceptual understanding in the longitudinal
study:
It feels different now because I know a lot more than I did before. If I were to solve the
same problems, it would be easier. I understand a lot better too the whole concept behind
14
161
each problem. Like, all the problems that have been given to us, I feel like, somehow, one
is related to each other. When youre little, you cant really understand that.
Romina emphasized the importance of building durable understanding. In particular, she explained that it involved the ability to recall as well as reconstruct
previously learned mathematical concepts:
Because everything I do I understand, because its more than just the numbers to me. If you
understand something from the beginning, youre going to always understand it. You cant
forget something like that. And like an equation, I dont really know any equations. Its like
things, I dont know any solid equations, but I could explain to you something and work
from there. And youre likely to forget an equation.
Jeff related understanding to the ability to explain what one knows to others:
The name really doesnt matter. Thats neither here nor there. I mean, just knowing how to
do it, thats the important part, thats what we learned. And thats being able to do it, being
able to teach it to somebody else, to explain it, to use it for what you need to use it for.
Thats what really matters, not being able to know the name of it, or how to draw it up, or
anything like that.
Brian emphasized the importance of developing the ability to look deeper than
just the surface and of always asking why, qualities which he asserts that he
gained in the longitudinal study:
When I look back at things, Im happy I got involved in this program. Because, I know
at times, I seem very frustrated with it. But if I think hard, I really have gained a lot of
knowledge, and I learned how to look into things deeper than just surface things like, Why
is it like this? Now, I start thinking like that. And it helps me compute things in my mind
better. Like, I really dont know how to put this, but it just helps me in doing things other
than math. I think more in-depth and very seriously about things.
Ankurs idea of understanding was not as explicit as that of others. However,

when recalling different mathematics experiences in different schools, he was clear
about why he liked the one in which the teacher did not teach out of the textbook:
it promoted understanding.
At Harding [Elementary School], my math experience was, Id say it was good. The teacher
would teach, Id understand, Id participate, and it was just, I enjoyed it. And then we went
over to the Springfield [Regional High School], and I did not enjoy geometry class at all. It
was one of the first times that we used the textbook. I dont remember in Harding using a
math textbook. And the teacher would just simply teach out of the book, and assign homework, straight problems, and it wasnt anything that I enjoyed. Then after Springfield, we
came here [to the local Kenilworth High School] and I had Mr. Pantozzi [his mathematics teacher who was also involved with the longitudinal study as a researcher and graduate
student] for 2 years. And I enjoyed that. His teaching style was like none other, and it works.
In particular, if Rominas statement above further suggests that knowing or

understanding mathematics has a personal dimension, Jeffs statement suggests that
understanding has a social or interpersonal dimension to knowing mathematics.
162
J.M. Francisco
Fig. 14.1 Ankur Interview
14.2.3 Mathematics as a Discovery Activity

Ankurs statement in the previous section suggests that he favored a mathematical
environment where students [not teachers] came up with their mathematical ideas
or knowledge as opposed to merely receiving them from teachers or textbooks. The
implicit idea of learning as a discovery activity was present in the reflections of
all students, albeit articulated in slight different ways. Mike argued that discovery
learning helped the majority of students understand mathematics and emphasized
exploration of concepts over time and group work during the discovery process:
Kids can learn new things if they discover them themselves, and not if somebody tells them,
I think that is a better way of learning. Like Mr. Pantozzi, he gives us some information,
but basically, he lets us discover the things that a normal teacher would just tell us. Like
we were learning about e [base of natural logarithms], and he told me that when he was
in school, the teacher told them, e is this, 2.7, whatever. The teacher told him what it
is. In our class, all we did was just explore e. We took days at a time, and I have a good
understanding of it. I guess, in a normal class, only selected kids might understand it. But in
a class where everybodys working together, everybodys a part of the teaching, everybody
or at least the majority of kids will understand it.
Jeff also made reference to working on tasks over time and to group work, but
he stressed the importance of mathematical arguments or discussions during group
work. He asserted that participating in discussions was a better way than listening
to teachers for students to build durable mathematical understanding:
14
163
Fig. 14.2 Jeff Interview
Well when the teacher just comes out and tells you the answer, you find you can study it,
you can get it for that test, but a few weeks later, a few days later, it doesnt matter anymore,
you dont need to know it, and youre onto the next thing. And thats wasting your time.
Because you spend the whole year running through this, you learn, say, twenty different
things, but by the end of the year, youve forgotten them all, and you have nothing. If you
would, say, argue for a couple days or weeks or whatever on different topics, you cover ten
things, but when you walk away, you still know those ten things at the end. And thats why
its important to do that, and not just get the answers.
Romina also singled out mathematical discussions during mathematical activities. In particular, she added that disagreements during the discussions were an
important cognitive mechanism by which students built new knowledge and how
she, personally, learned mathematics:
Because if youre, like, passive, and Im like, This is what I think it is, and everyone is,
Okay, thats what it is, we all sit back and we all take that and we never go any further.
But if I disagree with someone, theyll have to explain it to me, and if theyre explaining it,
theyre either going to find something right, or theyre going to find something more. So,
if I dont agree with it, theyre going to explain it to me, but if they find something wrong,
maybe I can help, and then someone else may disagree with me. And thats how we get
through everything. We just disagree. Ive always had to argue to get somewhere, because
they never actually told me where we were heading with anything. So, through arguing,
thats the only reason I know math.
Brian put the emphasis on hands-on experiences during mathematical activities.

He argued that hands-on activities motivated students to do mathematics and helped
them build durable understanding of mathematics:
164
J.M. Francisco
Fig. 14.3 Brian Interview
If I could change courses, I would make everyone hands-on because kids get tired of sitting
there. But when youre up doing things, time flies, and you have fun and you learn, which
you retain more.
Ankur favored problem-solving activities involving interesting and challenging

tasks and collaborative work, as opposed to teaching by the textbook:
Right now, in my current math class, the teacher doesnt use a book. Id say he comes
up with problems, and most of the problems are interesting problems, and a lot of them
are challenging, and all the students participate. We enjoy working in groups. We help one
another, and that helps out a great deal.
14.2.4 Mathematics as an Activity Involving Discourse

Some of the students statements in the previous section, particularly those by Jeff
and Romina about mathematical arguments, suggest the idea of doing mathematics
as a discursive activity. The statements assign mathematical arguing the cognitive
role of fostering knowledge acquisition. Jeffs statement about understanding as
explaining ideas also suggests the idea of arguments as a way of proving or establishing the validity of mathematical claims. In the statement below, Jeff is even more
explicit about the arguments as way of proving mathematical claims:
14
165
We didnt know if we were right or wrong. You only knew so much, but I would have my
idea about how to get to a certain point and you might have the same idea about how to get
to it. But getting there was the hardest part. That is what we were arguing about, the right
way to get there, the right way to make sure we covered the basis, how to make sure, how
to prove what we needed to accomplish.
Mike suggests a similar idea, but puts it in the context of probability. However,
he does not claim that arguing as proving only takes place in probability. Rather, he
suggests that because uncertainty is more common in probability than in any other
area of mathematics, arguments are more likely to take place in probability:
The reason we argued about math, because math is like, when we do about probability,
probability is an iffy subject. Like, sometimes, I mean the math says its right, but do you
believe its right, and sometimes that influences your decision. Thats probably why we
argue. I remember the problem with the World Series Problem [see Appendix A]. We had
two different answers. I still dont know which one is correct.
The students had a response for those who might claim that in group work, some
students might not be engaged and so might not learn. Above, Mike suggested that
discovery learning with group work can help the majority of students learn mathematics. Ankurs statement below suggests a similar idea and tries to illustrate how
it happens:
Usually, you think that only one person in the group is learning, but if the group fully participates and everyone is involved, everyone in the group learns. When the Rutgers group
comes over here, we all learn. I dont think there is a case when someone doesnt understand. Because if one person doesnt understand, theyll say something, or even if theyre
quiet, someone else will suggest something, will ask them if they understand, or say Could
you explain it back to me And thats how everyone learns.
14.2.5 Mathematics and Other Disciplines

The students had different responses on whether the practices that they associated
with learning or doing mathematics were specific to mathematics or applicable to
other subjects or real life.
Ankur suggested that he thought that teaching out of the textbook, as opposed
to the discovery method, was more suited to other subjects such as history than to
mathematics:
In ninth grade, I was in a different school, and the teacher there taught me differently. She
[the ninth grade math teacher] taught more like a history teacher. A history teacher would
simply teach out of the book, just go right down through the years, and youd learn like that.
But the math teacher, I wouldnt think, a math teacher should teach like that. A math teacher
usually teaches differently. I dont know how to explain it, but it just seems that way. This
teacher taught straight out of the textbook, you wouldnt learn anything more, just simply
what the book stated.
Romina was the most categorical of all the students in her response. She claimed
that arguing about ideas was a learning practice specific to mathematics and could
not be implemented in other subjects such as English and History:
166
J.M. Francisco
Fig. 14.4 Romina Interview
Well, math is where the most arguing is. Like, you cant do this in other classes. Its not like,
in English, you read. You dont argue. Its there. Its written. And in history, you dont do
the same. In math, its like, well especially the way Ive been taught, because I have never
actually had a math teacher thats said, This is the equation, put in the numbers, and do it.
Ive always had to argue to get somewhere.
Brian and Mike, however, had different responses. They suggested that learning
mathematical practices were also relevant to other disciplines. For example, Brian
suggested that his history teacher also used a problem-solving approach as opposed
to just telling students what to do:
Well, the closest thing to my math class would have to be my history class. My history
teacher is an incredible teacher. He always, like for instance, were doing the Cuban Missile
Crisis thing, he set the class up into countries, and we had to all deal with the problems,
instead of just sitting there and telling us. Next to Mr. Pantozzi, he gets us involved just as
much as he does.
Mike argued that his longitudinal study experiences were relevant to other subjects. I think its relevant to a lot of other subjects, like science, history; I guess you
could apply it to, basically, all subjects and claimed that he used his experiences
in the longitudinal study, which he called a type of thinking in real life and other
subjects:
I guess I use the type of thinking in, like other subjects in school; I dont know how you can
apply it to life. Its not hard to recognize what style of thinking youre thinking of. I cant
compare it with someone elses because I dont know what theyre thinking. So, I think,
yeah, I probably do use it in life, and other subjects in school.
14
167
Fig. 14.5 Mike Interview
A closer analysis of the statements suggests that differences among the students
responses regarding the applicability of mathematical learning practices to other
subjects reflect differences in interpretation of the question asked. Jeff, Romina, and
Ankur seem to have answered the question of whether the mathematical learning
practices were actually taking place in other subjects. Mike and Brian, however,
seem to have understood the question as asking whether they believed that those
practices were applicable to other subjects.
14.3 Conclusions
The analysis of the interviews with the five students who participated in this study
suggests that the students (1) are confident in their mathematical ability, (2) emphasize mathematical understanding over memorization, and view mathematics as
(3) a discovery activity and (4) a discursive activity. The results also suggest an
agreement that that (5) the practices were not being implemented in the regular
schools, except in a few isolated cases (a history teacher and Mr. Pantozzi). The
findings suggest a few insights.
The students emphasis on the importance of learning as a discovery activity
suggests that they view themselves as learners as active participants in the construction and justification of their mathematical knowledge, and not as mere receivers
of knowledge and truth from experts or textbooks written by experts. Within the
domain of personal epistemological beliefs, such a view is held by individuals
holding sophisticated or powerful personal epistemological beliefs. As a result, the
findings of the present study challenge the aforementioned widespread belief that
students below college hold nave epistemological beliefs based on research that
show that to be the case among freshmen college students. Given that the conditions
of the longitudinal study were consistent with a constructivist approach to learning,
168
J.M. Francisco
the results also support findings from exploratory studies suggesting that students
who experience constructivist learning environments tend to develop more sophisticated epistemological views than students who experience teaching approaches
based on showing and telling students what to do.
The students views about mathematics are also consistent with the nontraditional approach to mathematical learning and teaching, advocated by the
research community and promoted through publications such as the 2000 Principles
and Standards for School Mathematics of standards of the National Council of
Teachers of mathematics (NCTM). This is evident in the students emphasis on
durable mathematical understanding as opposed to memorization of concepts or
procedures, discovery learning, convincing or explanatory arguments, and collaborative work. The students articulate the merits of such practices in enhancing
learning for the majority of students and point out that these practices also motivate
them to learn. This suggests powerful beliefs within the particular field of personal
epistemological beliefs and within the larger field of mathematics education.
Another dimension of the depth of the students mathematical views is reflected
in the nature of the students articulation of cognitive process involved in doing or
learning mathematics. The students provide different characterizations of mathematical understanding concepts with qualifiers such as conceptual, operational, durable,
personal, and interpersonal. Understanding is also defined not only as recall but also
as the ability to reconstruct previously learned ideas. Arguing is associated with
knowledge acquisition as well as justification or proof for mathematical claims.
There are also rich descriptions of the conditions in which learning, particularly
discovery learning, takes place: hands-on activities, explorations, problem solving, interesting and challenging tasks, collaborative work as arguing or discussing
ideas, work on tasks over time, and so on. The students ability to articulate in
detail different cognitive aspects involved in learning is another measure of depth
of the students mathematical beliefs. In particular, Rominas idea about knowing or understanding mathematics as personal is particularly insightful. A great
deal has been written about the issue under the idea of personal representation
(Francisco & Maher, 2005; Maher, 2005; diSessa & Sherin, 2000; Davis, 1992b;
Davis & Maher, 1990). The idea has been to encourage teachers to attend to and promote the conception between formally defined mathematics and students personal
conceptualizations to promote understanding.
Finally, it is particularly interesting that the students mathematical views mirror
the particular conditions within which they engaged in mathematical activities in
the longitudinal study. Under the idea that the longitudinal study was more of a
learning experiment rather than teaching experiment, researchers encouraged
the students to work collaboratively with other students; justify their reasoning
to classmates; be the arbiters of whether or not a solution was correct based on
whether it made sense; work on the same tasks over an extended period of time;
and revisit similar or same task and refine their ideas and mathematical reasoning.
Such conditions are reflected in the students thoughts about their experiences in the
longitudinal study. This suggests that the importance of construing mathematical
14
169
beliefs within the particular experiences in which students engage in mathematics. In particular, this highlights the importance of teachers paying closer attention
to the kind of beliefs that they might be promoting in their students through their
conscious or unconscious practices or beliefs in mathematics classroom. This is
an area which remains largely unexplored, as few studies have examined the relation between mathematical beliefs and particular settings, whether within or across
cultural settings.
In summary, we provide here an existence proof that students below college
level are capable of building powerful mathematical beliefs and insights about the
cognitive processes and conditions involved in doing mathematics. However, we
also emphasize the importance of examining mathematical beliefs within particular
learning conditions.
In this and preceding chapters, we followed students in the longitudinal study
through elementary school, middle school, and high school, working on problems
in combinatorics. In the following chapter, we look at a group of college students
working on the towers and pizza problems, and we see how their work compares to
that of the younger students.
Chapter 15
Adults Reasoning Combinatorially

Barbara Glass
Date and Grade:

Tasks:
Participants:
Researcher/Instructor:
19982000; College Freshman

Pizzas and Towers
Danielle, Donna, Errol, Jeff C., Linda, Lisa, Mary,
Melinda, Mike C., Penny, Rob, Stephanie C., Samantha,
Steve, Tim, Tracy, and Wesley. (We use the initial C for
college for Jeff, Mike, and Stephanie to distinguish them
from the elementary students of the same names discussed
in other chapters.)
Barbara Glass
15.1 Introduction
In the preceding chapters of this book, we have provided considerable evidence
showing elementary and secondary school students success in solving open-ended
problems, over time, under conditions that encouraged critical thinking. In this
chapter, we address the question as to whether similar results can be achieved by
liberal-arts college students within a well-implemented curriculum that includes a
strand of connected problems to be solved over the course of the semester. From a
perspective of conceptualizing reasoning in terms of solving open-ended problems,
it was of interest to learn whether students in a liberal-arts college mathematics
course could be successful in providing arguments to support their reasoning and in
making connections in a problem-solving-based curriculum.
Students enrolled in college-level mathematics courses might be expected to have
already developed effective reasoning skills. Unfortunately this is too often not the
case. This may be explained, in part, by a history of mathematics instruction in
settings that devalue thinking and focus on rote and procedural learning.
B. Glass (B)
Sussex County Community College, Newton, New Jersey, USA
e-mail: bglass@sussex.edu


171
172
B. Glass
Often, in traditional mathematics classrooms, the answer key or the teacher is

the source of authority about the correctness of answers; unfortunately, quick, correct answers are often valued more than the thinking that leads to the answer. Too
often, teachers ask students to explain their thinking only when answers are wrong,
emphasizing the product rather than the process of problem solving. Sanchez and
Sacristan (2003) offer data to support this from studying students written work.
They report that students are not accustomed to expressing mathematical ideas, and
they offer as an explanation that the emphasis in schools is mainly on producing
correct solutions. One consequence is that students tend to develop the belief that
all problems can be solved in a short amount of time. Students often stop trying to
build a solution if they are unable to solve a problem immediately. For example,
in a survey, high-school students were asked to respond to the question What is a
reasonable amount of time to work on a problem before you know its impossible?
Schoenfeld (1989) reported that the largest response was 20 min and the average
time was 12 min. Further, students view school mathematics as a process of mastering formal procedures. These rules are often removed from real-life experience and
application. As a result, students can feel that answers need not make sense. It is
not surprising, then, that students accept and memorize what they are told without
making any attempt to deal with meaning (Schoenfeld, 1987).
Since many of the students in this study were previously taught mathematics
in this fashion, it would not be entirely surprising if they were unable to apply
knowledge from previous mathematics courses to novel situations. Moreover, since
a students willingness to think about a problem is influenced by notions about what
mathematics is and what should be expected of students, it is not surprising when
students do not display the level of reasoning of which they are capable.
In this chapter, we examine how a small group of community college students
enrolled in a liberal-arts mathematics class solved open-ended non-routine problems in which they had to build and justify a solution. The tasks were the towers
and pizza problems and extensions of these tasks. Our questions were (1) How do
college students solve non-routine mathematical investigations? (2) How do college students representations and level of reasoning contrast with those of younger
students from a longitudinal study engaged in the same investigations? (3) What
connections, if any, do the college students make to analogous problems and to the
rules learned in previous classes? (4) To what extent, if any, do the college students
justify and generalize their results?
15.2 The Study

The study was conducted in a mathematics class for liberal-arts majors called
Mathematical Concepts. The curriculum includes algebra and problem solving.
Liberal-arts students also took a second mathematics course called Contemporary
Mathematics that introduces logic, counting methods including combinations and
permutations, probability and statistics, geometry, and a cluster of applications
15
173
called consumer math. The two liberal-arts mathematics courses can be taken
in either order, so some students in Mathematical Concepts had already taken
Contemporary Mathematics and others had not. Most of the students in the
Mathematical Concepts class take the course to fulfill the mathematics requirement,
although a few take the course as an elective.
The mathematical background of the students in this study varied widely. Some
had taken college preparatory mathematics in high school, while others took only
general mathematics courses. Some had already taken other college-level mathematics courses, while for others this was their first college-level mathematics course.
When asked on a questionnaire about their mathematics background, many students described themselves as being very poor mathematics students who disliked
and feared mathematics, while others stated that they liked mathematics and had
always done well in mathematics classes. There was also a wide range of ages,
some students having recently completed high school, with others not having taken
any mathematics for many years.
The study took place in a relatively new community college of moderate size
in an area of New Jersey that ranges from rural to suburban with very little racial
or cultural diversity. In the fall semester of 2000 there were 929 full-time and 1,357
part-time students enrolled. As with other community colleges, some of the students
attend because poor academic records prevent them from being accepted elsewhere.
Others are excellent students who attend the college for a variety of reasons including lower costs and the convenience of being close to their homes and places of
employment.
Nine classes ranging in size from 6 to 25 students were studied between 1998 and
2000. Sections of the course met for 15 weeks for two 75-min classes each week or
three 50-min classes each week. The students spent approximately half of the class
time working on various non-routine problems in a small group setting. After they
worked together on these problems the students were encouraged to present their
solutions to the class. In addition, a weekly problem-solving homework assignment
was given. As a part of the assignment, students were required to give a written
explanation of their solution method and a justification of how they knew that their
solution was correct. Students also submitted write-ups of the problems done in
class.
Two groups from each class were videotaped as they worked on the towers and
pizza problems. In addition, task-based interviews with ten representative students
were videotaped. Students were selected because they were willing to be videotaped while participating in problem-solving sessions and willing to participate in
videotaped follow-up interviews.
15.3 Student Solutions

The students worked on the towers problem during the 8th or 9th week of the
semester. By this time, they had become accustomed to working on problems and to
justifying their solutions. The students began by working on the four-tall towers
174
B. Glass
problem. They then were asked to consider the five-tall towers problem. Some
groups also worked with three-color towers problems.
The students worked on the pizza problem during the 13th or 14th week of each
semester, first on the four-topping pizza problem and then on the five-topping pizza
problem. After they solved the basic problems, some groups were asked to consider
the pizza with halves problem, in which a topping could be placed on either a whole
pizza or a half pizza.
15.3.1 Towers Problems

Most of the college students used patterns or some other form of local organization immediately, and some immediately imposed a global organization scheme. An
organization by cases according to the number of cubes of one color was the method
chosen by six students. One group, which started the problem by randomly generating towers using a build and check method, switched to this organization by cases
at the suggestion of Jeff C. He said,
Here, put the ones that have three yellows and a red all together. [Danielle rearranges the
towers.] Okay. So now we do three yellows and a red at the bottom, cause you dont have
that. [Jeff C. builds YYYR and hands it to Danielle.] And the ones that have two and two,
put those together. [Danielle rearranges the towers.] Now the ones that have three reds and
the other.
The cases were no cubes of the selected color, then one, two, three, and four of
the selected color. All six students who selected organization by cases determined
that there were two solid-color towers (one all of one color and one all the other
color). All six used a staircase pattern to show that they had found all towers with
three cubes of one color and one cube of the other color; refer to Fig. 15.1 for an
example staircase pattern.
Fig. 15.1 Danas family of

one red cube and four yellow
cubes
RED
YELLOW
15
175
The case with two cubes of one color and two cubes of the other color was more
problematic. The students used a variety of methods to demonstrate that they had
found all towers in this group. Two groups, Melindas group and Donnas group,
stated that they had found all towers because they were unable to find any more.
But as the students in these groups spoke to the instructor, they began to organize
their towers and move toward a proof by cases. However, both groups still stated
that their justification for the claim that they had all towers with two of each color
was that they could not find any more.
Three of the students, Lisa, Errol, and Wesley, tried to argue that the number of
towers is sixteen because four times four is sixteen. The instructor responded that
they needed a reason why the answer should be four times four. Lisa then produced
a proof by cases, although she had difficulty justifying the case of two cubes of each
color. During her interview 7 weeks later, Lisa found an organization that accounted
for all of the towers with two cubes of each color.
Wesley rearranged his towers, but he offered no explanation for why his arrangement produced all possible towers. About 7 weeks later, during the interview,
Wesley produced a similar arrangement and used it to account for all possible combinations with a proof by cases where his cases were (1) towers with four cubes of
the same color together, (2) three cubes of the same color together, (3) two cubes of
the same color together, and (4) no cubes of the same color together.
Errols partner, Mary, offered a proof by cases. However, Errol wanted a proof
that his numerical argument worked. As he continued to think about the problem,
he rearranged the towers in a way that he thought showed that four times four was
the correct answer. This arrangement grouped towers with a red on top together and
towers with a yellow on top together. When the instructor continued to question him
as to why this showed that the answer should be four times four, Errol turned to
simpler cases in an attempt to verify his numerical argument. He then noticed the
doubling pattern and used that to develop an argument by induction, abandoning the
four-times-four argument.
Five of the college students did a proof by cases for the five-tall towers. Each
of these proofs referred to opposites (pairs of towers with opposite colors in the
same positions). They also all used a staircase pattern to account for the towers with
one cube of one color and four cubes of the other color. They used a variety of
methods to justify the cases with two cubes of one color and three cubes of the other
color.
After Rob and his group had organized their towers by cases, they noticed that the
number of five-tall towers was double the number of four-tall towers. They extended
the doubling pattern to predict how many towers they would get if the towers were
three tall, two tall, and one tall. They then built the one-tall and two-tall towers in
order to test their theory. While justifying their answer to the five-tall towers problem, they referred both to their doubling pattern and to a proof by cases (Glass,
2001). The instructor asked the students to think of a reason why the number of
towers doubled. After a few minutes, Rob explained to Steve that the number doubled because you could add either a red cube or a yellow cube to the bottom of each
tower. He explained as follows, building from a generic original tower he called X.
176
B. Glass
Okay, lets say the top of our tower is X, X. [Rob writes an X on his paper.] Then were
putting one on the bottom. For every X we can have a Y [yellow] down here, or for every
X we can have a red [R] down here. So for each block we have, there are now two more
things it could be. So before we just had X. This is X. [Rob picks up the solid red tower of
four as an example.] Now we have XR and XY derived from this. XY and XR. [Rob holds
up RRRRY and RRRRR.]
Steve demonstrated that he understood Robs explanation by using Robs procedure to build two-tall towers by adding cubes to the bottom of one-tall towers.
Wesley built his five-tall towers by adding a red cube to the top of each of his
towers of four. He then built the opposites of these towers to find all towers with
a yellow cube on top. He justified that he had found all five-tall towers with an
inductive argument, but he was unable to extend this reasoning to predict how many
six-tall towers there are. However, during an interview 7 weeks later, Wesley correctly extended the doubling pattern beyond the case that went from four-tall to
five-tall and predicted that there would be 64 six-tall towers.
Jeff C. applied the fundamental counting principle to predict that there would
be 32 five-tall towers. After Jeffs group had produced those 32 towers, he used an
inductive argument to show that they had found all possible towers by pairing each
of the five-tall towers with the corresponding four-tall tower that would generate it.
Errol used the inductive argument that he had developed while working with
four-tall towers to predict that there would be 32 five-tall towers. Even though he
did not build the five-tall towers in class, he used an inductive method to produce
a list of all five-tall towers on his written assignment. Figure 15.2 shows Errols
method.
There are sixteen possibilities. To justify my answer we will start with the possibilities if
the towers were two cubes high (R-red, W-white)
W-W
W-R
We have 4 possibilities.
R-W
R-R
Now, if we want to go to towers three blocks high, we simply take the 4 towers we have
and add a white block to the top and do the same with the red block (8 towers)
W-W-W
R-W-W
R-W-R
W-W-R
&
R-R-W
W-R-W
R-R-R
W-R-R
Now, for 4-cube high towers, we do the same thing: add a white block to the top of all
eight 3-cube high towers and add a red to each of the eight towers. (This would also
work if you put them on the bottom instead). (16 towers)
W-R-W-W
W-R-W-R
W-R-R-W
W-R-R-R
W-W-W-W
W-W-W-R
W-W-R-W
W-W-R-R
&
R-R-W-W
W-R-W-R
R-R-R-W
R-R-R-R
Fig. 15.2 Errols written justification of four-tall towers
R-W-W-W
W-W-W-R
R-W-R-W
R-W-R-R
15
177
b
Fig. 15.3 Pennys written justification of her answer to the four-tall towers problem
Penny, who had been absent the day that the class worked on the towers problem, did the problem at home. She invented a tree diagram strategy to produce an
inductive argument for the four-tall and five-tall towers. Her written work is shown
below. Her diagram is shown in Fig. 15.3.
The answer is 24 = 16 because my first cube will be either blue or brown (2 choices) my
second cube will be either blue or brown matched to a blue or brown first cube (4 choices).
For each of those combinations I can use either a blue or a brown cube, doubling my
possibilities to 8, and for each of those eight combinations I can add a blue or a brown
cube which finally doubles my answer giving me 16 possibilities. (I am using big B for
brown and a little b for blue.)
Tim used a binary coding system to justify his conclusion that he had found all
possible combinations. This is the same method used by tenth-grade Mike from
the Rutgers longitudinal study to justify his assertion that there were 32 different
five-topping pizzas.
Several groups also had time to work on the three-color towers problem. Mike
C.s group and Robs group worked on four-tall towers, while Jeffs group worked
on three-tall towers. Rob and his partners applied the inductive method that they
had developed for towers with two colors to solve the problem quickly. Jeff C. used
the fundamental counting principle to calculate the number of towers, but he did not
use an inductive method to build the towers. Mikes group divided the problem into
two cases: (1) towers with at most two colors and (2) towers with all three colors.
Each student in the group used two of the three colors to build all 16 possible fourtall towers that contained those two colors. After eliminating the three duplicate
one-color towers that they had produced using this method, they had a total of 45
towers. Then they worked on the second case, towers that contained at least one
cube of each color. They each chose one of the colors and built all 12 towers with
two cubes of that color and one cube of each of the other colors. Their solution for
the three-color four-tall towers problem was thus 81.
15.3.2 Pizza Problems

All the college students used a justification by cases approach to the pizza problem.
The students created their two-topping lists systematically; they held one topping
178
B. Glass
fixed and paired it with each of the other toppings. Then they moved to the next
topping on the list. Some students paired each additional topping only with toppings that were below it on the list, reasoning that they had already accounted for
the other combinations. Other students considered all pairs and then eliminated the
ones that they already had. For the three-topping pizzas, some of the students again
systematically went down the list of toppings, while others failed to exhaust both
major and minor items. For example, after some students combined pepperoni and
green peppers with all other possibilities, they moved to the pepper and mushroom
toppings instead of exhausting all possibilities that contained pepperoni.
Two students used a chart that was similar to the chart that fourth-grader Brandon
had created when he did the pizza problem (Maher & Martino, 1998). Instead of the
ones and zeroes that Brandon had used, these students made a check to indicate that
a topping was on a pizza and left a blank space to indicate toppings that were not on
the pizza. Interestingly, this is the same method that Brandons partner Colin used.
Stephanie C., who was simultaneously enrolled in a statistics class, hypothesized that she could calculate the number of pizzas using combinations formulas.
She and her partner Tracy systematically created a list of pizzas and compared the
numbers to those that Stephanie C. had conjectured using her formulas, confirming
Stephanies prediction.
Several other students who had previously studied combinations tried to calculate the number of pizzas using combinations formulas. However, they did not
understand combinations well enough to apply them to the problem correctly. For
example, Melinda stated that the problem could be solved either by combinations or
by permutations, but she could not remember which to use. Also, students who tried
to use formulas attempted to do a single calculation instead of doing separate calculations for each number of toppings. In short, most students were not successful
at using combinations. The correct combinations formula for the number of pizzas
having exactly r toppings when there are n toppings to select from is
n Cr =

n!
n
=
r
r! (n r)!
To find the total number of four-topping pizzas, therefore, it is necessary to sum

C
4 0 through 4 C4 (the number of pizzas with exactly 04 toppings).
15.3.3 Connections Between Problems

During the pizza problem session, several students noticed a relationship between
the towers problem and the pizza problem. For example, Rob and Samantha
explained as follows:
ROB:
INSTRUCTOR:
So we decided the toppings are the block positions.

Okay. So you have pepperoni at the top.
15
ROB:
INSTRUCTOR:
SAMANTHA:
ROB:
INSTRUCTOR:
ROB:
INSTRUCTOR:
ROB:
SAMANTHA:
179
Right. Because it was convenient. So onion would be the second

block, sausage would be the third block, and mushroom would
be the bottom block.
Okay. What would your colors be?
Orange and yellow.
Yeah.
What would orange be?
Orange means its on the pizza.
And yellow?
Its off.
Off the pizza.
Jeff C. was not able to explain the isomorphism during class, but he did explain
it a week later during an interview.
JEFF C.:
INSTRUCTOR:
JEFF C.:
INSTRUCTOR:
JEFF C.:
INSTRUCTOR:
JEFF C.:
INSTRUCTOR:
What you could do to relate that to the topping problem. Is that,

you could say, you could designate each spot for a topping. This
is the pepperoni. This is the green pepper spot, and this is the
sausage spot. [Jeff C. writes toppings by the drawing of the first
tower.]
Okay.
Or onion, or whatever you want to have it.
Whatever.
So with nothing on it, with no toppings, theres only one possibility. [Jeff C. points to the first tower he drew.] Now for
one topping, theres three possibilities for a one topping pizza.
Theres pepperoni. [Jeff C. marks the first block of the second
tower.] Theres green pepper [Jeff C. marks the second block of
the third tower], and theres sausage. [Jeff C. marks the bottom
block of the fourth tower.]
Okay.
Then for a two topping pizza [Jeff C. draws three more towers.], therere three possibilities. [Jeff C. moves the actual towers
with one white cube.] Pepperoni and green pepper [Jeff C. marks
the first and second blocks on the first tower.], pepperoni and
sausage [Jeff C. marks the first and third blocks.], and green
pepper and sausage. [Jeff C. marks the second and third blocks.]
Okay. All right. Which is where the blue, the blue blocks are.
[Jeff C. points to the towers with one white cube.] So in other
words, if you took, you couldnt flip that one around [Jeff C.
takes WBB and turns it upside down and puts it next to BBW.]
because then youd have two of the same combination. Youd
have two pizzas with pepperoni and green pepper.
Okay.
180
JEFF C.:
B. Glass
So those are the three possibilities. The ones with two toppings.
And then for three toppings, theres only one possibility pepperoni, green pepper, and sausage [Jeff C. draws another tower
on paper and marks all the blocks.], or blue, blue, and blue. [Jeff
C. indicates the solid blue tower.]
Robs group tried to relate the pizza toppings to colors in the towers problem;
they conjectured that each topping corresponded to a cube of a specific color. This
is the same explanation that the third-grader Meredith had originally given for the
relationship between the two problems. Meredith later revised her explanation to
note correctly that the blocks in the towers corresponded to the toppings and the colors in the towers corresponded to the presence or absence of each topping (Maher &
Martino, 1999). Unlike Meredith, however, the students in Robs group did not take
their exploration any further and so did not revise their explanation.
Melinda, Stephanie C., Wesley, and Lisa mentioned the relationship between the
two problems during their interviews, approximately 1 week after the class session on the pizza problem. At this time, each student displayed an understanding of
the isomorphism between the two problems. During her interview, Lisa first built
the four-tall towers and then she produced a pizza problem chart similar to the
one that she had made in class. As she completed the chart, she discovered that
the rows of the chart looked like the towers. As a consequence, she was able to
match the towers with the rows in her chart. A portion of Lisas chart is shown in
Fig. 15.4.
Pep.
Onion
Mush.
Bk.
0
0
0
0
0
0
0
0
0
0
0
Fig. 15.4 Lisas chart for the

pizza problem
15
181
Lisa remarked that the entries in her chart were like the binary system that the
class had studied earlier in the semester.
15.3.4 Connections with Pascals Triangle

While working on the pizza problem, Mikes group noticed the doubling pattern
and verified that it continued to hold for smaller numbers of available toppings. As
Mike C. thought about why the number of pizzas should double, he discovered that
the numbers from the pizza problem matched the rows of Pascals triangle. He was
unable to explain, however, why the addition rule of Pascals triangle applied to
pizzas.
After Rob and Steves group completed the four-topping pizza problem, Steve
remarked that they should look for a pattern before moving on to the five-topping
problem. Rob noticed that the pizzas from the four-topping problem formed a row
of Pascals triangle. After figuring out how the addition rule of Pascals triangle
worked with pizzas, Rob used an inductive method and Pascals triangle to create a
list of pizzas for the five-topping pizza problem. Robs explanation of the addition
rule for the Pizza Problem was similar to that used by Mike and the other students
in the Night Session, discussed in Chapter 12. Rob wrote,
Zero
toppings
Zero row
One
topping
1st row
Two
toppings
2nd row
Three
toppings
0
1
P
0
1
0
1
P
G
M
3
0
1
3rd row
Four
toppings
P
G
M
O
4th row
Five
toppings
P
G
M
O
S
5th row
1
P
G
2
PG
1
PM
GM
PG
3
PO
GO
MO
PM
GM
PG
6
PS, GO
GS, MO
MS, PM
OS, GM
PO, PG
10
Fig. 15.5 Robs drawing of Pascals triangle
PGM
1
PMO
GMO
PGO
PGM
PGMO
POS, PGO
GOS, PMO
MOS, GMO
PMS, GMS
PGS, PFM
10
PMOS
GMOS
PGOS
PGMS
PGMO
PGMOS
182
B. Glass
The reason this [Pascals triangle] works is because every time we add another topping we
are increasing the possibility of choice, without losing the old ones. In other words, all the
two topping pies in the two topping total still apply when there are three total toppings. Also
all those that had two toppings, by adding a third topping, are now three topping pizzas.
In this way we can see absolutely, positively, without a doubt, that we have all possible
combinations each new row is built by adding the old columns with the new topping.
This once again is the Pascals triangle principle of adding old combinations with new
possibilities to find new combinations.
Robs drawing of Pascals triangle is shown in Fig. 15.5.
15.4 Discussion
The college students solved problems and justified their solutions in many of the
same ways as the Kenilworth students had done in elementary school and high
school. However, the college students solved the problems more quickly than the
third and fourth grade children and, unlike the elementary school children, the college students did not generally rely on random checking as a primary strategy. All
college students were able to solve the four-tall towers problem and start the fivetall towers problem within a 50-min or a 75-min period. Most also finished the
five-tall towers problem within the same period. Several students also had time to
work on extensions of the problem. However, the college students also showed less
inclination than the Kenilworth students to think about problems for extended periods of time. Many stopped thinking about the problem after they had arrived at an
answer, even when the instructor asked them to spend some more time considering
the problem in order to reveal their thinking.
After Rob and his group had built their towers and organized them by cases,
they noticed the doubling pattern and they developed a proof by induction. This
is similar to what Milin had done in the fourth grade. (Refer to Chapter 5 for
details.) Mike C. also noticed that the number of towers and pizzas was doubling, but
Mike C. was unable to provide a reason for the doubling pattern. It is interesting to
note that in fourth grade both Stephanie and Milin had noticed the doubling pattern
in the towers problem. Stephanies progression from pattern recognition to development of an inductive argument took about 8 months while Milins understanding
developed more quickly (Alston & Maher, 2003; Maher & Martino, 2000). Perhaps
Mike C. would also have recognized the reason for the doubling pattern if, like
Stephanie, he had been given an extended time frame in which to develop his ideas.
Mike C., however, was limited to about 8 weeks to develop his ideas about the
problem.
The college students made very few connections with previous mathematical
knowledge, and most of these connections were trivial connections. Several students recognized that the problems were related to permutations or combinations,
but most did not understand these concepts well enough to correctly apply them
to the solutions of novel problems. This would suggest that learning about mathematics in an atmosphere in which students are told what to do does not enable
15
183
them to develop genuine understanding. In contrast, the students in this study did
demonstrate a high level of reasoning as they thought about the problems, justified
their solutions, and made connections between problems and the mathematics they
learned within the course.
Some of the conditions of the Rutgers University longitudinal study, such as
extended classroom sessions and revisiting the same problem several times within
an extended time frame, could not be replicated because of the time constraints of
a college course. However, many of the conditions that enabled the elementary and
secondary students to become thoughtful problem solvers were duplicated. Both
groups were given rich mathematical tasks and were encouraged to explain their
reasoning and methods of solution and to justify their solutions to the problems.
Both groups were engaged in thoughtful mathematics. They found patterns, developed methods of justifications, and provided justifications that their patterns were
reasonable. It cannot be disputed that the students in the Rutgers longitudinal study
benefited from exposure to rich mathematical experiences over an extended period
of time. It is also significant that the students in this study, who had previously experienced a variety of generally traditional mathematics instruction, demonstrated that
it is not too late to introduce rich mathematical experiences in a collegiate level
mathematics class. The level of reasoning that these students demonstrated provides
evidence that it is possible to experience thoughtful mathematics within a traditional
15-week college semester.
In this chapter, we have shown how college students worked on the towers and
pizza problems. In the following chapter, we will follow up on the work on these
college students and other college students, comparing the work of these college
undergraduates to the work of the longitudinal study high-school students (see
Chapter 8) on the extension of the towers problem called Ankurs Challenge (finding
the number of four-tall towers, built when choosing from three cube colors, having
at least one cube of each color).
Chapter 16
Comparing the Problem Solving of College

Students with Longitudinal Study Students
Barbara Glass
16.1 Introduction
In this chapter we consider a variety of solutions to Ankurs Challenge from students
ranging from high school to graduate level study of mathematics. The problem was
created by Ankur and first posed to tenth-grade classmates in an after-school session
of the Rutgers longitudinal study (see Chapter 8).
Find all possible towers that are four cubes tall, selecting from cubes available in three
different colors, so that the resulting towers contain at least one of each color.
Since 1998, when Ankur first presented this problem to four high school classmates, we have given his problem to several cohorts of students enrolled in liberal
arts mathematics classes and in graduate mathematics-education courses. Students
presented their written work and gave further verbal explanations and clarifications
of their solutions. Researcher notes provided the data for the oral explanations. We
classified their forms of reasoning into four categories: (1) justification by cases,
(2) inductive argument, (3) elimination argument, and (4) analytic method (use of
formulas). We present here some representative solutions from the high school (H),
undergraduate (U), and graduate (G) students, listed according to the arguments that
were used.
16.2 Justifications by Cases

We list below nine ways that students used justification by cases; one from high
school students, five from undergraduates, and two from graduate students. All those
who used justification by cases began with the observation that one color would
B. Glass (B)
Sussex County Community College, Newton, New Jersey, USA
e-mail: bglass@sussex.edu


185
186
B. Glass
appear twice and each of the other colors would appear once. Many then proceeded
by selecting one color and listing all possibilities for that color in some organized
fashion and then arguing from symmetry that the other cases would be the same.
Details are given below for these justifications by cases.
16.2.1 Romina, Jeff, and Brians Solution (H)

Rather than using particular colors, these high school students used the general
codes 1, 0, and X to indicate the three colors, and they focused on the placement
of the duplicate color, using 1 to indicate the color that was duplicated and X and
O to indicate the other two colors. As shown in Fig. 16.1, they found six possible
ways to place the two 1s in a four-tall tower. They observed that each tower pictured in Fig. 16.1 represents two towers, one with the X and O as shown on the
top of the diagram and the other with the X and O in the bottom position, giving
12 towers. Finally, arguing from symmetry, they noted that X and O could also be
the duplicate color, and so there are 12 towers for the duplicate X case and 12 for
the duplicate O case. They concluded that there are therefore 36 towers that fulfill
Ankurs condition.
Fig. 16.1 Romina, Jeff, and
Brians six prototype towers
O
x
x
O
O
x
x
O
O
x
x
O
O
x
x
O
O
x
x
O
O
x
x
O
16.2.2 Joanne and Donnas Solution (U)

Joanne and Donna used red, blue, and green cubes. They found six ways to place
two blocks of the same color in a tower containing four blocks, counting from the
top: positions 1 and 4, 1 and 3, 1 and 2, 2 and 4, 2 and 3, and 3 and 4. Each
16
Comparing the Problem Solving of Students
187
Fig. 16.2 Joanne and Donnas listing of towers
of those six possibilities gives six towers because there are three possibilities for
the same color blocks and two possibilities for the remaining color. Six towers for
each of six possibilities gives 36 towers. Refer to Fig. 16.2 for their enumeration of
towers.
16.2.3 Rob and Jessicas Solution (U)

Rob and Jessica used yellow, red, and blue cubes. For their first case (which we
designate as Case I), they chose yellow as the duplicated color. For Subcase A, they
fixed the blue cube in the top position and then they moved the red cube into the
second, third, and fourth positions. This gives a total of three towers for this subcase.
For Subcase B, they fixed the red cube in the top position and moved the blue cube
into the second, third, and fourth positions, giving another group of three towers.
For Subcase C, they placed one yellow cube in the top position of the tower and the
second yellow cube in the second, third, and fourth positions in turn. They noted
that each position of the second yellow cube produces two towers, because the red
and blue cubes can be reversed. Therefore, there are six towers for Subcase C. The
total number of towers for Case I (yellow as the duplicated color) is therefore 12.
They repeated this process for the case of two red cubes and two blue cubes, giving
a total of 36 towers. Refer to Fig. 16.3 for Rob and Jessicas list of 12 towers in
Case I.
Fig. 16.3 Rob and Jessica: 12 towers with two yellow cubes
188
B. Glass
16.2.4 Maries Solution (U)

Marie used blue, red, and yellow cubes. She started by assuming that the blue cube
was the color that appears twice. For the three cases, under this assumption, Marie
controlled for variables. For the first case (Case 1) under this assumption, she fixed
the position of the first blue cube on the top of the tower and then she moved the
second blue cube to the second, third, and fourth positions. Reasoning from symmetry, Marie noted that there are two towers for each of those positions because the red
and yellow cubes can be reversed. So for Case I, she found six towers. For Case 2,
she fixed the first blue cube in the second position and moved the second blue cube
into the two remaining possible positions. As noted above, reasoning from symmetry, Marie knew that there are two towers for each position; therefore, there are four
towers for Case 2. For Case 3, Marie placed the two blue cubes in the third and
fourth position; that gives two towers for this case. Refer to Fig. 16.4 for a diagram
of the 12 towers that Marie found under the assumption that the duplicated color
is blue. Again reasoning from symmetry Marie noted that this process could be
repeated for red and yellow as the cube appearing twice, and so there are 36 towers.
Fig. 16.4 Maries three cases for blue
16.2.5 Bobs Solution by Cases (U)

Bob used red, blue, and yellow cubes. He wrote, There has to be one color that
appears twice, while the other two colors appear once. If the blue cube appears
twice, keep the two blue cubes together and move to all possible positions. There
are two towers for each position because the other two colors can be reversed. Next
separate the two blue cubes by one and move into all possible positions. Again each
position will give two towers. Finally place the two blue cubes in the first and fourth
position, separated by two cubes, to give two more towers. This process can be
repeated for each of the other colors. Figure 16.5 shows Bobs three cases for blue.
16.2.6 Aprils Solution (U)

April used blue, purple, and white (B, P, and W) cubes. She started by considering
the case where blue is the top cube in the tower. April wrote, Start with blue on
the top. If there is also a blue in the second position, the third and fourth position
16
189
Fig. 16.5 Bobs three cases for blue
Fig. 16.6 Aprils three cases: Blue on top, white on top, and purple on top
must be PW or WP in order to have all three colors in the tower. If the second
cube is purple, the other two cubes must have at least one white cube. They can be
WW, BW, WB, WP, or PW. If the second cube is white, the other two cubes must
have at least one purple cube. They can be PP, PB, BP, PW, or WP. This gives 12
combinations with blue on top. There are also 12 combinations with white on top
and 12 combinations with purple on top for a total of 36 towers. Figure 16.6 is a
diagram of Aprils f three cases of blue on top, white on top, and purple on top.
16.2.7 Bernadettes Solution (U)

Bernadette used blue, purple, and white cubes. Her enumeration by cases was similar to Aprils, except she worked from the bottom up; her three cases were: blue
cube on the bottom of the tower, white cube on the bottom, and purple cube on the
bottom. Figure 16.7 shows Bernadettes three cases.
190
B. Glass
Fig. 16.7 Bernadettes three

cases: blue, white, and purple
on the bottom
Bernadette made an exhaustive list of all possibilities for towers with a blue cube
on the bottom and the repeated the procedure for the other two colors. She wrote of
her attempts to organize her work by finding a pattern:
After a lot of trouble in class trying to figure this problem out, and with your help, I finally
came to the conclusion that there are 12 possibilities with each color block on the bottom,
with 36 possibilities in all. I know my answer is correct because of the pattern I found.
For example, the blue block was the concrete tower for 12 towers. The blue block was
positioned in every possible way, including one blue block at least always on the bottom.
Then, the purple and white blocks would alternate positions. I showed this on my work. I did
the same with the white and purple blocks, replacing the blue, getting only 36 possibilities
of towers.
16.2.8 Tims Solution (G)

Tim used the colors red, yellow, and green (R, Y, and G); he described three cases of
towers with the required conditions and reasoned from symmetry that all three cases
would have the same number of towers. He built the required towers by starting
with four-tall towers built from two colors with exactly two cubes of each color.
His diagram, shown in Fig. 16.8, illustrates the organization (holding the top cube
fixed and moving the second cube of that color into all possible other positions)
that shows that there are exactly six such towers. Once he had those six towers, he
Fig. 16.8 Tims diagram for

the four-tall towers with two
red and two green
R
R
G
G
R
G
R
G
R
G
G
R
G
R
R
G
G
R
G
R
G
G
R
R
16
191
exchanged each of the red cubes for a yellow cube in order to fulfill the conditions
of the problem; since this can be done in two ways (first exchange one red cube and
then exchange the other red cube), this gave him 12 towers with two green cubes.
Reasoning from symmetry, he concluded that there are 12 towers for each of the
other two cases. He described his reasoning as follows: Given three colors Red
Yellow and Green, towers 4 tall containing at least one cube of every color will
yield towers with 1R, 1Y, 2G; 1 R, 2Y, 1G; and 2R, 1Y, 1G. All these cases will be
equal in number. Consider 2R and 2G. There are six towers that are four tall with
2R and 2G. Now exchange a Y for one of the two Rs in each tower. There are two
ways to do this for each tower. Therefore there are 2 6 = 12 towers of [each of
the cases] 1R, 1Y, 2G; 1R, 2Y, 1G; and 2R, 1Y, 1G, for a total of 36.
16.2.9 Tracis Solution (G)

Traci used colors A, B, and C. First she found all towers that fulfilled the condition
with color A on the bottom (Case 1). For her first subcase (AB), she placed a cube
of color A on the bottom and a cube of color B next. Within this subcase, Traci first
placed C in the third position; the top position can thus be any of the colors, giving
three towers ABCA, ABCB, and ABCC. Continuing the AB subcase, Traci kept AB
in place and determined that there were two more towers that started with AB and
that did not have C in the third place; these are ABAC and ABBC. Thus for the
AB subcase, she had a total of five towers. The second subcase was AC. Within this
case, Traci first assigned B to the third position. As with subcase AB, the top position
can be any one of the three colors, giving three towers ACBA, ACBB, and ACBC.
To finish subcase BC, Traci left the first positions as AC and found two additional
towers, ACAB and ACCB. Thus for subcase AC, there are again five towers. For
the last subcase (AA), Traci assigned AA to the bottom two cubes; the two towers
that can be made with AA are AABC and AACB. These three subcases cover all
possibilities for the case in which color A is on the bottom; they give a total of 12
towers with A. Arguing from similarity, Traci found that there are also 12 towers
with B on the bottom and 12 with C on the bottom, for a total of 36 towers. Tracis
written work is shown in Fig. 16.9.
16.3 Inductive Arguments

Four students (three undergraduates and one graduate student) used inductive
arguments, either in whole or as part of their justifications of their solutions.
16.3.1 Errols Solution (U)

Errols solution combined proof by cases with an induction argument. He used red,
blue, and white cubes, with a tree diagram and a horizontal code (for example,
RBWR for redbluewhitered). He fixed the first cube as red and worked with three
192
B. Glass
Find all permutations with A on bottom
ABC
CC
ABC
BB
CB
CCC
AB
BBB
AC
BC
BBB
BB
CCC
CC
AA
AAA
AA
AAA
AA
AA
Use all
three
colors
once, then
add extra
A
bottom, B
next
remaining
A
bottom C
next all
colors
once, then
alt.
A
bottom, C
next
remaining
A
bottom, C
next
remaining
12
perm, w/A
on the
bottom
There will be 12 permutations w/B on the bottom, and another 12

permutations w/C on the bottom.
12 X 3 = 36
Fig. 16.9 Tracis justification by cases
subcases (second cube white, second cube blue, and second cube red). If the second
cube is white, then Errol said that the third and fourth cubes would be one of the
four combinations bluewhite, whiteblue, bluered, or redblue; Errol missed the
fifth combination (blueblue). Similarly, Errol noted that if the second cube is blue,
then the third and fourth cubes would be one of the four combinations whitered,
redwhite, whiteblue, or redblue; here he missed the whitewhite combination.
For his third subcase, Errol noted that if the first two cubes are red, then the third
and fourth cubes would have the other two colors (whiteblue or bluewhite), for a
total of two combinations. He thus found 10 of the 12 possible towers, and so when
he multiplied this case by three (since there are three colors involved, we can find
out how many possibilities there are with one specific color on top, and multiply the
answer by 3), his answer of 30 was off by 6. Figure 16.10 shows Errols written
work.
16.3.2 Christinas Solution (U)

Using colors A, B, and C, Christina built up the four-tall towers by starting with
all nine possible two-tall towers that can be built when selecting from three colors.
Using a strategy of controlling variables, she built these towers by adding A, B,
and C to each of the three one-tall towers. (Fig. 16.11 shows her set of two-tall
towers.) Christina then placed a cube of color A on the top of each two-tall tower,
giving nine three-tall towers, as shown in Fig. 16.12. Each of the nine three-tall
towers produced three four-tall towers when she added a fourth cube of each color
to the bottom, for a total of 27 towers. After Christina crossed out the 15 towers that
did not fulfill Ankurs condition (see Fig. 16.13), she was left with 12 towers. She
16
Fig. 16.10 Errols written justification
Fig. 16.11 Christinas set of

two-tall towers
Fig. 16.12 Christinas set of

three-tall towers with an A on
top
Fig. 16.13 Christinas set of 27 four-tall towers, with 12 meeting Ankurs condition
193
194
B. Glass
followed the same procedure for colors B and C and produced 24 more towers for a
total of 36.
16.3.3 Bobs Inductive Solution (U)

Working with colors red, blue, and yellow, Bob started by making the six three-tall
towers that have all three colors, as shown in Fig. 16.14. He demonstrated that all
such towers were accounted for by controlling for variables: he held the top color
fixed and moved the other two cubes in both possible positions.
Fig. 16.14 Bobs starting set
of towers
Then Bob built four-tall towers by adding a red, yellow, or blue cube to the bottom of each of the six three-tall towers. He noted that this would give all towers with
two of the same color on the bottom and the other colors in all possible positions.
Then Bob added a red, yellow, or blue cube to the top of the six three-tall towers,
giving all towers with the of the same color on the top and the other colors in all
possible positions. But with this procedure, Bob missed the towers with the duplicated color in the middle. When the instructor asked him to justify his solution, Bob
abandoned this method and returned to a proof by cases. Bobs solution by cases is
discussed earlier in this chapter.
16.3.4 Frances Solution (G)

Frances used a tree diagram as shown in Fig. 16.15, with red, yellow, and blue cubes
(R, Y, and B). Her strategy was a combination of proof by cases and induction.
Case I was to start with color R, Case II was to start with color Y, and Case III was
to start with color B. Describing Case I on the tree diagram, she said, (1) I started
on the first block as R (color 1). Then the second could be R (color 1), Y (color 2),
or B (color 3). (2) If the second is R, the third could only be Y or B. If the third is
Y, then the fourth must be B. If the third is B, the fourth must be Y [2 towers]. (3) If
the second is B, then the third could be R, Y, or B. If R, the fourth could only be Y.
If B, the fourth could only be Y. If Y, the fourth could be R, Y, or B [5 towers]. (4)
If the second is Y, then the third could be R, Y, or B. If R, the last could only be B.
If Y, the last could only be B. If B, the last could be R, Y, or B [5 towers, for a total
of 12].
Frances made the same inductive arguments for Cases II and III, for a total of 36
towers.
16
195
R
R
Y
Y
B
12
B R Y
B Y R Y B Y
R
R
B R Y B B R
R B
R Y BR
12
36
R
R
R
Y
R
Y R Y B YR Y
12
B R R
Fig. 16.15 Frances tree diagram showing the generation of towers for three cases
16.4 Elimination Arguments

Four students (two undergraduates and two graduate students) used elimination
arguments. All started with the fact that 81 (34 ) towers can be made when building
four-tall towers and selecting from three colors. Then they eliminated the towers
that did not meet Ankurs criteria. Penny listed all the towers and crossed off the
ones that did not meet the criteria. Robert, Liz, and Mary calculated the number of
towers that did not have all three colors and subtracting them from 81. Robert, Liz,
and Mary also used formulas; they used the fact that the number of n-tall towers
when selecting from three colors is 3n .
16.4.1 Pennys Solution (U)

Pennys argument was a combination of inductive reasoning and elimination. She
used a tree diagram to list all 81 four-tall towers with three colors, and then she
crossed off the towers that did not meet Ankurs criteria.
196
B. Glass
16.4.2 Roberts Solution (U)

Robert used colors red, blue, and green. Robert started with the 81 four-tall towers
that can be built when selecting from three colors; then he subtracted the towers that
did not meet Ankurs criteria; these are the towers with exactly one color and the
towers with exactly two colors. Robert gave the number of towers with exactly one
color as 314 , or three (one all red, one all blue, and one all green). He found the
number of four-tall towers with exactly two of the colors by calculating the number
of all possible four-tall towers (which is 24 ) and subtracting the towers with only
one of the colors (which is 214 ). Since there are three combinations of two colors
(red/blue, red/green, and blue/green),
number by three.
two-color

he multiplied
the
His complete calculation is: 34 3 24 2 14 3 14 = 81 45 = 36
Refer to Fig. 16.16 for Roberts written work.
Problem: Want to build a tower 4 high with 3 uniquer colors (I chose

red, green and blue) with the restriction, that at least 1 of each color is
used.
I started by first calculating the total number of towers that are 4 tall
w/3 colors and restrictions. So I drew a tower and mapped out
possiilities
3
Each segment of the tower had three

possible choices for color so total amount
was 3333 = 34
Red, Blue or Green

Next, I decided to subtract the towers that did not meet the clause of
at least one of each color - I started with the easiest onesones
which were 4 tall, but used only one color.
R
Which I found to be 3, but decided to write as 3 (14). Which helped

me see something better. Then I decided to subtract the ones that are
4 tall, 2 colors, at least one of each which 24.
But I had to subtract 2 (or2(14)) because that left possibility of
getting all one color or another. Since there are three combinations of
two colors (red/blue, red/green, blue/green) I multiplied by 3 giving
total number of 2 of 3 colors, 3 high, at least one of each color to
3(242(14)). That gave me the total number of 4 tall, 3 colors, at
least 1 of each color to
34 3(24 2(14)) 3(14) = 34 45 = 36
Fig. 16.16 Roberts written justification
16
197
16.4.3 Lizs Solution (G)

Lizs strategy was similar to Roberts; she started with the 81 four-tall towers that
can be built selecting from three colors, and then she subtracted the towers that
do not meet Ankurs criteria, which are the towers with at most two colors. She
observed that there are 16 towers with just red and green cubes (including the all
red and all green towers). Similarly, there are 16 towers with just red and blue cubes
(including the all red and all blue towers), and 16 towers with just blue and green
cubes (including the all blue tower and the all green tower). Since the three towers
of all one color each appeared twice in the three groups of 16, Liz concluded that
there are 316 3 or 45 towers that do not fit Ankurs criteria. Subtracting 45 from
81 gives 36, the number of towers with at least one of each color.
Liz wrote:
There are 34 = 81 towers four-tall when choosing from three colors. But not all of
them have at least one of each color. Subtract out the ones that dont have at least
three colors. These are the ones that have at most three colors.
Say the colors are r, g, b.
There are 24 = 16 with just r & g, including 1 all r, 1 all g.
There are 24 = 16 with just r & b, including 1 all r, 1 all b.
There are 24 = 16 with just b & g, including 1 all b, 1 all g.
There are three duplicates here
There are 3 16 3 = 45 with at most two colors. 81 45 = 36 with at least one
of each color.
16.4.4 Marys Solution (G)

Like Liz, Mary started with the fact that 81 four-tall towers can be built when
selecting from three colors, and she also calculated that there are three sets of
16 four-tall towers that can be built when selecting from two colors. She subtracted 48 from 81, giving her 33 towers that she though fulfilled Ankurs condition.
Then she noticed that she subtracted three too many (the three single-color towers), and she added back the three, giving 36. Figure 16.17 shows Marys written
explanation.
16.5 Analytic Method

Only one student used a completely analytic method, relying exclusively on
formulas, although others used formulas in individual pieces of their solutions.
198
B. Glass
Fig. 16.17 Marys written work
16.5.1 Leanas Solution (G)

Leana used her knowledge of combinatorics formulas to calculate that the number
of ways to arrange two As, one B, and one C is 4 factorial (4!) divided by 2 factorial
(2!), which is 12. The same formula applies when B is the repeated color and again
when C is the repeated color, giving 36 towers fulfilling Ankurs criteria. Figure
16.18 shows Leanas written analysis.
16.6 Discussion
We placed the forms of reasoning displayed by these students into four major categories: elimination, inductive, controlling variables, and recursive. However, there is
16
199
4tall 3 colors at least 1 of each color

Colors A B C
CASE I
CASE II
A is the color repeated
B is the color repeated
C is the color repeated
CASE III
AABC
BBAC
CCAB
Tower
How many different
ways can you arrange
AABC?
How many different

BBAC?
How many different

CCAB?
4!
2!
Same
Same
eliminates repeats
12
12
4.3.2.1
= 12
2
12 + 12 + 12 = 36
Another way:
A1A2BC
4.3.2.1
2
24
2
A is the color being repeated

123 = 36
= 12
Works for each color
Gets rid of repeats
Fig. 16.18 Leanas analysis of Ankurs Challenge
considerable variation within each category as well as overlap among the categories.
All but one of the students who used an elimination method used formulas to calculate the total number of towers. That one student (Penny) used an inductive method
to generate her list of all 81 towers. All but two of the students who chose to do
a justification by cases did so by controlling for variables. The other two students
(Marie and Bob) instead used a recursive argument in which they focused on a fixed
cubed and rotated it exhaustively for particular cases. The approaches to arguing
by cases varied. Students chose different cases into which to separate the towers
and different variables for which to control as they built their justification. There
were also variations within the other approaches. For example, students started their
inductive argument at different tower heights. Errol and Francis started at height one
and Christina started at height two. Bob started at height three, but he missed some
of the towers as a result. He eventually resolved the discrepancy when he used a
200
B. Glass
solution by cases. Bob was the only student who used two different methods, and
that strategy helped him to find the towers that he missed when using induction.
The sharing of ideas was an important component in students problem solving.
It provided them with the opportunity to review their work, reflect on their ideas,
and sometimes to modify their results. While the written work does not show the
interchange of ideas that came about as students discussed their work with others,
the invitation to students to share their ideas resulted in a more careful review of the
work and thus a greater confidence in the reasoning offered. For example, it was only
after sharing her justification with the instructor that April became confident that
she had indeed found all possible towers. Also, Mary found her error (subtracting
out some towers twice) when reviewing her work for presenting to others. In some
cases, the discussion revealed to students flaws in their reasoning, resulting in a reexamination of the solution. As an illustration, the process of justifying that he had
the correct number of towers enabled Bob to realize that his inductive method had
caused him to miss several combinations. In this context, we can observe how the
process of justifying their answers can enable students to reflect upon what they
have done and on whether their answer is reasonable.
While the forms of reasoning generally fell into the four categories, the distribution of correct solutions was not uniform according to category. Few students
(mostly graduate students) used formulas, and most of those students also used an
elimination argument. Also, only graduate students and one senior undergraduate
student used formulas correctly. Undergraduates successfully used arguments by
cases and induction, and their primary method of solution was reasoning by cases.
Rich problems can be challenging and engaging for students at a wide range of
levels. Ankurs challenge, a problem initially proposed to a group of high school students, has turned out to be of interest to students at many levels, and it has resulted
in multiple kinds of thoughtful arguments. An important feature of this problem
was that students were required to account for all of the towers and then to build
arguments that were convincing to themselves and to others. It may be that problems that call for explanation and justification trigger sense making in students.
We suggest that multiple opportunities for students to express ideas, revise them,
and share them both in writing, and verbally are important contributors to students
sense making. Therefore we recommend that problems that invite students to explain
and justify their ideas in writing and in the verbal sharing of results be included in
mathematics courses at all levels.
Chapter 17
Closing Observations
Arthur B. Powell
In the previous 16 chapters, we have witnessed ordinary students develop extraordinary mathematical ideas, forms of reasoning, and heuristics. Extraordinary are these
students accomplishments since their mathematical behaviors emerged not from
quickly parroting rules and formulae but rather from deliberately engaging their own
discursive efforts. As Speiser (Chapter 7, this volume) notes, these students built
fundamental mathematical understanding, over time, through extended task-based
explorations. They created models, invented notation, and justified, reorganized, and
extended previous ideas and understandings to address new challenges. That is, they
performed mathematics: created mathematical ideas and reasoned mathematically.
These behaviors ideating and reasoning are fundamental human activities and
how they occur in the realm of mathematics, specifically elementary combinatorics,
is what this book contributes.
Internationally, a community of mathematics education researchers has recognized this how question as substantially important. In January 1983, David H.
Wheeler (19252000), the founding editor of the international journal, For the
Learning of Mathematics, sent a letter to 60 or so mathematics educators inviting them to engage a daunting task: suggest research problems whose solution
would make a substantial contribution to mathematics education (Wheeler, 1984,
p. 40). The varied and thought-provoking responses of more than 15 educators were
published, some in each of the three issues of the fourth volume of the journal. On
Wheelers mind was the famous example of the 23 problems from various branches
of mathematics that David Hilbert (18621943) announced in an address delivered
to the Second International Congress of Mathematicians in 1900 at Paris (p. 40)
and predicted that from the discussion of which an advancement of science may
A.B. Powell (B)

Department of Urban Education, Rutgers University, Newark, NJ, USA
e-mail: powellab@andromeda.rutgers.edu


201
202
A.B. Powell
be expected (Hilbert, 1900, p. 5).1 Of all the published responses to Wheelers

challenge, Tall (1984) offered the briefest list of what he considered to be the central questions: (1) how do we do mathematics? and (2) how do we develop new
mathematical ideas? (p. 25, emphasis added).
Hilberts 23 problems contributed to more than a century of vigorous, fruitful
research activity in physics and mathematics.2 Similarly, considered responses to
Talls two questions require substantial research efforts in different environments
over extended periods of time. For researchers in mathematics education to entertain these questions, we must find ways to observe what learners do as they do
mathematics as well as to describe and analyze how they develop their mathematical
ideas.
It bears noting that 8 years after Tall issued his central questions, Davis et al.
(1992) similarly challenged mathematics education researchers to study the emergence among learners of what lies at the core of mathematics: mathematical ideas.
Expanding on Talls second question, Davis noted that very little research in mathematics education has focused on the actual ideas in students minds or on how
well teachers are able to identify these ideas, interact with them, and help students
improve on them (p. 732). The chapters of this book have presented rich descriptions and analyses of actual ideas that students built in the realm of combinatorial
reasoning. This work has implications for teaching both in the design and sequence
of effective tasks and in demonstrating how teachers could productively interact
with student ideas.
The global picture depicted in the chapters of this book underscores the need
for time to think deeply and discursively. A special issue of Educational Studies in
Mathematics collected several analyses concerning discourse in mathematics classrooms. Commenting on these studies, Seeger (2002) wonders about the possibility
of a grand, panoramic theory of learning. In arguing for a comprehensive theory
of mathematics education, he suggests that such a theory needs to embrace four
metaphors of learning that form the axes social individual and construction
acquisition, and represents them in a two-by-two grid (p. 289).3 In addition, Seeger
further suggests that theoretical work has to be balanced by the systematic development of focal problems for practice, theory, and research in mathematics education
1 According to Gray (2000), Hermann Minkowski (18641909), whose metric concept (order-p
geometry) provided the theoretical foundation for non-Euclidean, taxicab geometry, was a close
friend of Hilbert and urged him to accept the invitation to speak at the Congress: Most alluring
would be the attempt to look into the future, in other words, a characterisation of the problems to
which the mathematicians should turn in the future. With this, you might conceivably have people
talking about your speech even decades from now (as quoted in Gray, 2000, p. 1).
2 See Grattan-Guinness (2000) for a critical appraisal of the range of Hilberts problems against
the panoply then evident in mathematics.
3 Here Seeger (2002) differs from Sfards (1998) theorization in which she argues for two
metaphors acquisition and participation or construction that conceptualizes perspectives on
learning and in which she claims that though complementary they are mutually not amenable to
critique.
17
Closing Observations
203
(p. 289). He proposes two focal problems for mathematics education, one concerning ecological validity and representation and the other referring to the question of
time and change.
Besides epistemological concerns, the question of time and change also concerns
methodological issues. Building ideas and understanding are certainly temporal and
unbounded. Consequently, there are complex judgments an investigator has to make
when inquiring into what learners build, understand, or acquire from a discussion or
lesson on a particular issue. When does an investigator examine what learners say,
do, and write? Should these actions be examined in the immediate proximity of the
discussion or lesson, in some other, more distinct time, or in some combination of
these times? Ball and Lampert (1999) raise somewhat similar questions in a study
of teaching practice.
Epistemologically and methodologically, this book contributes to understanding
the relation between time and development. As outcomes of individual and collective constructive actions over the course of the longitudinal study, the participants
build ideas, reason, and employ heuristics to resolve various tasks. They reveal
and make salient the important relationship between time and development. The
processes by which the participants build their ideas evidence an epistemological
reality: knowledge construction is often a slow process. Mathematical ideas do not
develop instantaneously and robustly but rather emerge slowly and in their nascent
state are rather fragile. Ideas dawn and mature over time. To loose fragility, among
other things, ideas need to be reflected on deeply, presented publicly, submitted to
challenge, available for negotiation, and subject to modification. That is, the essence
of developing and understanding mathematical ideas is often a protracted, iterative,
and recursive phenomenon, occurring over more time than is usually appreciated
or acknowledged in practice in classrooms and in reports in the literature (Pirie &
Kieren, 1994; Seeger, 2002). If learners are to develop deep understandings that are
less fragile and more durable than is often witnessed by teachers in schools, they
need to be offered extended periods of time to wrestle with a problem as well as
to debate and negotiate heuristics, to articulate and justify their results, and to have
their ideas challenged and then defend or modify their ideas.
If we agree that students must be actively and purposely engaged in their learning
so that they can take ownership and be proud of their accomplishments, we need to
create opportunities for this to occur. For example, strands of investigations can be
integrated into the regular curriculum as enrichment. When we eliminate the pressure of testing and grades, students can invest in thinking and reasoning for its own
sake and for the intrinsic rewards that knowing deeply entails. Perhaps every several
weeks, within particular strands, investigations can be revisited, and students can
bring their more recent, accumulated knowledge to a more sophisticated examination of earlier solutions, and thereby extend their knowing. A focus on justification
as a strand of school mathematics has great potential for building a solid foundation
for the later study in many fields and certainly of mathematics. A focus on reasoning and sense making is an important requirement for a productive, responsible
citizenry. Questioning, challenging, analyzing, revisiting all lead to better ways of
knowing. Can we as educators meet the challenge of educating thoughtful students
204
A.B. Powell
who are motivated by sense making and the critical review of ideas? The challenge
awaits us.
The Video Mosaic Collaborative at Rutgers University provides a mechanism
for the ongoing building and sharing of knowledge. We invite readers to visit
our website at http://www.video-mosaic.org/, view the videos and accompanying
objects that provided the data for this book and join our expanding community of
researchers by providing additional study and analysis.
Appendix A
Combinatorics Problems
Listed here are the major combinatorics problems the students encountered from
elementary school through high school, along with brief discussions of solutions.
1. Shirts and Jeans (May 1990, Grade 2; October, 1990, Grade 3) Stephen has
a white shirt, a blue shirt, and a yellow shirt. He has a pair of blue jeans and a
pair of white jeans. How many different outfits can he make?
He can make six different outfits; each of the two pairs of jeans can be
matched with each of the three shirts. The outfits are: blue jeans/white shirt,
blue jeans/blue shirt, blue jeans/yellow shirt, white jeans/white shirt, white
jeans/blue shirt, and white jeans/yellow shirt.
2. Shirts and Jeans Extended (October, 1990, Grade 3) Suppose Stephen had
another pair of jeans, a black pair. How many different outfits can he now make?
He can make 12 different outfits: number of shirts times number of jeans.
3. Four-Tall Towers (October 1990, Grade 3; December 1992, Grade 5) Your
group has two colors of Unifix cubes. Work together and make as many different towers four cubes tall as is possible when selecting from two colors. See if
you and your partner can plan a good way to find all the towers four cubes tall.
At each position in the tower, there are two color choices. Therefore, there are
2222=16 possible towers that are four cubes tall. This can be generalized
to an n-tall tower with two colors to choose from; there are 222. . . 2=2n
possible towers that are n cubes tall, when there are two colors to choose from.
This can also be generalized to an n-tall tower with m colors to choose from;
there are mmm. . . m=mn possible towers that are n cubes tall with m
colors to choose from. In the following discussions, we will call the first generalization (the n-tall tower with two colors) the towers problem, and we will call
the second generalization (the n-tall tower with m colors) the generalized tower
problem.
4. Cups, Bowls, and Plates (April 1991, Grade 3) Pretend that there is a birthday
party in your class today. Its your job to set the places with cups, bowls, and
Education Library 47, DOI 10.1007/978-0-387-98132-1,

205
206
Appendix A: Combinatorics Problems
plates. The cups and bowls are blue or yellow. The plates are blue, yellow,
or orange. Is it possible for ten children at the party each to have a different
combination of cup, bowl, and plate? Show how you figured out the answer to
this question.
Each of the two cup choices can be matched with each of the two bowl choices,
and each cup-bowl pair can be matched with any of the three different plate
choices. Therefore, there are 223 = 12 possibilities. Therefore, yes, it is
possible for ten children at the party each to have a different combination of
cup, bowl, and plate.
5. Relay Race (October 1991, Grade 4) This Saturday there will be a 500-m
relay race at the high school. Each team that participates in the race must have a
different uniform (a uniform consists of a solid colored shirt and a solid colored
pair of shorts). The colors available for shirts are yellow, orange, blue, or red.
The colors for shorts are brown, green, purple, or white. How many different
relay teams can participate in the race?
There are four choices for shirts and four choices for shorts, so there are 44
= 16 ways to make uniforms. Sixteen different relay teams can participate.
6. Five-Tall Towers (February 1992, Grade 4; December 1997, Grade 10) Your
group has two colors of Unifix cubes. Work together and make as many different towers five cubes tall as is possible when selecting from two colors. See if
you and your partner can plan a good way to find all the towers five cubes tall.
There are 25 = 32 towers five cubes tall.
7. Four-Tall Towers with Three Colors (February 1992, Grade 4) Your group has
three colors of Unifix cubes. Work together and make as many different towers
four cubes tall as is possible when selecting from three colors. See if you and
your partner can plan a good way to find all the towers four cubes tall.
Since there are three choices for each of four positions, there are 34 = 81
possible towers that are four cubes tall when selecting from three colors.
8. A Five-Topping Pizza Problem (December 1992, Grade 5; December 1997,
Grade 10) Consider the pizza problem, focusing on the number of pizza combinations that can be made when selecting from among five different toppings.
There are 25 = 32 different pizzas.
9. Guess My Tower (February 1993, Grade 5) You have been invited to participate in a TV Quiz Show and the opportunity to win a vacation to Disney
World. The game is played by choosing one of four possibilities for winning
and then picking a tower out of a covered box. If the tower you pick matches
your choice, you win. You are told that the box contains all possible towers that
are three tall that can be built when you select from cubes of two colors, red,
and yellow. You are given the following possibilities for a winning tower:
207
All cubes are exactly the same color.

There is only one red cube.
Exactly two cubes are red.
At least two cubes are yellow.
Which choice would you make and why would this choice be better than any
of the others?
In order to decide which is the best choice, we need to find the probability of
each choice. The total number of 3-tall towers is 8. The probabilities are:
All cubes are exactly the same color: There are two ways (all red or all
yellow). The probability is 28 = 0.25.
There is only one red cube: There are three ways; the red cube can be on the
top, in the middle, or on the bottom. The probability is 38 = 0.375.
Exactly two cubes are red: This is the same as saying exactly one cube is
yellow. The probability is the same as for exactly one red cube: 38 = 0.375.
At least two cubes are yellow: This is equivalent to saying that either exactly
two cubes are yellow or exactly three cubes are yellow. As discussed above,
the probability that exactly two cubes are yellow (the same as the probability
that exactly two cubes are red) is 0.375. Since there is one way for exactly
three cubes to be yellow, that probability is 18 = 0.125. The probability of
either event is therefore 0.375 + 0.125 = 0.5. (We can add because the two
events are mutually exclusive.)
At least two cubes are yellow is the most likely event.
Assuming you won, you can play again for the Grand Prize which means you
can take a friend to Disney World. But now your box has all possible towers
that are four tall (built by selecting from the two colors yellow and red). You
are to select from the same four possibilities for a winning tower. Which choice
would you make this time and why would this choice be better than any of the
others?
The total number of four-tall towers is 24 = 16. The probabilities are:
All cubes are exactly the same color: There are two ways (all red or all
yellow). The probability is 216 = 0.125.
There is only one red cube: There are four ways; the red cube can be on the
top, second from the top, second from the bottom, or on the bottom. The
probability is 416 = 0.25.
Exactly two cubes are red: The number of ways to accomplish this is C(4,2)
= 6. The probability is therefore 616 = 0.375.
208
At least two cubes are yellow: This means that exactly two cubes are yellow,
exactly three cubes are yellow, or exactly four cubes are yellow. As discussed
above, the probability that exactly two cubes are yellow (the same as the
probability that exactly two cubes are red) is 616 = 0.375. The probability
that exactly three cubes are yellow is the same as the probability that one
cube is red: 416 = 0.25. Since there is one way for exactly four cubes to
be yellow, that probability is 116 = 0.0625. The probability of any one of
the three events is therefore 0.375 + 0.25 + 0.0625 = 0.6875.
At least two cubes are yellow is the most likely event.
10. The Pizza Problem with Halves (March 1993, Grade 5) A local pizza shop
has asked us to help them design a form to keep track of certain pizza sales.
Their standard plain pizza contains cheese. On this cheese pizza, one or two
toppings could be added to either half of the plain pizza or the whole pie. How
many choices do customers have if they could choose from two different toppings (sausage and pepperoni) that could be placed on either the whole pizza
or half of a cheese pizza? List all possibilities. Show your plan for determining
these choices. Convince us that you have accounted for all possibilities and that
there could be no more.
With two topping choices, there are four possibilities for the first half pizza,
because each topping can be either on or off that half of the pizza. The four
choices are: plain (sausage off, pepperoni off), sausage (sausage on, pepperoni
off), pepperoni (sausage off, pepperoni on), and sausage/pepperoni (sausage on,
pepperoni on). Consider each of the four possibilities in turn.
Case 1: Plain. There are four possibilities for the other half of the pizza, the
four listed above (plain, sausage, pepperoni, and sausage/pepperoni).
Case 2: Sausage. There are three possibilities for the other half of the pizza:
sausage, pepperoni, and sausage/pepperoni. (We omit plain, because we
already accounted for the plain-sausage pizza in Case 1.)
Case 3: Pepperoni. There are two possibilities remaining for the other half of
the pizza: pepperoni and sausage/pepperoni. (Plain and sausage are already
accounted for.)
Case 4: Sausage/pepperoni. There is only one possibility left for the other
half of the pizza; that is sausage/pepperoni.
There are 4+3+2+1 = 10 possible pizzas with halves.
11. The Four-Topping Pizza Problem (April 1993, Grade 5) A local pizza shop
has asked us to help design a form to keep track of certain pizza choices. They
offer a cheese pizza with tomato sauce. A customer can then select from the
following toppings: peppers, sausage, mushrooms, and pepperoni. How many
different choices for pizza does a customer have? List all the possible choices.
209
Find a way to convince each other that you have accounted for all possible
choices.
There are 2222 = 16 possible pizzas.
12. Another Pizza Problem (April 1993, Grade 5) The pizza shop was so pleased
with your help on the first problem that they have asked us to continue our
work. Remember that they offer a cheese pizza with tomato sauce. A customer
can then select from the following toppings: peppers, sausage, mushrooms, and
pepperoni. The pizza shop now wants to offer a choice of crusts: regular (thin)
or Sicilian (thick). How many choices for pizza does a customer have? List all
the possible choices. Find a way to convince each other that you have accounted
for all possible choices.
Each of the 16 four-topping pizzas has two choices of crust, so there are 32
pizzas.
13. A Final Pizza Problem (April 1993, Grade 5) At customer request, the pizza
shop has agreed to fill orders with different choices for each half of a pizza.
Remember that they offer a cheese pizza with tomato sauce. A customer can
then select from the following toppings: peppers, sausage, mushroom, and pepperoni. There is a choice of crusts: regular (thin) and Sicilian (thick). How many
different choices for pizza does a customer have? List all the possible choices.
Find a way to convince each other than you have accounted for all possible
choices.
The first half of the pizza can have 24 = 16 possible topping configurations,
as described above. Consider each of those configurations in turn. Following
the procedure described above for the two-topping half-pizza problem, we find
that there are 16+15+14+. . . +3+2+1 possible pizzas; this sum is given by
16172. Since each pizza can have a thick or thin crust, we multiply by 2.
The number of possible pizzas is 161722 = 272.
14. Counting I and Counting II (March 1994, Grade 6) How many different twodigit numbers can be made from the digits 1, 2, 3, and 4? Each of four cards is
labeled with a different numeral: 1, 2, 3, and 4. How many different two-digit
numbers can be made by choosing any two of them?
Counting I: Assuming that you are not permitted to reuse digits, there are four
choices for the first digit and three for the second digit, giving 12 two-digit
numbers. (They are 12, 13, 14, 21, 23, 24, 31, 32, 34, 41, 42, and 43.)
Counting II: There are four choices for the first digit and four choices for the
second digit. This makes 16 different two-digit numbers. (They are 11, 12, 13,
14, 21, 22, 23, 24, 31, 32, 33, 34, 41, 42, 43, and 44.)
15. Towers-Binomial Relationship (March 1996, Grade 8) In an interview,
Stephanie discusses the relationship between the towers problems and the
binomial coefficients.
210
Binomial coefficients arise in connection with the binomial expansion formula

(a+b)n . The following can be shown by induction:
(a + b) =
n
n

n
r=0
anr br
The coefficient of anr br is given by:

n!
n
=
r
r!(n r)!
This number is the rth entry in the nth row of Pascals Triangle, and it gives
the number of towers with exactly r cubes of one color, when building towers
that are n-tall and there are two colors to choose from. Hence, the binomial
expansion and the towers problem are isomorphic, with the number of instances
of a in the rth term being equal to the number of towers having exactly r cubes.
16. Five-Tall Towers with Exactly Two Red Cubes (January 1998, Grade 10) You
have two colors of Unifix cubes (red and yellow) to choose from. How many
five-tall towers can you build that contain exactly two red cubes?
You are selecting two items (the positions of the two red cubes) from five
choices (the number of cubes in the tower); there are ten ways to do this:

5!
5
= 10
=
2
2!(5 2)!
17. Ankurs Challenge (January 1998, Grade 10) Find all possible towers that are
four cubes tall, selecting from cubes available in three different colors, so that
the resulting towers contain at least one of each color. Convince us that you
have found them all.
Suppose the colors are red, blue, and green. We are counting the towers in three
cases: (1) those with two red cubes, one blue cube and one green cube, (2) those
with one red cube, two blue cubes, and one green cube, and (3) those with one
red cube, one blue cube, and one green cube. The following equation gives the
number of ways of selecting m groups of objects of size r1 through rm :
n
r1 , r2 , ..., rm

=

n!
, where
ri = n
r1 ! r2 ! ... rm !
So the number of four-tall towers containing exactly two red cubes, one blue
cube, and two green cubes is:
4
2, 1, 1

=
4!
= 12
2! 1! 1!
211
Similarly for the other two cases:

4
1, 2, 1

=
4
1, 1, 2

= 12
Hence the number of towers with the required condition is 12+12+12 = 36.
18. The World Series Problem (January 1999, Grade 11) In a World Series, two
teams play each other in at least four and at most seven games. The first team to
win four games is the winner of the World Series. Assuming that the teams are
equally matched, what is the probability that a World Series will be won: (a) in
four games? (b) in five games? (c) in six games? (d) in seven games?
The number of ways for a team to win the series (four games) in n games is
the number of ways it can win three times in n1 games (and then win the
last game). This is given by C(n1,3). The probability of any given set of outcomes for n games is 12n (since there are two equally likely outcomes for
each game). So the probability that one team wins the series in n games is
given by C(n1,3) 2n , and the probability of a win for either team is double
that: C(n1,3) 2n1 . The probabilities are:
(a) C(41,3) 241 = C(3,3) 23 = 18 = 0.125.
(b) C(51,3) 251 = C(4,3) 24 = 416 = 0.25.
(c) C(61,3) 261 = C(5,3) 25 = 1032 = 0.3125.
(d) C(71,3) 271 = C(6,3) 26 = 2064 = 0.3125.
19. The Problem of Points (February 1999, Grade 11) Pascal and Fermat are
sitting in a caf in Paris and decide to play a game of flipping a coin. If the
coin comes up heads, Fermat gets a point. If it comes up tails, Pascal gets a
point. The first to get ten points wins. They each ante up 50 francs, making the
total pot worth 100 francs. They are, of course, playing winner takes all. But
then a strange thing happens. Fermat is winning, eight points to seven, when he
receives an urgent message that his child is sick and he must rush to his home
in Toulouse. They carriage man who delivered the message offers to take him,
but only if they leave immediately. Of course, Pascal understands, but later, in
correspondence, the problem arises: how should the 100 francs be divided?
We can list all the circumstances where Fermat gets two points before Pascal
gets three points. He can do this in two flips, three flips, or four flips. (The game
cannot proceed past four flips. As soon as both players get to nine points, the
next flip will produce a winner. It takes three flips for this to happen.)
(a) Two flips: Fermat wins both. Probability =122 = 14.
(b) Three flips: Fermat wins one of the first two and the last one. Probability
= C(2,1)23 = 14.
212
(c) Four flips: Fermat wins one of the first three and the last one: Probability
= C(3,1)24 = 316
Probability of any of these events = 14 + 14 + 316 = 1116. Therefore
Fermat should get 100 1116 Francs 69 Francs and Pascal should get 31
Francs.
20. The Taxicab Problem (May 2002, Grade 12) A taxi driver is given a specific
territory of a town, shown below. All trips originate at the taxi stand. One very
slow night, the driver is dispatched only three times; each time, she picks up
passengers at one of the intersections indicated on the map. To pass the time,
she considers all the possible routes she could have taken to each pick-up point
and wonders if she could have chosen a shorter route. What is the shortest route
from a taxi stand to each of three different destination points? How do you
know it is the shortest? Is there more than one shortest route to each point? If
not, why not? If so, how many? Justify your answer.
Using Powells et al. (2003) notation to denote coordinates on the taxicab grid,
(n,r) indicates a point n blocks away from the taxi stand and r blocks to the
right. So the blue dot is at (5,1), the red dot is at (7,4), and the green dot is at
(10,6). Taking the shortest route means going in two directions only (down and
to the right). Finding the number of shortest paths from the taxi stand (0,0) to
any point (n,r) involves the number of ways to select r segments of one kind of
movement in a path that includes two kinds of movements; i.e., the number of
shortest paths to (n,r) is C(n,r). For the specific cases given above, the shortest
paths are:
Blue: C(5,1) = 5.
Red: C(7,4) = 35.
Green: C(10,6) =210.
Appendix B
Counting and Combinatorics Dissertations
from the Longitudinal Study
Franciso, J. M. (2004). Students reflections on their learning experiences: Lessons

from a longitudinal study on the development of mathematical ideas and
reasoning. Unpublished doctoral dissertation, Rutgers University, Newark, NJ
Glass, B. H. (2001). Mathematical problem solving and justification with community college students. Unpublished doctoral dissertation, Rutgers University, New
Jersey.
Kiczek, R. D. (2001). Tracing the development of probabilistic thinking: Profiles
from a longitudinal study. Unpublished doctoral dissertation, Rutgers University,
New Jersey.
Martino, A. M. (1992). Elementary students construction of mathematical knowledge: Analysis by profile. Unpublished doctoral dissertation, Rutgers University,
Newark, NJ.
Muter, E. M. (1999). The development of student ideas in combinatorics and proof:
A six year study. Unpublished doctoral dissertation, Rutgers University, Newark,
New Jersey.
OBrien, M. (1994). Changing a school mathematics program: A ten-year study.
Unpublished doctoral dissertation, Rutgers University, New Jersey.
Powell, A. B. (2003). So lets prove it!: Emergent and elaborated mathematical
ideas and reasoning in the discourse and inscriptions of learners engaged in a
combinatorial task. Unpublished doctoral dissertation, Rutgers University, New
Jersey.
Muter, E. M. (1999). The development of student ideas in combinatorics and proof:
A six year study. Unpublished doctoral dissertation, Rutgers University, Newark,
New Jersey.
Sran, M. K. (2010). Tracing Milins development of inductive reasoning: A case
study. Unpublished doctoral dissertation, Rutgers University, Newark, NJ.
Steffero, M. (2010). Tracing beliefs and behaviors of a participant in a longitudinal
study for the development of mathematical ideas and reasoning: A case study.
Unpublished doctoral dissertation, Rutgers University, Newark, NJ.
Tarlow, L. D. (2004). Tracing students development of ideas in combinatorics and
proof. Unpublished doctoral dissertation, Rutgers University, New Jersey.
Uptegrove, E. B. (2005). To symbols from meaning: Students investigations in
counting. Unpublished doctoral dissertation, Rutgers University, New Jersey.
213
References
Alston, A. S., & Maher, C. A. (2003). Modeling outcomes from probability tasks: Sixth graders
reasoning together. In N. A. Pateman, B. J. Dougherty, & J. T. Zilliox (Eds.), Proceedings
of the 27th annual conference of the International Group for the Psychology of Mathematics
Education (vol. 2, pp. 2532). Honolulu, HI: CRDG, College of Education, University of
Hawaii.
Ball, D. L., & Lampert, M. (1999). Multiples of evidence, time, and perspective. In E. C.
Lagemann & L. S. Schulman (Eds.), Issues in education research: Problems and possibilities
(pp. 371398). San Francisco, CA: Jossey-Bass.
Bellisio, C. W. (1999). A study of elementary school students ability to work with algebraic notation and variables. Unpublished doctoral dissertation, Rutgers, The State University of New
Jersey, New Brunswick, NJ.
Bruner, J. (1960). The process of education. Cambridge, MA: Harvard University Press.
Carey, S., & Smith, C. (1993). On understanding the nature of scientific knowledge. Educational
Psychologist, 28(3), 235251.
Case, R. (1992). Neo-Piagetian theories of child development. In R. J. Sternberg & C. A. Berg
(Eds.), Intellectual development. New York: Cambridge University Press.
Cathcart, W. G., Pothier, Y. M., Vance, J. H., & Bezuk, N. S. (2006). Learning mathematics in
elementary and middle schools: A learner-centered approach. Upper Saddle River, NJ: Pearson
Prentice Hall.
Ceci, S. J. (1989). On domain specificity. . .More or less general and specific constraints on
cognitive development. Merril-Palmer Quarterly, 35(1), 131142.
Cobb, P., Wood, T., Yackel, E., & McNeal, B. (1992). Characteristics of classroom mathematical
traditions: An interactional analysis. American Educational Research Journal, 29, 573604.
Creswell, J. W. (1998). Qualitative inquiry and research design: Choosing among five traditions.
Thousand Oaks, CA: Sage.
Davis, R. B. (1984). Learning mathematics: The cognitive science approach to mathematics
education. Hillsdale, NJ: Lawrence Erlbaum Associates.
Davis, R. B. (1992a). Reflections on where mathematics education now stands and on where it may
be going. In D. A. Grouws (Ed.), Handbook of research on mathematics teaching and learning
(pp. 724734). New York: Macmillan.
Davis, R. B. (1992b). Understanding Understanding. Journal for Research in Mathematics
Education, 11, 225241.
Davis, R. B., & Maher, C. A. (1990). The nature of mathematics: What do we do when we do
mathematics? [Monograph]. Journal for Research in Mathematics Education, 4, 6578.
Davis, R. B., & Maher, C. A. (Eds.). (1993). Schools, mathematics, and the world of reality.
Needham, MA: Allyn & Bacon.
Davis, R. B., & Maher, C. A. (1997). How students think: The role of representations. In L. D.
English (Ed.), Mathematical reasoning: Analogies, metaphors, and images (pp. 93115).
Hillsdale, NJ: Lawrence Erlbaum Associates.
215
216
References
Davis, R. B., Maher, C. A., & Martino, A. M. (1992). Using videotapes to study the construction of mathematical knowledge by individual children working in groups. Journal of Science,
Education and Technology, 1(3), 177189.
De Corte, E., Opt Eynde, P., & Verschaffel, L. (2002). Knowing what to believe": The relevance
of students mathematical beliefs for mathematics education. In B. K. Hofer & P. R. Pintrich
(Eds.), Personal epistemology: The psychology of beliefs about knowledge and knowing.
Mahwah, NJ: Lawrence Erlbaum Associates.
diSessa, A. A., & Sherin, B. L. (2000). Meta representation: An introduction. Journal of
Mathematical Behavior, 19(4), 385398.
Franciso, J. M. (2004). Students reflections on their learning experiences: Lessons from a longitudinal study on the development of mathematical ideas and reasoning. Unpublished doctoral
dissertation, Rutgers University, Newark, NJ.
Francisco, J. M., & Maher, C. A. (2005). Conditions for promoting reasoning in problem solving:
Insights from a longitudinal study. Journal of Mathematical Behavior, 24(2/3), 361372.
Freudenthal, H. (1991). Revisiting mathematics education: China lectures. Dordrecht, The
Netherlands: Kluwer Academic Publishing.
Grattan-Guinness, I. (2000). A sideways look at Hilberts twenty-three problems of 1900. Notices
of the American Mathematical Society, 47(7), 752757.
Gray, J. (2000). The Hilbert problems 19002000. Newsletter 36 of the European
Mathematical Society. Retrieved February 10, 2003, from http://www.mathematik.unibielefeld.de/kersten/hilbert/gray.html
Hammer, D., & Elby, A. (2002). On the form of a personal epistemology. In B. K. Hofer &
P. R. Pintrich (Eds.), Personal epistemology: The psychology of beliefs about knowledge and
knowing. Mahwah, NJ: Lawrence Erlbaum Associates.
Harvard-Smithsonian Center for Astrophysics. (2000). Private universe project in mathematics.
Retrieved July 4, 2008, from http://www.learner.org/channel/workshops/pupmath
Hilbert, D. (1900). Mathematical problems. Retrieved February 7, 2003, from http://babbage.
clarku.edu/djoyce/hilbert/problems.html
Hofer, B. K. (1994). Epistemological beliefs and first-year college students: Motivation and cognition in different instructional contexts. Paper presented at the annual meeting of the American
Psychological Association, Los Angeles, CA.
Hofer, B. K. (2002). Personal epistemology as a psychological and educational construct: An introduction. In B. K. Hofer & P. R. Pintrich (Eds.), Personal epistemology: The psychology of
beliefs about knowledge and knowing. Mahwah, NJ: Lawrence Erlbaum Associates.
Konold, C., Pollatsek, A., Well, A., Lohmeier, J., & Lipson, A. (1993). Inconsistencies in students
reasoning about probability. Journal for Research in Mathematics Education, 24(5), 392414.
Lampert, M. (1990). When the problem is not the question and the solution is not the answer:
Mathematical knowing and teaching. American Educational Research Journal, 27, 2963.
Landis, J. H. (1990). Teachers prediction and identification of childrens mathematical behaviors:
Two case studies. Unpublished doctoral dissertation, Rutgers, The State University of New
Jersey, New Brunswick, NJ.
Landis, J. H., & Maher, C. A. (1989). Observations of Carrie, a fourth grade student, doing
mathematics. Journal of Mathematical Behavior, 8(1), 312.
Leder, G., Pehkonen E., & Trner, G, (Eds.) (2002). Beliefs: A hidden variable in mathematics
education? Dordrecht: Kluwer.
Lyons, N. (1990). Dilemmas of knowing: Ethical and epistemological dimensions of teachers
work and development. Harvard Educational Review, 60(2), 159180.
Maher, C. A. (1988). The teacher as designer, implementer, and evaluator of childrens mathematical learning environments. The Journal of Mathematical Behavior, 6, 295303.
Maher, C. A. (1998). Constructivism and constructivist teaching: Can they coexist? In
O. Bjorkqvist (Ed.), Mathematics teaching from a constructivist point of view (pp. 2942).
Finland: Abo Akeademi, Pedagogiska fakulteten.
References
217
Maher, C. A. (2002). How students structure heir own investigations and educate us: What we have
learned from a fourteen year study. In A. D. Cockburn & E. Nardi (Eds.), Proceedings of the
twenty-sixth annual meeting of the International Group for the Psychology of Mathematics
Education (PME26) (Vol. 1, pp. 3146). Norwich, England: School of Education and
Professional Development, University of East Anglia.
Maher, C. A. (2005). How students structure their investigations and learn mathematics: Insights
from a long-term study. Journal of Mathematical Behavior, 24(1), 114.
Maher, C. A. (2008). The development of mathematical reasoning: A 16-year study (Invited Senior
Lecture for the 10th International Congress on Mathematics Education, published in book with
electronic CD). In M. Niss (Ed.), Proceedings of ICME 10 2004. Roskilde, DK: Roskilde
University, IMFUFA, Department of Science, Systems and Models.
Maher, C. A., & Davis, R. B. (1990). Building representations of childrens meanings. In R. B.
Davis, C. A. Maher, & N. Noddings (Eds.), Constructivist views on the teaching and learning of
mathematics: Journal for Research in Mathematics Education monograph (vol. 4, pp. 7990).
Reston, VA: National Council of Teachers of Mathematics.
Maher, C. A., & Davis, R. B. (1995). Childrens explorations leading to proof. In C. Hoyles &
L. Healy (Eds.), Justifying and proving in school mathematics (pp. 87105). London:
Mathematical Sciences Group, Institute of Education, University of London.
Maher, C. A., & Martino, A. (1991). The construction of mathematical knowledge by individual children working in groups. In P. Boero (Ed.), Proceedings of the 15th conference of the
International Group for the Psychology of Mathematics Education (vol. 2, pp. 365372). Assisi,
Italy.
Maher, C. A., & Martino, A. M. (1992a). Teachers building on students thinking. The Arithmetic
Teacher, 39, 3237.
Maher, C. A., & Martino, A. M. (1992b). Individual thinking and the integration of the ideas of
others in problem solving situations. In W. Geeslin, J. Ferrini-Mundy, & K. Graham (Eds.),
Proceedings of the sixteenth annual conference of the International Group for the Psychology
of Mathematics Education (pp. 7279). Durham, NH: University of New Hampshire.
Maher, C. A., & Martino, A. M. (1996a). The development of the idea of mathematical proof:
A 5-year case study. Journal for Research in Mathematics Education, 27(2), 194214.
Maher, C. A., & Martino, A. M. (1996b). Young children inventing methods of proof: The gang of
four. In L. Steffe, P. Nesher, P. Cobb, G. Goldin, & B. Greer (Eds.), Theories of mathematical
learning. Hillsdale, NJ: Erlbaum.
Maher, C. A., & Martino, A. M. (1997). Conditions for conceptual change: From pattern recognition to theory posing. In H. Mansfield (Ed.), Young children and mathematics: Concepts and
their representations. Durham, NH: Australian Association of Mathematics Teachers.
Maher, C. A., & Martino, A. M. (1998). Brandons proof and isomorphism can teachers help
students make convincing arguments? (pp. 77101). Rio de Janeiro: Universidade Santa Ursala.
Maher, C. A., & Martino, A. M. (1999). Teacher questioning to promote justification and generalization in mathematics: What research practice has taught us. Journal of Mathematical
Behavior, 18(1), 5378.
Maher, C. A., & Martino, A. M. (2000). From patterns to theories: Conditions for conceptual
change. Journal of Mathematical Behavior, 19, 247271.
Maher, C. A., Martino, A. M., & Alston, A. S. (1993). Childrens construction of mathematical
ideas. In B. Atweh, C. Kanes, M. Carss, & G. Booker (Eds.), Contexts in mathematics education: Proceedings of the 16th annual conference of the Mathematics Education Research Group
of Australia (MERGA), pp. 1339. Brisbane, Australia.
Maher, C. A., & Speiser, R. (1997a). How far can you go with block towers? Proceedings of the
International Group for the Psychology of Mathematics Education (PME 21), Lahti, Finland,
4, 174181.
Maher, C. A., & Speiser, R. (1997b). How far can you go with block towers? Stephanies
intellectual development. Journal of Mathematical Behavior, 16(2), 125132.
218
References
Martino, A. M. (1992). Elementary students construction of mathematical knowledge: Analysis

by profile. Unpublished doctoral dissertation, Rutgers, the State University of New Jersey, New
Brunswick, NJ.
Martino, A. M., & Maher, C. A. (1999). Teacher questioning to promote justification and generalization in mathematics: What research practice has taught us. Journal of Mathematical
Behavior, 18(1), 5378.
Miles, M. B., & Huberman, M. A. (1994). Qualitative data analysis: An expanded sourcebook
(2nd ed.). Thousand Oaks, CA: Sage.
Muter, E. M. (1999). The development of student ideas in combinatorics and proof: A six year
study. Unpublished doctoral dissertation, Rutgers University, Newark, NJ.
National Council of Teachers of Mathematics. (2000). Principles and standards for school
mathematics. Reston, VA: National Council of Teachers of Mathematics.
OBrien, M. (1994). Changing a school mathematics program: A ten-year study. Unpublished
doctoral dissertation, Rutgers, the State University of New Jersey, New Brunswick, NJ.
Papert, S. (1980). Mindstorms: Children, computers and powerful ideas. New York: Basic Books.
Perry, W. G. (1970). Forms of intellectual and ethical development in the college years: A scheme.
New York: Holt, Rinehart and Winston.
Pintrich, P. R., & Hofer, B. K. (1997). The development of epistemological theories: Beliefs about
knowledge and knowing and their relation to learning. Review of Educational Research, 67(1),
88140.
Pirie, S. E. B., & Kieren, T. (1992). Creating constructivist environments and constructing creative
mathematics. Educational Studies in Mathematics, 23, 505528.
Pirie, S. E. B., & Kieren, T. (1994). Growth in mathematical understanding: How can we
characterise it and how can we represent it? Educational Studies in Mathematics, 26(23),
165190.
Powell, A. B. (2006). Socially emergent cognition: Particular outcome of student-to-student
discursive interaction during mathematical problem solving. Horizontes, 24(1), 3342.
Powell, A. B., Francisco, J. M., & Maher, C. A. (2003). An evolving analytical model for
understanding the development of mathematical thinking using videotape data. Journal of
Mathematical Behavior, 22(4), 405435.
Roth, W.-M., & Roychoudhury, A. (1994). Physics students epistemologies and views about
knowing and learning. Journal of Research in Science Education, 31(1), 530.
Sanchez, E., & Sacristan, A. S. (2003). Influential aspects of dynamic geometry activities in the
construction of proofs. Proceedings of the 27th annual conference of the International Group for
the Psychology of Mathematics Education and the 25th annual meeting of the North American
Chapter for the Psychology of Mathematics Education (vol. 4, pp. 119126). Honolulu, HI.
Schoenfeld, A. H. (1987). Whats all the fuss about metacognition. In A. H. Schoenfeld (Ed.),
Cognitive science and mathematics education (pp. 189215). Newark, NJ: Lawrence Earlbaum
Associates, Inc. Hillsdale.
Schoenfeld, A. H. (1989). Explorations of students mathematical beliefs and behavior. Journal of
Research in Mathematics Education, 20(4), 338355.
Schommer, M. (2002). An evolving theoretical framework for an epistemological belief system.
In B. K. Hofer & P. R. Pintrich (Eds.), Personal epistemology: The psychology of beliefs about
knowledge and knowing. Mahwah, NJ: Lawrence Erlbaum Associates.
Schommer, M., & Walker, K. (1995). Are epistemological beliefs similar across domains? Journal
of Educational Psychology, 87(3), 424432.
Seeger, F. (2002). Research on discourse in the mathematics classroom: A commentary.
Educational Studies in Mathematics, 49(13), 287297.
Sfard, A. (1998). Two metaphors for learning and the dangers of choosing just one. Educational
Researcher, 27(2), 413.
Sfard, A. (2001). Learning mathematics as developing a discourse. In R. Speiser, C. Maher, &
C. Walter (Eds.), Proceedings of 21st conference of PME-NA (pp. 2344). Columbus, OH:
Clearing House for Science, Mathematics, and Environmental Education.
References
219
Speiser, R. (1997). Block towers and binomials. Journal of Mathematical Behavior, 16(2),
113124.
Sran, M. K. (2010). Tracing Milins development of inductive reasoning: A case study.
Unpublished doctoral dissertation, Rutgers University, Newark, NJ.
Sternberg, R. (1989). Domain-generality versus domain-specificity: The life and impending death
of a false dichotomy. Merrill-Palmer Quarterly, 35(1), 115130.
Steffero, M. (2010). Tracing beliefs and behaviors of a participant in a longitudinal study for
the development of mathematical ideas and reasoning: A case study. Unpublished doctoral
dissertation, Rutgers University, Newark, NJ.
Tall, D. (1984). Communications: Research problems in mathematics educationIII. For the
Learning of Mathematics, 4(3), 2229.
Tarlow, L. D. (2004). Tracing students development of ideas in combinatorics and proof.
Unpublished doctoral dissertation, Rutgers, The State University of New Jersey, New
Brunswick, NJ.
Torkildsen, O. (2006). Mathematical archaeology on pupils mathematical texts. Un-earthing of
mathematical structures. Unpublished doctoral dissertation, Oslo University, Oslo.
Uptegrove, E. B. (2004). To symbols from meaning: Students investigations in counting.
Unpublished doctoral dissertation, Rutgers, The State University of New Jersey, New
Brunswick, NJ.
Weil, A. (1984). Number theory, an approach through history. Boston, MA: Birkhauser.
Wheeler, D. (1984). Communications: Research problems in mathematics educationI (The
letter). For the Learning of Mathematics, 4(1), 4047.
Wilson, S. (1977). The use of ethnographic techniques in educational research. Review of
Educational Research, 47(1), 245265.
Yackel, E., & Cobb, P. (1996). Sociomathematical norms, argumentation, and autonomy in
mathematics. Journal for Research in Mathematics Education, 27(4), 458477.
Index
A
Abstraction, 10, 17, 24, 77, 85
Addition rule, 13, 73, 77, 82, 111114,
124125, 129130, 133, 136138, 141,
152, 181
Addition rule of Pascals Triangle, 181
Adults, 171183
Ali, 99100
Alston, Alice S., 56, 35, 45, 67, 97,
121, 182
Amy-Lynn, 59, 6263, 6667, 121122,
124126
Angela, 9798, 103, 121, 127129
Angelas Law of Towers, 98
Ankur, 59, 6264, 6671, 8990, 9294,
105107, 109119, 133135, 138140,
142, 157, 159160, 162, 164165,
167, 185
Ankurs Challenge, 8995, 110, 183, 185,
199200, 210
Another Pizza Problem, 70, 209
Argument, 13, 27, 31, 3334, 35, 37, 4043,
4557, 61, 6667, 72, 75, 8995,
97, 102103, 175176, 185, 191,
195, 200
B
Beliefs, 157159, 168169
Binary notation, 106, 108, 116, 118120, 135,
141, 143
Binary numbers, 106, 137
Binomial, 12, 14, 73, 7677, 104105,
108112, 114, 116, 121131, 134136,
143, 152, 209210
Binomial expansion, 12, 14, 108112, 114,
116, 134136, 210
Bobby (Robert), 45, 59, 62
Branches, 103, 201
Brian, 59, 6263, 66, 6871, 8992, 94,

105108, 110, 119, 133135, 138139,
142, 145150, 153, 157, 159, 161, 163,
166167
C
Cases, 12, 14, 3334, 3637, 3943, 4546,
49, 51, 56, 62, 66, 68, 74, 77, 85, 90, 94,
97100, 104, 110, 119120, 127128,
130, 143, 145, 153, 158, 174175, 177,
182, 185192, 194195, 200
Choices, 1112, 5960, 6971, 103, 107108,
116, 122, 124, 127128, 130, 143, 151,
177, 196
Choose, 12, 43, 59, 98, 107, 123, 127,
135140, 142
Choose notation, 135136, 140
Claims, 145153, 164, 168, 202
Coefficient, 77, 106, 135
Collegiate, 183
Combinations, 13, 1923, 25, 2830, 33,
3638, 40, 4243, 4546, 6067,
6971, 77, 79, 82, 85, 8990, 92, 94,
101103, 106, 110, 122, 124, 127128,
130, 137, 172, 175179, 182, 189, 192,
195196, 200
Combinatorial reasoning, 73, 202
Combinatorics, 3, 6, 911, 14, 1718, 2425,
73, 110, 112114, 120, 122, 131, 133,
140141, 144146, 169, 198, 201
Combinatorics problems, 11, 14, 1718,
112113, 131, 133, 144
Conditions, 3, 28, 142, 158, 168169, 171,
183, 190191
Conjecture, 99, 147
Connections, 34, 9, 18, 72, 74, 82, 105120,
130131, 141, 143, 171172,
178182
Consumer math, 173
221
222
Contemporary Mathematics, 172173
Contradiction, 14, 31, 40, 6667, 97, 99
Controlling for variables, 14, 35, 43,
97, 99
Convincing arguments, 4, 104
Counting I, 209
Counting II, 209
Counting methods, 172173
Cousin, 3133, 36, 42, 46
Critical events, 9
Cups, Bowls, and Plates, 205206
D
Dana, 11, 17, 19, 2125, 2732, 43, 60
Danielle, 171, 174
Discourse, 4, 74, 142, 146, 150, 159,
164165, 202
Discursive, 150151, 164, 167, 201
Donna, 171, 186187
Doubling rule, 3940, 42, 49, 56,
106, 130
Duplicates, 19, 2930, 33, 36, 56, 61, 72,
7882, 100, 122, 127, 197
Durability of ideas, 141
Dyadic choice, 151
E
Educational Studies in Mathematics, 202
Elevator, 31, 34, 37
Empirical, 82, 152
Epistemological beliefs, 157159, 168
Errol, 171, 175176, 192, 199
Euclidean geometry, 150
Explanation, 21, 40, 5155, 67, 70, 74, 9295,
97, 101, 106, 109, 112, 115117, 119,
123, 126, 135136, 141, 172175,
180181, 197, 200
F
Factorials, 122
Family tree, 5253
Fermat, 77
Fermats Recursion, 7778
Final Pizza Problem, 70, 209
Five-tall towers with exactly two red cubes, 33,
109, 119
Five-tall towers selecting from two colors, 31
Five-topping pizza problem, 13, 106, 112, 135,
174, 181
Forms of reasoning, 6, 11, 14, 39, 7273, 185,
200201
Formula, 7778, 82, 103104, 108109, 122,
125, 128, 130, 141, 143, 178, 199
For the Learning of Mathematics, 201
Index
Four-tall towers selecting from 2 colors, 28,
196197
Four-tall towers selecting from 3 colors, 195
Four-topping pizza problem, 6870, 123, 127,
174, 181, 208
Francisco, John M., 157169
G
Gang of Four, 38, 4243, 51, 75
Generalization, 3, 17, 91, 94, 98, 106, 126
Generalize, 3, 10, 18, 41, 90, 95, 104105,
133, 147, 151, 172
General rule, 9798, 101, 124, 135, 141, 143
Geometric, 108109
Glass, Barbara, 171183, 185200
Grade 1, 7
Grade 2, 2022, 24, 73
Grade 3, 3, 7, 2223, 28, 74
Grade 4, 56, 31, 3342, 77
Grade 5, 45, 59, 77
Grade 6, 7
Grade 7, 75
Grade 8, 33, 73, 75, 85
Group work, 25, 31, 36, 70, 103, 121, 162,
165, 177
Guess My Tower, 42, 45, 50
H
Harding School, 6
Heuristic, 1113, 25, 27, 43, 59, 67, 147,
201, 203
High school, 48, 14, 89, 97, 105, 133134,
142, 145146, 157158, 161, 169,
172173, 182183, 185186, 200
Hilbert, David, 201
I
Inductive argument, 4243, 4557, 67,
97, 102103, 176177, 182, 185,
191195, 199
Inductive reasoning, 14, 3839, 5056, 97,
100, 195
Interlocution, 151
Isomorphic relationship, 108, 118, 120
Isomorphism, 4, 14, 77, 109, 126, 130, 133,
147, 151153, 179180
J
Jaime, 17, 23
Jeff C., 171, 174, 176177, 179180
Junior year, 134
Justification, 1112, 14, 24, 34, 39, 42, 46,
59, 67, 90, 94, 97, 100, 104, 109110,
118119, 126, 147, 168, 173, 175177,
185193, 196, 199200, 203
Index
Justification by cases, 97, 104, 119, 177, 185,
192, 199
Justify, 3, 12, 19, 39, 4243, 49, 61, 6566,
73, 92, 130131, 146153, 168, 172,
175176, 183, 194, 200, 203
K
Kenilworth, 47, 12, 59, 6970, 161, 182
Kiczek, Regina, 121, 133
Knowing deeply, 203
L
Learning environment, 131
Lecture classes, 118120
Letter codes, 106, 122, 142
Liberal-arts, 171173, 185
Linda, 171
Lisa, 171, 175, 180
Logic, 39, 172
Logical reasoning, 10, 17
Longitudinal study, 38, 1011, 14, 18,
24, 27, 75, 86, 89, 9899, 105, 130,
157161, 167169, 172, 177, 183,
185200, 203
M
Magda, 9799, 103, 121, 127130
Maher, Carolyn A., 4, 4557, 5972, 8995,
105, 121, 133
Making sense, 131, 134, 141, 145
Martino, Amy M., 4, 9, 1721, 2728, 33, 35,
39, 41, 47, 49, 7475, 178, 180, 182
Mary, 171, 175, 195, 197198, 200
Mathematical beliefs, 157, 159, 168169
Mathematical concepts, 6, 119, 161, 172173
Mathematical ideas, 45, 911, 1719, 24, 57,
7475, 85, 95, 130131, 133, 142, 145,
158, 162, 172, 201203
Mathematical structure, 3, 147
Melinda, 171, 178, 180
Metaphor, 13, 85
Michael, 17, 1921, 23, 35
Michelle, 27, 38, 41, 45, 5054, 56, 5960, 62,
64, 97, 99103, 121122, 127
Michelle I., 27, 45, 5054, 56, 59, 6465
Michelle R., 45, 50, 5253, 56, 59, 62
Mike, 59, 6264, 6668, 7071, 8990, 9294,
105110, 114120, 133142, 145152,
157, 159160, 162, 165167, 171, 177,
181182
Milin, 25, 27, 3541, 4552, 5556, 5961,
6465, 71, 77
Muter, Ethel M., 59, 8995, 105120
223
N
National Council of Teachers of Mathematics
(NCTM), 5, 17, 144, 168
National Science Foundation, 5
Night session, 112, 116, 118, 133,
135143, 181
Notation, 1314, 1819, 2425, 27, 42, 5961,
65, 67, 7274, 77, 103, 105108, 110,
112, 116, 133142, 201
n-Tall tower, 12, 27
N-tall towers selecting from r-colors,
42, 77
O
Opposite, 29, 31, 3536, 42, 46, 54, 67,
99100, 150, 175
Order, 6, 12, 24, 70, 75, 8990, 106107, 113,
118, 124, 127128, 130, 133, 140,
143144, 153, 173, 175, 182, 189,
191, 202
Outfit, 11, 1921, 2324, 43
Over time, 73, 75, 160, 162, 168, 171,
201, 203
P
Pantozzi, Ralph, 121, 161162, 166167
Papert, 7475
Partial cases, 42
Pascal, 211212
Pascals Identity, 1314, 105, 111115, 118,
120121, 124125, 127130
Pascals Triangle, 1214, 7273, 77, 8183, 86,
104105, 110117, 120121, 123126,
128131, 133138, 140141, 143144,
147, 151153, 160, 181182
Pattern, 6, 10, 18, 24, 2728, 31, 3334, 3743,
46, 5054, 7374, 77, 8182, 85, 90,
99, 101, 104, 106, 119, 123, 128, 130,
147, 174175, 181183, 190
Penny, 171, 177, 195, 199
Permutations, 172, 178, 182, 192
Pirie, Susan, 97, 158, 203
Pizza
with halves problem, 1213, 6769, 174
plain, 12, 59, 106107, 113, 117118,
122123, 125, 127, 141, 208
problem, 1213, 59, 6870, 7273,
104107, 111113, 116123, 125, 127,
129130, 135136, 141, 145, 169,
172174, 177178, 180181, 183
Powell, Arthur B., 9, 145154
Power of 2, 122, 124
Probability, 165, 172
224
Problem of Points, 211
Proof
by cases, 41, 72, 119, 130, 175, 191, 194
by contradiction, 99
R
Reasoning, 34, 6, 911, 14, 17, 2425, 27, 31,
3839, 42, 45, 50, 52, 5455, 5972,
7576, 84, 9495, 97, 100, 112, 145,
147153, 158, 168, 171183, 185, 188,
191, 195, 201203
by contradiction, 97
Recursive argument, 72, 199
Relay Race, 206
Representation, 34, 1014, 1725, 27, 46, 51,
59, 62, 64, 73, 76, 9495, 97, 105120,
133144, 168, 172, 203
Rica, Fred, 45
Rob, 171, 175, 177178, 181182, 187
Robert, 121122, 124127
S
Samantha, 171, 179
Scheme, 21, 24, 94, 105106, 109, 174
Second International Congress of
Mathematicians, 201
Senior year, 145, 200
Sense making, 6, 1112, 24, 159162, 200,
203204
Shelly, 121127
Sherly, 99100, 121, 127129
Shirts and jeans, 11, 1724, 27, 29, 5960,
64, 73
Shirts and jeans extended, 205
Socially emergent cognition, 150
Sociomathematical norms, 153
Sophomore year, 119, 134135, 140, 143
Specialize, 147148, 151
Speiser, Robert, 33, 7386, 105112, 201
Sran, Manjit K., 2743, 4557, 5972
Index
Staircase, 33, 40, 42, 174175
Standard notation, 133144
Statistics, 172, 178
Stephanie, 11, 17, 1925, 2835, 4042,
4562, 71, 7385, 121127, 171, 178,
180, 182
Stephanie C., 171, 178, 180
Steve, 171, 175, 181
Strategy, 13, 21, 23, 29, 31, 3536, 43, 45, 51,
67, 69, 72, 92, 94, 9899, 100, 119,
147, 153, 177, 182, 192, 194, 197, 200
T
Tarlow, Lynn D., 59, 97104, 121131
Taxicab geometry, 150, 202
Taxicab problem, 14, 144147, 150153, 212
Theory, 51, 74, 102103, 175, 202
Tim, 171, 177, 190
Tower
binomial relationship, 209
families, 3132, 34
Tracy, 171, 178
Tree diagrams, 122, 127
Trial and error, 29, 31, 3542, 51
U
Understanding, 45, 9, 12, 18, 21, 24, 33,
41, 43, 4546, 5154, 57, 7376, 95,
101, 105106, 109, 121, 134, 140, 158,
160164, 180, 201, 203
Unifix cubes, 12, 27, 42, 50, 97, 126
Uptegrove, Elizabeth B., 914, 97104,
105120, 133144
W
Wesley, 171, 175176, 180
Y
Yankelewitz, Dina, 1725, 2743, 4557,
5972

Carolyn A. Maher, Arthur B. Powell Combinationatoric and Reasoning

Caricato da

Informazioni sul documento

Titolo originale

Copyright

Formati disponibili

Condividi questo documento

Condividi o incorpora il documento

Opzioni di condivisione

Hai trovato utile questo documento?

Questo contenuto è inappropriato?

Copyright:

Formati disponibili

Carolyn A. Maher, Arthur B. Powell Combinationatoric and Reasoning

Caricato da

Copyright:

Formati disponibili

Combinatorics

Mathematics Education Library

For further volumes:

Carolyn A. Maher Arthur B. Powell

Dr. Arthur B. Powell

Dr. Elizabeth B. Uptegrove

This book is dedicated to the Kenilworth students

Our research project on mathematical learning focuses on the accomplishments of

Introduction, Background, and Methodology

1 The Longitudinal Study . . . . . . . . . . . . . . . . . . . . . . . .

Foundations of Proof Building (19891996)

3 Representations as Tools for Building Arguments . . . . . . . . . .

4 Towers: Schemes, Strategies, and Arguments . . . . . . . . . . . .

5 Building an Inductive Argument . . . . . . . . . . . . . . . . . . .

6 Making Pizzas: Reasoning by Cases and by Recursion . . . . . . .

7 Block Towers: From Concrete Objects to Conceptual Imagination

Part III Making Connections, Extending, and Generalizing

Representations and Connections . . . . . . . . . . . . . . . . . . .

Pizzas, Towers, and Binomials . . . . . . . . . . . . . . . . . . . . .

Representations and Standard Notation . . . . . . . . . . . . . . .

So Lets Prove It! . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Extending the Study, Conclusions, and Implications

Doing Mathematics from the Learners Perspectives . . . . . . .

Adults Reasoning Combinatorially . . . . . . . . . . . . . . . . . .

Comparing the Problem Solving of College Students with

Appendix A Combinatorics Problems . . . . . . . . . . . . . . . . . . .

Appendix B Counting and Combinatorics Dissertations

Carolyn A. Maher, Arthur B. Powell, and Elizabeth B. Uptegrove

C.A. Maher (B)

a rectangular grid.) They recognized the isomorphism, used it to make conjectures

The Longitudinal Study

1.1 Theoretical View

C.A. Maher (B)

C.A. Maher et al. (eds.), Combinatorics and Reasoning, Mathematics

1.2 Background of the Study

The Longitudinal Study

Schools and Rutgers University. It should be noted that the RutgersKenilworth

1.2.1 Teacher Development Component

1.2.2 Intervention Design

1.3 Longitudinal Study: Grades 13

The Longitudinal Study

1.4 Longitudinal Study, Grades 48

1.5 Longitudinal Study: High School Years

1.6 Longitudinal Study: Beyond High School

C.A. Maher (B)

C.A. Maher et al. (eds.), Combinatorics and Reasoning, Mathematics

C.A. Maher and E.B. Uptegrove

2.2 Theoretical Perspectives

2.3 Selected Problems

sense making, argumentation, and justification in mathematics. As Yackel and Cobb

2.3.1 Shirts and Jeans

C.A. Maher and E.B. Uptegrove

The definition of a tower is an ordered sequence of Unifix cubes, snapped

A detailed description of the students work on Pascals identity is given in

C.A. Maher and E.B. Uptegrove

2.4 Concluding Remarks

Foundations of Proof Building (19891996)

Representations as Tools for Building

Date and Grade:

May 30, 1990 and October 11, 1990; Grades 2 and 3

C.A. Maher (B)

C.A. Maher et al. (eds.), Combinatorics and Reasoning, Mathematics

C.A. Maher and D. Yankelewitz