ANDREAS J.

STYLIANIDES

THE NOTION OF PROOF IN THE CONTEXT OF ELEMENTARY

SCHOOL MATHEMATICS

ABSTRACT. Despite increased appreciation of the role of proof in students’ mathematical

experiences across all grades, little research has focused on the issue of understanding and
characterizing the notion of proof at the elementary school level. This paper takes a step
toward addressing this limitation, by examining the characteristics of four major features
of any given argument – foundation, formulation, representation, and social dimension –
so that the argument could count as proof at the elementary school level. My examination
is situated in an episode from a third-grade class, which presents a student’s argument that
could potentially count as proof. In order to examine the extent to which this argument
could count as proof (given its four major elements), I develop and use a theoretical frame-
work that is comprised of two principles for conceptualizing the notion of proof in school
mathematics: (1) The intellectual-honesty principle, which states that the notion of proof in
school mathematics should be conceptualized so that it is, at once, honest to mathematics
as a discipline and honoring of students as mathematical learners; and (2) The continuum
principle, which states that there should be continuity in how the notion of proof is con-
ceptualized in different grade levels so that students’ experiences with proof in school have
coherence. The two principles offer the basis for certain judgments about whether the par-
ticular argument in the episode could count as proof. Also, they support more broadly ideas
for a possible conceptualization of the notion of proof in the elementary grades.

KEY WORDS: elementary school mathematics, mathematical argument, mathematical rea-

soning, proof, social dimension of proof

1. INTRODUCTION

Proof is fundamental to doing and knowing mathematics – it is the basis

of mathematical understanding and essential for developing, establishing,
and communicating mathematical knowledge (Kitcher, 1984; Polya, 1981).
The centrality of proof in mathematical practice has led several researchers
in many countries to emphasize the importance of proof in school mathe-
matics (e.g., Balacheff, 1988; Ball et al., 2002; Healy and Hoyles, 2000;
Mariotti, 2000). In North America in particular, the current research and
policy discourse is not simply tilted in favor of a widespread presence of
proof in school mathematics but it recommends also that proof be incorpo-
rated into the mathematical experiences of even elementary students (e.g.,
Ball and Bass, 2003; NCTM, 2000; Schoenfeld, 1994; Yackel and Hanna,
2003).1

Educational Studies in Mathematics (2007) 65: 1–20

DOI: 10.1007/s10649-006-9038-0 
C Springer 2006
2 A. J. STYLIANIDES

The idea of incorporating proof into students’ mathematical experiences

from the beginning of their schooling raises the important issue of what
could count as proof in the early grades. This issue is particularly relevant to
the current educational context of North America, but its investigation can
have implications on the teaching and learning of proof in other contexts
as well. Specifically, if research develops a conceptualization of proof in
the early grades that is both appropriate for young students and compatible
with how proof is typically understood in the upper grades, there will be a
solid foundation on which mathematics educators can develop a cohesive
program for proof instruction across all grades. The development of such a
program is important, because, nowadays, much elementary mathematics
teaching in many countries “focuses on arithmetic concepts, calculations,
and algorithms, and, then, as they enter secondary school, pupils are sud-
denly required to understand and write proofs, mostly in geometry” (Ball
et al., 2002, pp. 907–908). In other words, students’ transition from el-
ementary to secondary school mathematics is abrupt and is linked to a
‘didactical break’ represented by the introduction of the new requirement
for proof (Balacheff, 1988; Sowder and Harel, 1998). Several researchers
(Mariotti, 2000; Moore, 1994; Sowder and Harel, 1998; Usiskin, 1987)
have identified this abrupt introduction of students to proof at secondary
school as a possible explanation for the many difficulties that secondary
school students face with proof (e.g., Coe and Ruthven, 1994; Fischbein,
1982; Healy and Hoyles, 2000).
Despite the importance of research on conceptualizing the notion of
proof in the early grades, research on proof in these grades has paid little
attention to this topic. The present paper takes a step toward addressing
this limitation, with particular attention to the issue of understanding and
characterizing the notion of proof in the context of elementary school math-
ematics. Specifically, the paper aims to provide insight into the features that
a mathematical argument (or simply argument) – a connected sequence of
assertions intended to verify or refute a mathematical claim – needs to have
in order to count as proof in this context. To promote this goal, I analyze
an episode from a third-grade class that offers a rich setting for investi-
gating four major elements of an argument that are important to consider
when deciding whether an argument qualifies as proof. The four elements
are the argument’s foundation (i.e., what constitutes its basis: definitions,
axioms, etc.), formulation (i.e., how it is developed: as a logical deduction,
as a generalization from particular cases, etc.), representation (i.e., how
it is expressed: using everyday language, algebraically, etc.), and social
dimension (i.e., how it plays out in the social context of the community
wherein it is created). The selection of these four elements of an argument
finds support from how arguments and proofs are generally conceptualized
 PROOF IN ELEMENTARY SCHOOL MATHEMATICS 3

in the discipline of mathematics: Arguments and proofs generally build on

accepted statements such as definitions and axioms (cf. foundation), use
accepted canons of correct inference such as logical equivalencies (cf. for-
mulation), and are expressed using appropriate mathematical language (cf.
representation) (e.g., Kitcher, 1984). Furthermore, the process of accepting
an argument as proof relies heavily on the social mechanisms of the math-
ematical community (cf. social dimension) (e.g., Ernest, 1998; Tymoczko,
1986/1998).

2. A THEORETICAL FRAMEWORK FOR CONCEPTUALIZING THE NOTION

OF PROOF IN SCHOOL MATHEMATICS

In this section, I propose a theoretical framework that guides my analysis

of the episode from the third-grade class. The framework is comprised of
two guiding principles for conceptualizing the notion of proof in school
mathematics: the intellectual-honesty principle and the continuum princi-
ple. These two principles can be useful in conceptualizing other concepts in
school mathematics and also more broadly (e.g., the concept of ‘modeling’
in science).
The intellectual-honesty principle states that the notion of proof in
school mathematics should be conceptualized so that it is, at once, honest
to mathematics as a discipline and honoring of students as mathematical
learners. Put differently, the intellectual-honesty principle approaches the
problem of conceptualizing proof in school mathematics as an optimization
problem whose primary restriction is serious consideration of both the nor-
mative aspects of proof in mathematics (e.g., Kitcher, 1984; Polya, 1981;
Tymoczko, 1986/1998) and what is known or conceptually accessible to
the learners under teacher guidance or collaboration with peers (e.g., Ball
and Bass, 2003; Lampert, 1990; Maher and Martino, 1996; Yackel and
Cobb, 1996). The intellectual-honesty principle is inspired by the work of
Bruner (1960) who was one of the first to argue for the possibility of achiev-
ing a defensible balance between disciplinary and learner considerations,
hypothesizing also that “any subject can be taught effectively in some in-
tellectually honest form to any child at any stage of development” (p. 33).
Since then several mathematics educators and other educators have exem-
plified both the possibility and the merit of creating classroom environments
committed to honoring both disciplinary and learner considerations (e.g.,
Ball, 1993; Ball and Wilson, 1996; Lampert, 1992).
The continuum principle states that there should be continuity in how
the notion of proof is conceptualized in different grade levels so that stu-
dents’ experiences with proof in school have coherence. According to this
4 A. J. STYLIANIDES

principle, a conceptualization of proof in school should address the dis-

continuity problem that currently characterizes students’ experiences with
proof in school: Most students first – and sometimes only – encounter proof
in high school courses on Euclidean geometry, and, when this happens,
proof seems alien and unfamiliar to them, rather than being a natural ex-
pansion of their previous mathematical experiences (e.g., Ball et al., 2002;
Moore, 1994; Sowder and Harel, 1998; Usiskin, 1987). Also, it is imper-
ative that students engage with a consistent meaning of proof throughout
the grades: Students should not develop a conception of proof in the el-
ementary school that has to be undone or unlearned in high school (e.g.,
Martin and Harel, 1989).
The intellectual-honesty and continuum principles offer insight into
what arguments might qualify, or not qualify, as proofs in the elementary
school. For example, an empirical argument (Balacheff, 1988; Sowder and
Harel, 1998) cannot qualify as proof in the elementary grades, because, if
it did, both principles would be violated. The intellectual-honesty principle
would be violated because: (1) in mathematics, empirical evidence does
not constitute a proof, and (2) psychological findings and existing evidence
from elementary school settings show that young children can engage suc-
cessfully in deductive reasoning and proof (e.g., Carpenter et al., 2003;
Galotti et al., 1997; Light et al., 1989; Maher and Martino, 1996). The con-
tinuum principle would be violated because, “[i]f [elementary] teachers
lead their students to believe that a few well-chosen examples constitute a
proof, it is natural to expect that the idea of proof in high school geometry
and other courses will be difficult for the students” (Martin and Harel, 1989,
pp. 41–42).2 In analyzing the episode from the focal third-grade class, I
use the two principles to gain further insight into the notion of proof in
elementary school.

3. DATA AND ANALYTIC METHOD

3.1. Data
The episode for the paper is derived from a database of the Mathematics
Teaching and Learning to Teach Project at the University of Michigan
that documents an entire. The records collected across the year include
videotapes of the lessons, fieldnotes, student work, and the teacher’s journal
with her lesson plans and teaching reflections.
The class was socioeconomically, ethnically, and racially diverse with
22 students of multiple ability levels. The mathematics period in the class
was approximately one hour long, five days per week. During each period,
the class worked on one or two problems, carefully selected by the teacher to
 PROOF IN ELEMENTARY SCHOOL MATHEMATICS 5

be mathematically rich. The class often began with the students exploring
a problem individually or in pairs, then in small groups, and finally in
the whole group. The curriculum was organized around units on general
mathematical topics such as arithmetic and number theory.
One of Ball’s goals was to help her students become skilled mathe-
matical reasoners. To promote this goal, she modeled her “classroom as
a community of mathematical discourse, in which the validity for ideas
rest[ed] on reason and mathematical argument, rather than on the author-
ity of the teacher or the answer key” (Ball, 1993, p. 388). Inspired by
Bruner’s (1960) notion of intellectual honesty, Ball’s teaching was a con-
tinuous struggle to achieve a defensible balance between mathematics as a
discipline and children as mathematical learners:
I must consider the mathematics in relation to the children and the children in
relation to the mathematics. My ears and eyes must search the world around us,
the discipline of mathematics, and the world of the child with both mathematical
and child filters. (Ball, 1993, p. 394)

3.2. Analytic method

The classroom episode I analyze in this paper shows a third grader using
everyday language and definitions of even and odd numbers to create a
deductive argument for the conjecture ‘odd + odd = even.’ Her argument
fails to convince many of her classmates. My analysis focuses on this
argument, examining its four major elements – foundation, formulation,
representation, and social dimension – against the intellectual-honesty and
continuum principles. To what extent are the four elements of the argument
compatible with the two principles? And thus, what can one say about
whether the argument could count as proof? In analyzing the episode, I take
the perspective of the researcher who studies the classroom activity after
the fact and from the outside, trying to gain insight into the general issue
of the notion of proof in the elementary grades through close examination
of a third grader’s argument.

4. P RESENTATION AND ANALYSIS OF THE EPISODE

4.1. Background: A few days before the episode

On January 26, the third graders in Deborah Ball’s class are investigating
what happens when they add any two odd numbers. They check several
examples and formulate – with Ball’s help – the conjecture ‘odd + odd
= even.’ Ball then challenges the students to explain why their conjec-
ture should be accepted. This provokes many different arguments. Jeannie
6 A. J. STYLIANIDES

argues that the class cannot prove the conjecture for all pairs of odd num-
bers, because odd numbers and even numbers “go on for ever” and so one
“cannot prove that all of them work.” But other students disagree with this
argument. For example, Ofala asserts that the conjecture is true because she
verified it in 18 particular cases. The lesson continues with students sharing
their thoughts and the teacher pressing students to justify their thinking.
The next lesson, on January 30, begins with Ball reviewing the issue
raised by Jeannie about why the class could not prove the conjecture. Ball
explains to the students that mathematicians would address this issue by
trying to understand what property of odd numbers makes the combination
of any two of them an even number. She then helps the students remind
themselves of their definitions of even and odd numbers (the definitions
are formulated over the set of non-negative integers):
• Odd numbers: Numbers that if you group them by twos there is one left
over.
• Even numbers: Numbers that if you group them by twos there is nothing
left over.
Ball challenges the students to use their definitions to prove the conjec-
ture. The class meeting ends with students working in their groups on the
conjecture.

4.2. The episode

The class meeting on January 31 opens with Ball eliciting reports from
groups on their work the previous day. Tembe goes first up to the board
and illustrates the conjecture in the particular case 7 + 7 = 14. Then Betsy
says she has a proof for the conjecture and goes up to the board.

1a. Betsy: What we figured out how it’s always true is that we would have
seven dots, or lines plus seven lines [draws fourteen lines on the
board]

. . . and then [counts the lines] . . . we said that we had to circle

them by twos [she starts circling groups of two lines] –

and also we said that . . . just a second [finishes circling groups of

two lines] –
 PROOF IN ELEMENTARY SCHOOL MATHEMATICS 7

1b. [From now on she looks at the class, away from the board, as she
explains.] that if you added another even one to an odd number, or
another one to an odd number, then it would equal an even number,
‘cause all odd numbers if you circle them, what we found out, all
odd numbers if you circle them by twos, there’s one left over, so
if you . . . plus one, um, or if you plus another odd number, then
the two ones left over will group together, and it will make an
even number.

After a few questions from students, Betsy says:

2. Betsy: . . . if you added two odd numbers together you can add the
ones left over and it would always equal an even number.

Ball summarizes the discussion thus far and asks for other reactions from
students.
3. Ball: Tembe gave one example and Betsy tried to show that this would
always be true. . . . Do people think that does prove that an odd
plus an odd would always be even?
4. Riba: No.
5. Ball: [Mei shakes her head]. Why not, Mei?
6. Mei: I don’t think so, because you don’t know about like in the thou-
sands, and you don’t know the numbers, like, you don’t even
know how you pronounce it, or how you say it.
7. Ball: What do other people think? Does the thing that Betsy did, does
that show that any time you have an odd number plus an odd
number, you get an even number? Mei thinks it doesn’t because
you wouldn’t know about big, very big numbers. Jeannie, you
think it does show it?
8. Jeannie: I don’t think so, because she didn’t say it had to be those two
numbers, those two odd numbers, it could be any two odd num-
bers because, um, there’s always one left.
9. Mei: I know, but that is not, like, telling. . . I’m trying to say that that’s
only one example. You can’t really say that it will work for every
odd number, even – I don’t think it would work for numbers that
we can’t say or figure out what they are.
10. Ball: What do other people think about this argument that Jeannie and
Mei are having? Mei’s saying that she is still not convinced that
this could apply for all numbers, big, big numbers, and Jillian
is saying that it would always be true because there’d always be
one left over from each odd number. Riba, what do you think
about that?
11. Riba: How does, um, Jeannie know that it would always work? She
never tried all the numbers.
12. Betsy: Mathematicians can’t even do that. You would die before you
counted every number.
13. Riba: I know, that’s why I mean, that’s what I mean – you don’t know
if it always works.
8 A. J. STYLIANIDES

14. Ball: Anyone else have a comment? Sheena?

15. Sheena: I agree sort of with Riba, but you don’t have to try all the num-
bers, because you would die before you tried all the numbers
because there are some numbers you can’t even pronounce, and
some numbers you don’t even know that they’re there. But, still,
I still think it’s true because, um, we’ve tried like enough of, of
examples already. It proves it’s true.

The discussion continues for a while and then Ball asks students to work
further on the conjecture in their groups. When the class resumes, Ball calls
on Mark to report on his work with Nathan.

16. Mark: Well, first, me and Nathan, we were just getting answers and we
weren’t thinking about proof. And we were still getting answers
and we were thinking about, we were trying to prove, and Betsy
came and she had proved it, and then we all agreed that it would
work.
The class meeting continues with the students working on a different con-
jecture.

4.3. Betsy’s argument

The first part of Betsy’s argument was a follow-up on Tembe’s example;
she showed why ‘14’ is an even number, using the class’s definition of even
numbers (line 1a). Thereafter, Betsy looked at the class and expressed her
argument without any further reference to the figure she had drawn on the
board.
Betsy’s argument is summarized in the following segment:

[A]ll odd numbers if you circle them by twos, there’s one left over, so if you . . .
plus one, um, or if you plus another odd number, then the two ones left over will
group together, and it will make an even number (line 1b).

Betsy’s argument is comprised of two connected assertions: (1) when two

odd numbers are added together, the two ones left over from each odd
number (after the numbers are grouped by twos) can be combined to form
a new group of two (“if you plus another odd number [to an odd number],
then the two ones left over will group together”), and (2) the resulting
number is even (“and it will make an even number”). The first assertion is
based on the class’s definition of odd numbers as the numbers that when
grouped by twos there is one left over (“[A]ll odd numbers if you circle them
by twos, there’s one left over”). The second assertion is based (implicitly)
on the class’s definition of even numbers as the numbers that when grouped
by twos there is nothing left over.
 PROOF IN ELEMENTARY SCHOOL MATHEMATICS 9

Below I examine the three structural elements – foundation, formulation,

and representation – of Betsy’s argument; in the next section I examine the
argument’s social dimension.

4.4. The foundation, formulation, and representation of Betsy’s argument

The definitions of even and odd numbers reviewed by the class on January
30 constituted the foundation of Betsy’s argument. These definitions were
the basis of a connected sequence of assertions that concluded with the
derivation of the statement ‘odd + odd = even.’ Thus, the formulation of
Betsy’s argument was as a logical deduction from a set of definitions. The
logical deduction involved analysis of the properties (definitions) of even
and odd numbers (objects in question) and it rested on operations on these
objects (e.g., grouping by twos). These operations were detached from any
particular case. Betsy’s argument had a verbal representation, using every-
day language. The language used refers to the manipulation of a specific
representation of numbers close to a quantity. This quantity can be made in
groups of two ‘units’ and a person can consider whether one or no ‘unit’ is
left over. Yet, a quantity is not a number (abstraction from the unit). This
raises the subtle issue of expressing operations (grouping) on an abstract
object (number), which was not captured fully by the way Betsy’s argument
was represented. It is important to note, however, that Betsy’s argument was
consistent with the class’s definitions of even and odd numbers.
Given the analysis above of the foundation, formulation, and represen-
tation of Betsy’s argument, one can recognize this argument as an example
of what Balacheff (1988) calls thought experiment:
The thought experiment invokes action by internalizing it and detaching itself from
a particular representation. It is still coloured by an anecdotal temporal develop-
ment, but the operations and foundational relations of the proof are indicated in
some other way than by the result of their use. . . (p. 219)
Also, like it happens with Betsy’s argument, the linguistic construction of
a thought experiment entails the difficulty of “expressing the operations on
an abstract object (a class of objects)” (p. 225). The connection between
Betsy’s argument and Balacheff’s thought experiment shows the value of
Betsy’s argument, because the thought experiment represents the highest
level in Balacheff’s hierarchy of arguments that hold a privileged position in
the cognitive development of proof: naive empiricism, crucial experiment,
generic example, and thought experiment.
In light of the analysis of Betsy’s argument thus far (including its con-
nection to Balacheff’s thought experiment), could it count as proof at the
elementary school level? To address this question, I use the intellectual-
honesty and continuum principles.
10 A. J. STYLIANIDES

Beginning from the intellectual-honesty principle, Betsy’s argument

seems to have achieved a defensible balance between respecting third
graders as mathematical learners and honoring mathematics as a disci-
pline. Regarding the former, the fact that the argument was developed by a
third grader, coupled with the fact that students in other elementary class-
rooms are able to develop similar arguments for the same statement (see,
e.g., Carpenter et al., 2003), offer compelling evidence that the argument is
within third graders’ conceptual reach under teacher guidance. Regarding
the latter, Betsy’s argument is based on an analysis of definitions of even and
odd numbers and is not evidenced on any particular case. Rather, it follows
logically from the definitions and implicates the use of deductive reason-
ing. Thus, Betsy’s argument has some of the basic features that generally
characterize the notion of proof in the discipline as a sequence of assertions
that begins from statements accepted as first principles (definitions and ax-
ioms) and proceeds with logical deductions. Yet, Betsy’s argument is not
exemplar of the proof mathematicians would give for the statement ‘odd
+ odd = even’; Betsy’s argument uses everyday language, which creates
difficulty in expressing operations (grouping) on an abstract object (num-
ber). Important to note, however, is that the intellectual-honesty principle
does not require that students’ proofs be like the proofs of mathematicians;
such a requirement would violate the principle, as it would tilt the balance
between honoring the learner and respecting the discipline towards the side
of the discipline. The general character of Betsy’s argument and its classi-
fication as a thought experiment in Balacheff’s (1988) hierarchy indicates
further its respect to the discipline.
Turning now to the continuum principle, there are some sound corre-
spondences between Betsy’s argument and the algebraic proof often given
for the statement ‘odd + odd = even’ in the upper grades (see Figure 1). The
algebraic proof is essentially the codification in symbols of Betsy’s verbal
descriptions. Therefore, it is reasonable to assume that students who would
learn or produce Betsy’s argument in third grade would be well prepared
to learn or produce the standard algebraic proof in secondary school.
To conclude, the analysis of the foundation, formulation, and repre-
sentation of Betsy’s argument – filtered through the intellectual-honesty
and continuum principles – suggests that her argument has a good chance
to count as proof at the elementary school level. Also, any other thought
experiment (Balacheff, 1988) produced by an elementary student would
presumably satisfy the two principles to the same extent (when examined
across the same elements: foundation, formulation, and representation).
Thus, the analysis identifies a class of arguments of which Betsy’s argu-
ment is a representative – thought experiments – that have high potential to
be called proofs in the elementary grades. This does not suggest however
 PROOF IN ELEMENTARY SCHOOL MATHEMATICS 11

Betsy’s argument (line 1b) Standard algebraic proof

[A]ll odd numbers if you circle them All odd numbers are numbers of the form
by twos, there’s one left over, 2n + 1, where n is a whole number.

so, if you … plus one, um, or if you So, after adding two odd numbers we have:
plus another odd number, then the two (2k + 1) + (2m + 1) = (2k + 2m) + (1 + 1)
ones left over will group together, = 2k + 2m + 2
[or 2 • (k + m + 1)],

and it will make an even number. which is an even number.

[Implicit in the last step is the use of [Implicit in the last step is the use of the
the existing definition of even numbers definition of even numbers as the numbers
as the numbers that when grouped by of the form 2n, where n is a whole number.]
twos there is nothing left over.]

Figure 1. Correspondences between Betsy’s argument and the standard algebraic proof.
(Note: The proofs are formulated over the set of whole numbers.)

that lower-level arguments would necessarily lack this potential, but the
analysis does not illuminate possible boundaries between proofs and argu-
ments that do not qualify as proofs.

4.5. The social dimension of Betsy’s argument

In this section, I examine the social dimension of Betsy’s argument – that
is, how it played out in the context of the classroom community wherein it
was created. I begin by discussing briefly the social dimension of proof in
the discipline of mathematics. Next, I draw on this discussion, and on the
intellectual-honesty and continuum principles, to propose a conceptualiza-
tion of the social dimension of proof in school settings. Finally, I use this
conceptualization to analyze Betsy’s argument, trying to decide whether it
could count as proof.

4.5.1. The social dimension of proof in the discipline of mathematics

There is nowadays increased consensus among philosophers of mathemat-
ics that the process of knowledge construction and validation in mathemat-
ics rests on socially accepted rules of discourse that regulate what count as
acceptable mathematical truths and the methods of deriving new truths from
existing ones (e.g., De Millo et al., 1979/1998; Ernest, 1991; Lakatos, 1976;
Tymoczko, 1986/1998). Within this framework of knowledge construction
and validation in mathematics, proof is described as a social construct. For
example, Ernest (1998) notes that proofs in a mathematical community
12 A. J. STYLIANIDES

result from a social discourse in which participants agree to accept or deny

particular theories; therefore, proofs are outcomes of socially agreed upon
sets of rules and mathematical objects (such as definitions, axioms, and
theorems).
The process of accepting an argument as proof relies heavily on the so-
cial mechanisms of the mathematical community. In particular, it has been
remarked that “[a] proof becomes a proof after the social act of ‘accepting it
as a proof”’ (Manin, 1977; in Hanna, 1983, p. 71), and that proofs become
believable and “increase our confidence in the truth of mathematical state-
ments only after they have been subjected to the social mechanisms of the
mathematical community” (De Millo et al., 1979/1998, p. 275). The accep-
tance of an argument as proof often depends on whether mathematicians are
convinced by it. This is reflected, for example, in Hersh’s (1993) definition
of proof: “proof is convincing argument, as judged by qualified judges”
(p. 389; italics in original). It should be made clear though that conviction
in the mathematical community is not an arbitrary process; instead, it rests
on socially accepted rules of discourse that regulate its function and guar-
antee the quality of the accepted proofs (De Millo et al., 1979/1998).

4.5.2. The social dimension of proof in school settings

Given the discussion of the social dimension of proof in the discipline of
mathematics, how might we conceptualize the social dimension of proof
in school settings? The intellectual-honesty and continuum principles sup-
port specific ideas that, collectively, give rise to a possible conceptual-
ization. The intellectual – honesty principle, which calls for a defensi-
ble balance between mathematics as a discipline and school mathematics,
supports the following idea: An argument that could count as proof in a
classroom community should be accepted as proof by the community –
and, thus, it should be convincing to the students – on the basis of so-
cially accepted rules of discourse relevant to proving that are compatible
with those of wider society. The continuum principle, which calls for co-
herence in how proof is conceptualized along the trajectory of students’
schooling, supports a complementary idea: The rules of discourse relevant
to proving that are accepted by students should evolve with the devel-
opment of students’ knowledge across grades. Below I elaborate on this
conceptualization.
The conceptualization denotes that the convincing power of an argu-
ment is by itself not enough to capture the social dimension of proof in
school settings (and, by implication, it is insufficient to characterize what
could count as proof in these settings). If convincingness in school settings
is dissociated from certain rules of discourse that can regulate its function,
 PROOF IN ELEMENTARY SCHOOL MATHEMATICS 13

the process of proof acceptance may reduce to practices that have nothing
to do with authentic ways of thinking and operating mathematically: stu-
dents voting on whether given arguments should count as proofs, etc. Also,
unqualified arguments (such as empirical arguments) may be elevated to
the status of proof.
The limitations of equating proof acceptance with the convincing power
of an argument highlight the significant role that socially accepted rules
of discourse can play within a classroom community. If students accept
valid rules of discourse about what constitute accepted truths and what
methods can be used to derive new truths from existing ones, these rules
can serve as mechanisms for winnowing out the arbitrariness of conviction.
Examples of such rules of discourse include: agreement on definitions of
concepts relevant to a proving task, and shared understanding of what is
a legitimate way to use definitions in the development of arguments. The
rules of discourse should not be given to the students by use of authority;
instead, they should be interactively constituted by the teacher and the
students in the course of classroom activity (e.g., Balacheff, 1990; Lampert,
1990; Mariotti, 2000; Yackel and Cobb, 1996). Also, the rules of discourse
should evolve with the development of students’ knowledge across grade
levels, having as social reference the rules of the mathematical community
(Balacheff, 1991).
The teacher has an important role to play in the negotiation, acceptance,
and development of rules of discourse in a classroom community (Balach-
eff, 1990, 1991; Ball and Bass, 2003; Lampert, 1990; Mariotti, 2000; Simon
and Blume, 1996; Yackel and Cobb, 1996). According to Lampert (1990),
although teachers and students negotiate together what counts as knowl-
edge and the processes whereby knowledge is acquired, the teachers are in
a position of social intellectual authority with more power than the students
over the results of these negotiations. The study of Yackel and Cobb (1996)
of how a teacher and her second graders had continually regenerated and
modified, through their ongoing interactions, norms about what counted as
an acceptable mathematical reason highlights further the significant role
of the teacher as the representative of the mathematical community in the
classroom. That is, the teacher has responsibility to help students construct
mathematical ways of knowing that are compatible with those of wider
society. Nevertheless, the teacher can face several challenges in trying to
promote this goal. In his study on the benefits and limits of social interac-
tion in the domain of proof, Balacheff (1991) shows that a social situation
that fosters students’ commitment to developing a common argument for
a given proving task may favor the development of arguments and inter-
actional behaviors that deviate from desirable arguments and behaviors
relevant to the notion of proof.
14 A. J. STYLIANIDES

4.5.3. Betsy’s argument in the social context of the classroom community

Betsy’s argument provoked dispute in the classroom community. On the
one hand, some students were convinced by Betsy’s argument and helped
Betsy explain her argument to the class. For example, Jeannie clarified
the point that Betsy’s argument was about any two odd numbers: “[Betsy]
didn’t say it had to be those two numbers [on the board], it could be any two
odd numbers because, um, there’s always one left” (line 8). Also, the shift
recorded in Mark and Nathan’s proving behavior – from checking particular
cases to thinking about proof (line 16) – is an example of assimilation of
Betsy’s ideas after some time and struggle.
On the other hand, some students objected to Betsy’s argument and ei-
ther tried to show that it was invalid (Mei and Riba) or to replace it with
a different argument (Sheena). Beginning with Mei, there are two possible
reasons for her objection to Betsy’s argument. First, Mei appeared to think
that Betsy’s argument was about the specific example on the board (line 1a).
Mei noted: “I’m trying to say that that’s only one example. You can’t really
say that it will work for every odd number” (line 9). Second, Mei seemed
to have difficulty understanding how the idea of ‘grouping by twos’ could
apply to numbers one “[didn’t] even know how [to] pronounce” (line 6) or
“figure out what they [were]” (line 9). Riba was not convinced by Betsy’s ar-
gument either. She believed that, to prove that the statement “would always
work,” one had to check all possible cases; Betsy “never tried all the num-
bers” (line 11). Given that mathematicians “would die” before checking all
possible cases (lines 12–13), Riba concluded that one could not “know if
[the conjecture] always work[ed]” (line 13). Sheena was the third student
in the episode who was not convinced by Betsy’s argument. Unlike Mei
and Riba, though, Sheena did not claim that Betsy’s argument was invalid.
Rather, she proposed an empirical argument in the place of Betsy’s deduc-
tive argument: She referred to the statement mentioned earlier in the episode
that one could not check all possible cases (lines 6, 9, 11–13), but she did
not consider this to be a problem; according to her, “[one didn’t] have to try
all the numbers” (line 15). Thus, she argued that the conjecture was “true
because, um, [the class had] tried enough of, of examples already” (line 15).
I turn now to discuss some points related to how Betsy’s argument un-
folded in the social context of the class. The students seemed to share
important rules of discourse relevant to proving: They recognized the need
for convincing their peers based on mathematical reason, and they scru-
tinized the arguments proposed by using, in the most part, mathematical
rather than other kinds of criteria. Also, the students seemed to share a
common understanding of the definitions of even and odd numbers and of
the fact that one cannot enumerate all elements of an infinite set. Despite
these rules of discourse that the students seemed to share, Betsy’s argument,
 PROOF IN ELEMENTARY SCHOOL MATHEMATICS 15

which was a notable sample of student reasoning from the point of view of
an outside observer, did not convince many students in the class. This was
primarily because different students seemed to use different (sometimes
conflicting) rules of discourse about methods for deriving new truths from
existing ones. Betsy, for example, manipulated the definitions of even and
odd numbers to deduce logically the desired statement. Mei and Riba had
difficulty understanding (or accepting) Betsy’s method of logical deduction
from definition. Sheena had a different method for deriving the desired
statement: empirical verification. Finally, students’ engagement with the
conjecture ended without the students reaching an agreement on a proof
and without the teacher ratifying a purported proof or explaining why other
arguments could not count as proofs. Thus, the episode ended without a
situation for institutionalization (Brousseau, 1981), that is, a situation that
would “aim at pointing out, and giving an official status to, some piece of
knowledge that has been constructed during the classroom activity” (Bal-
acheff, 1990, p. 260).3
Given the above discussion of how Betsy’s argument unfolded in the
social context of the class, and also the earlier discussion about the social
dimension of proof in school settings, could a researcher who studies the
classroom activity from the outside consider Betsy’s argument a proof?
The conceptualization of the social dimension of proof in school settings
that I proposed earlier denotes that an argument that could count as proof
in a classroom community should be accepted as proof by the community
– and, thus, it should be convincing to the students – on the basis of socially
accepted rules of discourse that are compatible with those of wider society.
Betsy’s argument was not accepted as proof by the classroom community.
Many students were not convinced by it and also different students seemed
to use different (sometimes conflicting) rules of discourse about methods
for deriving new truths from existing ones. Although Betsy’s method was
compatible with the methods of wider society, many students did not seem
to understand (or accept) her method. Accordingly, one can draw two con-
clusions. First, Betsy’s argument cannot count as proof, because it was not
accepted as such by the classroom community. Second, the classroom com-
munity had not yet developed, or socially shared, all the necessary rules
of discourse that would support the elevation of Betsy’s argument to the
status of proof within the community.
But then, how could Betsy’s argument unfold differently in the social
context of her class so that her argument could count as proof? The most
obvious scenario would be for the class to share all necessary rules of
discourse that would lead students to agree that Betsy’s argument was a
proof. However, this scenario is hard to realize not only because of the
limits of social interactions in the domain of proof (Balacheff, 1991) but
16 A. J. STYLIANIDES

also because students in the early grades are just beginning to interactively
constitute with their teachers certain rules of discourse; the development
and acceptance of these rules is a long and demanding process (Yackel and
Cobb, 1996). So, what might be another scenario, perhaps less desirable
but more realistic?
One such scenario would be for the teacher to ratify Betsy’s argument
as a proof at the end of the episode, explaining to the students why this
argument qualified as proof while the other arguments that were proposed
did not (this explanation could refer to the arguments’ foundation, formu-
lation, and representation). The teacher’s act would initiate a situation for
institutionalization (Brousseau, 1981) that would both legitimize Betsy’s
argument as proof and guarantee that the methods used in its construction
would be genuinely considered as knowledge to be retained for future work
(Balacheff, 1990). This scenario is consistent with the conceptualization of
the social dimension of proof proposed above, because, by the end of the
episode, (1) Betsy’s argument would be accepted as proof by the commu-
nity, and (2) the teacher’s act would help establish within the community
certain rules of discourse that would support the acceptance of this argu-
ment as proof. Also, this scenario highlights both the power of the teacher
over what rules of discourse get privileged in classroom activity (Lampert,
1990) and the teacher’s special role to help students construct mathematical
ways of knowing that are compatible with those of wider society (Yackel
and Cobb, 1996).

5. C ONCLUSION

In this paper, I focused on the issue of understanding and characterizing

the notion of proof in the context of elementary school mathematics. To
gain insight into this issue, I analyzed an episode from a third-grade class,
which presented an argument by a student – Betsy – for the statement ‘odd
+ odd = even’ that could potentially count as proof at the elementary school
level. My analysis centered on this argument, probing its major elements –
foundation, formulation, representation, and social dimension – and trying
to decide whether it qualified as proof. To facilitate my analysis, I devel-
oped and used a theoretical framework that is comprised of two guiding
principles for conceptualizing the notion of proof in school mathematics:
the intellectual-honesty principle and the continuum principle. The two
principles offered the basis for certain judgments about whether Betsy’s
argument could count as proof. Also, they supported more broadly ideas
for a possible conceptualization of the notion of proof in the elementary
grades.
 PROOF IN ELEMENTARY SCHOOL MATHEMATICS 17

From the foundation, formulation, and representation points of view,

the application of the two principles identified a class of arguments of
which Betsy’s argument is a representative – thought experiments (Balach-
eff, 1988) – that have high potential to be called proofs in the elementary
grades. This does not suggest however that lower-level arguments would
necessarily lack this potential, but the analysis does not illuminate possible
boundaries between proofs and arguments that do not qualify as proofs.
From the social dimension point of view, the application of the two prin-
ciples gave rise to the following conceptualization of the social dimension
of proof in school settings: An argument that could count as proof in a
classroom community should be accepted as proof by the community –
and, thus, it should be convincing to the students – on the basis of socially
accepted rules of discourse relevant to proving that are compatible with
those of wider society. These rules of discourse should evolve with the de-
velopment of students’ knowledge throughout the grades. Teachers have an
important role to play in the negotiation, acceptance, and development of
these rules of discourse among their students. Based on this conceptualiza-
tion, Betsy’s argument could not count as proof, primarily because it was not
accepted as such by the classroom community wherein it was developed.
In analyzing the episode, I took the perspective of the researcher who
studied the classroom activity after the fact and from the outside, trying
to gain insight into the notion of proof in the elementary grades. But what
if I took, for example, the perspective of the teacher as she was trying –
in real time – to decide for herself whether Betsy’s argument could count
as proof in order to design accordingly her instruction? I hypothesize that
an analysis of the episode from the teacher’s perspective would suggest
aspects of a slightly different conceptualization of the notion of proof in
school mathematics that would, additionally, shed some light on the role of
teachers in cultivating proof among their students. With regard to the three
structural elements of an argument – foundation, formulation, and repre-
sentation – this conceptualization would agree with the conceptualization
I sketched in the previous paragraph. However, the two conceptualizations
would differ in the treatment of the fourth element of an argument: social
dimension. This element would no longer be determinant of whether an
argument counts as proof. Rather, it would be a central part of the teach-
ers’ agenda in trying to cultivate proof among their students. Specifically,
teachers would first examine the three structural elements of given student
arguments to decide for themselves whether they qualify as proofs (argu-
ments like Betsy’s would therefore qualify as proofs), and then engineer
appropriate instructional interventions to help their students develop and
socially share the rules of discourse that would allow them to accept or
reject these arguments as proofs.
18 A. J. STYLIANIDES

To conclude, further research is needed to develop a fine grained con-

ceptualization of the notion of proof that is appropriate in the context of
elementary school mathematics and that offers some insight into the role
of teachers in cultivating proof among their students. The development of
such a conceptualization will support the design of a coherent program
for proof instruction across all grades, thus addressing the ‘discontinuity
problem’ currently manifested in students’ experiences with proof as they
pass from elementary to secondary school mathematics.

A CKNOWLEDGEMENTS

This paper is based in part on my dissertation thesis, which was com-

pleted at the University of Michigan under the supervision of Deborah
Ball. My work in preparing the paper was supported in part by funds from
the National Science Foundation to the following Centers for Learning and
Teaching: Center for Proficiency in Teaching Mathematics and Diversity
in Mathematics Education Center. The opinions expressed in the paper
are mine and do not necessarily reflect the position, policy, or endorse-
ment of the National Science Foundation or of either of the two Centers. I
wish to thank Heinz Steinbring and three anonymous reviewers for useful
comments on an earlier version of the paper.

NOTES

1. Some researchers in other regions of the world have advanced the same argument. See,
for example, Ball, Hoyles, Jahnke, and Movshovitz-Hadar (2002, p. 907).
2. By proposing two principles that prevent empirical arguments from qualifying as proofs,
I do not mean to devalue the importance of empirical (inductive) explorations in iden-
tifying patterns, generating conjectures, and offering insight into the development of
proofs.
3. It is beyond the scope of this paper to discuss the reasons for which the teacher decided
to conclude the episode this way.

REFERENCES

Balacheff, N.: 1988, ‘Aspects of proof in pupils’ practice of school mathematics’, in D.

Pimm (ed.), Mathematics, Teachers and Children, Hodder and Stoughton, London, pp.
216–235.
Balacheff, N.: 1990, ‘Towards a problématique for research on mathematics teaching’,
Journal for Research in Mathematics Education 21(4), 258–272.
Balacheff, N.: 1991, ‘The benefits and limits of social interaction: The case of mathematical
proof’, in A.J. Bishop (ed.), Mathematical Knowledge: Its Growth through Teaching,
Kluwer Academic Publishers, Dordrecht, pp. 175–192.
 PROOF IN ELEMENTARY SCHOOL MATHEMATICS 19

Ball, D.L.: 1993, ‘With an eye on the mathematical horizon: Dilemmas of teaching elemen-
tary school mathematics’, The Elementary School Journal 93(4), 373–397.
Ball, D.L. and Bass, H.: 2003, ‘Making mathematics reasonable in school’, in J. Kilpatrick,
W.G. Martin and D. Schifter (eds.), A Research Companion to Principles and Standards
for School Mathematics, National Council of Teachers of Mathematics, Reston, VA, pp.
27–44.
Ball, D.L., Hoyles, C., Jahnke, H.N. and Movshovitz-Hadar, N.: 2002, ‘The teaching of
proof’, in L.I. Tatsien (ed.), Proceedings of the International Congress of Mathematicians,
Vol. III, Higher Education Press, Beijing, pp. 907–920.
Ball, D.L. and Wilson, S.M.: 1996, ‘Integrity in teaching: Recognizing the fusion of the
moral and intellectual’, American Educational Research Journal 33, 155–192.
Brousseau, G.: 1981, ‘Problèmes de didactique des décimaux’, Recherches en Didactique
des Mathématiques 2, 37–128.
Bruner, J.: 1960, The Process of Education, Harvard University Press, Cambridge, MA.
Carpenter, T.P., Franke, M.L. and Levi, L.: 2003, Thinking Mathematically: Integrating
Arithmetic and Algebra in Elementary School, Heinemann, Portsmouth, NH.
Coe, R. and Ruthven, K.: 1994, ‘Proof practices and constructs of advanced mathematics
students’, British Educational Research Journal 24(4), 333–344.
De Millo, R., Lipton, R. and Perlis, A: 1979/1998, ‘Social processes and proofs of theorems
and programs’, in T. Tymoczko (ed.), New Directions in the Philosophy of Mathematics,
Princeton University Press, Princeton, NJ, pp. 267–285.
Ernest, P.: 1991, The Philosophy of Mathematics Education, The Falmer Press, London.
Ernest, P.: 1998, Social Constructivism as a Philosophy of Mathematics, State University
of New York Press, Albany, NY.
Fischbein, E.: 1982, ‘Intuition and proof’, For the Learning of Mathematics 3(2), 9–18, 24.
Galotti, K.M., Komatsu, L.K. and Voelz, S.: 1997, ‘Children’s differential performance on
deductive and inductive syllogisms’, Developmental Psychology 33(1), 70–78.
Hanna, G.: 1983, Rigorous Proof in Mathematics Education, OISE Press, Toronto.
Hanna, G. and Jahnke, H.N.: 1996, ‘Proof and proving’, in A.J. Bishop, K. Clements,
C. Keitel, J. Kilpatrick and C. Laborde (eds.), International Handbook of Mathematics
Education, Kluwer Academic Publishers, Dordrecht, Netherlands, pp. 877–908.
Healy, L. and Hoyles, C.: 2000, ‘Proof conceptions in algebra’, Journal for Research in
Mathematics Education 31(4), 396–428.
Hersh, R.: 1993, ‘Proving is convincing and explaining’, Educational Studies in Mathemat-
ics 24, 389–399.
Kitcher, P.: 1984, The Nature of Mathematical Knowledge, Oxford University Press, New
York, NY.
Lakatos, I.: 1976, Proofs and Refutations: The Logic of Mathematical Discovery, Cambridge
University Press, Cambridge.
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, 29–63.
Lampert, M.: 1992, ‘Practice and problems in teaching authentic mathematics’, in F.K.
Oser, A. Dick and J. Patry (eds.), Effective and Responsible Teaching: The New Synthesis,
Jossey-Bass Publishers, San Francisco, CA, pp. 295–314.
Light, P., Blaye, A., Gilly, M. and Girotto, V.: 1989, ‘Pragmatic schemas and logical rea-
soning in 6- to 8-year-old children’, Cognitive Development 4, 49–64.
Maher, C.A. and Martino, A.M.: 1996, ‘The development of the idea of mathematical proof:
A 5-year case study’, Journal for Research in Mathematics Education 27, 194–214.
20 A. J. STYLIANIDES

Mariotti, M.A.: 2000, ‘Introduction to proof: The mediation of a dynamic software envi-
ronment’, Educational Studies in Mathematics 44, 25–53.
Moore, R.C.: 1994, ‘Making the transition to formal proof’, Educational Studies in Math-
ematics 27, 249–266.
National Council of Teachers of Mathematics: 2000, Principles and Standards for School
Mathematics, National Council of Teachers of Mathematics, Reston, VA.
Polya, G.: 1981, Mathematical Discovery: On Understanding, Learning, and Teaching
Problem Solving, Wiley Combined Edition, New York, NY.
Schoenfeld, A.H.: 1994, ‘What do we know about mathematics curricula?’ Journal of
Mathematical Behavior 13, 55–80.
Simon, M.A. and Blume, G.W.: 1996, ‘Justification in mathematics classrooms: A study
of prospective elementary school students’, Journal of Mathematical Behavior 15(1),
3–31.
Sowder, L. and Harel, G.: 1998, ‘Types of students’ justifications’, The Mathematics Teacher
91(8), 670–675.
Tymoczko, T. (ed.): 1986/1998, New Directions in the Philosophy of Mathematics, Princeton
University Press, Princeton, NJ.
Usiskin, Z.: 1987, ‘Resolving the continuing dilemmas in school geometry’, in M.M.
Lindquist and A.P. Shulte (eds.), Learning and Teaching Geometry, K-12, National Coun-
cil of Teachers of Mathematics, Reston, VA, pp. 17–31.
Yackel, E. and Cobb, P.: 1996, ‘Sociomathematical norms, argumentation, and autonomy
in mathematics’, Journal for Research in Mathematics Education 27, 458–477.
Yackel, E. and Hanna, G.: 2003, ‘Reasoning and proof’, in J. Kilpatrick, W.G. Martin
and D. Schifter (eds.), A Research Companion to Principles and Standards for School
Mathematics, National Council of Teachers of Mathematics, Reston, VA, pp. 227–236.

University of California-Berkeley
Education, EMST, Tolman Hall #1670
Berkeley, CA 94720-1670, USA
E-mail: astylian@berkeley.edu