ARGUMENTATION AND PROOF:

A CONTRIBUTION TO THEORETICAL PERSPECTIVES

AND THEIR CLASSROOM IMPLEMENTATION
Paolo Boero*, Nadia Douek**, Francesca Morselli*, Bettina Pedemonte***
*DIMA, Genoa University **IUFM, Nice University ***DiDiMa, ITD-CNR, Genoa
with a reaction by Leonor Camargo, Patricia Perry, Carmen Samper,
Universidad Pedagógica Nacional in Bogotá, Colombia
The complexity of proving and the relationship between argumentation and proof are
subjects of major concern in Mathematics Education. In this Research Forum we will
propose the integration of Toulmin's model of argumentation with Habermas'
elaboration of rational behavior, both adapted to proof and proving. After a short
presentation of Toulmin's and Habermas' tools at work, we will provide some
theoretical reasons for their integration. Then we will show how this construct allows
us to frame the planning, management and analysis of some classroom activities
aimed at students' approach to relevant aspects of proving and proof. Finally, we will
suggest some further developments. A reaction by Carmen Samper and her
colleagues (related to their own research work) will introduce the discussion.

1. INTRODUCTION
Since the late nineteen eighties, mathematical proof and proving have been one of the
main subjects of research in mathematics education. Different strands of research
have developed since that time. In particular, several mathematics educators (starting
with Gila Hanna, Nicolas Balacheff, Raymond Duval and others: see Balacheff,
1987; Hanna, 1989; Duval, 1991) have considered, within an educational perspective,
the relationships, tensions, and potential oppositions between: formal proof, and
semantic or informal argument; proof as a cultural product subject to logical and
communicative (textual) constraints, and proving as the process aimed at that
product; mathematical proof, and ordinary argumentation. In spite of the apparent,
broad differences, all these relationships deal with a crucial dichotomy: on the one
hand, we have to consider mathematics output as fitting a set of rules, constraints,
logical and textual models; on the other hand, there is the creative and constructive
side of mathematicians' (and students') activity when they are engaged in
understanding and validating mathematical statements, using cultural processes.
The theoretical background consists of two main constructs (Toulmin's model for
argumentation, see Toulmin, 1974; and Habermas' construct of "rational behavior"-
see Habermas, 2003 - as adapted in Boero & Morselli, 2009); other constructs will be

1- 1
2010. In xxxx (Eds.). Proceedings of the 34th Conference of the International Group
for the Psychology of Mathematics Education, Vol. 1, pp. XXX-YYY. Belo
Horizonte, Brazil: PME.
Last names of authors, in order on the paper

discussed in specific instances (Peirce's abduction; cognitive unity and structural

continuity - see Pedemonte, 2007, 2008; phases of proving - see Arzarello, 2007;
Douek, 2009). Our research provides evidence that:
- Toulmin's model of argumentation can be used to analyze and compare the
organization of arguments between the exploratory, creative phases and the deductive
chaining of the resulting arguments (see Section 2).
- Habermas' construct of rational behavior can be used to analyze the components of
the proving process (and the tensions between these): the conscious assumption of the
epistemic constraints (related to mathematical deduction) inherent in the desired
product; the teleological component (i.e. the conscious choice of the tools to achieve
the aim of the activity); and the communicative component (i.e. the conscious choice
of tools in order to conform the product to the textual constraints of proof or to be
understood by others) (see Section 3).
The Research Forum will focus on the need to integrate the two models and on a
proposal for integrating them. Indeed, if considered separately, Toulmin's model is
suitable to analyze the organization of arguments, not the subject's intentions nor the
tensions between the different "components" of the proving process; conversely,
Habermas' construct is suitable to take the subject's intentions and consciousness into
account, but it does not offer the possibility of modeling and comparing different
kinds of productions (in particular, the composition of arguments in the exploratory
phases and in the final proof) and eliciting possible continuities and discontinuities
(and related obstacles) between them.
The main goals of the Research Forum are:
- to present the integrated theoretical frame (Section 4);
- to show how such a frame can be used for the design, management and analysis of
teaching and learning activities regarding proof (Section 5).
In a Vygotskian perspective (Boero, 2006), the contribution made by this Research
Forum concerns some aspects of the students' approach to scientific knowledge in the
case of theorems: in particular, awareness and intentional use of knowledge are
promoted through the direct mediation of the teacher in the interplay between
mathematical discussion and the story narration of proof construction (see Section 5).
Some research and educational developments will be discussed in Section 6.

2. TOULMIN'S MODEL
In the educational literature, Toulmin’s model has been already used for analyzing
and documenting both how learning progresses in a classroom (Krummehuer, 1995)
and how to create a context for arguing in the class (Wood, 1999). The use of this
model in our previous research was critical for comparing students’ argumentations
with their proofs, from a structural and cognitive point of view (Pedemonte 2005,
2007, 2008). This comparison is based on the hypothesis that proof is a particular

1- 2 PME 34 - 2010
Last names of authors in order as on the paper

type of argumentation in mathematics (Pedemonte, 2007). According to linguistic

theories (Plantin, 1990; Toulmin, 1974) proof is a set of rational justifications
expressed as inferences. These inferences are analyzed and compared using
Toulmin’s model .
2.1.Toulmin’s model: a methodological tool to analyze argumentation and proof
In Toulmin’s model an argument consists of three elements (Toulmin, 1974):
C (claim): the statement of the speaker D (data): data justifying claim C, W
(warrant): the inference rule, which allows data to be connected to the claim.
In any argument, the first step is expressed by a viewpoint (an assertion, an opinion).
In Toulmin's terminology the standpoint is called the claim. The second step consists
of providing data to support the claim. The warrant provides the justification for
using the data as a support for the claim. The warrant, which can be expressed as a
principle a rule, acts as a bridge between the data and the claim. Three other elements
that describe an argument can be taken into account: B (backing) the support of the
rule; Q (qualifier) the strength of the argument; Re (rebuttal) the exception to the
rule. The force of the warrant would be weakened if there were exceptions to the rule:
in that case conditions of exceptions or rebuttal should be inserted. The claim must
then be weakened by means of a qualifier. Backing is required if the authority of the
warrant is not accepted straight away. Overall, Toulmin's model of argumentation
contains six related elements as shown in figure 1.

Q : Qualifier Re : Rebuttal

D : Data C : Claim

W: Warrant

B: Backing

Fig. 1: Toulmin's model of argumentation

To understand how Toulmin’s model is applied, consider the following argument

given by a student to answer the question1:
“What can you say about -a2 if a is an integer number different from 0? Is it a positive or
a negative number?”
“-a2 is a negative number (claim) because the square of each number is a positive
number, but with minus it becomes a negative number (warrant)… unless the square is

1
This answer was given in a test designed to evaluate algebraic competencies in a class of 13-14 year-olds.

PME 34 - 2010 1- 3
Last names of authors, in order on the paper

made for the whole number and the minus… in this case -a2 is a positive number
(rebuttal). No… this is impossible because -a2 is different than the square of (–a)2”
The argument can be illustrated as follows:

Q : probably unless Re: -a2>0 because -a2 = (-a) 2>0

D : -a2, a∈Z, a≠0 C : -a2<0

since W: n2≥0 ∀n∈Z; if n>0 then –n<0

on account of B: Algebraic rules and in particular -a2 ≠ (-a) 2

Fig. 2: Toulmin's model example

2.2. Why is it important to compare argumentation and proof?
The relationship between conjecturing and proving has been analyzed in mathematics
education from different points of view and with different educational aims. On the
one hand, some research suggests major differences between argumentation and
proof both from a social and epistemological point of view (Balacheff, 1987) and
from a cognitive and logical point of view (Duval, 1991). On the other hand, some
Italian studies highlight the continuity that may exist between argumentation as a
process of producing a conjecture and constructing its proof (Boero, Garuti, Mariotti,
1996). This continuity is called cognitive unity. During the problem-solving process,
argumentation is usually required in order to produce a conjecture. The hypothesis of
cognitive unity is that, in some cases, this argumentation can be exploited by the
student in the construction of a proof by organizing some of the previously produced
arguments into a logical chain. Research on cognitive unity (Boero & al., 1996;
Garuti, Boero, Lemut, Mariotti, 1996) shows that proof is more “accessible” to
students if an argumentation activity leads to the construction of a conjecture, which
is then proved by building on the argumentation. Following these research studies,
Pedemonte, (2007) has shown that the analysis of cognitive unity does not cover all
aspects of the relationship between argumentation and proof. In particular, it seems
very important for a cognitive analysis of argumentation and proof to consider two
points of view:
• the referential system, made up of the representation system (the language,
heuristics, and drawings) and the knowledge system (conceptions and
theorems) of argumentation and proof (Pedemonte, 2005). The analysis of
cognitive unity takes into account the referential system.
• the structure intended to allow logical cognitive connection between
statements (deduction, abduction, and induction structures) (Pedemonte, 2007).

1- 4 PME 34 - 2010
Last names of authors in order as on the paper

There is continuity in the referential system between argumentation and proof if some
expressions, drawings, or theorems used in the proof have been used in the
argumentation supporting the conjecture. There is structural continuity between
argumentation and proof when inferences in argumentation and proof are connected
through the same structure (abduction, induction, or deduction). For example, there is
structural continuity between argumentation and proof if some abductive steps used
in the argumentation are also present in the proof. It is important to observe that if
continuity in the referential system between argumentation and proof supports the
construction of proof, this is not generally the case for the structural continuity. As a
matter of fact, to produce a deductive proof, a structural distance is in many cases
necessary because the structure of argumentation is usually not deductive. Sometimes
students are unable to construct a proof because there is “spontaneous” continuity
between argumentation and proof (e.g. from abductive argumentation to a sort of
“abductive” proof: see the example in section 2.3). Toulmin’s model is an important
and effective tool to analyze argumentation and proof because the two kinds of
analysis - the structural analysis and the referential system analysis - can be
performed using it.
Toulmin’s model as a way to analyze and compare the referential system and the
structure between argumentation and proof
A warrant in a proof is an axiom, definition, or theorem. Backing is the theoretical
system which justifies the warrant. In the argumentation supporting the conjecture,
these elements do not necessarily belong to a theoretical system. If the warrant is a
mathematical rule there will probably be continuity between argumentation and proof
because this rule can be replaced in the proof with a theorem (Pedemonte 2005). In
contrast, if the rule is not correct, it cannot be replaced by a theorem in the proof. In
this case, three possibilities can be identified: the proof is not constructed by the
student (the cognitive unity is broken), an incorrect “proof” is constructed and is
based on the incorrect rule used in the argumentation (cognitive unity); the incorrect
argumentation is abandoned, another argumentation is constructed and turned into a
correct proof (the cognitive unity is broken but a successful argumentation is
constructed).
In Toulmin's model a step appears as a deductive step: data and warrants lead to the
claim. However, other argumentative structures can be represented by this model
(abduction, induction, etc.). For example, an abductive step can be expressed as
follows (the question mark means that data are to be sought in order to apply the
inference rule justifying the claim).
D:? C:B

W: A ⇒ B

Fig. 3: Abduction in Toulmin's model

PME 34 - 2010 1- 5
Last names of authors, in order on the paper

2.3. Example
In this section, we describe the resolution process (taken from a set of data collected
from a teaching experiment carried out in some traditional 12th and 13th grade
classes in France and Italy) for an open-ended geometry problem.
Problem: ABC is a triangle. Three exterior squares are constructed along the triangle's
sides. The free points of the squares are connected, defining three more triangles.
Compare the areas of these triangles with the area of triangle ABC.
The area of each of these triangles is equal to the area of triangle ABC. The solution
to this problem is not obvious to students. In order to find a successful strategy, an
abductive argumentation is usually constructed, as we show in this example. Note
that, in the proof, abductive steps are present because students were not able to
transform argumentation in a deductive proof. This example is representative of the
vast majority of analyzed resolution processes. The analysis start at claim C6 (the
conjecture); at this point the students affirm that the area of triangle ABC and the
area of triangle ECL (see figure) are equal. So far, they have calculated these areas.
103. C : The areas are always equal The figure as represented by the students
… with the calculator the areas are equal using Cabri Geometry
104. N: Now we have to see why!

C6: The triangles’ areas are equal

Students have to justify the claim C6, which is constructed as a “fact”. Two abductive
steps are developed in the argumentation. The first is needed to justify the equality of
the triangles' areas. The students look for a relationship between bases and heights
that make a constant area (step 6). The second step is needed to justify that this
relationship always holds. The triangle bases are congruent, so the students tend to
justify the congruence of the heights (step 7).
105. C : We need to find how the base D6 : ?
and the height change …if there is a C6
relationship that makes the area W: Formula of the area
constant… The area is constant… but I If the relationship between bases and heights
don’t understand…so we have to find has to be constant, the heights have to be
base by height congruent to base by congruent because the bases are congruent.
height of the other triangle D7 : congruent bases C7 : D6: CB*AN
106. N: If we take the constant bases and ? (congruent heights) =CL*ED
we change the heights...
W: transitivity of congruence

1- 6 PME 34 - 2010
Last names of authors in order as on the paper

The comparison between heights leads explicitly to a comparison between the two
small triangles ANC and EDC constructed on the heights. Students try to establish
their congruence. If the two triangles are congruent, their heights are congruent. This
step is an abductive step (step 8). As in the previous example, students look for data
to apply the congruence criterion and justify the congruence between the two small
triangles ANC and EDC (step 9). Again the argumentation structure is abduction.
115. C: But why are the heights
congruent? ?D8 : ANC=EDC C8=D7
116. N: We have…. we see that this side ’
W: Inheritance of congruence
is congruent to this side of triangle ABC
…
117. C : Then the small triangle is ?D9 C9 : D 8
congruent to the other small triangle …
118. N : … yes it’s true, two sides are
congruent W: SAA congruence criterion
119. C : So there is a 90° angle
120. N: We need another side or
another angle…
Once students find data to justify the congruence between the two small triangles
ANC and EDC, they can construct the proof.
Proof: The protocol appears to be an example of cognitive unity. But there is no
structural change between argumentation and proof. The transcript of this proof
describes the students’ abductive reasoning. The students are unable to completely
cover the distance between argumentation and deductive proof. Argument 7 is still an
abductive step (congruency between heights has to be proved). This is the reason why in
this case we can observe a structural continuity between argumentation and proof.

Students consider the triangles ABC D : congruent bases

7 C7 : The areas
and ELC ? (congruent heights) of triangles are
We know that this base is congruent to
the base of the triangle. Now we have to equal
prove that the heights are congruent. W: area formula
We have verified this fact by means of
the congruence criterion proved on the D8: ANC=EDC C8: congruent
sheet with the drawing. heights
On the sheet with the drawing: W: inheritance of congruence
Triangle ANC = Triangle EDC
EC=AC D9: EC=AC C9 : D8
EDC=ANC=90° EDC=ANC=90°
ACN=ECD because ACE=90°, ACN=ECD
DCN=90° and if the angle DCA is
removed from the two other angles we W: SAA congruence criterion
have the same angle.

PME 34 - 2010 1- 7
Last names of authors, in order on the paper

2.4. Some comments and further considerations

Some experiments in geometry classes (Pedemonte, 2007) highlighted that this
“spontaneous” structural continuity between abductive argumentation and “abductive
proof” can be an obstacle for students in the construction of a deductive proof.
However, this continuity between abductive argumentation and proof seems to be
absent when students produce algebraic proofs. As a matter of fact, unlike in the
geometry example, the structural distance between argumentation and proof (from an
abductive argumentation to a deductive proof) is not one of the common difficulties
met by students in solving problems involving properties of numbers. Since
algebraic proof is characterized by a strong deductive structure, abductive steps in the
argumentation activity can be useful in linking the meaning of the letters used in the
algebraic proof with numbers used in the argumentation. In fact, the strength of the
deductive structure in algebraic proof prevents, at least partially, the occurrence of
structural continuity between argumentation and proof (Pedemonte, 2008). Moreover,
in this case analysis performed through Toulmin’s model has shown that, unlike in
the geometrical case, abductive steps in the argumentation can be useful for the
construction of proof if they favor continuity in the “content” of argumentation and
proof (sometimes abductive steps assume an important role in the argumentation
because, through them, students’ reasoning maintains the connection between
construction of a conjecture and proof).

3. PROVING AS A RATIONAL BEHAVIOUR: HABERMAS' MODEL

Balacheff (1982) points out that the teaching of proofs and theorems should have the
double aim of making students understand what a proof is and making them learn to
produce proofs. Accordingly, we think that proof should be dealt with in mathematics
education by considering both the object aspect (a product that must meet the
epistemic and communicative requirements established in today mathematics - or in
school mathematics) and the process aspect (a special case of problem solving: a
process intentionally aimed at a proof as product) of proof. We have tried (Boero,
2006; Morselli, 2007) to match these considerations with Habermas’ elaboration
about rationality in discursive practices; we will present here a unified synthesis of
our previous work. Habermas (2003, ch.2) distinguishes three interrelated
components of rational behaviour: the epistemic component (inherent in the control
of the propositions and their chaining), the teleological component (inherent in the
conscious choice of tools to achieve the goal of the activity) and the communicative
component (inherent in the conscious choice of suitable means of communication
within a given community). With an eye to Habermas’ elaboration, in the discursive
practice of proving we can identify: A) an epistemic aspect, consisting in the
conscious validation of statements according to shared premises and legitimate ways
of reasoning (cf. the definition of “theorem” by Mariotti & al. (1997) as the system
consisting of a statement, a proof, derived according to shared inference rules from
axioms and other theorems, and a reference theory); B) a teleological aspect,

1- 8 PME 34 - 2010
Last names of authors in order as on the paper

inherent in the problem-solving character of proving, and the conscious choices to be

made in order to obtain the desired product; C) a communicative aspect, consisting
in the conscious adhering to rules that ensure both the possibility of communicating
steps of reasoning and the conformity of the products (proofs) to standards in a given
mathematical culture.
Our point is that considering proof and proving according to Habermas’ construct
may provide the researcher with a comprehensive frame within which to situate a lot
of research work performed in the last two decades (see below for some examples),
to analyze students’ difficulties concerning theorems and proofs (see the two
examples in the next subsection), and to discuss some related issues and possible
implications for the teaching of theorems and proof (see Boero, 2006; Morselli &
Boero, 2008; 2009).
Regarding the epistemic aspect, i.e. in the analysis of proofs and theorems as objects,
mathematics education literature offers some historical analyses (e.g. Arsac, 1988)
and surveys of epistemological perspectives (e.g. Arzarello, 2007). These help us
understand how theorems and proofs originated and how they were perceived in
different historical periods and how, even today there is no fully shared agreement
about what constitutes a “mathematical proof” (cf. Habermas' comment about the
historically and socially situated character of epistemic rationality). Also relevant to
epistemic rationality is Duval’s (2001) focus on the distance between mathematical
proof and ordinary argumentation(by referring proof to the model of formal
derivation). He also considers how to make students aware of that distance and able
to manage the construction and control of a deductive chain. Harel (2008) uses the
DNR theoretical framework to frame the classification of students’ proof schemes
(they concern proof as a final product). We note that, in terms of Habermas’
components of rationality, Harel’s ritual and non-referential symbolic proof schemes
may be attributed to the dominance of the communicative aspect in an educational
context where students' intentions are driven by the necessity of conforming to the
supposed requests of the teacher, with lacks inherent in the epistemic component (cf.
Harel’s N, “intellectual Necessity”).
Concerning the proving process, some analyses of its relationships with arguing and
conjecturing suggest possible ways to enable students to manage the teleological
rationality. In particular, Boero, Douek & Ferrari (2008) focus on the existence of
common features between arguing and proving processes, and they present some
activities (designed for grade 1 on), based on those commonalities, that may prepare
students to develop effective proving processes.
3.1 Two examples of analysis within the frame of rationality
Morselli (2007) investigated the conjecturing and proving processes carried out by
different groups of university students (7 first-year and 11 third-year mathematics
students, or 29 third-year students preparing to become primary school teachers). The
students were given the following problem: What can you tell about the divisors of

PME 34 - 2010 1- 9
Last names of authors, in order on the paper

two consecutive numbers? Motivate your answer in general. Different proofs can be
carried out at different mathematical levels (by exploiting properties of the
remainder, examining multiples, or using algebraic tools). The students worked on
the problem individually, writing down their process of solution. Afterwards,
students were asked to reconstruct their process and comment on it; these interviews
were audiotaped. Morselli (2007) provided several examples of individual solutions
and related interviews, in particular showing how students’ failures or mistakes were
due to lacks in some aspects of rationality and/or the dominance of one aspect over
the others. We will consider here only portions of two very similar examples,
concerning students that are preparing to teach in primary school, in order to show
how Habermas’ construct works as a tool for in-depth analyses.
Example 1: Monica
Monica considers two couples of numbers: 14, 15 and 24, 25. By listing the divisors,
she discovers that “Two consecutive numbers are odd and even, hence only the even
number will be divisible by 2”. Afterwards, she lists the divisors of 6 and 7 and
writes: “Even numbers may have both odd and even divisors”. After a check on 19
and 20, she writes the discovered property, followed by its proof:
Property: two consecutive numbers have only one common divisor, the number 1. In
order to prove it, I can start by saying that two consecutive numbers certainly cannot
have common divisors that are even, since odd numbers cannot be divided by an even
number. They also cannot have common divisors different from 1, because between the
two numbers there is only one unit; if a number is divisible by 3, the next number that is
divisible by 3 will be greater by 3 units, and not by only one unit. Since 3 is the first odd
number after 1, there are no other numbers that can work as divisors of two consecutive
numbers.
Monica carries out reasoning that is intentionally aimed (teleological aspect): first, at
the production of a good conjecture; then, at its proof. Proof steps are justified one by
one (epistemic aspect) and communicated with appropriate technical expressions
(communicative aspect). The only lack in terms of rationality concerns the short-cut
in the last part of the proof: Monica realizes that something similar to what happens
with 3 (the next multiple is “greater by three units”) will happen a fortiori with the
other odd numbers that are bigger than 3 (“Since 3 is the first odd number after 1”),
but she does not make it explicit. Her awareness (cf. epistemic rationality) is not
communicated in the explicit mathematical form (lack of communicative rationality).
Monica’s later comments on her text confirm the analysis:
Monica: (...) and then I thought that 3 was the first odd number after 1 and so if 3 does
not enter there, also the bigger ones do not enter there [from the previous
text, “there” means: between two consecutive numbers on the number
line].
Observer: To make more general what you said with 3, what would you write now?

1- 10 PME 34 - 2010
Last names of authors in order as on the paper

Monica: ehm... I have tried to go beyond the specific case of 3, but I do not know if I
have succeeded.
Example 2: Caterina
Starting from the fact that two consecutive numbers are always one odd and one even, we
may conclude that the two numbers cannot be both divided by an even number.
Afterwards, we focus on odd divisors; we start from 1, and we know that all numbers
may be divided by 1; the second one is 3. We have two consecutive numbers, then the
difference between them is 1, then they will not be multiples of 3, since it will be
impossible to divide both of them by a number bigger than 1.
Caterina is able to justify all the explicit steps of her reasoning (epistemic aspect), she
develops goal-oriented reasoning (teleological aspect) and illustrates her process with
appropriate technical expressions (communicative aspect). In spite of good intuition,
there is a lack in her reasoning: she does not consider divisors greater than 3 (unlike
Monica). Later, after having seen also the production of her colleagues, Caterina
comments:
My reasoning is not mistaken: indeed, I reach the conclusion giving a general
explanation, saying that, since there is no more than one unit between the two numbers,
the only common divisor is 1. Nevertheless, I can not create a mathematical rule.
Observing the other solutions, I think that the correct rule is the following: along the
number line we note that a multiple of 2 occurs every two numbers, a multiple of 3
occurs every three numbers, hence a multiple of N occurs every N numbers. Then, two
consecutive numbers have only 1 as common divisor.
From the objective point of view of epistemic rationality, Caterina’s argument was
not complete, though her comment reveals that she is not aware of it. From her
subjective point of view, Caterina is convinced to have found a cogent reason for the
validity of the conjecture (“not mistaken reasoning”, “general explanation”), thus to
have achieved her goal (teleological rationality). Some colleagues’ solutions induced
her to reflect on the lack of a “mathematical rule.” However, from her comment it
seems that this lack is not considered by her as a lack in the reasoning but as a lack in
the mathematical communication.
3.2. Habermas' construct and the use of algebraic language in proving
In this Subsection we briefly consider the use of algebraic language in proving and
show how Habermas' construct can be further specialized to frame this. Algebraic
language can play the role of a tool for proving through modelling (see Norman,
1993, and Dapueto & Parenti, 1999) within mathematics (e.g. when proving theorems
of elementary number theory).
Our interest for considering the use of algebraic language in the perspective of
Habermas’ construct depends on the fact that our previous research (Boero, 2006;
Morselli, 2007) suggests that some of the students’ main difficulties in conjecturing
and proving depend on specific aspects (already pointed out in literature) of the use
of algebraic language. Difficulties with these aspects make conjecturing and proving

PME 34 - 2010 1- 11
Last names of authors, in order on the paper

complex and demanding activities for students. In particular, we refer to: an

epistemic component, inherent in the need to check the validity of algebraic
formalizations and transformations and to correctly and purposefully interpret
algebraic expressions in a given context of use; a teleological component, inherent in
the goal-oriented character of the choice of formalisms and of the direction of
transformations; a communicational component, inherent in the restrictions that come
from the need to follow taught communication rules, which may contradict private
rules of use or interfere with them. In Morselli & Boero (2009) we have tried to show
how framing the use of algebraic language with the perspective of Habermas’ theory
of rationality: first, provides the researcher with an effective tool to describe and
interpret some of the main difficulties met by students when using algebraic language
in proving; second, provides the teacher with some useful indications for the teaching
of algebraic language; third, suggests new research developments.

4. THE INTEGRATION OF HABERMAS' AND TOULMIN'S MODELS

As we have seen in the previous sections, we think that Toulmin's model for
argumentation and Habermas' model for rational behavior, suitably adapted to the
specificity of "mathematical proving", are valuable analytical tools to deal with some
relevant aspects of the complexity of the conjecturing and proving activity. These
models can be useful to analyze students' performances and difficulties in developing
proving skills. In particular, those analytical tools allow one to evaluate the distance
between the actual behavior of students and the behavior that teachers would like to
promote, such as the case of the relationships between argumentation and proof (in
the delicate transition from the argumentative search for reasons of validity of a
statement, to their organization and chaining in a mathematical proof - see Section 2)
and the case of the rational behavior inherent in proving, with its epistemic,
teleological and communicative components (see Section 3). Educational
implications include pointing out some negative features of the traditional school
approach to mathematical proof: for instance, Habermas' model reveals that teachers'
pressure on communicative and epistemic rationality (according to teacher's culture)
usually prevails over the need for teleological rationality (see Morselli & Boero,
2009). The models also suggest some guidelines for preparing teachers and planning
teaching (see Boero 2006, Morselli & Boero, 2009) with sensitivity to students'
conceptions. However, neither model on its own seems to offer a comprehensive
frame for classroom implementation of teaching of mathematical proof, especially
concerning practices that should allow students to develop consciousness about the
requirements of proving in mathematics and the ways of managing proving. The
contribution that we would like to bring in this Section is a proposal to integrate
Toulmin's and Habermas' analytical tools, in order to get a frame suitable for (1)
combining them in the analysis of students' behavior, and (2) inspiring and planning
innovative classroom practices aimed at developing students' awareness of the nature
of proving. By their nature, the integration of the two models can work as an

1- 12 PME 34 - 2010
Last names of authors in order as on the paper

integration of two different levels of detail and focus of the analysis: when looking at
the (oral and written) texts produced by students through Toulmin's lens, the unit of
analysis is the fundamental step of argumentation which connects data to claim based
on a warrant (with possible expansions in the context of the argumentation, and links
of different kinds - deductive chaining, opposition, etc. - with other steps). When
looking at the behavior with Habermas' lens, the unit of analysis is related to the
motive of one specific phase of the activity (e.g. the production of the final text of the
proof, or the organization of some arguments into a reliable chain), or to a specific
strategy of proving (e.g. that related to the use of algebraic language). Here we can
see how the components of Habermas' rationality are useful to focus on the
legitimacy of reasoning steps, on the specificity of the intention, and on the
communicational constraints.
The two models complement each other in the following sense. In the case of the
expert, the process underlying the discursive behavior of proving develops under
(more or less conscious) constraints of epistemic validity, efficiency related to the
goal to achieve, and communication according to shared rules. Those constraints
(particularly, but not only, the first one!) result in two levels of argumentation: the
level (we could say, a meta-level) inherent in the awareness of the constraints on the
three components of rational behavior in proving; and the level inherent in the
specific nature of the three components. The protagonists of the classroom scene (the
teacher and the students) develop different levels of argumentation. Initially, the
teacher guides students' activity, applying his/her meta-level of argumentation. Thus,
students become aware of these components and try to conform to them when
working out their argumentation, until they can justify their steps of reasoning from
the perspective of these components. Meta-level is not a goal for students, it is a
teaching means.
The proposed integration of Toulmin's and Habermas' models allows us to look at the
enculturation in the "culture of theorems" managed by the teacher as a process in
which the teacher (through suitable tasks like the "story narration" of proof
construction that will be presented in the next Section) offers to students the
opportunity to enrich the level of argumentation used in the justification of a
statement (on shared epistemic bases), with the meta-level concerning the awareness
of the epistemic, teleological and communicative requirements of proving. Warrants
spontaneously referred to by students mainly concern the epistemic aspect of proving
and can be visual evidence (in the case of geometry), properties obtained through
algebraic transformations, and properties for which no doubt exists. At the meta-level
that must be promoted by the teacher, warrants concern the reliability and the nature
of the "epistemic warrants" ("the conclusion is not yet legitimate because we can use
only agreed properties and already proved statements"), the efficiency of the
strategies and the choices adopted to achieve the proof ("we use algebraic language
because it allows us to express ideas precisely..."), the rules of communication in
mathematics ("this text is not yet satisfactory because all the steps of reasoning must

PME 34 - 2010 1- 13
Last names of authors, in order on the paper

be made explicit with the correct words"). We can observe that in the case of
epistemic rationality, backing at the meta-level refers to the reference theory and the
related rules of inference.
According to the "integrated" model of Toulmin and Habermas, the organization of
reasoning in terms of warrants must be consciously assumed by the student as one of
the main aims of the activity. The mathematical discussion orchestrated by the
teacher (Bartolini Bussi, 1996) appears as an important organization of the collective
classroom activities aimed at promoting the argumentation of students at the meta-
level with the goal of developing rational behavior in proving. In interplay with
classroom mathematical discussion, the story-narration of proof construction will aim
to produce argumentative texts in which the search for warrants and their presentation
at the meta-level become the main goals of students' activity.
In order to frame the educational choices and the teaching experiments that will be
presented in the next Section, we need to focus on phases of proving activity and
some specific features of argumentation in proving.
Three modes of reasoning
Inspired by Lolli's analysis of proof production (see Arzarello, 2007), we consider
proving as a cognitive, culturally situated activity engaging three modes of reasoning:
• Mode 1: exploration and production of reasons for validity of the statement;
• Mode 2: organisation of reasoning into a cogent argumentation;
• Mode 3: production of a deductive text according to specific cultural
constraints concerning the nature of propositions and their chaining.
For educational aims, these reasoning modes can be considered as successive phases
of a proof construction. But in the process of proving, they do not appear separately,
rather they are deeply connected, and their succession may vary and loop. We will
speak of a phase of exploration (or of a phase of organization of reasoning, …) when
it is the main reasoning mode. The different reasoning modes develop based on
different cultural rules, follow different criteria of validity, and may use different
semiotic registers. When moving from one mode to another, the student should come
to know what is allowed or needed in the new mode. For instance, abductive
reasoning is allowed only in Mode 1, and it is not easy for students to move from it to
the deductive reasoning needed for the other modes (see Pedemonte, 2007, and
Section 2). Similar considerations may concern the use of examples.
From the above description, it is evident that argumentation is an intrinsic component
of proving in its different modes of reasoning; moreover, we argue that
argumentation (at the meta-level) is a didactical tool that allows the teacher to guide
students to manage the different modes of reasoning and the relationships between
them in a conscious way. In Section 5 we will present ways to foster both kinds of
argumentation.

1- 14 PME 34 - 2010
Last names of authors in order as on the paper

Lines of argumentation
In a proof, elementary argumentations may form a linear chain in which each
conclusion is an input for the following step, thus forming one whole "line of
argumentation". However, in many cases argumentation may contain "blocks", or
side argumentation branches that meet the main line to input a supplementary data or
argument. A block might be considered as a secondary line of argumentation. The
hierarchical relations between various argumentations involved in a proof (see
Knipping, 2008) may be a source of difficulty for students.
The teacher may use this “lines relating blocks” structure to organize student's
proving activity (when necessary) into parts in order that: firstly, they can handle the
blocks, and secondly, they can reorganize them according to a main line (see the
example below for Pythagoras' Theorem). When planning and managing a collective
discussion, the teacher may use this structure as a reference to question locally the
arguments according to epistemic rationality, and to link steps according to epistemic
and teleological rationality. Globally, the teacher may use this structure to illustrate
the pertinence of steps, inaccordance with teleological rationality.

5. FROM THEORETICAL CONSIDERATIONS TO CLASSROOM WORK

In this Section we present educational choices related to our theoretical framework
and the analysis of some excerpts from teaching experiments that were planned and
performed accordingly. The teaching experiments were performed in two middle
school classes (age of the students: 12-13 in Italy and 13-14 in France) in spring
2009. Their aims were twofold: as concerns students, we aimed to introduce them to
mathematical argumentation and proof and to make them aware of some relevant
features of proof; as concerns research, we wanted to ascertain the potential of our
theoretical constructs to plan and analyze classroom work.
5.1 The teaching experiments: educational choices
When moving from theoretical framing to educational implementation, we needed to
make some general and specific choices regarding the methodology of work in the
classroom and the organization of the activities. We now present the educational
choices that shaped our teaching experiments, showing their coherency with the
theoretical assumptions presented in the previous Sections. As mentioned above,
argumentation is crucial both as a part of the proving process and as a means for
fostering reflection on the practices of mathematical proof related to different modes
of reasoning and to the components of rationality. Consequently, in our teaching
experiments we devoted special attention to argumentative activities (at the content
level and at the meta-level): giving reasons for the validity of a statement
(argumentation with a strong epistemic component), but also unpacking the reasons
behind a strategic choice, a way of presenting a solution etc (where also the
teleological and communicational components are deeply involved). We hypothesize
that argumentation is fostered by an interplay between classroom discussions and

PME 34 - 2010 1- 15
Last names of authors, in order on the paper

individual productions, which can follow each other according to a cyclic

organization. Next, we briefly present and discuss our didactical choices, stressing
the importance of those cycles of activities.
Mathematical discussion
Mathematical discussions orchestrated by the teacher (Bartolini Bussi, 1996)
stimulate efforts of expression and explanation and allow a focus on the three
components of rationality, favouring the consciousness of logical rules and their
range of validity. Moreover, during the discussion a student may discover others’
positions and conceptions and learn how to position herself in regard to these, thus
evolving her conceptions and/or processes. Discussing a statement may bring
students to methodological and meta - level reflections on issues such as the different
role and value of an example in the three reasoning modes (producing an example to
support the statement can be an effective step in the exploratory phase, but it is not a
valid argument when organising a general mathematical justification). Also, the
relation between arguments and the construction of lines of argumentation (mode 2)
can be discussed, thus drawing students' attention to the teleological goal of the line
of argumentation in relation to its steps.
Individual text production
Based on a Vygotskian perspective, we advocate the importance of text writing in the
development of conscious handling of ideas and their organization. Among the
individual text productions dealt with in the literature, we consider “narrations of
research” (Bonaffé, 1993) useful for exploratory phases, and we advocate “report of
what I have learnt after our discussion” to foster learning evolution and
consciousness (Assude & Paquelier, 2005). We propose a specific form of individual
written production for the learning of proof: the story narration of the proof
construction.
Story narration of proof construction
After a classroom discussion, students are asked to write down an individual “story”
about the organization of reasoning that was the purpose of the discussion. We
hypothesize that this task may lead students to grasp the rationale of the proof, with
special attention paid to teleological aspects. The idea of such a task was inspired by
the following quotation from Toulmin (1974, p. 6):
Logic is concerned not with the manner of our inferring, or with questions of technique:
its primary business is a retrospective, justificatory one - with the arguments we can put
forward afterwards to make good our claim that the conclusions arrived at are acceptable,
because justifiable, conclusions.
Writing down a story is also in line with developments in education in recent
decades, which recognise the potential of narratives (Bruner, 1990): to narrate a story
means to organize facts so as to highlight their (possibly implicit) causal and
temporal links; thus, narrative may become a form of thought organization and

1- 16 PME 34 - 2010
Last names of authors in order as on the paper

content understanding (Dettori & Morselli, 2008). We believe it is important to lead

students to create a story that connects steps and propositions and fragments of
argumentation with reasons of validity that refer explicitly to epistemic warrants. In
the story narration, it should become clear that students recognize the involved lines
of argumentation, their possible hierarchical relations, and their role in the logical
combination that produces the proof. Thus, narration should make clear students'
teleological aims. At this level, communication efforts should not yet be subject to
production rules for mathematical texts, but instead should respond to the need of
mutual understanding.
Interplay between discussion and individual story narration
Following Vygotsky's internalisation principle, we consider that story narration tasks
must be combined with classroom discussions according to a cyclic organization: a
discussion having a collective construction role should be followed by an individual
written story narration having the role of internalisation and personal reorganization
of ideas. Finally, a discussion on the produced stories should close the "cycle" by
fostering reflection on the crucial elements of a proof and on the strategic choices
needed to achieve it.
During early stages of proof learning, story narration should be prepared by suitable
tasks to gradually build a specific didactical contract, elicit essential requirements,
and develop related competencies. For instance, a narration of a performed
exploration may be followed by a discussion starting with teleological matters ("why
did you decide to calculate...?"). After students produce their first story narration, the
class should compare those texts. Later, the story narration may follow a collective
oral organisation of reasons in the style of the expected final text.
The idea of the interplay between mathematical discussion and story narration can be
compared to Duval's model of proof teaching (Duval, 1991). He proposes the
production of a "proof graph" avoiding linguistic organisation of the propositions,
followed by the production of a text based on the interpretation of the graph. His goal
is to teach deductive reasoning by drawing the learner's attention to the operative
status of propositions (excluding their content). within contrast to Duval's model, our
aim is the organization of reasoning (mode 2) based on the content of propositions. In
fact, we situate ourselves in a different epistemological perspective of proof
construction: Duval's model of deductive reasoning is formal derivation, while for us
it is only a model for the final product, not adequate for the school approach to
theorems and proof (cf. Thurston's position on the priority of proposition content in
proof construction and proof checking: Thurston, 1994).
5.2 The teaching experiments: some examples
Next, we will present three activities that concern important stages of the students'
path towards the culture of theorems. All activities combine exploration, conjecture
and proof. Our aim is to show how the integrated theoretical framework is used both

PME 34 - 2010 1- 17
Last names of authors, in order on the paper

in the planning and in the a posteriori analysis of classroom activities. The examples
also illustrate argumentation at the meta-level.
First example: explorations in elementary number theory
We will show how, within a mathematical discussion, teacher's interventions are
crucial in injecting into the debate elements of meta-mathematical argumentation
concerning teleological or communicational reasons. The whole sequence was
developed over a period of three months and encompassed exploration, conjecture
and justification of 8 open problems. For each problem, a cycle of individual or group
activities and mathematical discussions was planned. The students mainly
encountered situations of cognitive unity: they could, in the proving phase, exploit
the arguments that had previously emerged during the conjecturing phase. The aims
for students were to build a suitable didactical contract and develop an argumentative
attitude. In the first activity introducing algebra as a proving tool, students were given
the following task:
The teacher proposes the following game: Choose a number, double it, add 5, take away
the chosen number, add 8, take away 2, take away the chosen number, take away 1.
Without knowing the number that you initially chose, is it possible for the teacher to
guess the result of the game? If yes, in what way?
The students worked individually, and afterwards they shared and compared their
solutions, first in small groups and then within a class discussion. All the groups
stated that the teacher can guess the results, because the result is always 10,
independent of the chosen number; some groups even tried to find out reasons why
the result is always 10, as evidenced by Group B's text:
Group B: With any chosen number, the result is 10 because multiplying by 2 is
equivalent to adding twice the chosen number, the same number that after
must be taken away twice, which gives zero, and doing the other
calculations, even in a different order, you always get 10.
In the subsequent class discussion, students were led to shift from answering that the
teacher can guess the result because it is always 10, to understanding why the result is
always 10. This means, in our perspective, to consciously search for warrants for the
conclusion “the teacher can guess the result without knowing the chosen number”.
During the first group comparison, some students had been able to find a reason why
but were not able to communicate the reason to their classmates. Students had
realized that solutions in natural language were not always effective in
communicating the reasons to the others. This paved the way to the subsequent task,
aimed at proposing algebra as a proving tool: "Write the game in the form of an
expression, using a different colour for the chosen number. Write an expression that
works for any number you choose". The students solved the task individually, and
afterwards they shared and compared their solutions in a class discussion. Two main
representations of the game were singled out: the representation given by Ric:

1- 18 PME 34 - 2010
Last names of authors in order as on the paper

N*2+5-N+8-2-N-1=10
and a sequential representation of the game, by Tor (originally arranged vertically):
N*2=A; A+5=B; B-N=C; C+8=D; D-2=E; E-N=F; F-1=10
A significant discussion comparing Ric's and Tor's representations of the problem
was fostered by the teacher's question: "In your opinion, which of the two
representations would be chosen by a mathematician?" The teacher wanted to
highlight the importance of strategic choices, such as the choice of the representation,
thus emphasizing the teleological dimension. This is an example of argumentation on
a meta level, where focus is not on the task itself, but rather on the way of solving the
task and the reasons for choosing such a way.
During the discussion, the students shared their motivations for choosing one of the
two representations. For instance, Mus’s choice (“For me, it is worthwhile to use
Ric’s representation because it is more schematic and more mathematical”) relies on
communicational warrants, while Alex’s choice (“First of all because [Ric’s] follows
the text more, and after because it is more correct”) relies on intertwined epistemic
and communicational warrants. Some students, like Giam, support Tor’s
representation on the basis of communicative warrants: “Tor’s representation,
anybody can do it, and he can follow all the steps, while in that of Ric, yes you do it,
but you don’t really realize what you are doing”.
Indeed, both representations are correct from a mathematical point of view (thus
meeting the requirements of epistemic rationality) and are also perceived as effective
from a communicational point of view. The point is that Ric’s expression is more
efficient for the original aim of the task (to understand why it is possible to guess the
result of the game). This means, in terms of the model, that the Ric’s expression
better meets the teleological requirements. Reference to the task is a teleological
warrant. For the teacher and the external observer, it is important not only that the
students choose Ric’s representation, but also that they understand the reasons why
Ric’s expression is more suitable (since it allows us to understand why the result is
always 10). The observer’s and teacher’s interventions during the discussion focus on
these reasons:
Observer: you all said a lot of good things, actually doing one or the other is the same,
and in both cases you get the result, OK? But do you remember the
question of last session? The question was not “what is the result”, but
“will the teacher be able to guess the result? [...]
Observer: you said: yes, because you always get 10, and some of you also explained
something more, we also had some motivations why you always get 10.
Teacher: do you remember? Brac, you said it, because you said that doing NX2
means…
Brac: I mean… it is like doing… yes, it is like doing N+N.

PME 34 - 2010 1- 19
Last names of authors, in order on the paper

Teacher: N+N. in the expression written by Ric, then… there is NX2, Brac, please go to
the blackboard and write N+N under NX2. Do we all agree that it is the
same thing? And after you write all the expression: +5-N+8… And you
already noticed that… after N+N, what do I have?
Ash: -N.
Voices: two times.
Teacher: and so?
Brac: they all disappear.
Teacher: can I understand this, in Tor’s representation?
Voices: no.
Fag: but, at the end there is +8, so, the two representations are equivalent, but Ric’s is…
easier.
Teacher: but why is it easier?
Giam: because you understand that the chosen number disappears.
Teacher: because I can answer…
Observer: to the original question. Ric’s representation helps us understand why it is not
necessary to know the chosen number to get the result.
Summing up, we have seen an example of the distinction between argumentation to
solve the problem and argumentation on a meta level: the former refers to the search
for reasons for the validity of the statement, the latter is an argumentation on the
proof itself. The discussion of the strategic choices made (for instance, the choice of a
suitable representation to carry out the proof), as well as the discussion on the way of
presenting the final proof, are part of such an argumentation. Teleological rationality
is emphasized, then communicational rationality intervenes.
Second example: conjecture and proof with no cognitive unity, in geometry
We present this example to discuss elements of argumentation on a meta level that
should characterize individual story narration. The students were asked to conjecture
about a triangle (say A,B,D are its vertices) inscribed in a circle (diameter [AB]). The
exploration, carried out within a DGS environment, was rather poor, but all the
students conjectured that the triangle would be right-angled. Afterward, they were
asked to prove the conjecture. The information provided by the DGS (measurement
of the angles) allowed students to produce the conjecture without any theoretical
considerations. In order to prove the conjecture, it was helpful to create a symmetric
point to D (the third point of the triangle) and to consider the rectangle completing
the triangle. The lack of cognitive unity is due to this supplementary construction. If
we analyze the expected proof in terms of blocks and lines of argumentation, we can
see three blocks: a preparatory block (“let us construct the point D', and consider the
quadrilateral ADBD'”), related to the mathematical practice of supplementary or

1- 20 PME 34 - 2010
Last names of authors in order as on the paper

intermediate construction; a second block (“to prove that the built quadrilateral is a
rectangle”); and a final block (“since the quadrilateral is a rectangle, the triangle is
right-angled”).
When the students tried to prove the statement, the teacher helped most of the
students by suggesting: “Wouldn't a right-angled triangle be half a rectangle?” (in
French a right-angled triangle is called “triangle rectangle”).In this way, the teacher
wanted to lead the students to visualize the triangle as part of a rectangle, thus
suggesting the supplementary construction. During the individual work, most
students found useful arguments, but not ones sufficient to form a proof. Many
argued that the triangle was right-angled because the measure of the angle provided
by the DGS was 90°. Few students produced a proof like the intended one.
As a starting point for the subsequent mathematical discussion, the teacher selected
some arguments previously provided by the students (both incorrect and correct ones)
and presented them to the class. For each argument, the class was asked to answer a
series of questions: is it true? Under which conditions? How can we prove it is true?
What will this argument be useful for? Is it expressed well? If we need to express it
differently, then how? The aim was to foster the explicit emergence of the three
components of rationality. After the discussion, students were asked to narrate the
story of the reasoning involved in the proof of the conjecture. Here are two texts:
Text 1:
We produced the following conjecture: it seems that, given that [AB] is a diameter of the
circle and D a point on the circle, the angle must be right. First we can say that the angle
measures 90°, when we move the circle or when we move D while proving that ABD is
always half of a rectangle. To prove this we should draw the symmetric of D (D' on the
figure), to prove that the quadrilateral is a rectangle we use the diagonals of the
quadrilateral [AB] and [DD']. Because thanks to a property of the rectangle, if the
diagonals have the same measure then it is a rectangle. Moreover, we can also say that a
rectangle has four angles and that they are all right. So simply, all the angles are right,
and the angle is also right. In fact this angle is one of the right angles of the rectangle
ADBD', this is why it is 90°.
Text 2:
We want to prove that the angle is right. We can see that the triangle ADB is half of a
rectangle, to prove this we tried to take the angle and move it on the circle, then to widen
and reduce the circle, concluding the angle remains always 90°. To prove it we draw the
symmetric of the point D in relation to O. We can see that [AB] and [DD'] cross each
other in their middle and have same length, thus [AB] and [DD'] are diagonals because
they are diameters. If we draw [AD'], [BD'] and [DD'] we get a rectangle AD'BD, this
quadrilateral has four right angles, and we know that if in a quadrilateral there are four
right angles, then it is a rectangle. Conclusion: (no conclusion was written).
Before discussing briefly the two texts, we provide some comments on our criteria of
analysis. We were interested in the explicit presence of epistemic reasons, and these

PME 34 - 2010 1- 21
Last names of authors, in order on the paper

were expected to be backed, as in the previous discussion. Communicational and

teleological reasons addressed by the teacher during the previous discussion were
expected to guide the production of the stories and possibly to be mentioned. Even if
present, they were not expected to be backed (in a “normal” mathematical proof text,
the teleological choices are not expressed, and certainly not backed!).
In text 1, Habermas' components of rationality all seem to be present: they are
weaker at the beginning of the text, stronger at the end. The student enchains two
arguments of different natures: moving the circle or point D; and proving that the
triangle is half a rectangle, thus seeming unaware of the fact that they have different
teleological uses at the level of proof creation and conjecturing. In exploration,
drawing conclusions from observations is allowed and may be used to produce a
conjecture; when producing and organizing arguments, observations cannot be used
to draw conclusions, but proving that we have a rectangle can be used. These
teleological alternatives reveal the student's epistemic reasons. The fact that her proof
seems to rely on the results of DGS exploration, as well as on the reference to
mathematical properties, reveals a fragile epistemic rationality. However, she seems
conscious of the need for a strategy to prove based on mathematical reference
statements (a teleological perspective to organize proof arguments), since she begins
her next step by “to prove this...”. In the central part, the narration seems to structure
the text; teleological reasons guide her when referring to the diagonals. These reasons
are rapidly ascertained by the epistemic reason explicitly backing this linking of
arguments. Afterward, epistemic reasons allow her to draw a conclusion about the
right angles, but the argumentation is not explicit (backing is not exposed). The
repetition of arguments and the familiar language used by the student reveal that she
is gradually understanding the strategy, and she is conscious that she is now able to
make conclusions.
As regards text 2, we may say that the student absorbed from the previous discussion
the need to make a strategy explicit, but he cannot handle these teleological purposes.
The first sentence shows a teleological posture (We want to prove that the angle is
right). The epistemic reasoning is not acceptable in mathematics (the backing is taken
from observation and experience). Moreover, the student wants to give a “second
proof”, unaware of another proving rule (one argumentation is sufficient) and of the
uses of epistemic arguments. He is aware of the need to show that constructing D' is a
step in the proof, and this again suggests teleological rationality in action. The
subsequent sentence (“we can see that [AB] and [DD']... because they are
diameters”) provides a wrong reason of validity for the properties of [AB] and [DD'].
The next assertion (AD'BD is a rectangle), which is not justified, is followed by a
deductive reasoning relying implicitly upon mathematical backings; afterwards, he
turns back from this deduction (it has 4 right angles) to the original assertion (it is a
rectangle).

1- 22 PME 34 - 2010
Last names of authors in order as on the paper

Both texts were selected and given to all students, who were asked to compare them
during a class discussion. All the students, despite their difficulties in reading and
understanding their classmates' production, recognized many similarities and
differences. They preferred text 1 (“this is good, she explains it well”), but nobody
addressed the mathematical validity of the arguments or the validity of the
mathematical enchainment. Many students noted that there was no conclusion in text
2.
Third example: Pythagoras theorem
Finally, we present a sequence designed to introduce Pythagoras’ theorem. During
the last school year, the sequence was tested for the first time, and a second teaching
experiment is currently underway. Once again, the situation lacks cognitive unity. It
is not difficult to get the conjecture through a loosely-guided path, but constructing a
proof requires strong guidance by the teacher: changes of frames are needed (from
considering lengths to considering surfaces; from geometrical to algebraic register).
Moreover, there are side argumentation blocks that must be hierarchically organized
within the main argumentation line (and this needs a teleological handling). Teachers'
guidance, classroom discussions and story making should allow students to approach
the rationale of the proof and offer occasions for learning about proof and proving.
To this aim, we planned a series of activities organized into two sequences. The first
sequence consists of two different tasks. Its final aim is to conjecture and make sense
of Pythagoras' theorem. Task 1 is preliminary: the students are guided to check and
justify the triangle inequality, and afterwards they are asked to narrate the story of
their reasoning. Its aim is to focus on change of frame (geometric- algebraic), and on
the additive relations involved. It is followed by classroom discussion.
Task1: Consider the statement: "In a triangle with sides a, b and c, a+b is always smaller
than c". Is it true? always? Why? Prepare yourself to explain how you checked it and
why you think it is true, or it is not, or what makes you doubt.
This task also tries to foster exploration through examples, and (especially in the
discussion) to lead students to express the rationale of the activity and see the
generality of the proposition. Expressions such as “we wanted to see if it is true that...
so we tried to verify it with four examples” are encouraged: such simple narrations
reflect an ability to reconstruct the logical skeleton of the activity they went through.
This connects a Mode 1 reasoning with a Mode 2, and prepares Task 2.
Task 2 (individual): If we consider the squares of the lengths, instead of the lengths
themselves, the situation is different. See if a relation between the squares of the lengths
of the sides of a triangle exists. Once you think you produced a valid statement (a
"conjecture"), put it clearly in words to explain it to your classmates.
Together with the text, the worksheet contains also drawings of right-angled, acute
and obtuse triangles, presented with measures of their sides. After the individual
solution, students are invited to share and discuss their conjectures within a
mathematical discussion. Attention is paid to the different ways of performing the

PME 34 - 2010 1- 23
Last names of authors, in order on the paper

exploration and to the different acceptable expressions of the conjecture(s) (according

to mathematical standards). Incomplete or erroneous conjectures may offer fine
opportunities to make explicit the important elements of the theorem (in particular,
the condition of validity of Pythagoras' theorem, i.e. an angle being right).
Task 3 (individual): Write down the conjecture as you now think it should be. Explain it
and illustrate it with some examples.
The teacher concludes with the standard formulation of Pythagoras' theorem. This
first sequence aims at involving students in Modes of reasoning 1 (drawing,
measuring, calculating, …) and 2 (exploiting the gathered data, noticing regularities,
expressing a general result, discussing and justifying propositions, and organizing the
steps of exploration in relation to a goal).
The second sequence aims at guiding the students to prove Pythagoras' theorem.
Given that cognitive unity is not possible, students are guided by means of individual
and collective activities through the different blocks of the proof; afterward, they are
asked to reconstruct the lines of argumentation.
Task 4 (individual): Here we study the proof of the theorem we have conjectured, you
will be guided towards this proof. Consider a right angled triangle with sides a, b, c. We
use it to build the square A (see below). The area of its central square S is c2.

I) Can you describe how A can be obtained by using only our right triangle? explain why
S is a square (of area c2)?
II) Try to write the area of A in two different ways (you may need to arrange the four
identical right triangles differently). Find and explain the two ways.
III) How can this help us to validate our conjecture?
Geometrical reasoning is expected to intertwine with algebraic reasoning in order to
demonstrate the equality between the areas. If needed, some supplementary tasks can
be inserted either for the whole group or for some students.
After the individual work, the teacher orchestrates a mathematical discussion.
Attention is paid to the reasoning behind the steps of argumentation and calculation,
to the necessity of such steps, and to the connection between geometrical and
algebraic arguments. The discussion is intertwined with methodological reflection
about the validity of the reasoning, its communicability, and its acceptability by an
external reader.

1- 24 PME 34 - 2010
Last names of authors in order as on the paper

We may notice that the formulation of task 4 is designed to pave the way to the
subsequent story narration of the proof. The subsequent discussion is meant to
prepare students to write a “story” (task 5).
Task 5 (individual): Write down how you organized your steps of reasoning to reach a
general justification of the conjecture, and justify why those steps are important.
This task should allow students to grasp and reconstruct the rationale of the proof.
After the individual story narration, a selection of stories is provided by the teacher to
the whole class and discussed.

6. DISCUSSION AND FURTHER DEVELOPMENTS

The complexity of proving from cognitive, epistemological and educational points of
view stresses the need for comprehensive frameworks to analyze students' behavior
and guide teachers' classroom activities regarding proving. In recent years, our
adaptations of Toulmin's model for argumentation and of Habermas' construct of
rational behavior to the case of proving have represented partial contributions in that
direction. By integrating the two constructs, we aim at getting a more powerful tool
that will address the needs for a more complete analysis of students' behavior and, at
the same time, will better frame the role of the teacher as responsible for students'
enculturation in the culture of theorems. The integration is based on the idea that
those elements of awareness that characterize the expert's rational behavior in
proving are inherent in an argumentation at the meta-level (with its specific
warrants). This awareness drives and controls the argumentation at the specific level
of proving (epistemic validity of arguments, effectiveness of tools and methods,
efficiency of communication). Argumentation at the meta-level must be passed over
to students through specific educational devices (like the interplay between classroom
mathematical discussion and individual story-narration of proof construction). Our
examples outline a long-term teaching intervention intended to promote students'
enculturation in the culture of theorems (including those aspects of awareness that
characterize experts' rational behavior in proving).
The integration, and the resulting experimental activities concerning the crucial phase
of the organization of reasoning, suggest further research questions that need to be
dealt with in different ways. In particular, on the theoretical side we need:
- to characterize the warrants (and the argumentation at the meta-level) for the other
phases of the proving process (the exploration, and the revision of the proof);
- to establish more precise relationships between the warrants of argumentation at the
meta-level and the components of Habermas' rationality for the different phases of
the proving process;
- to study the relationships between the mastery of meta-mathematical knowledge (in
terms of warrants for argumentation at the meta-level) and meta-cognition: indeed,

PME 34 - 2010 1- 25
Last names of authors, in order on the paper

meta-cognition seems to largely depend, in the case of proving activities, on meta-

mathematical knowledge.
Further insight into these issues would not only represent an advancement in our
theoretical perspectives on the activity of proving, but also would allow us to design
better tasks for students and suggest more effective teaching interventions in
classroom discussions.
In terms of the educational side, the management of the didactical contract in the plan
aimed at promoting students' maturation within the meta-level of argumentation
reveals an issue that is difficult for teachers to deal with, because focusing on the
meta-level of argumentation may "distract" students from the main goal of justifying
the validity of statements (a crucial goal not only for proving, but in all kinds of
activities that enhance students' intellectual growth).
Finally, preparing teachers to manage the activities illustrated in the previous Section
will be a challenge. If appropriately guided by the instructor, proving activities at the
adult level might provide an opportunity to see the meta-level of argumentation
(through the instructor’s guide). Also, analyzing students' performances in ordinary
as well in innovative activities of proving might enhance teachers' competencies in
identifying students' difficulties and finding ways to overcome them.

7. A REACTION TO BOERO’S ET AL. DOCUMENT

Our research lies in student argumentation during activity carried out to formulate a
conjecture, find reasons it must be true, and construct a proof in similar ways as
experts would. Through the following analysis of an episode within an experiment
involving pre-service math teachers, we aim to provide elements for the study of the
utility of the integrated model. Specifically, we want to illustrate two aspects: some
imprecision in the description of Habermas’ model which, we think, requires more
elaboration, and the difficulties in using the integrated model when analyzing
ongoing argumentation.
The students were in their third undergraduate geometry course when the experiment
was carried out. This course and the previous one had three distinguishing
characteristics: the axiomatic system used was collectively constructed by the group
through an inquiry community of practice, the theoretical elements of the system
arose from the conjectures students formulated as solutions to specially designed
situations, and a dynamic geometry program was used as a mediation tool for the
process of learning to prove within that axiomatic system. At one point, the students
were given the following problem:
With Cabri, construct a circle with center C and a fixed point P in its interior. For which
chord AB of the circle, that contains point P, is the product AP × PB maximum?
They modeled the situation, dragged and discovered that the product does not vary.
We use the integrated model to analyze an excerpt of the interaction within a group

1- 26 PME 34 - 2010
Last names of authors in order as on the paper

of three students when they formulated and outlined the path to construct a proof for
their conjecture: Given a circle with center C, a fixed point P which belongs to the
interior and a given chord AB which contains P, then the product AP × BP is
constant. Once the conjecture was written, Nancy suggested substituting the phrase
“a given chord AB” by “any chord AB”. This incident brings forth the
communicative aspect of the Habermas model: we see a conscious use of the
universal quantifier when they change a particular statement to a general one, and
thus an adherence to rules established for the formulation of conjectures. Later in the
intervention [line 193 below], Nancy’s reference to the word “any” seems to acquire
an epistemic sense because through it she manifests that obtaining an equal product
for the second chord validates the claimed invariance. This and our analysis of
excerpts where our students discuss whether words such as “let”, “exist”,
“determine”, “localize” or “choose” are the adequate ones to use when expressing an
idea leads us to ask: Is this a part of the communicative or epistemic aspect? Is using
schemes to organize a proof, such as two columns, considered a communicative
aspect of an argument?
162 Alejandro: In the proof we can construct another chord, right? [Assenting murmurs.] To have
similar triangles.
168 Fabián: In every case we get similarity? Or can we achieve congruency at some point,
with two chords?
169 Nancy: It would be better to use congruency to prove it is always equal.
170 Alejandro: No, because…what we need to prove is a ratio.
171 Fabián: Yes, a product.
172 Obs: The what?
173 Alejandro: The ratio between [Moving index finger over the chord in the calculator
construction.] segments AP and PB.
174 Fabián: Then better similarity.
176 Alejandro: Because the only moment in which they can be congruent is when P is the center.
177 Nancy: Also, yes.
178 Fabián: That is why I ask if in some point they would be congruent but then when they are
congruent they will be similar.
179 Nancy: Yes, it’s the same.
180 Fabián: Anyway it’s the same similar as… but it is better similar than congruent.
181 Obs: I want to ask a question: you Alejandro said something about creating another
chord? … And why did you think of that?
186 Alejandro: Because since we see the product will always be the same, right? [With index
finger points to the constructed chord on the screen.] Then another chord [points
to a chord that hasn’t been constructed.] can give us similarity or ratio between
this side, this segment that we would create new and this [Points to the imaginary

PME 34 - 2010 1- 27
Last names of authors, in order on the paper

chord..], that… [Starts constructing the other chord.], wait, we’ll create it
190 Alejandro: [Constructing the other chord on the computer.] This point… on the circle… then,
here we create similar triangles [moves index sketching a triangle on the screen.]
191 Nancy: We would have that the ratio of … or that by ratios we get that AP times BP is
going to be the same for both chords …
192 Fabián: It will always be the same.
193 Nancy: And that way we would confirm that it would be for any chord [Pause.]
195 Nancy: Then it wouldn’t be only for the one we use; we also compare it with another one.
The students have set their goal: proving that the product is constant for any chord.
This leads them to construct two chords that contain P with the purpose of obtaining
similar triangles, which allow them to work with ratios that will lead to equal
products. This goal motivates an auxiliary construction without which it is practically
impossible to prove the conjecture. Therefore, we recognize the teleological aspect in
their argumentation because they sketch a plan to reach their goal and propose an
auxiliary construction as a tool to obtain it. They have at the same time evoked
similar triangles as the idea within which they feel sure to find the crucial step of
their proof. This leads all their latter actions in a process towards constructing the
justification, and we identify epistemic rationality in these actions. These
observations suggest that the difference between the teleological and epistemic
aspects is that the first one lies in simply suggesting a path and the latter in justifying
the choice. We ask ourselves two questions: Besides referring to the conscious and
appropriate validation of the propositions, does the epistemic aspect include
considerations about the theoretical status of the statements mentioned? In which of
the aforementioned rationalities do we set the identification of the theoretical field
that is used to sketch the plan for a proof?
We have two different ways of interpreting Toulmin’s model in this excerpt. For the
first one, we recognize that using Toulmin’s model requires that the three elements of
the argument be explicit. Due to the fact that this is an ongoing argument, we think it
is possible to take the liberty to see the process as follows: asking whether another
chord can be constructed in the proof is asking whether it is useful for the proof. In
that sense, it is the expression of missing data and therefore of abductive reasoning.
Although the warrant is not given explicitly, the claim (170) is equal ratios. Being
able to convert this argument into a deductive one is difficult because the data is not a
statement but an idea for an auxiliary construction. This is why, when the students
report to the observer how they thought of constructing another chord, the deductive
argument uses three other elements: having similarity (data) gives ratios and, through
these, products (claim). The problem is that there is no warrant. If we look at this
situation in a more rigorous form, we can say that they are evoking theory and not
constructing an argument.

1- 28 PME 34 - 2010
Last names of authors in order as on the paper

Another possible analysis is that, initially, they seem to assert that the two chords that
contain P determine similar triangles (claim) [162]; no warrant is given, and instead
the generality of the statement is questioned (rebuttal) [168]. Up to that point, a
mathematical statement is formulated but no argumentation given. Nancy’s
contributions detour the conversation because, in fragment [169-180], she refers to
the relevancy of similar/congruent triangles for the proof. Taking into account the
discovered invariance of the product (data), it is asserted/proposed that the use of
similar triangles will allow completion of the proof (claim) because, from it, equal
ratios can be deduced and from these equal products (warrant). The following
questions arise: what is the structure of this argument? Can this argument be outlined
by Toulmin’s model? Would it be of interest to outline arguments such as the one
presented? With respect to the use of Toulmin’s model, would it be convenient to
specify with greater precision the types of arguments that can be modeled? Does the
difficulty in using the model lie in the fact that the students are simultaneously trying
to find what is pertinent and trying to construct a justification?
Notes: Sections 1, 4, 6 are by P. Boero, N. Douek, F. Morselli, B. Pedemonte; Section 2 is by B.
Pedemonte; Section 3 is by P. Boero and F. Morselli; Section 5 is by N. Douek and F. Morselli;
Section 7 is by L. Camargo, P. Perry, C. Samper.
References
Arsac, G. (1988). L'origine de la démonstration: essai d'épistémologie didactique.
Recherches en Didactique des Mathématiques, 8(3), 267-312.
Assude, T & Paquelier, Y (2005). Souvenirs et approche temporelle des apprentissages
mathématiques. Revue Canadienne de l’Enseignement des Sciences, des Mathématiques
et des Technologies. Vol.5.2, 223-239.
Arzarello, F. (2007). The proof in the 20th century. In P. Boero (Ed.), Theorems in school
(pp. 43-63). Rotterdam: Sense Publishers.
Balacheff, N. (1987). Processus de preuves et situations de validation. Educational Studies
in Mathematics, 18(2), 147-176.
Bartolini Bussi, M. G. (1996), Mathematical Discussion and Perspective Drawing in
Primary School. Educational Studies in Mathematics, 31 (1-2), 11-41.
Boero, P. (2006). Habermas’ theory of rationality as a comprehensive frame for
conjecturing and proving in school. Proceedings. of PME 30, vol.2, pp. 185-192. Prague,
Czech Republic: PME.
Boero, P.; Douek, N. & Ferrari, P. L. (2008). Developing mastery of natural language. In L.
English (Ed.), International Handbook of Research in Mathematics Education (pp. 262-
295). New York: Routledge.
Boero, P. & Morselli, F. (2009). Towards a comprehensive frame for the use of algebraic
language in mathematical modelling and proving. Proceedings of PME 33, vol. 2, pp.
185-192. Thessaloniki, Greece: PME.
Bonafé, F. (1993). Narration de recherche. Repères IREM n°12.

PME 34 - 2010 1- 29
Last names of authors, in order on the paper

Bruner, J. (1990). Acts of meaning. Harvard University Press, Cambridge, MA.

Dettori, G. & Morselli, F. (2008). Accessing knowledge through narrative context. Learning
to Live in the Knowledge Society, 253-260
Douek, N. (2009). Approaching proof in school: from guided conjecturing and proving to a
story of proof construction. Proceedings of the ICMI Study 19 Conference, vol. 1, pp.
142-147.
Duval, R. (1991). Structure du raisonnement déductif et apprentissage de la démonstration.
Educational Studies in Mathematics, 22(3), 233-261.
Garuti, R.; Boero, P.; Lemut, E. & Mariotti, M. A. (1996). Challenging the traditional
school approach to theorems: a hypothesis about the cognitive unity of theorems.
Proceedings of PME 20, vol. 2, pp. 113-120. Valencia, Spain: PME.
Habermas, J. (2003). Truth and justification. Cambridge (MA): MIT Press.
Hanna, G. (1989). More than formal proof. For the Learning of Mathematics 9(1), 20-23.
Harel, G. (2008). DNR Perspective on mathematics curriculum and instruction: focus on
proving, part I. ZDM, 40, 487–500.
Krummheuer G. (1995) The ethnography of argumentation. In P. Cobb & H. Bauersfeld
(Eds.), The emergence of mathematical meaning: Interaction in classroom cultures (pp.
229-269). Hillsdale, NJ: Lawrence Erlbaum Associates.
Morselli, F. & Boero, P. (2009). Habermas' construct of rational behaviour as a
comprehensive frame for research on the teaching and learning of proof. Proceedings of
the ICMI Study 19 Conference, vol. 2, pp. 100-105.
Pedemonte B. (2005): Quelques outils pour l’analyse cognitive du rapport entre
argumentation et démonstration. Recherches en didactique des mathématiques 25, 3,
313-348
Pedemonte, B. (2007). How can the relationship between argumentation and proof be
analysed? Educational Studies in Mathematics, 66(1), 23-41.
Pedemonte, B. (2008). Argumentation and algebraic proof. ZDM, 40(3), 385-400.
Plantin C. (1990). Essais sur l'argumentation, Kimé (ed.), Paris.
Toulmin, S. (1974). The uses of arguments. Cambridge: Cambridge U. P.
Thurston, W. P. (1994). On proof and progress in mathematics. Bulletin of the American
Mathematical Society, 30, 161-177.
Wood, T. (1999). Creating a Context for Argument in Mathematics Class. Young Children's
Concepts of Shape. Journal for research in mathematics education, 30(2), 171-191.

1- 30 PME 34 - 2010