Vous êtes sur la page 1sur 16

Towards a Problematique for Research on Mathematics Teaching

Author(s): Nicolas Balacheff


Source: Journal for Research in Mathematics Education, Vol. 21, No. 4 (Jul., 1990), pp. 258-
272
Published by: National Council of Teachers of Mathematics
Stable URL: http://www.jstor.org/stable/749524
Accessed: 02/10/2008 21:22

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at
http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless
you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you
may use content in the JSTOR archive only for your personal, non-commercial use.

Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at
http://www.jstor.org/action/showPublisher?publisherCode=nctm.

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed
page of such transmission.

JSTOR is a not-for-profit organization founded in 1995 to build trusted digital archives for scholarship. We work with the
scholarly community to preserve their work and the materials they rely upon, and to build a common research platform that
promotes the discovery and use of these resources. For more information about JSTOR, please contact support@jstor.org.

National Council of Teachers of Mathematics is collaborating with JSTOR to digitize, preserve and extend
access to Journal for Research in Mathematics Education.

http://www.jstor.org
Journalfor Researchin MathematicsEducation
1990, Vol. 21, No. 4, 258-272

TOWARDSA PROBLEMATIQUE
FOR RESEARCH
ON MATHEMATICSTEACHING

NICOLAS BALACHEFF, Laboratoire IRPEACS, CNRS

This articlepresentsthe main featuresof the theoreticalframeworkof Frenchresearchknown


as recherches en didactique des mathematiques. The foundation of this approach consists
mainly of the relationshipsbetween two hypotheses and two constraints,which are presented
togetherwith some specific key words.Outlinesaregiven of Brousseau'sthdoriedes situations
didactiques(theoryof didacticalsituations).An example is given that presentsin some detail
the rationalefor the constructionof a didacticalsituationand its analysis.This articleends with
some questions addressedto researchon mathematicsteaching.

Kilpatrick(1981) pointedout some years ago that"one of our greatestneeds in


researchideas on mathematicallearningand thinkingis for conceptual, theory-
building analyses of the assumptionswe are using in our research"(p. 370). It
could be addedthatthatsearchfor theoriesis not sufficient,insofaras theoriesare
of no use if they are not relatedto precise problems.To say thatour problemis to
improvemathematicsteachingor even the teachingof algebra,or thatone of our
problems is pupils' difficulties in thinkingmathematically,is too vague. First of
all, theoriesare tools eitherto solve problemsor to clarifythem and improvetheir
formulation.Inversely,to solve researchproblemsvery often leads to the improve-
ment of theories,or at least it puts them underquestion;and sometimesit leads us
to considerthe need for new theories.This fundamentaldialecticbetween theories
and researchproblems is at the core of the approachto researchon mathematics
teachingI would like to presenthere.
A prioblmatique is a set of researchquestions relatedto a specific theoretical
framework.It refers to the criteriawe use to assert that these researchquestions
are to be consideredand to the way we formulatethem. It is not sufficientthatthe
subjectmatterbeing studied is mathematicsfor one to assert that such a study is
researchon mathematicsteaching.A problembelongs to a probldmatiqueof re-
search on mathematicsteaching if it is specifically related to the mathematical
meaningof pupils' behaviorin the mathematicsclassroom.In this articleI present
the main featuresof such a prioblmatique and an example of a researchdone in
this framework.

The materialin this article was an invited addressat the researchpresession of the 65th
annualmeeting of the National Council of Teachersof Mathematics,Anaheim, CA, April
1987, writtenwhen I was a memberof the Equipede Rechercheen Didactiquedes Math6ma-
tiqueset de l'InformatiquefromGrenoble;its contenthas had the benefitof discussions with
many of my colleagues there. I deeply appreciatediscussions with Jere Confrey and her
comments on the earlierversion of this article.I would also like to thankJeremyKilpatrick
for his comments and editing remarksthat helped me to carryout this final version.
259

ANDTWOCONSTRAINTS
A PROBLEM,
TWOBASICHYPOTHESES,

Our theoreticalframeworkis groundedon two hypotheses:The constructivist


hypothesis and the epistemological hypothesis.
The constructivisthypothesisis thatpupilsconstructtheirown knowledge,their
own meaning.The fact thatpreviousknowledgeis questioned,the disequilibration
in the Piagetiansense, resultsin the constructionof new knowledge as a necessary
response to the pupils' environment.
The epistemologicalhypothesis(Vergnaud,1982) is thatproblemsarethe source
of the meaningof mathematicalknowledge, but also intellectualproductionsturn
into knowledge only if they prove to be efficient and reliable in solving problems
that have been identified as being importantpractically(they need to be solved
frequently and thus economically) or theoretically (their solution allows a new
understandingof the relatedconceptualdomain).
These two hypotheses imply thatpupils' learningdependson their recognition
and re-constructionof problems as being their own. It is not sufficient that the
teacherproposesa problemfor this problemto become thatof the pupils, because
usually the responsibilityfor what is true in the mathematicsclassroom depends
on the teacher.A problem is a problem for a studentonly if she or he takes the
responsibilityfor the validity of its solution.This transferof the responsibilityfor
truthfrom teacherto pupils must occur in orderto allow the constructionof mean-
ing. Here is our fundamental problem: What are the conditions for the devolution
of the responsibility for truthfrom teacher to pupils in the mathematics classroom?
If this devolutionprocess is achieved, then we can considerthatpupils' intellec-
tual activity is intrinsicallyjustified by the problemand not by what they thinkis
expected by the teacher.As far as learningis a personalprocess, its productwould
be privateknowledge, the pupils' conceptions.But this process conflicts with two
constraintsspecific to the teachingprocess, which has to guaranteethe socializa-
tion of pupils' conceptionsfor the following reasons:
1. Mathematicalknowledgeis a social knowledge.Pupils shouldmaketheirown
the knowledge thatexists outside the classroom.It has a social statusin society or
in smaller social groups under whose control it is used. For example, the com-
munity of mathematiciansor that of engineers can be taken as a social frame of
reference.
2. The mathematicsclass exists as a community.The teacher has to obtain a
certainhomogeneity in the meaningof the knowledge constructedby pupils, and
she or he has to ensure its coherence. Otherwise,the functioningof the class will
hardlybe possible. This constraintis quite evident if one considers the language
or the means of representationspecific to a given piece of mathematicalknowl-
edge. Because of the constructivisthypothesis,the use of authorityis not desirable.
Thus the homogenizationcan only be the result of a negotiationor of other spe-
cific social interactionssuch as the one Brousseau(1986a, 1986b) has described
to frame his thdorie des situation didactiques.
260 Towardsa Probldmatiquefor Researchon Teaching

ELEMENTS DIDACTICAL
OFBROUSSEAU'S THEORY

We considerthe aim of teachingto be carryingpupils from theirinitial concep-


tions relatedto a given item of mathematicalknowledge to resultantconceptions
throughwhat we call a didactical process. The control and the design of this di-
dacticalprocess constitutethe heartof our approach.
It follows from the two hypotheses we mentionedthat the fundamentalmeans
to initiate this process are mathematicalproblems. Mathematicalproblems are
fundamentalinsofaras they constitutemeans to challenge the pupils' initial con-
ceptions and to initiate their evolution. Also, they are fundamentalbecause they
convey the meaning of the mathematicalcontent to be taughtmainly by making
explicit the epistemologicalobstacles thatmust be overcome for the construction
of thatmeaning.
Pupils' behaviorsin the context of a classroom situationcannot be understood
only throughan analysis of the mathematicalcontent involved or its relatedpsy-
chological complexity.The problemsoffered to pupils in a didacticalsituationare
set in a social context dominatedby both explicit and implicit rules thatpermitit
to functionbut also thatgive meaningto pupils' behaviors.Forexample,consider
the case of a pupilproposinga solutionto a given problemwithoutevidence of any
attemptto base its solutionon a proof. Before one makes any diagnosisof concep-
tual understanding,cognitive level, or ability level, one has to examine whether
therewas any necessity for the pupil to give a proof in such a situation(Balacheff,
1982, 1988b).The rulesof social interactionin the mathematicsclassroominclude
such issues as the legitimacyof the problem,its connectionwith the currentclass-
room activity,and the responsibilitiesof both the teacherand pupils with respect
to what constitutesa solutionor to what is true.We call this set of rules a didacti-
cal contract.A rule belongs to the set, if it plays a role in the pupils' understand-
ing of the relatedproblemand thus in the constitutionof the knowledge they con-
struct.
Thus, the pupils' behaviorand the type of controlspupilsmay exert on the solu-
tion they producestronglydepend on the feedback given duringthe situation.If
there is no feedback, then the pupils' cognitive activity is differentfrom what it
could be in a situationin which the falsity of the solution could have seriouscon-
sequences.In this last type of situation,pupils will searchfor a proof, the level of
which could dependon both the natureof the knowledge they have availableand
the pressureof the situation.Perhapsthey will even reconsidertheirown knowl-
edge before producinga definitive answer.
Brousseau (1981) differentiatestypes of situationswith respect to the kind of
cognitive functioningthey imply. First,thereare situationsimposed by the social
constraintsI have mentioned. Brousseau calls them situationsfor institutional-
ization. They aim at pointing out, and giving an official status to, some piece of
knowledge that has been constructedduringthe classroom activity. In particular
they concernthe knowledge, symbolic representation,and so on, to be retainedfor
furtherwork. A new mathematicalconcept has to be recognized as somethingto
Nicolas Balacheff 261

be kept for furtheractivities. Otherwisepupils may soon forget it. On the other
hand, althoughmany new intellectualconstructionsmight appearduringa prob-
lem-solving process, not all of them will reach the status of knowledge to be re-
tained.Thatshows the importanceof this kindof situationwithinwhich the teacher
gives the status of knowledge to be retainedto some new intellectualconstruct.
Also, the new knowledge has to be proveduseful: It has to functionin orderto es-
tablishits practicalinterestandalso to stabilizethe new cognitive stateof the pupil.
But the processes for institutionalizationand furtheractivities (e.g., systematic
problemsand exercises) are not neutral,in the sense thatthey do not maintainthe
initial meaning constructed;throughthese activities pupils' conceptions evolve
(Boschet, 1983; Robert, 1982). The control of this evolution is a didactical
problem.
But such a statuscan be given to a piece of knowledge only if it has been con-
sidered as an object explicitly recognized and not just as a tool implicitly used in
problem-solving activities (Douady, 1985). Brousseau distinguishes two main
types of situationsthat allow one to elicit the formulationof pupils' intellectual
productions: situations for validation and situations for formulation.
Situationsfor validationrequirepupils to offer proofs and thus to formulatethe
relatedtheoriesand means underlyingtheirproblem-solvingprocesses. Situations
for decision (Balacheff, 1987) are situationsfor validationwithin which there is
an intrinsicneed for certaintybut a proof is not explicitly requested.
Situations for formulationinvolve the constructionand the acquisition of ex-
plicit models andlanguage.Situationsfor communicationare situationsfor formu-
lation with explicit social dimensions. The problemof formulationis not a mere
problem of encoding ready-madeknowledge. In a situation that specifically re-
quires a formulation(i.e., whose success depends on the quality of the formula-
tion), it appearsfrom experimentalstudies thatthe process engaged is dialectical
(Laborde,1982):The failureof a formulationchosen for the purposeof a problem-
solving strategycauses a reconsiderationof the underlyingknowledge itself, its
components,and its relationships.That is quite clear in a situationfor communi-
cation because of its social dimension.
Pupilscannotenterdirectlythe situationscharacterizedabove;beforetheknowl-
edge becomes an object of discourse, it has to exist as a tool. At that initial point,
Brousseau considers anothertype of situation:situationsfor action. These situ-
ations favor the developmentof conceptions-as models for action-necessary to
initiate the teaching-learning process or the search for a solution to a given
problem.
A KEYISSUE:PUPILS'ERRORS
Pupils' errorsare the most obvious indicationof their difficulties with mathe-
matics. The problemof the meaningof these errorsis one of the key issues in the
field of researchon mathematicsteaching.
Let us take the case of decimal numbers:To the question"Does thereexist any
real numberbetween 2.746 and 2.747?" Izorche (1977) found that about40% of
262 Towardsa Probldmatiquefor Researchon Teaching

16-year-oldpupils answeredthat "it is not possible." This type of errorhas also


been shownat the level of primaryschool andat the beginningof secondaryschool
(Perrin-Glorian,1986). In a task in which pupils had to orderdecimals, Grisvard
and Leonard(1981) found thatthe proceduresused by pupils can be describedby
the following rules: (a) The decimal that has a bigger numberto the right of the
decimal point is bigger (62% of their sample);(b) the smallernumberis the one
thathas the longerdecimalpart(16%of theirsample).The existence of these rules
has been confirmedby otherreseachers(Nesher & Peled, 1986).
The problemis not only to eliminatesuch errorsbut to identifywhattheirorigin
might be. The basic hypothesisof our theoryis thatthese errorsare not mere fail-
uresbut symptomsof specific pupils' conceptions.In the case of decimalnumbers,
a hypothesis is that pupils' conceptions can be related to the errorsmentioned
above in the following way: (a) Decimals are integers with a decimal point that
sharesome propertieswith the integers;(b) decimalsarepairsof integersseparated
by a decimal point, a conceptionthatcan also explain errorslike (2.4)2= 4.16. If
we claim thatsuch conceptionsarepartof the pupils' knowledge,we have to show
thatthey allow the pupils to solve some problemscorrectly.
For Conception(a), we considerproblemsof calculation.To succeed in learn-
ing how to calculatewith decimals,it is efficient to considerthem as integerswith
a decimalpoint.Pupilshave thenonly to learnhow to cope with the decimalpoint,
having addedor multipliedthe numbersas if they were mere integers.At a deeper
level we note thatdecimals areoften introducedto pupils in a context of measure-
ment, in which they appearto be integers with the decimal point as information
abouta chosen unit.
For Conception(b), we consider one of the algorithmsat hand for comparing
decimals: Firstyou comparethe integers writtenon the left of the decimal point,
and if they are equal you then comparethe integerswrittenon the right,provided
that they have the same numberof digits. But some pupils forget this constraint
when they comparetwo decimals. It could be arguedthat in this case pupils will
not succeed in performingcomparisontasks, so their errorswill be apparentto
them. But more often thannot the pupils do not really need to pay attentionto the
constraint,because the exercises thatareofferedto themfrequentlyhave the same
numberof digits to the rightof the decimal.Theirconception is reinforcedby the
fact that in everyday life decimals used to code a price are in fact understoodas
being a pairof integers:francsandcentimesin France,dollarsandcents in the U.S.
These descriptionsof pupils' conceptionsof decimals are hypotheticaldescrip-
tions proposedby researchers.They are validatedby experimentalmeans and by
the fact that they allow us to foresee what the pupils' productionswill be for a
given task. It is not possible to make a direct observationof pupils' conceptions
relatedto a given mathematicalconcept;one can only infer them from the obser-
vation of pupils' behaviors in specific tasks, which is one of the more difficult
methodologicalproblemswe have to face.
So if pupils' conceptions have all the propertiesof an item of knowledge, we
have to recognize that it might be because they have a domain of validity.These
Nicolas Balacheff 263

conceptionshave not been taughtas such, but it appearsthatwhathas been taught


opens the possibility for theirexistence. Thus, the question is to know whetherit
is possible to avoid a priori any possibility of pupils constructing unintended
conceptions.
THEDIDACTICAL TRANSPOSITION
I have suggestedthatpupils' unintendedconceptionscan be understoodas prop-
erties of the contentto be taughtor of the way it is taught.To overcome thatdiffi-
culty, a first idea could be to searchfor a new definitionof that contentwithin the
framework of mathematics as a science. In such an approach the meaning of
mathematicsis likely to be reducedto the text of its presentation.In that context,
the content to be taughtmight appearessentially as being more elementarythan
its scientific reference. Such an approachdoes not take into account that mathe-
matics is firstof all a tool to solve problemsor thatproblemsfor which mathemati-
cal concepts have been forged arepartof theirmeaning,just as partof theirmean-
ing resides in the context of theirdiscovery.
Mathematicalconcepts cannot be fully understoodif we do not know the type
of problemthey allow one to solve. Also, we have to know thattheirconstruction
is not only a deductiveprocess but also the resultof dialecticalconfrontationsof
differentpoints of view, togetherwith differentmetaphysicalconceptions.Finally,
we should know thatthis process of constructionis still not finished.
More often thannot, this historicalcontext of discovery cannot be carriedinto
the mathematicsclassroom. Because it is not by means of the same mathematical
activity or within the same epistemological context, the meaning constructedby
pupils may be qualitativelydifferentfrom thatof mathematicsas a scholarlytopic.
Consequently,we can no longerconsiderthe relationshipbetween mathematicsas
a science and mathematicsas a content to be taughtas being the result of a mere
process of elementarization.
On the otherhand,mathematiciansare not solely responsiblefor deciding what
is to be taughtand how. At least in France,thatis the resultof a social interaction
within which teachers and mathematiciansas well as parents, politicians, and
industrialistsare involved. All these interactionscontribute to constituting the
specific epistemology of the content to be taught. The concept of didactical
transpositioncoined by Chevallard(1985) aims at giving a theoreticalframework
to the study of this process, a process by which some mathematicalknowledge is
transformedto become teachable.The most general constraintsthat call for the
didacticaltranspositionprocessandat the same time assign its formareas follows:
compatibilityconstraintsbetween the didactical system and society, ideological
constraintspertainingto the different sociological groups involved, and finally,
constraintsspecific to didacticalfunctioningin the strictsense (Chevallard,1982).
It shouldbe emphasizedthatthe didacticaltranspositionis unavoidable,because
of constraintsspecific to mathematicsteaching. I would like to mention two of
them:
1.Any contenthas to be embeddedin a contextin orderto be teachable.But later
264 Towardsa Problematiquefor Researchon Teaching

it has to be takenout of thatcontext to exist as genuine mathematicalknowledge.


This context, togetherwith the way it is moved aside, becomes partof the mean-
ing for the learnerof the mathematicalcontenttaught.
2. Any contenthas to be supportedby the pupils' previousknowledge. But this
old knowledge can turn into an obstacle to the constitutionof new conceptions,
even thoughit is a necessaryfoundation.But moreoften thannot, to overcomethis
obstacle is partof the constructionof the meaningof the new piece of knowledge.
Forthis reason,following Bachelard(1938), we call it an epistemologicalobstacle.
Before searching for hypothetical good didactical transposition-even if we
supposethatan optimalone exists-we must describeits function,answeringthe
question: How can the didactical transpositionbe characterizedso that we can
predictwhich meaningsit might allow learnersto construct?

AN EXAMPLE
OFTHEDESIGNOFA DIDACTICAL
SEQUENCE:
THESUMOFTHEANGLESOFA TRIANGLE
I illustratethis approachto researchon mathematicseducationwith an example
takenfrommy own field of research,which is thatof problemsrelatedto the learn-
ing and teachingof mathematicalproof (Balacheff, 1988a, 1988b, 1988c). I pres-
ent in detail the constructionof a didacticalprocess designed to allow pupils to
formulate a conjecture and then to prove it; I then present one of the results
obtained.
It is well known thatpupils have greatdifficultiesin learningwhata mathemati-
cal proof is. Very often, teachersand researchersmentionthe insufficientlogical
maturityof pupilstogetherwith theirlack of awarenessof the necessity for proofs.
To some extent I agree with these statements,but ourprobldmatiqueleads me to
go a bit beyond these remarksto addressthe following questions:
1. What is a mathematicalproof for mathematiciansas professionals,and what
is it as a contentto be taught?What is a mathematicalproof as partof the mathe-
maticalactivity within the classroom?
2. On whatbasis can pupils constructa meaningfor the notion of mathematical
proof?
3. Whatare the contexts in which mathematicalproof can appearas an efficient
or relevanttool for solving problemspupils have recognizedas such?
I will here concentrateon the last question. Since it is usually forgottenthat as
children,pupilsarelogical enoughto cope with most of the problemsthey encoun-
ter in everydaylife, this problemis often discussed as a linguistic/formalgap be-
tween the logic of common sense and mathematicallogic. But thisprobl"matique
misses a key point: Mathematics,unlike everyday life, is concernedwith theory.
The key word in mathematicsis rigor;in everydaylife it is efficiency. Thatmeans
that the teachingprocess should allow for this shift in pupils' interestfrom being
practitionersto becoming theoreticians(Balacheff, 1987).
Thus, to raise the problem of proof in the mathematicsclassroom, we need to
Nicolas Balacheff 265

shift to pupilsthe responsibilityfor the truthof some mathematicalstatement.Such


a statementhas first to be recognizedas a conjecture;thatmeans thatit is not mere
speculationbut thatpupils considerit plausibleand also sharea sufficientinterest
in knowing whetherit is trueor not. The formulationof the proof shouldbe justi-
fied not by an injunctionof the teacherbut by an intrinsicneed, which could stem
from a debate among pupils aboutthe validity of the conjecture.
Our study of the characteristicsof such a situationfor validationwas based on
the constructionandanalysisof a didacticalprocessin which pupils about 12 years
old discover, formulateas a conjecture,and then try to prove that the sum of the
measuresof the threeangles of a triangleis 180'.
Outlines of the Situation
Pupils' conceptions of the notion of angle are likely to lead them to assert that
the largerthe triangle,the largerthe sum of its angles (Close, 1982). Because of
this conception, the value of a proof proposed by the teacher, even after some
manipulations,are doubtful,because (a) the assertionitself might appeararbitrary
insofar as results like 182' or 178' are pragmaticallyas good candidatesas 180',
and (b) the pupils will be left with an open conflict between theirintuition(Fisch-
bein, 1982) and the authorityof the proposedproof.
Let us try to solve this didacticalproblem,using these initial,wrongconceptions
of the pupils in orderto lead them to the intendedconjecture-and thus to a new
conception-and then to cope with the problemof its truth.For that purposewe
identify four main constraints:
1. It is not possible to tell the pupilsbeforehandthatthe purposeof the sequence
will be to establish that the sum of the angles of a triangle is 180'. That would
destroy the problem, because the assertion would no longer be considered as a
conjecture;the studentknows the teacheralways tells the truth.This is a classic
example of one of the basic beliefs held in the didacticalcontract.
2. The validity of the measurementof a particularset of trianglesas a means to
establishthe conjectureshouldbe dismissed. But this decision shouldbe takenby
the pupils on theirown and not imposedby the teacher;otherwisethey will seek a
proof thatis acceptableto the teacher.
3. The situationwe design should elicit the pupils' conceptions about the rela-
tions between the size of a triangleand the value of the sum of its angles, because
it is from the contradictionbetween this conceptionand the fact thatthe sums are
around180' thatthe conjecturecould stem. This requiresa situationfor action.
4. We should provide the classroom with a situation for validation oriented
towardthe constructionof a proof of the conjecture.That supposes a didactical
contractin which the pupils have the responsibilityfor the truthof the conjecture.
This is possible only if they have had the responsibilityfor formingthe statement
of the conjectureitself.
It is under these constraintsthat a sequence of didactical situations has been
designed. Note that the following teaching setting is not the solution to the prob-
266 Towardsa Problematiquefor Researchon Teaching

lem I have formulated.Whatis importantis the relationshipsbetween the theoreti-


cal analysis and the constructionof this didacticalprocess.

Conditionsfor the Genesis of a Conjecture


It is possible to ask the pupils to measurethe threeangles of a trianglewithout
giving specific reasons:The teacherasks each pupil to carryout the task and then
to propose his or her result. The possible variety of the results has no specific
meaning for the pupils with respect to their conceptions because almost all tri-
angles aredifferent.I do not considerthis a situationfor actionbecause it does not
provide the mobilization of the conceptions specifically related to the intended
conjecture.
The following activityallows the pupilsto discriminate,fromthe variationin the
obtainedresults, between what is due to the measurementand what is explained
by theirconceptions.We confrontthe class with the computationof the sum of the
angles of a unique triangle.Each pupil gets a copy of the same triangle,and we
ask them to predict the sum of its angles. The predictions are recorded by the
teacherbefore the pupils startmeasuringand computing.We have proposeda tri-
angle large enough to activatethe expected conceptions.
After this task has been completedtwo things are done:
1. Each pupil is confrontedwith his or her predictionand asked for a comment
about a discrepancybetween the predictionand the result obtained.This request
should elicit a formulationof the possible conceptionsunderlyingthe prediction.
Thereis not necessarilya cognitive conflict,for as far as the pupil is concernedthis
discrepancycan be regardedas uniquefor the chosen triangle.This situationfor
action preparesfor the coming of the conjecture.
2. The teacherrepresentsthe collected resultson the chalkboardby means of a
histogramand then asks for comments.That leads to the problemof the determi-
nationof the exact value of the sum of the angles of a given triangle;it will appear
thatmeasurementis not a reliablemeans to an answer.

Towardsthe Birth of a Conjecture


To raise the questionof the invarianceof the sum of the angles, we need to have
pupilsmeasurethe angles andcomputethe sum in morethanone triangle.Because
the numberof trianglesmanipulatedwill not be very large, the set chosen is very
important.Takinginto account the pupils' conceptions, we use the shape of the
trianglesas a didactical variable: Pupils are likely to focus on the size of the tri-
angle andthe type of angle withinthe triangle.Thuswe choose threetriangleswith
shapes, and contrastbetween these shapes, sufficiently unusualto challenge pu-
pils when they are asked to predictthe sum of the angles (Figure 1).
The pupils work in teams of three or four, each team being asked to make one
predictionfor each trianglebefore any measurementand computation.The debate
necessary to make a decision elicits the underlyingconceptions and initiates the
constructionof argumentsfor or againstthe assertionthatthe sum of the angles of
Nicolas Balacheff 267

Figure 1. Triangleswith contrastingshapes.

a triangledependson its shapeor on its size. Because of the social interaction,this


situationhas the characteristicsof a situationfor decision.
After the task has been completed,each team is confrontedwith its prediction.
The teacherasks for a comment about a possible gap between the predictionand
the resultobtainedfor each triangle.The teacherrepresentsthe set of resultson the
chalkboardby meansof a histogramandasks for comments.Issues concerningthe
value of the sum of the angles for each triangleare discussed.
Actually, all that activity is not sufficient to ensure that the conjecturewill be
formulatedandrecognizedcollectively by the class. Two possible cases shouldbe
considered:
1. The sum of 1800 seems to be evident from a comparisonbetween the predic-
tions and the results of the measurements.But some pupils may still assert that it
is because of the particularchoice of triangles.In thatevent, the teacherchallenges
the class to find a triangle in which the sum of the angles is quite differentfrom
1800. The confrontationbetween the robustness of pupils' conceptions and the
difficulty in finding a triangle in which the sum of the angles is different from
180' leads to a formulationof the conjecturetogetherwith the problemof its proof.
2. The class supportsthe statement"Thesum of the angles of a triangleis 180'"9
as a conjecture.But because an appeal to measurementhas been dismissed, the
problemof constructinga proof on the groundof rationalargumentscan be stated.
Whateverthe case, the situationnow has the characteristicsof a situationfor
validation,because the class has the responsibilityto producea proof of the con-
jecture.The teacherstays aside; she or he has managedthe situationbut has never
offered any opinion aboutthe validity of the resultsproducedor of the conjecture.
268 Towardsa Problematiquefor Research on Teaching

Closure
To show thatthe sum of the angles of a triangleis 180', or to refute it, is now a
problemfor the class. It is an open-endedproblemfor which there is no evidence
that pupils will find any solution within the time constraints of the traditional
school context.
We have then to considerpossible scenariosfor a conclusion:
1. The pupils agree on a proof of the conjecture.Then the teacherjust has to
ratify it, providedthatit is acceptable.If it is not acceptable,then thereis a nego-
tiation to either reject it, suggest a modification of it, or even begin to develop
anotherproof.
2. The pupils do not agree on a single proof of the conjecture.Then the teacher
should managethe negotiationin orderto accept some proofs and reject others.
3. The pupils do not find any solution.Then the teacherhas the following alter-
natives:(a) to proposea solutionthatis consistentwith the pupils' conceptions,the
strategiesthey unsuccessfullyinitiated,and the level of proof they have revealed
(Balacheff, 1988b);or (b) to proposethatthey admitthe truthof the conjectureand
delay the productionof a proof.
Even if the conjecturehas not been provedby the pupils themselves,the knowl-
edge constructedthroughoutthis sequence should be quite different from what
they might have constructedafter merely observing some trianglesand having a
proof presentedto them. Here the propositionhas been developed as a conjecture
by pupils on their own. It has been discussed and settled as a genuine problem.
Even if the productionof a proof is now delayed, a real attempthas been made to
solve the problem.The proposalof the teacherhas practicalreasonsbut does not
rely on a prioriprinciplespupils do not know. This situationfor institutionaliza-
tion guaranteesthat what has been producedduringthe sequence is valid and is
genuinelyconsideredas knowledge.It implies thatpupilsandteacherrecovertheir
own place and responsibilitywithin the teachingsituation.
A Few Words About the Results Obtained
This didactical process has been developed in seven 7th-grade mathematics
classroomsin France,two of which have been videotaped(Balacheff, 1988a).The
mainresultI would like to presentis the one thatis specifically relatedto ourtheo-
reticalframework:the robustnessof pupils' conceptions.
In all the classroomsobserved, 180' appearedto be dominantrightfromthe first
activity,but the pupils' measurementsrangedfrom 160' to 260'. The pupils' pre-
dictionson the secondactivityconfirmedthe dominanceof 180', but the rangewas
quite large:from 160' to 770'. For the measurementof the common triangleal-
most all pupils found 180'. That is possible only if the resultof the measurement
has been correctedtowards180'. It mightbe proposedthatit is possible to end the
didacticalprocess at this point. The pupils seemed, from their behavior,ready to
acceptas truethat"thesum of the angles of a triangleis 180'." But in drawingsuch
a conclusion one would mistakeconformistbehaviorfor genuine knowledge. The
Nicolas Balacheff 269

thirdactivity evidenced the robustnessof the pupils' initial conceptions:Despite


the fact that the pupils were working collectively, half of them predicteda value
ratherdifferentfrom 180 for at least one of the given triangles.
The debate among the pupils, first within each team and then within the class,
initiatedthe constructionof the conjectureas such, insofaras it was challengedby
the claim that a trianglecould be sharpenedenough to have the sum of its angles
very small. Only at this stage of the didacticalprocess were the pupils' conceptions
called into question;the evolution towarda correctconceptioncould start,having
been activatedby the debate aboutthe validity of the conjecture.
The constructivisthypothesisis clearlysupportedby this experiment,which also
brings to light the implicationof social interactionsin the learningprocess. The
existence of a conformistcognitive behavioris probablyone source of the diffi-
culties that characterize effective teaching, for this phenomenon might allow
teaching to progressdespite the absence of real learning.

CONCLUSION
What I have presentedgives an idea of the probldmatiqueand its relatedtheo-
retical framework,on which are based what we in Francecall the recherchesen
didactique des mathe'matiques. As I have tried to show, the key word of this
probldmatiqueis meaning.Some basic questions,which have not been considered
in this article,are as follows:
*Whatmathematicalmeaningof pupils' conceptionscan we inferfrom an obser-
vation of their behavior?
*Whatkinds of meanings can pupils construct in the context of mathematics
teaching?
*Whatis the relationbetween the meaningof the contentto be taughtand thatof
the mathematicalknowledge chosen as a reference?
*Whatdeterminesthe transformationof mathematicsto constituteit as a content
to be taught?As a contenttaught?
*Beyond definitions, how can one characterizethe meaning of mathematical
concepts?
This researchis essentially experimental,which means that it relies on the ob-
servationof experimentalsettings specifically designed to answer precise ques-
tions. Ouraim is to constructa fundamentalbody of knowledgeaboutphenomena
andprocesses relatedto mathematicsteachingand learning.The social purposeof
such an enterpriseis to enable teachers themselves to design and to control the
teaching-learningsituation,not to reproduceready-madeprocesses. This knowl-
edge shouldallow teachersto solve the practicalproblemsthey meet, to adapttheir
practiceto theiractualclassroom.
But for practicalreasons this experimentalapproachis very difficult. Because
of time constraints,the observationof a sequencelike the one aboutthe sum of the
angles of a trianglecan be done only one to threetimes a year.Given what teach-
270 Towardsa Problematiquefor Researchon Teaching

ers are planningto do in theirclassroom,the periodduringwhich the experiment


can be conductedis quite short.Furthermore,to delegateothersto do the observa-
tion is very difficult,since at presentit is not well known whatmustbe said in order
to allow otherresearchersto repeatan experiment.Let me emphasize that one of
the mainobstacles we meet is the communicationwithinour researchcommunity.
That is stronglyrelatedto two essential open questions that concern researchon
mathematicsteachingas a scientific domain:
1. Whatdoes a researchresultconsistof? Whenwe design a teachingexperiment
with respect to some mathematicalcontent, the result is not the teaching setting
itself but the answerto the initial researchquestionor a new formulationof it, or
the evidence of intrinsiclinks between pupils' behaviorand some set of variables
whose controlconditionsthe teachingprocess, or even the principlesof the teach-
ing design.
2. Whatis a proof in our field of research?
Othertypes of researchexist, for example,the observationof real teachingsitu-
ations. This researchis not as well developed in Franceas it is in other countries.
Such researchmust be developedbecause it will be of crucialimportancein mak-
ing an effective relationshipbetweenresearchand practice.The confrontationand
the discussionof bothtypes of researchprojectscould be organizedarounda meta-
probldmatiqueabout which I would like to add a few words as a conclusion.
During the observationphase of an experiment,facts and events are recorded
andthenreportedwith an accuratedescription.But two majorquestionsoccurwith
respect to observation:
1. Not all the facts are relevantto researchin the didacticsof mathematics.But
which ones areto be retained?On the basis of which criteria?Indeedthe way rele-
vant facts are recognized is stronglyrelated to the theoreticalbackgroundof the
research.The discussion on this point could be organizedaroundthe concept of
didacticalfact: Withina teachingprocess what facts are relevantfor the purpose
of a didacticalanalysis?Fromwhat theoreticalbasis can the criteriafor recogniz-
ing didacticalfacts be derived?
2. Whena fact occursat a given momentwithinthe didacticalprocess,it implies
thatothershave not occurredat thatmoment.That seems quite clear.But it raises
an importantquestionfor ourresearch.Can we guess the set of possible didactical
facts to appearundercertainconditions?This a priorianalysis should be a meth-
odological principlefor researchbased on observation.It leads us to discuss the
necessity of the occurrenceof an event. For such an analysis we need a theoretical
background-some model to predictas precisely as possible, in a given situation,
what will be the pupil's behavior,the teacher'sbehavior,the interactionbetween
them, and so on. The meaningof an observedfact stems from both its occurrence
and the nonoccurrenceof otherpossible facts.
Finally, researchhas not been completed, whateverit is, since we have not ex-
amined the problemof the conditions for its reproducibility.What kind of infor-
Nicolas Balacheff 271

mationdo we have to communicateto enableotherresearchersto repeatan experi-


ment, to observe the same facts?To give an accuratedescriptionof the experimen-
tal setting and of the facts observedis not sufficient.We need also all the informa-
tion aboutthe theoryand the relatedpiroblmatiquethathas led to this hypothesis,
particularlyin the case of furtherfalsifications.

REFERENCES
Bachelard,G. (1938). Laformation de l'esprit scientifique.Paris:Vrin.
Balacheff, N. (1982). Preuveset demonstrationsau collbge [Proofsanddemonstrationsin high school].
Recherchesen Didactique des Math matiques,3, 261-304.
Balacheff, N. (1987). Processus de preuve et situationsde validation [Proof processes and situations
for validation].EducationalStudies in Mathematics,18, 147-176.
Balacheff,N. (1988a). Une dtudedes processus de preuve en mathematiqueschez des leves de college
[A study of proof processes in mathematicsof high school students].Th6se d'etat, Universit6Jo-
seph Fourier,Grenoble.
Balacheff, N. (1988b). Aspects of proof in pupils' practiceof school mathematics.In D. Pimm (Ed.),
Mathematics,Teachersand Children(pp. 216-135) Hodder& Stoughton.
Balacheff, N. (1988c). Treatmentof refutations.:Aspectsof the conmplexity
of a constructivistapproach
of mathematicslearning. Unpublishedmanuscript.
Boschet, F. (1983). Les suites num6riquescomme objet d'enseignement [Numericalsequences as a
content to be taught].Recherchesen Didactique des Mathdmatiques,4, 141--164.
Brousseau, G. (1981). Problemes de didactiquedes d6cimaux [Problemsin teaching decimals]. Re-
cherches en Didactique des Math matiques,2, 37-128.
Brousseau,G. (1986a). Theorisationdes phInomnnesd'enseignementdes mathematiques[Theoriza-
tion of the phenomenaof mathematicsteaching]. These d'6tat, Universit6de Bordeaux.
Brousseau,G. (1986b). Basic theoryand methods in the didacticsof mathematics.In P. F. L. Verstap-
pen (Ed.), Proceedings of the Second Conferenceon SystematicCo-operationBemteenTheoryand
Practice in MathematicsEducation. (pp. 109-161). Enschede, The Netherlands:NICD, 1988.
Chevallard,Y. (1982). Un exemple d'analyse de la transpositiondidactique[An exampleof the analy-
sis of the didacticaltransposition].Recherches en Didactiquedes Mathematiques,3, 157-239.
Chevallard,Y. (1985). La transpositiondidactique[Thedidacticaltransposition].Grenoble:La Pens6e
Sauvage.
Close, G. S. (1982). Children's understandingof angle at the primarylsecondarytransferstage. Un-
publishedmaster'sof science thesis: Polytechnicof the South Bank, London.
Douady, R. (1985). The interplaybetween differentsettings, tool-objectdialectic in the extension of
mathematicalability. In L. Streefland(Ed.), Proceedings of the NinthInternationalConferencefor
the Psychology of MathematicsEducation. (Vol. II, pp. 33-52) Utrecht,The Netherlands:State
Universityof Utrecht.
Fischbein,E. (1982). Intuitionand proof. For the Learningof Mathematics,3 (2), pp. 9-18, 24.
Galbraith,P. L. (1979). Pupils Proving. Nottingham: University of Nottingham, Shell Centre for
MathematicsEducation.
Grisvard, C., & L6onard, F. (1981). Sur deux r6gles impicites utilis6es dans la comparaison des
nombresd6cimaux [On two implicit rules used in comparingdecimal numbers].BulletinAPMEP,
340,450-460.
Izorche, M. L. (1977). Les rnels en classe de seconde [The real numbersat the tenthgrade](M6moire
de DEA) Bordeaux:IREMet Universit6de Bordeaux.
Kilpatrick,J. (1981). Researchon mathematicallearningandthinkingin the UnitedStates.Recherches
en Didactique des Mathdmatiques,2, 363-379.
Laborde,C. (1982). Langue naturelleet ccrituresymbolique[Naturallanguageand symbolic writing].
These d'6tat, Universit6Joseph Fourier,Grenoble.
272 Towards a Probldmatique for Research on Teaching

Nesher, P., & Peled, I. (1986). Shifts in reasoning.EducationalStudies in Mathematics,17, 67-80.


Perrin-Glorian,M. J. (1986). Repr6sentationdes fractionset des nombresd6cimauxchez des dl6ves de
CM2et du college [Representationof fractions and decimal numbersby pupils in CM, and high
school]. Petit X, 10, 5-29.
Robert,A. (1982). Acquisitionde la notionde convergencedes suites numeriquesdans l'enseignement
supirieur [Acquisitionof the notion of the convergence of numericalsequences in post secondary
school]. These d'dtat,Universit6de ParisVII, Paris.
Vergnaud,G. (1982) Cognitiveanddevelopmentalpsychology andresearchin mathematicseducation:
Some theoreticaland methodologicalissues. For the Learningof Mathematics,3 (2).

AUTHOR
NICOLAS BALACHEFF, Directeur de recherche CNRS, LaboratoireIRPEACS,CNRS, BP 167
69131 Ecully Cedex, France

Vous aimerez peut-être aussi