Solving Equations: The Transition from Arithmetic to Algebra

Author(s): Eugenio Filloy and Teresa Rojano

Source: For the Learning of Mathematics , Jun., 1989, Vol. 9, No. 2 (Jun., 1989), pp. 19-
25
Published by: FLM Publishing Association

Stable URL: http://www.jstor.com/stable/40247950

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms

is collaborating with JSTOR to digitize, preserve and extend access to For the Learning of
Mathematics

This content downloaded from

176.30.187.92 on Fri, 22 Sep 2023 11:39:19 +00:00
All use subject to https://about.jstor.org/terms
 Solving Equations: the Transition from Arithmetic
to Algebra
EUGENIO FILLOY, TERESA ROJANO

Recent researches have pointed to certain conceptual

immediate. It is necessary to construct, or acquire, some
and /or symbolic changes which mark a difference be-
elements of an algebraic syntax, properly speaking. The
tween arithmetical and algebraic thought in the construction
individ- of these syntactic elements is based on
ual. Some concern the different interpretations arithmetical
of lettersknowledge which has worked well up to a
[Booth, 1984], the notion of equality [Kieran, 1980, 1981],
certain point, but it must also break with certain arithmet-
and graphic or symbolic conventions for coding opera-
ical notions - hence the presence of a cut.
tions and transformations in the solution of equations
Consider the concept of equation. In arithmetical
[Matz, 1982]. These observations make it feasible to the
terms, hy-left side of the equation corresponds to a
pothesize certain lines of evolution from arithmetical
sequence to
of operations performed on numbers (known or
algebraic language which correspond to the notions and
unknown); the right side represents the consequence of
the forms of representation of the objects and operations
having performed such operations. This is what we might
involved in the changeover. The changes which the
call the "arithmetical" notion of equality. From such a
learner has to make to gain access to algebraicnotion,
languagean equation such as Ax + B' C can be solved by
can then be visualized, on each of these lines, as cut-points
merely undoing, one by one, the operations given in the
separating one kind of thought from the other. left hand sequence, starting with the number C. We shall
One of these cuts is particularly interestingcallfor this the
type of equation "arithmetical."
theme of problem solving: it is suggested by an analysis of
The arithmetical notion does not apply to an equation
the strategies and methods of solving systemsofof theequa-
form Ax + B = Cx + D; its resolution involves
tions found in the pre-symbolic algebra textbooks of the drawn from outside the domain of arithmetic
operations
13th, 14th and 15th centuries [Boncompagni, 1954;
- that is, operations on the unknown. For these opera-
Arrighi, 1974; Arrighi, 1970; Hughes, 1981; Benedetto
tions to&become meaningful to the learner it is necessary,
Arrighi, 1967]. We can observe, for instance, in turn,
that thefor such an equation (which we call "non-
solution strategies for equations such as x2 + c = arithmetical")
2bx and to be provided with some meaning. This, in
x2 = 2bx + c are completely different from each other
its turn, implies a change in the concept of equation or of
[Hughes, 1981; Filloy & Rojano, 1984]. This difference
the equality of numbers. With respect to the "meaning" of
would not exist if the authors had had recourse to the rule a non-arithmetical equation, the learner must at least
of transposing terms from one side of an equation to the understand that the expressions on both sides of the
other for, at a syntactic level, the two equations would equals sign are of the same nature (or structure), and that
then be similar. But this facility would already imply an there are actions which give meaning to the equality of the
advanced ability to operate on the unknowns in the expressions (for instance, the action of substituting a
equations. numerical value for the unknown).
The operational insufficiency exhibited at the pre- Profound changes in arithmetical habits and concepts
symbolic stage of algebra suggests the presence of a cut, do not occur spontaneously at the moment the learner is
i.e. a break in the development concerning operations on faced with the need for such a change. Children in our
the unknown, now at the level of individual thought. In study, when faced for the first time with equations of the
our clinical study with 12-13-year-old children, "Operat- form Ax+ B= Cx approached them with trial-and-error
ing on the unknown", as part of our more general study, methods, giving no indication that they spontaneously
"The acquisition of algebraic language", we show that attempted to operate on the terms containing x [Filloy &
operating on the unknown appears as an action which is Rojano, 1984, 1985]. Suitable interventions from a teach-
required for the solution of certain first degree equations er at the point of transition may be crucial for students
with at least two appearances of the unknown. To solve learning algebra for the first time.
these equations it is not enough to invert the operations On the other hand, although some modification of the
on the co-efficients. Examples of such equations are arithmetical notions must take place in order that the new
38jc + 72 = 56* ; 3x + 20 = x + 164 algebraic notions may be acquired, it is also necessary for
the arithmetical knowledge to be preserved. Even in the
According to our study, the transition from the opera- case of the single example we have presented, it is neces-
tional resolution of equations such as x + 27 = 58 or 4(x + sary that arithmetical operations continue to be recog-
1 1) = 52 to the resolution of the above equations is not nized as such so that all the previously acquired opera-

For the Learning of Mathematics 9, 2 (June 1989)

FLM Publishing Association, Montreal, Quebec, Canada 19

This content downloaded from

176.30.187.92 on Fri, 22 Sep 2023 11:39:19 +00:00
All use subject to https://about.jstor.org/terms
tionality in the resolution of equations is preserved. What In order to solve these it is sufficient to invert or to "undo"
is required is an operational level of knowledge which can the indicated operations. // is not necessary to operate on
be placed between the arithmetical and the algebraic one, or with the unknown.
i.e. a level of pre-algebraic knowledge.
C After") Students have received no instruction on how to
The transition stage
solve equations of the types
As we have said above, the students' conceptions of oper- Ax ± B = Cx
ations performed on numbers have to change in order Ax±B~Cx ±D.
that the concept of operating on objects other than To solve these it is not sufficient to invert the indicated
numbers (such as unknowns) and the conception of these
operations. // is necessary to operate on what is repre-
"new" objects themselves (what they represent, or may sented
come to represent) may be developed. The teaching of
algebra at this point must have recourse to suitable didac-
The population involved in our study consisted of three
tic means which bring into play all the related elements
consecutive classes in the second year of an experimental
which are involved in the change.
secondary school in Mexico City. A written "pre-algebra"
One of two opposed positions about the type of didac-
test was given to all the students containing three types of
tic resource to be used proposes modelling in some con-
questions: arithmetical equations with letters (e.g. 5jc + 3 =
crete context (i.e. familiar to the students) the new opera-
90), arithmetical equations without letters (e.g. □ - 95 =
tions and objects so that they become endowed with
23), and word problems leading to simple arithmetical
meaning. Taking this as a starting point, the first elements
equations. Children whose scores on all three parts of the
of an algebraic syntax are constructed on the basis of the
test were comparable, falling into either the bottom third,
behaviour of the model. The contrary didactic position
middle third or top third of possible scores, were selected
proposes beginning at the syntactic level, learning the
for interview and the interviews videotaped.
appropriate syntactical rules and applying them, later on,
During the interviews the students were presented with
to the resolution of equations. The latter is the traditional
sequences of five or more items to work on, illustrated by
teaching approach based either on the Viète model
the following examples.
(transposition of terms from one side of the equation to
the other) or the Eulerian model (operating on both sides Sequence E. (verification of the pre-test)
of the equation with additive and multiplicative inverses). e.g. x + 5 = 8; 13.x = 39; (jc + 3) 6 = 48
If we adopt the first position, we find it necessary to
Sequence C. (the equation as equivalence)
know something about the processes which intervene e.g. x + 5 = 5 + 2;x + ;c/4 = 6 + jc/4; x + 5 = x + x
between the actions performed at the concrete level (i.e.
on the model) and the corresponding elements of alge- Sequence I. (operating on the unknown)
braic syntax which are obtained from them. These pro- e.g. 3 + 2x = 5x; Ix + 2 = 3x + 6; 10* -18 = 4*
cesses, which we call "abstractions from the operations", Sequence A. (word problems of "find a number
show certain standard characteristics in the course of
type)
their development by individuals, but they can also be
The analysis of the interviews, which confirmed the pres-
greatly influenced by differences between subjects and
they also vary with the model. ence of the didactic cut (especially in the work of the
To clarify some of the ideas we are putting forward we students in the top stratum), is described in detail in Filloy
and Rojano [1984]. The analysis also brings out the char-
go on to illustrate them by reference to our clinical study.
acteristic approaches of each stratum to the problem
"Operating on the unknown" posed by the cut, i.e. their spontaneous approaches to the
solution of the non-arithmetical equations in Sequences
Our analysis of texts that antedate Viète's The analytical
C and I.
art, together with our own development of experimental
teaching sequences, suggested to us the existence of In a the next section we will look at our attempt to study
didactic cut along the child's evolutionary line of thought
the effects of instruction using concrete contexts to model
from arithmetic to algebra. This cut corresponds to non-arithmetical
the equations. We focus on students with
major changes that took place in the history of symbolic
high scores on the "pre-algebra" test in order to reduce the
algebra in connection with the conception of the "un- chance that the "transition phenomena" are due to short-
known" and the possibility of "operating on the un- comings in the students' basic arithmetical knowledge.
known".
In terms of the curriculum, the cut is located at the Concrete models for operating on
transition between: the unknown
C Before") The students know how to solve arithmetical Although there are theoretical reasons for believin
equations of the types semantic approach to learning algebra is more l
Ax±B=C lead to good algebraic performance in later years
A(Bx ±C)=D purely syntactic one, this does not mean that the c
x/A = B tion of algebraic syntax can be immediately an
x/A = B/C derived from such an approach. Certain proce

20

This content downloaded from

176.30.187.92 on Fri, 22 Sep 2023 11:39:19 +00:00
All use subject to https://about.jstor.org/terms
abstraction have to be gone through; for example, of the transference to the more difficult ca
operations which are performed on the elements of the abstracted the processes from the
"concrete" situation, where new objects and new opera- carried out on the models.
tions are being modelled. Such processes presuppose in
turn others, such as the process of generalising the actions Analysis of the abstraction processes
performed on the model, and the process of discriminat- Over the course of the interviews, processes of abstraction
ing the various cases that can be modelled. of the operations on the "new" objects (the unknown, in
We used two models in our research: the balance and a this case) become manifest. We detected a number of
geometrical model. phenomena.
1 . The temporary loss of previous abilities, coupled with
The geometrical model
behaviors fixated on the models. The most frequent
As applied to Ax + B = Cx, where A, B and C are given
instances were the loss of operativity in solving the
positive integers and in this case, C> A. (The children
simplified (arithmetical) equation, i.e. a failure to rec-
were presented with particular numerical instances.)
ognize (C - A)x = B as an equation which the students
Step 1: translating the equation into the model could solve syntactically on the basis of their earlier
i- A - i- C- j experience. We diagnose a "centration" or "fixation"
on the model which precludes the students detaching
the simplified equation from the concrete meaning that
« i°i ii the model confers on it.

Step 2: Example. VT, age 13, who showed

comparing considerable profi-
areas
ciency in solving arithmetical equations, including those

fe tin
with negative solutions, is working on the equation 8x +
30 = 5x + 9. Using the geometric model, VT derives the
simplified equation 3x + 30 = 9.
•- A-«C-A VT: "How . . . ?"
I: "Do
Step 3: producing the new equation, (C you
- A)x - think
B you can solve this one?"
VT: "Nine minus . . . nine minus a number larger than
Step 4: solving the new equation nine is going to give a negative number, divided by
Step 5: verifying the solution three . . /

There is a long pause, and when the interviewer requests

The balance model
an answer, VT begins to reexamine the geometrical con-
Again, as applied to an equation of the type AxEventually
figuration. + B- Cx the interviewer asks her to try and
Step 1: translating the equation intosolve
the themodel
equation without using the diagram.

1 QDQ..D ■■ ■ I I on a., a
I: "Can you solve it now?"
VI: "No. . .(pause). . . Well, if you want me to, I'll. . ."
(A) objects with equal (B) objects with equal (C) objects with same
(unknown) weight (known) weight (unknown) weight
She takes out her calculator, inverts operations and
obtains x= -7.

Step 2: repeated removal in pairs of objects of

VT resorts to the geometric model, though its application
unknown weight, maintaining equilibrium,
until none remain in the left hand pan not pertinent for this example, and in spite of the fact
is
that she already knows that the solution will be negative.
| The automation of her actions on the elements of the

(B) objects (C-A) objects model hides its meaning; even the performance of the
actions, which do not work, is not sufficient to make her
voluntarily
Step 3: writing the new equation, (Cabandon
- A)x the
- solution
B path. Since VT has
handled the model swiftly and fluently in many previous
Step 4: solving the equation items, her adherence to it here looks like inertial behavior.
Step 5: verifying the solution
2. The modification of the arithmetical notion of equa-
tion, (a) The modification arises because examples are
encountered whose structure does not correspond to
Those children who displayed a high level of pre-algebraic
that of the examples used in the instructional phase of
proficiency were provided with only the first step of either
modelling.
model, leaving to them the development of the subse-
Example. MT, also
quent steps with as little help as possible from age 13, isthe
working on a sequence of
inter-
viewer. Once they had mastered the items.use
She hasofalready
the abstracted
models from actions
foron the
the type Ax + B = Cx, they were given increasingly
model and ignores it in the resolution ofcom-
the equations. At
plex types: Ax + B = Cx+D;Ax~B= first Cx
she does+not D;A
perceivex-B= Cx of the equa-
the equivalence
- D, etc., so that we could observetions 2xhow
+ 3 = 5jc; they
3 + 2x = 5x;made
5x = 2x + 3; the
5x = 3 + 2x.

21

This content downloaded from

176.30.187.92 on Fri, 22 Sep 2023 11:39:19 +00:00
All use subject to https://about.jstor.org/terms
Item 16: 2* + 3 = 5* will give me 51?" Since such a number is not an integer,
MT: Writes: 2* + 3 = 5* - 2* = 3*. "So x equals one." she anticipates that it will be negative. It is not until the
Item 17: 3 + 2* = 5* interviewer focuses her attention on the multiplicative
MT seems unsure what movement of x must be effected. operation linking * and its coefficient that she regains her
former operativity and evokes the inverse operation to
The interviewer recalls the equation in Item 16. MT now
reach a solution. In other words, it was not sufficient for
recognizes their equivalence but admits she had not pre-
her to have reached an effectively algebraic statement; it
viously notice it.
was also necessary for her to have meanings for the opera-
Item 18: 5* = 2* + 3 tions embedded in the statement.
MT modifies it to solve it as in the previous item. She
writes: 4. The fixation on the model which may persist under-
5x = 2x + 3 neath an apparent algebraic operativity on the equa-
2x + 3 = 5x tion elements.
2* + 3 = 5*-2* = 3*.
Example. VT, after 28 items in the non-arithmetical
Item 20: 5x = 3 + 2x sequence, is still using the geometric model and shows no
In spite of the fact that MT recognizes the equation as the intention of giving it up.
"same" as the one in item 19, she does not assign it the
Item 29:8*~ 10 = 6* -4
same solution, only the same method of resolution:
5x = 3* + 2x 3x x = 1 VT: "Perhaps if I remove these parts first. . ."(indicatin
the "gaps" representing the -10 and the -4).
2. (b) Another way in which the arithmetical notion of
8 6
equation becomes modified is through the endowment
of the "new equations" with meaning. One way of ho I U
providing them with sense is through the verification
process.
x
Example. After solving 25 items in Sequence I, the first
few with the model and the rest at a syntactical level, MT
spontaneously gives a "more algebraic" interpretation ofVPs second attempt:
the equation 10* - 18 = Ax + 6 (Item 26).
8 6
MT: "In other words, they are equivalent."
I: "What do you mean by equivalent?"
MT: "If I find what x is and do this to it (pointing to the
left side of equation) and get a result, it has to be
equal to this result (pointing to the right side).

■ID -D
3. The use of personal codes to indicate actions to be
performed in the solution process. This is an inter-
mediate stage on the path to the development of a fully
algebraic syntax as the personal code becomes inade-
quate when the examples become more complex. 8 6

Example: MT rapidly gives up the use of the concrete

model and generates her from codes to indicate actions on
the elements of the equations.
Item 13: 129* + 51 = 231*
MT writes: 129* + 51 = 231* - 129* = 102*

"Therefore * equals two." (sic!)

Item 15:* + 5 = 2*
MT writes: * + 5 = 2*-* = 1* ■
"One * must be equal to five, * equals 5."
In the verification process relating to Item 13, MT also There
exhibits the "temporary loss of previous abilities" that weVT: "Well, * ... minus 10 ..."
have already referred to. In spite of the fact that she hasAfter some help from the interviewer.
the simplified equation (51 = 102*) in front of her, sheVT: "Two times *, minus 10, plus 4 (writes 2* - 10 + 4)
loses her previous facility and begins to ask herself (revert-must be equal to . . ."
ing to a more primitive level) "What number times 102I: "Nothing!"

22

This content downloaded from

176.30.187.92 on Fri, 22 Sep 2023 11:39:19 +00:00
All use subject to https://about.jstor.org/terms
Item 30: 23* - 7 = 14* + 2 Taking account of individual students' preferences or
Without any help VT produces: tendencies to choose certain solution methods helps us
clarify the interactions between the syntax and the seman-
23 14 tics of algebraic language. By differentiating the inter-
action phenomena that are strongly linked to a preference

x *- ~^i.
- 1 x - J *-
for algorithmic or analytic approaches, we minimize the
risk of making false generalizations about the evolution
of certain operations from concrete to syntactical form.
23 9 iu We have analyzed in detail the interview protocols of two
girls after the instructional intervention - two girls with
extreme and antagonistic tendencies: MT, in the high
stratum, age 13, with a marked algorithmic tendency, and
VT, in the same stratum, the same age, with a strong
semantic tendency. For both students, the same model
was used in the instructional phase on operating on the
unknown.
With the interviewer's hel
remaining area
Our mostis not
relevant findings were: zero.
VT: "Nine times * minus 7."
1 . The spontaneous development of the use of a concrete
I: "What is this equal to?"
model in order to operate on the unknown in an equa-
VT: "It equals two." Writes 9* + 7 = 2.
tion is not uniform even in students with similar levels
Item 32: 10*- 3 -Ax of pre-algebraic proficiency. The form of the develop-
Without any help, VT says, "That would be six times *, ment is strongly influenced by the individual student's
minus 3, equals zero", and writes 6x - 3 = 0. tendency to choose a particular approach. Extreme
cases were detected in which developments in the use of
5. A breaking away from the model, transferring the the same model were quite dissimilar. In one case (VT)
operativity on the coefficients to the terms containingwith an operative tendency, the development adhered
the unknown. At the time of transition this can lead toto the use of the model context even when the equation
the familiar error of combining terms of different types required very complicated modelling procedures.
degrees. In the other case (MT) there was a constant search for
Item 11: 15* +13=16* the syntactic elements present in the actions on the
model as they were repeated in equation after equation
MT:"16* minus 15* equals one. So one and 13 equals
14." and in type after type. The subject broke away from the
semantics of the model and started associating actions
Her attention is drawn to the geometric model.
on the model with a more abstract language through
MT proceeds again: 15* + 13 = 16* - 15*
the creation of personal codes, belonging neither to the
MT: "One * ... h'm. There should be one times * model nor to algebra.
»♦

2. Obstacles to the abstraction of the model operations

I: "And what does that have to be equal to?"
MT: "Thirteen." Writes 1*= 13 exist which do not depend on the particular model
being used nor upon the student's personal prefer-
ences. They arise from the degree of emphasis on the
6. The presence of obstacles peculiar to each model
modelling aspects which permit the use of the subject's
7. The recognition, through the use of models, of the previously acquired knowledge and operations. The
reduction of the new to the already known carries with
diversity of first degree equation types.
it the risk of hiding the difficulties of arriving at and
We discuss Results 6 and 7 in the following sections. using the new elements that must be learned. In the
process of abbreviating and automating actions on the
two models in use here, the operation on the unknown
Semantics versus syntax
can become hidden. In the geometric model, abbrevia-
Results 1-7 given in the previous section arise from the
tion leads to an ignoring of the linear dimension which
description and analysis of certain general patterns
represents the unknown; the operations on the areas
observed in the performance of the students in the high
stratum before and after they received instruction in reduced to operations on the "data" of the equa-
are
tions, so preventing the unknown from playing any
operating on the unknown. This analysis ignored individ-
ual differences in the children's attitudes or tendencies. role in the solution process.
These differences, however, are interesting when we come
to the "semantics vs. syntax" theme as we find that indi-
A
vidual students exhibit preferences towards certain solu-
tion methods, ranging from the most operative and algo-
rithmic to the most semantic and analytic. ÏI

23

This content downloaded from

176.30.187.92 on Fri, 22 Sep 2023 11:39:19 +00:00
All use subject to https://about.jstor.org/terms
In the balance model, due to the discrete nature of the plified equation is automatically expressed in terms of
coefficients of the unknown "x" as well as those of the the model and does not require further translation into
constant terms, operations of the same type can readily be an algebraic code.
performed on both kinds of numbers.
In the geometric model, on the other hand, it is not a
AX B C JC_ P trivial matter to realize that the resolution of the type
O D 0.0 ■•...■ OOO...O ■■...■ Ax + B = Cx + D requires the overlapping of the
rectangles representing the first degree terms Ax and
Cx).
Automation (in both models) leads studen
later on the errors typically associated w
syntax, such as attempting
ll
to add and su
cients of different degree. Even subjects

I
attachment to the model commit these er
their introduction of personal codes.
codes, incidentally, may be produced by

Mu llQs
actions:

Ax + B = Cx + D
►- A- ♦- C-A-H

or by introducing "artificial" parentheses when substitut-

H owever the transition to types A x - B = Cx and A
ing a numerical value for the unknown,
e.g. B + Ax-+>(A + B)x. B = Cx + D, while impossible for the balance mo
can be accomplished in the geometric model with t
Comparing the two models introduction of the operation of removing areas w
A third phase of our clinical study undertook a compara- negative term are involved. This extension does no
tive analysis of the use of the balance model in the resolu-violence to the semantics of the model.
tion of non-arithmetical equations and, for the geometric Of course, in the case of equations with negative
model, of the processes of abstraction from the opera- solutions, there is no way, with either concrete model,
tions on the model to the syntactic level.The comparison of producing a meaningful reduction of the equation to
allowed us to identify the phenomena relating to thean arithmetical one.
abstraction processes which do not depend on the particu-
lar model being used. We were also interested in varia-
tions between the models which we should take into Concluding discussion
The results we have obtained allow us to assert that the
account in recommending teaching strategies.
The most relevant results we found were the following:correction of syntactic algebraic errors, and of the opera-
tional difficulties that occur in resolving complex prob-
1. There are ways, specific to each model, of translating lems or equations, cannot be left to be spontaneously
equations into terms of the model that become obsta- resolved by children on the basis of their initial grasp of
cles to the further use of the model. In the case of theoperational algebraic behaviour. This is because the path
geometric model, the obstacle stems from the breakingof such spontaneous developments does not go in the
down of the constant term (B) into two linear factors direction of what algebra is intended to achieve. The
(the dimensions of the rectangular area representingcorrection of these errors is a task for education and, if we
B), i.e. one finds a b and an h such that b * h- B andwant to introduce certain algebraic notions by means of
so that either b=C-Aorh= C- A. This method fails
models (including purely syntactic models), we would do
when B is not a multiple of C - A and becomes an well to bear in mind the main components of modelling.
obstacle to the extension of the solution method to
Modelling has two fundamental concepts. One of them
non-integral equations. is "translation", by means of which objects and opera-
In the case of the balance model, assigning values to tions in abstract situations are endowed with meanings
the unknown weights can be an obstacle to the devel- and senses by being given more "concrete" manifesta-
opment of the "natural" strategy of cancelling pairs of tions. That state of things at the concrete level represents
identical weights. Also, in the context of the concrete another state of things at the more abstract level. In our
model, the notion of the unknown may be weakened: geometric model, for example, the equality of two areas
with an example such as 4jc + 6 = 2x, the value of the corresponds to the equivalence of two algebraic expres-
first x must be less than the value of the second.
sions. Starting from our knowledge of how problems are
2. Some transferences in the uses of the model are more resolved at the concrete level, we introduce operations
"natural" in one form of the model than the other. which have analogues at the more abstract level and
With the balance model, the step from solving the Ax + which will also lead to a resolution there. For this reason,
B = Cx to solving the type Ax + B = Cx + D is a small the translation must be a two-way process so that it is
possible to identify operations at the abstract level with
one since "iterated cancellation" will reduce both types
operations at the concrete level.
to arithmetical equations. In this model, too, the sim-

24

This content downloaded from

176.30.187.92 on Fri, 22 Sep 2023 11:39:19 +00:00
All use subject to https://about.jstor.org/terms
 A second component of modelling is the "separation" point to the need for instructional interventions at key
of the new objects and operations introduced by means of points in the processes which are unchained during the
the model from the details of the meanings appropriate to initial stages of algebraic language acquisition.
the concrete context. This is what MT attempts to do in
the example we have discussed: to detach herself from the
semantics of the concrete model since, ultimately, what is
sought is not the solution of a situation she already knows
how to solve, but the way to solve a more abstract situa- References
tion by means of more abstract operations. This second Boncompagni, B. Intorno ad Alcune Opere di Leonardo Pisan
component is what drives modelling towards the con- matico del secoio decimoterzo. Tipografia delle Belle Arti
1954
struction of an extra-modular syntax.
(Calandri, P.M.) Arrighi G. Tractato D'Abbacho. Domus G
The study we are reviewing here shows that mastery of Pisa, 1974
the first component can weaken or inhibit the second. Eco, U. Tratado de Semiotica General. Editorial Lumen, Espa
Subjects like VT achieve good control of the concrete (Francesca, Piero della) Arrighi, G. Trattato D'Abbaco. Dom
model but, because of this, develop a tendency to stay and leana, Pisa, 1970
progress within the concrete context. Fixation on the Filloy, E. History-Epistomelogy-Mathematical Didactics. Pap
sented at the Ninth Conference for the Psychology of Math
model can delay the construction of an algebraic syntax Education, International Group. Noordwijkerhout, The N
since this requires breaking away from the semantics of lands, 1985
the concrete model. Filloy, E. Teaching Strategies for Elementary Algebra and the Int
In cases of a more syntactic tendency, like MT's, we tionship between the Development of Syntactic and Semanti
note that obstacles may be generated in the course of ties. Proceedings of the Eighth Annual Conference for the P
ogy of Mathematics Education, North American Chapte
abbreviating the actions and producing intermediate Lansing, Michigan, 1986, pp. 108-113
codes (intermediate between the concrete level and the
Filloy, E. and Rojano, T. La Aparicion del Lenguaje Arit
fully syntactic algebraic level). These obstacles hinder the Algebraico. L'Educazione Matematica, anno V (3), Cagliari
1984, pp. 278-306
abstraction of the operations performed at the concrete
Filloy E. and Rojano, T. From an Arithmetical to an Algebraic
level and are due to a lack, in the transition period, of (A clinical study with 12-13 year olds). Proceedings of th
adequate means of representing the states to which the Annual Conference for the Psychology of Mathematics Edu
various operations lead. North American Chapter. Madison, Wisconsin, 1984, pp. 51
The obstacles arise from a sort of "essential insuffi- Filloy, E. and Rojano, T. Obstructions to the Acquisition of El
Algebraic Concepts and Teaching Strategies. Proceedings
ciency" in the sense that modelling (when it is sponta-
Ninth Conference on the Psychology of Mathematics, Intern
neously developed by children) tends to hide what it is Group. Noordwikjerhout, The Netherlands, 1985
meant to teach. When either of the two components is Filloy, E. and Rojano, T. Operating the Unknown and Mo
strengthened at the expense of the other, the new objects Teaching (A clinical study among children 12-13 with high
ciency in pre-algebra). Proceedings of the Seventh Confere
and operations become harder to see.
the Psychology of Mathematics Education, North American
The dialectic between the two components of model- ter. Columbus. Ohio. 1985. dd. 75-79
ling should be taken into consideration by teachers and an Filloy, E.; Rosshandler, R. and Trujillo, M. Reporte parcial del
attempt made to develop the two processes harmoniously Proyecto Adquisicion del Lenguaje Algebraico: Diagnostico sobre
so that they do not obstruct each other. From our analysis pre-algebra. 82/83. Seccion de Matematica Educative, CIN-
VESTAV del IPN, 1985
of the cases discussed here, it is clear that this is indeed a
Hughes, B. Jordanus de Nemore, De Numéris Datis (A critical edition
task for education. The second aspect of modelling - the and translation). University of California Press, Berkeley, Los
breaking away from previous objects and operations - is Angeles, London, 1981
a process which negates part of the semantics of the Kieran, C. The Interpretation of the Equal Sign: Symbol for an Equi-
model. These partial negations take place during trans- valence Relation vs. an Operator Symbol. Proceeds of the Fourth
Conference of the International group for the Psychology of
ferences of the model from one problem situation to Mathematics Education. Berkeley, California, University of Cali-
another (in our geometric model they occur in the transfer fornia, 1980, pp. 163-169
from one type of equation to another), but when the Kieran, C. Concepts associated with the equality symbol. Educational
generalisation of the model is left to spontaneous devel- Studies in Mathematics. 12, 1981, pp. 317-320
Matz, M. Towards a process model for high school algebra errors. In D.
opment by the children, they are just as likely to negate
Seeman and J.S. Brown Intelligent Tutoring Systems. Academic
essential parts of the model itself (in the geometric model, Press, 1982, pp. 25-50
they may negate the presence of the unknown, or the (Mazzinghi, M° Antonio De) Benedetto, M°, Arrighi, G. Trattato di
operation on the unknown, for instance). The interven- Fioretti. Domus Galileana, Pisa, 1967
tion of teaching is necessary to the development of the Rojano, T. De la aritmética al algebra (estudio clinico con ninos de 12 a
13 anos de edad). Doctoral dissertation. Centro de Investigacion y
processes of detachment from, and negation of, the de Estudios Avanzados del IPN, 1985, Mexico
model, in order to lead towards the construction of the Rojano, T. Learning and Usage of Algebraic Syntax: Its Semantic
new notions. Aspects. Proceedings of the Eighth Annual Meetings for the Psy-
The transfer of the problem situation (semantics versus chology of Mathematics Education, North American Chapter. East
Lansing, Michigan, 1986, pp. 121-126
algebraic syntax) to a level of actions on a model permits
Springer Series in Cognitive Development
the closing of gaps between teaching and the problem Springer Series in Language and Communication
situation. Through the analysis of the interactions at this (Viète, F.) Witmer, T.R. (tr.) The Analytical Art. The Kent State
new level, didactical phenomena come to light which University Press, U.S., 1981

25

This content downloaded from

176.30.187.92 on Fri, 22 Sep 2023 11:39:19 +00:00
All use subject to https://about.jstor.org/terms